how to write an essay on hate speech

• Table of Contents
• Random Entry
• Chronological
• Editorial Information
• About the SEP
• Editorial Board
• How to Cite the SEP
• Special Characters
• Advanced Tools
• Support the SEP
• PDFs for SEP Friends
• Make a Donation
• SEPIA for Libraries
• Entry Contents

Bibliography

Academic tools.

• Friends PDF Preview
• Author and Citation Info
• Back to Top

Hate Speech

Hate speech is a concept that many people find intuitively easy to grasp, while at the same time many others deny it is even a coherent concept. A majority of developed, democratic nations have enacted hate speech legislation—with the contemporary United States being a notable outlier—and so implicitly maintain that it is coherent, and that its conceptual lines can be drawn distinctly enough. Nonetheless, the concept of hate speech does indeed raise many difficult questions: What does the ‘hate’ in hate speech refer to? Can hate speech be directed at dominant groups, or is it by definition targeted at oppressed or marginalized communities? Is hate speech always ‘speech’? What is the harm or harms of hate speech? And, perhaps most challenging of all, what can or should be done to counteract hate speech?

In part because of these complexities, hate speech has spawned a vast and interdisciplinary literature. Legal scholars, philosophers, sociologists, anthropologists, political theorists, historians, and other academics have each approached the topic with exceeding interest. In this current article, however, we cannot hope to cover how these many disciplines have engaged with the concept of hate speech. Here, we will focus most explicitly on how hate speech has been taken up within philosophy, with particular emphasis on issues such as: how to define hate speech; what are the plausible harms of hate speech; how an account of hate speech might include both overt expressions of hate (e.g., the vitriolic use of slurs) as well as more covert, implicit utterances (e.g., dogwhistles); the relationship between hate speech and silencing; and what might we do to counteract hate speech.

1.1 The Harms of Hate Speech

2. religious hatred and anti-semitism, 3.2 dogwhistles and coded language, 4. pornography, hate speech, and silencing, 5.1 the case for bans, 5.2 objections to bans, and some responses, 5.3 the supported counterspeech alternative, other internet resources, related entries, 1. what is hate speech.

The term ‘hate speech’ is more than a descriptive concept used to identify a specific class of expressions. It also functions as an evaluative term judging its referent negatively and as a candidate for censure. Thus, defining this category carries serious implications. What is it that designates hate speech as a distinctive class of speech? Some claim the term ‘hate speech’ itself is misleading because it wrongly suggests “virulent dislike of a person for any reason” as a defining feature (Gelber 2017, 619). That is not, however, the way in which the term is understood among most legal theorists and philosophers. Perhaps it would be useful to start with some examples.

Bhikhu Parekh (2012) lists the following instances as examples different countries have either punished or sought to punish as hate speech:

• Shouting “[N-words] go home,” making monkey noises, and chanting racist slogans at soccer matches.
• “Islam out of Britain. Protect the British people.”
• “Arabs out of France.”
• “Serve your country, burn down a mosque.”
• “Blacks are inherently inferior, lecherous, predisposed to criminal activities, and should not be allowed to move into respectable areas.”
• “Jews are conspiratorial, devious, treacherous, sadistic, child killers, and subversive; want to take over the country; and should be carefully watched.”
• Distribution by a political party of leaflets addressed to “white fellow citizens” saying that, if it came to power, it would remove all Surinamese, Turks, and other “undesired aliens” from the Netherlands.
• A poster of a woman in a burka with text that reads: “Who knows what they have under their sinister and ugly looking clothes: stolen goods, guns, bombs even?”
• Speech that either denies or trivializes the holocaust or other crimes against humanity.

Robert Post’s four bases for defining hate speech might help us organize the features of Parekh’s list:

In law, we have to define hate speech carefully to designate the forms of the speech that will receive distinctive legal treatment. This is no easy task. Roughly speaking, we can define hate speech in terms of the harms it will cause—physical contingent harms like violence or discrimination; or we can define hate speech in terms of its intrinsic properties—the kinds of words it uses; or we can define hate speech in terms of its connection to principles of dignity; or we can define hate speech in terms of the ideas it conveys. Each of these definitions has advantages and disadvantages. Each intersects with the first amendment theory in a different way. In the end, any definition that we adopt must be justified on the ground that it will achieve the results we wish to achieve. (Herz and Molnar 2012, 31)

The four definitional bases are in terms of: (1) harm, (2) content, (3) intrinsic properties, i.e., the type of words used, and (4) dignity. One could also attempt a hybrid definition by combining the ways mentioned. But, as is made clear in Post’s remarks, definitions of this sort are relative to the interests of the definer; “We must evaluate the status of ‘hate speech’ so defined in order to determine whether it achieves what we wish to accomplish and whether the harms of the definition will outweigh its advantages” (Herz and Molnar 2012, 31). The upshot is a rejection of a univocal definition that captures “the essence” of hate speech as a phenomenon.

It is important to note that many definitions of hate speech will not fall squarely within the categories Post outlines. For instance, the UN’s International Convention on the Elimination of All Forms of Racial Discrimination identifies hate speech both in terms of its content and its harmful consequences. Most definitions tend to characterize hate speech in multiple ways.

Harm-based definitions conceive of hate speech in terms of the harms to which targets are subjected. Things like discrimination or linguistic violence are candidates, though some (Gelber, 2017) argue that hate speech can harm one’s ability to participate in democratic deliberation. Susan Brison (1998a) offers a disjunctive definition that centers on a kind of abuse to targets. She defines hate speech as “speech that vilifies individuals or groups on the basis of such characteristics as race, sex, ethnicity, religion, and sexual orientation, which (1) constitutes face-to-face vilification, (2) creates a hostile or intimidating environment, or (3) is a kind of group libel” (313). ‘Harm’ as used by Brison refers to what Joel Feinberg describes “as a wrongful setback to (or invasion of) someone’s interests” (Brison, 1998b, 42).

Perhaps an immediate reaction to disjunctive definitions of the sort Brison offers is skepticism about the definitiveness of the purported list. When we go to test the definition’s application, we invariably find contestable inclusions and exclusions. Recall the examples from Parekh at the start of this section. Something like “Arabs out of France” might be included as an instance of hate speech on Brison’s account on the grounds that it creates a hostile or intimidating environment. Should statements that communicate a similar message in a less abrasive manner also be included? Suppose “Only French Nationals should occupy France” is roughly equivalent content-wise to “Arabs out of France.” If the former is indeed a less abrasive presentation though communicating the same content as the latter, what are we to make of its status? Many will find the statement odious; many will not. And since it is certainly not a face-to-face vilification or form of group libel, classifying it as hate speech will depend on how likely it is to create an intimidating or hostile environment.

The previous objection might entice one to opt for a content-based view. Content-based views define hate speech as that which “expresses, encourages, stirs up, or incites hatred against a group of individuals distinguished by a particular feature or set of features such as race, ethnicity, gender, religion, nationality, and sexual orientation” (Parekh, 2012, 40). This version makes it easier to conceive of semantically equivalent statements that differ in manner of presentation as instances of hate speech.

Content-based accounts face the challenge of determining which contents meet this standard. If the content that distinguishes hate speech from other types of speech must express, encourage, or incite hatred towards groups or individuals based on certain features, then the proponent of this view will need an account of expression. Is the speech in view that which signals the presence of a particular mental state in the speaker (i.e., hate) or that which is likely to prime feelings of animosity in a specific audience?

Another issue facing content-based approaches concerns distinguishing between speech that “respects ‘the decencies of controversy’” and that “which is outrageous and therefore hate inducing” (Post, 2009, 128). The ability to express a wide range of views, even contentious ones, is a cherished aspect of democratic societies. Failure to observe this distinction would broaden the scope of what counts as hate speech perhaps too much. In order to make this distinction, one could follow Post in tying it to “ambient social norms” that distinguish outrageous and respectful behavior. One challenge though is in determining the content of those social norms. For instance, a minority group whose opinions have little impact on the makeup of norms are unjustifiably excluded from influencing the shape of their society’s civility norms.

Definitions of hate speech based on intrinsic properties generally refer to those that emphasize the type of the speech uttered. What is at issue is the use of speech widely known to instigate offense or insult among a majority of society. Explicitly derogatory expressions like slurs are paradigmatic examples of this type of view. In general, the type of speech identified on this account is inherently derogatory, discriminatory, or vilifying.

Though attractive at first glance, classifying hate speech along these lines might prove to fall short in two ways. First, defining hate speech in this way might be too constricting. Some of the examples in our initial list would seem not to count as hate speech since they arguably lack the intrinsic features. “Arabs out of France,” for example, does not contain explicitly slurring terms. And second, this definition might prove too expansive. In cases where slurs are reappropriated by members of the target group or where artists incorporate them into a creative work, it would appear odd to count these as instances of hate speech. The concern is tied specifically to locating the issue in the terms themselves, as opposed to the use to which the terms are put.

Perhaps a final challenge to intrinsic property views can be derived from the work of Judith Butler (1997). On Butler’s account, hate speech is a kind of performative that is “always delivered twice-removed, that is, through a theory of the speech act that has its own performative power” (96). More specifically, “[w]hat hate speech does … is to constitute the subject in a subordinate position” (19). Butler locates the trouble with hate speech in its perlocutionary effects, a concept introduced by J.L. Austin that refers to the effects a speech act can have on its audience. An example of a perlocutionary effect is feeling amused at a joke or frightened from the telling of a ghost story. Unlike with intrinsic property definitions, Butler shifts focus to the nature of the acts performed rather than the terms in use. (For a critical look at Butler’s account, see Schwartzman (2002).)

Lastly, dignity-based conceptions focus primarily on the role of harms to the dignity of targets of hate speech. For instance, both Steven Heyman (2008) and Jeremy Waldron (2014) appeal to dignity in their accounts. Broadly speaking, hate speech on this kind of conception amounts to speech that undermines its target’s “basic social standing, the basis of [their] recognition as social equals and as bearers of human rights and constitutional entitlements” (Waldron, 2014, 59). This conception of hate speech will also include characterizations in terms of group defamation or group libel. Section 130 of Germany’s penal code is an example of legislation that incorporates a dignity-based conception of hate speech, prohibiting “attacks on human dignity by insulting, maliciously maligning, or defaming part of the population” (see Waldron, 2014, 8).

Worries about application follow dignity-based conceptions as well. Firstly, there may be questions about how we, in particular instances, are to distinguish between false statements about a group as a whole and those about a particular member of a group (Brown, 2017a). Presumably, only the former is consistent with an understanding of hate speech as a group-based phenomenon. Secondly, an implication of the view appears to be that it expands the range of things that would count as hate speech. Any speech that calls into question the basic standing of certain groups falls under this notion, which may make it more difficult to distinguish between contentious political speech and hate speech.

Perhaps a lesson to draw from the profusion of disjunctive definitions is a general skepticism about a definitive description of hate speech. We might concur with Alexander Brown that ‘hate speech’ is an equivocal term denoting a family of meanings (Brown 2017b, 562). According to Brown, ‘hate speech’ isn’t just a term with contested meanings, but rather, it is “systematically ambiguous; which is to say, it carries a multiplicity of different meanings” (2017b, 564). Because the expression is what is typically referred to as an essentially contested term, the hunt for a univocal or universal definition is futile.

The harms that have been attributed to hate speech comprise a long and varied list, ranging from the immediate psychological harms experienced in the moment by the person(s) targeted by an instance of hate speech, to much more long-term impacts that affect not only those targeted but whole communities, and even the strength of an entire nation.

A distinction between “assaultive hate speech” and “propagandistic hate speech” is helpful when discussing these harms (Langton 2012; 2018a; see also Gelber and McNamara (2016) who discuss “face-to-face encounters” and “incidences of general circulation”). Hate speech yelled at an individual on the street, or from a passing car, is a face-to-face encounter, and an assaultive speech act. This is, moreover, most often inter-group hate speech, where the speaker(s) are, for example, white, and the targets are non-white. On the other hand, propagandistic hate speech is often intra-group speech, spoken by members of one group to fellow ingroup members (e.g., a white person to other white people). The newsletter of the KKK, therefore, would fit into this category.

While this distinction is helpful to keep in mind, it should also not be overstated. Summarizing the results of their study which surveyed the experiences of the victims of hate speech, Katharine Gelber and Luke McNamara conclude that “the distinction between face-to-face encounters and general circulation hate speech is not always clear in the everyday experiences of racism endured by targets” (2016, 326). Any one instance of hate speech might fall into both categories. For example, it may occur in its first instance as an assaultive speech act, and then reports of the event may then take on a propagandistic aspect, as it is spread among the community. Similarly, even if an instance of hate speech is intended as a piece of propaganda, it may, when encountered by a member of the community it disparages, be akin to assaultive speech.

Still, this distinction helps reveal the wide range of the types of speech acts that are plausibly harmful, and also offers insight into how they harm. For example, Waldron (2014) focuses mainly on hate speech in its propagandistic mode, which he argues undermines the public assurance of equal social standing that members of non-dominant communities are entitled to—in his terms, their assurance of dignity. On this view, public hate speech—e.g., flyers that read ‘Muslims Out!’—is “an environmental threat to social peace, a sort of slow-acting poison, accumulating here and there, word by word” (2014, 4). Its harm is therefore one that attacks the broader society, and not just individuals targeted by hate speech.

On the other hand, the essays in the classic Words that Wound tend to focus more on what its authors term “assaultive speech,” that is, “words that are used as weapons to ambush, terrorize, wound, humiliate, and degrade” (Matsuda et al. 1993, 1). This leads them to focus more on hate speech’s ability to produce “direct, immediate, and substantial injury” (Lawrence, 1993, 57), such as “immediate mental or emotional distress” (Delgado, 1993, 93–94). On this approach, the most evident harms of hate speech are psychological. These psychological injuries scale up, however, when hate speech is endemic, and so result in the types of community or social harms highlighted by authors like Waldron. For this reason, the distinction between these approaches may be thought of as more a matter of emphasis.

This relationship between individual harms and broader social harms is also evident once we acknowledge the long-term effects of hate speech on victims, in addition to its more immediate impacts (Delgado and Stefancic, 2004, 14). Victims of hate speech may first experience “psychological symptoms and emotional distress” like heightened stress and fear in the immediate aftermath of assaultive hate speech, but they may also experience far-ranging consequences if they “modify their behavior and demeanor” to avoid receiving further hate messages, limiting their ability to participate fully in society (Matsuda, 1993, 24). Gelber and McNamara’s interview subjects confirm this complex web of effects that hate speech may cause, highlighting how “harms are often enduring and not ephemeral” (2016, 336). In this way, hate speech is both an immediate attack on one’s health and dignity, along with a threat to their community’s position in society. The cumulative effect of hate speech events, therefore, is a collection of harms located both in individuals and communities, which blurs the distinction between assaultive and propagandistic hate speech events.

Constitutive and Consequential Harms

Another distinction which is similarly helpful, but also fraught, is the distinction between constitutive and consequential harms—that is, harms that occur in the saying of some utterance of hate speech, and those that are its downstream results (see Maitra and McGowan, 2012, 6). This distinction draws on the speech act theory of J.L. Austin (1962) and has served an important role in the examination of hate speech from feminist philosophers of language (see, e.g., Langton, 1993; 2012; Maitra and McGowan, 2012; Maitra, 2012; McGowan, 2004; 2009; 2012; 2019; and others). Constitutive harms are those that correspond to what Austin called the illocutionary act , the act performed in saying X , while consequential harms correspond to perlocutionary effects , the results brought about by saying X . Most (though not all) of the harms surveyed above comprise consequential harms, as items such as psychological injury, feelings of fear, and societal withdrawal all most naturally fall into the perlocutionary effects category.

However, philosophers have also drawn attention to how hate speech can injure in a different way by indirectly affecting the positions of the social groups targeted by hate in a social hierarchy. That is, “by fixing facts about the distribution of social power, including facts about who has this power, and who lacks it” hate speech harms in a way not captured in the above account of individual injuries and their cumulative effects (Maitra and McGowan, 2012, 7). This is an immediate harm that occurs in the saying of the speech act, which (given appropriate circumstances and uptake) produces a shift in the normative landscape. It is in this way that an instance of hate speech may not only cause the injuries surveyed above but may also, for example, rank Indigenous Peoples as inferior, legitimate discriminatory behavior towards them (perhaps via incitement), or potentially silence them. (We return to the notion of silencing as an illocutionary harm of hate speech in Section 4 below.)

One reason to direct our attention towards the constitutive harms of hate speech is its potential to productively advance the debate over the legitimacy of potential restrictions. Mary Kate McGowan (2009) has made this case most explicitly. “Rather than focus on what a certain category of speech causes,” she writes, we ought to be “interested in what such speech actually does, in and of itself” (2009, 389–90). The idea here is that by focusing only on the harms caused by hate speech, we are inevitably drawn into a debate about balancing the costs and benefits of permitting or regulating speech, which often leads to an impasse. Alternatively, turning our attention to the acts hate speech constitutes can reveal features that help us avoid question of balancing harms, and opens the door to regulation. On this approach, some instances of hate speech can be seen to constitute acts of (verbal) discrimination, and should be considered analogous to other acts of discrimination—like posting a ‘Whites Only’ sign up at a hotel—that US law recognizes as illegal. As a speech act, hate speech can enact discriminatory rules in much the same way the physical sign does, and so ought to similarly be restricted (McGowan, 2012). This argument proceeds by a development of Austin’s notion of “exercitives,” which are speech acts that enact rules in a given domain, and is one example of the fruitful use of speech act theory to the philosophy of hate speech.

At the same time, however, it’s worth acknowledging that the distinction that this analysis relies on—between illocutionary acts and perlocutionary effects—is one that some argue is untenable (for one example, see Kukla, 2014). As illocutionary acts are indeterminate or incomplete without some form of audience uptake, it is difficult to articulate precisely how we ought to distinguish a speech act’s effects from its inherent qualities. Furthermore, the testimonials of victims of hate speech “suggests that there is a close and complex relationship between constitutive and consequential harms, and the harms are experienced cumulatively” (Gelber and McNamara 2016, 336–37). As such, any attempt to draw too neat of a distinction between these two types of harm risks misrepresenting victims’ experiences, and might tie the attempt to restrict hate speech unhelpfully to a philosophically contested distinction.

As a result, some caution must be applied when marking too stark of a contrast between these harms. Much like the distinction between assaultive and propagandistic hate speech, then, we can consider the distinction between consequential and constitutive harms to be analytically helpful in exploring the variety of harms attributable to hate speech, while recognizing that it is at the same time an abstraction from the on-the-ground realities of hate speech.

Religious belief is sometimes the source of putative cases of hate speech, and sometimes its target. In both cases, assessing the conceptual addition of religion to hate speech is a difficult task. Speech rooted in religious conviction is sometimes subjected to scrutiny to determine whether instances should count as hate speech or not. For instance, the Westboro Baptist Church’s demonstrations often make use of slurs and other explicitly defamatory language. This is an extreme case, which can be accommodated by extant hate speech legislation. Other cases, however, involve religious leaders making contentious statements—for instance, questioning the legitimacy or recognition of LGBT+ individuals, while claiming these are statements of love, not hate. Questions about religious speech of this sort concern whether it is simply contentious speech liberal democratic societies must tolerate or speech that runs afoul of deeply held norms that ought to be proscribed.

Some wonder whether religious sensibilities should be afforded special protection from offense. Amnon Reichman (2009), for instance, notes that some Israeli scholars have argued that providing special protection for religious beliefs is a good idea “so as not to push [religious] believers into having to choose between the authority of the state and the authority of their religion (namely, the authority of God)” (338). This relies on an assumption that religion is an institutionalized normative regime in competition with a legal regime where clashes over religious beliefs threaten the social fabric of society. It is in turn prudent to mitigate such clashes in order to avoid situations of unrest like the incidents involving comedic cartoons of Mohammed in the Dutch newspaper Jyllands-Posten and the French publication Charlie Hebdo .

It is not clear, however, that religious beliefs warrant special protection over other forms of belief that may be just as strongly held. Clashes over deeply held political beliefs can pose a similar threat to the social fabric as religious beliefs. Thus, there is no reason to think the same concern should not apply quite broadly. Providing certain types of speech special protection on these grounds would threaten to introduce quite repressive legislation on speech in general.

Holocaust denial, denial of the Armenian Genocide, and the denial of other crimes against humanity have also been the subject of special legislation, especially in Europe. As Michael Whine (2009) notes, 16 European states, as well as Israel, have criminalized Holocaust denial (543). In these contexts, at least one rationale for banning speech that denies or trivializes the Holocaust concerns its role in inciting hatred (Altman, 2012). One possible justification for such legislation rests on claims about what denial speech is. According to Martin Imbleau (2011), denial speech poses as an historical endeavor but is really propaganda. The denier’s aim is to “eradicate the awareness of the truth that prevents the resurgence of past criminal ideologies” (2011, 238). But if this is the rationale, it potentially opens up justifications for much broader application since similar claims might be made of other forms of propaganda. (For a general overview on Holocaust denial, see Robert Wistrich (2012) and Behrens et. al (2017).)

3. Slurs, Code Words, and Dogwhistles

As Parekh, Brison, and others have noted, hate speech can be expressed both explicitly and subtly. We can identify a few different expression-types that map onto the explicit and subtle instances, i.e., slurs , code words , and dogwhistles . The subtler forms may fall outside the scope of narrower conceptions of hate speech.

Perhaps the type of expression most often cited as the paradigm case of hate speech is slurs. Slurs are typically characterized as a type of insult that targets race, gender, sexual orientation, nationality, ability, politics, immigrant status, geographic region, and other categories. Much of the literature on slurs focuses primarily on the semantic and pragmatic properties of this linguistic class, with the expectation that such analyses also provide an account of how they in fact derogate their targets. There are, of course, competing accounts, some of which may be better suited than others for the purposes of legal and ordinary concepts of hate speech.

Before delving into competing accounts, it is good to put a working definition of ‘slur’ on the table. Typically, slurs are understood as conventionalized ways of demeaning and derogating individuals or groups of individuals and are contrasted with a co-referring neutral counterpart (Jeshion, 2013a; 2013b; Camp, 2013; Cepollaro, 2015). For instance, the following differ in regard to offense but are otherwise taken to make similar claims,

For many, (3.1) is regarded as offensive whereas (3.2) is simply a descriptive statement. The expression ‘cracker’ in (3.1) is a slur, while ‘white people’ in (3.2) is its purported neutral counterpart.

Though there seems to be widespread consensus that slurs have or could have neutral counterparts, not everyone shares this sentiment. Lauren Ashwell (2016), for example, denies that neutral counterparts (which she refers to as ‘neutral correlates’) play an essential role in identifying slurs. Ashwell claims that gendered slurs like ‘bitch,’ ‘slut,’ and ‘sissy’ derogate in ways similar to racial and ethnic slurs like ‘n***er,’ ‘k*ke,’ ‘cracker,’ and ‘sp*c,’ yet lack neutral counterparts. As a result, a definition need not include reference to neutral counterparts. In fact, making neutral counterparts central to defining slurs renders one incapable of accounting for terms that function similarly to slurs yet lack this purportedly central feature.

Ashwell makes a compelling case for the claim that gendered slurs lack neutral counterparts. Her larger claim that counterparts’ inessentiality for defining slurs has implications for pragmatic and semantic accounts that are also worth taking seriously. According to Ashwell, both sorts of accounts depend on the existence of neutral counterparts in their explanations of slurs.

Existing pragmatic accounts of slurs’ derogating capabilities are in particular trouble, for they tend to hold that a slur’s semantic contribution to a sentence is identical to the contribution that its neutral correlate would have had if it were used instead. This kind of account also leaves open the possibility that the slur could be sanitized—cleansed of its derogatory aspect—without semantic meaning change. … Existing semantic accounts, however, are not much better off—they are also structured to require the existence of a neutral correlate. (2016, 229)

For Ashwell, pragmatic and semantic accounts of slurs structurally require neutral counterparts, and so simply cannot jettison them. One response proponents of these kinds of accounts could give is that the gendered insults Ashwell highlights might exhibit properties that call their status as slurs into question. It could be open to these theorists to suggest that the terms they have identified as a matter of fact do carry neutral counterparts, that this is part of what distinguishes them as a class. And while the expressions Ashwell identifies seem to pattern in some ways like slurs, they also exhibit features that make them dissimilar. Thus, there is no need to wedge all insulting expressions into one class; there is room to expand our classifications in a way that preserves clarity.

Another important issue about slurs is their power to offend. Part of what makes them prime candidates for paradigmatic instances of hate speech is a widespread belief in their offensive potency. Indeed, much of the literature on slurs simply assumes they are offensive without offering much (if any) defense of that claim. It is not always clear whether the reader is supposed to understand offense as the provocation of a disliked mental state or as the violation of widely-accepted public norms.

Renée Bolinger (2017) discusses three ways to understand the claim that slurs are offensive:

• An audience actually took offense at a slurring utterance;
• The utterance warranted offense;
• Whether or not offense was warranted, it was rational for the audience to take offense.

The sense of ‘offense’ in (1) tracks how audiences actually respond at the moment of utterance. This could not be the sense in which offense is understood for at least two reasons. First, doing so would make the claim ‘slurs are offensive’ too strong. Since we would be tracking cases of actual offense, we would be focusing on particular uses of slurs, explaining what makes those utterances offensive rather than explicating the offensiveness of a linguistic class. As a result, the most natural interpretation of the claim would be that slurring utterances are invariably offensive, i.e., the use of slurs always provokes disliked mental states.

This, of course, raises a couple of questions. To begin with, does the strong claim deny the existence of non-offensive slur uses? Given things like linguistic reappropriation, some instances of indirect reports, and even instances of direct reports—especially by members of the slur’s targeted group—in which it is possible to utter slurs without provoking a disliked mental state in the speaker’s audience, the claim is obviously false. Further, there are also questions about who constitutes the audience . Is the relevant audience the one intended by the speaker? Everyone who witnesses the utterance? Only those who are present in the utterance situation? Because the claim must now be understood to be about particular slur utterances rather than the linguistic type, the claim must reflect the diversity of reactions provoked by different tokenings of slurs. A second reason is related to the questions about the audience: does everyone in the audience have to be offended, or is it sufficient if one, some, or a few are? What is the scope of the claim with respect to offended reactions? The answers to these questions will likely render the strong version of the claim untenable and weaker versions suspect. Thus, it is probably not the sense of ‘offense’ one should start with.

The sense expressed in (2) concerns moral justification for taking offense. Bolinger identifies three grounds for warranted offense at an utterance: intention, inappropriateness, and associations. A speaker may intend to offend, often doing so with expressions that are taboo, and thus considered inappropriate. Vulgar expletives like ‘fuck,’ ‘dick,’ or ‘shithead’ are typically viewed as inappropriate terms, at least in certain “polite” settings. Some expressions, like slurs, are not only inappropriate, but also carry associated attitudes and/or practices that amplify their offense. The swastika and confederate flag, for example, are both deeply associated with oppressive and genocidal practices towards Jewish people and African Americans, respectively.

This sense of offense still concerns one’s response to something, though it is not simply about how one reacts but one’s warrant to do so: “An utterance may warrant, but fail to actually generate offense merely because either there is no hearer, or the hearer fails to find the utterance offensive (perhaps because she shares the offensive attitude, fails to take it seriously, or misinterprets the utterance)” (Bolinger, 2017, 441). Bolinger notes that the associational offense category in particular is the one that is often the subject of hate crime legislation (ibid., 442). Such terms are often backed by formal social institutions, “adequately visible practices,” or a combination of both.

In (3), Bolinger uses ‘rational’ or ‘license’ to refer to the epistemic justification an audience has in taking offense at a slurring utterance. Here a gap opens up between what an audience member may be warranted in taking offense at as opposed to when it may be rational to do so. For instance, if a non-native speaker used a slur to refer to someone and we come to find out they were ignorant of the expression’s status as a slur, the target would still have been rational to take offense even if unwarranted. Undoubtedly, any of the three senses discussed may factor into an explanation of a given slur’s offense. However, theories of slurs are more appropriately aimed at capturing warranted and rational offense.

Consider again the following pair of statements:

The most straightforward explanation of the difference between (3.1) and (3.2) is that ‘cracker’ differs in some semantic respect from ‘white people’. Two of the most well-known versions of this approach are from Chris Hom (2008) and Elisabeth Camp (2013). On Hom’s account, ‘cracker’ as opposed to ‘white people,’ contains derogatory content. Slurs’ derogatory content is determined by the social institutions that undergird them, which consists of two components: an ideology and a set of practices . Hom defines an ideology as “a set of (usually) negative beliefs about a particular group of people” (431). As for the set of practices, these are racist practices that “can range from impolite social treatment to genocide” (ibid.). The two components combine to produce slurs’ semantic content, which contains a normative claim about the way individuals ought to be treated, because of possessing certain characteristics in virtue of being a member of an identifiable social group. (For alternative accounts of the relationship between slurs and ideology, see Kukla (2018) and Swanson (2015, Other Internet Resources).)

The pair of sentences in the example used here is illustrative of an observation many will have noticed when considering different examples. The slur in (3.1) is typically experienced as less offensive than ones that target members of marginalized groups. Language users recognize variation in offensive potency among slurs, some being more offensive than others. Hom refers to this as derogatory variation . Difference in the virulence of those backing racist institutions explains variation in offense on Hom’s account. Thus, ‘cracker’ is less offensive than slurs like ‘n***er,’ ‘sp*c,’ and ‘f*g’ because the racist and homophobic institutions backing them are much more virulent. (One might also wonder if there is any racist institution backing slurs for members of dominant groups at all.)

One objection raised against Hom’s view is that the semantic content he proposes of slurs is overwrought (Jeshion, 2013b). Robin Jeshion argues that Hom’s view “attributes highly specific sets of ideologies and modes of treating the group, yet it is doubtful that anything so semantically rich and well defined is semantically encoded in the slur” (318). That is, it is doubtful the racist means anything this racialized. Jeshion denies that slurs express anything as robust as Hom claims.

Camp offers an alternative semantic account in which slurs bear a close relationship to a perspective , which are “open-ended ways of thinking, feeling, and more generally engaging with the world and certain parts thereof” (2013, 335–336). According to Camp, a speaker’s slur use “signals a commitment to an overarching perspective on the targeted group as a whole” (ibid., 337). The perspective is a negative one that highlights certain characteristics or properties specifically associated with particular groups, ones that are presumed to warrant certain affective and evaluative responses.

What makes slurring perspectives a semantic feature for Camp is that they do not “merely signal … allegiance to a certain perspective,” but do so “in an overt and nondefeasible way, precisely in virtue of employing that expression” (ibid., 340). The use of a slur inserts a willful and noncancelable way of thinking about the target into a conversation. This is codified in the expression itself, and not something audiences “figure out” through the use of pragmatic mechanisms. This appears to be bolstered by the fact that one typically cannot erase a slur’s derogation by following up with a statement intending to do so, e.g.,

The tension of the contrast is one an audience might generally think finds its source in the meaning of the slur itself, rather than from features that emerge from the way language is used in a particular context. Further, as we saw in Jeshion’s objection to Hom’s view, the information slurs manage to convey isn’t very specific. This point is consistent with the open-ended nature of the perspectives Camp associates with slurs.

Though Camp’s account represents a marked improvement, critics still see shortcomings they believe should give us pause. Geoff Nunberg (2018), for instance, argues that Camp’s characterization of perspectives is too vague to capture the more specific colorings of slurs for specific groups: “Whatever distinguishes redskin from injun or nigger from coon , it’s more precise and richer than simply a disposition to think about the referents in certain ways” (Nunberg, 2018, 260–261). According to Nunberg, what is central for how slurs work is not the perspective the user employs to think about their target, but the allegiance it signals to a group or community disposed to think negatively of the target.

To take an obvious case, when you call a woman a shiksa you’re not just allying yourself with a disposition to think about gentile women in certain ways, but with the people who have that disposition. That group affiliation is primary and prior to the perspective it evokes: you can use shiksa appropriately without having any specific views of gentile women at all, but not without identifying with Jews. (ibid., 261)

For some theorists, the accounts offered by Hom and Camp leave out what they regard as an important aspect of slurring, namely, the role attitudinal expression plays in their derogatory power. These views agree that the difference between slurs and their purported counterparts is located in the realm of semantics; the previous accounts just leave out an important aspect. Jeshion (2013a) identifies three components of slurs’ semantics: (i) a truth-conditional component, (ii) an expressivist component, and (iii) an identifying component. The truth-conditional component of slurs corresponds to the same group referenced by its purported neutral counterpart. The expressivist component captures slurs’ ability to express contempt towards members of socially relevant groups in virtue of their group membership. Finally, the identifying component ascribes a property to the group that is seen as central to its identity. Mark Richard (2010) also proposes a view in which negative attitudes are included in the explanation of what slurs express. Jeshion and Richard’s accounts are typically referred to as expressivist views.

One issue expressivist views have been thought to have trouble with is derogatory variation . Derogatory variation refers to the sociolinguistic datum that slurs vary in their offensive potency. If we represent degrees of offense on a scale, slurs like ‘n***er’ and ‘k*ke’ are higher up on the scale than slurs like ‘cracker’ and ‘wop.’ Expressivist views have typically attributed one sort of attitude to slurs— contempt —which seems inadequate to capture the complexity of their offense profiles. For instance, consider co-referring slurs that vary in offense. Expressivist accounts appear to lack the resources to account for this variation. Thus, expressivism fails as an account of slurs for this reason.

Jeshion (2013a) attempts to answer this objection by arguing that her expressivist view “is only incompatible with versions of derogatory variation that stipulate that the variation derives from the semantics” (243). Jeshion maintains that focusing on slurring terms rather than particular utterances of those terms causes us to reflect on various factors at play that contribute to their power to offend. In effect, such focusing obscures the various factors brought to bear on judgments of offensiveness. Thus, Jeshion claims we ought to think our intuitive judgments about varying offense support the following thesis:

Derogatory Variation-Utterance : Utterances of different slurring terms engender different degrees of intensity of offensiveness. (2013a, 244)

Jeshion argues that this thesis is compatible with her account because weaponized uses of slurs are offensive for several reasons: semantic, pragmatic, sociocultural, and historical. As a result, there should be no expectation that a semantic view like hers need explain derogatory variation semantically.

Inferentialism describes slurs in terms of the kinds of inferences they license. Proponents of this kind of view include Robert Brandom (1994), Michael Dummett (1993), Lynne Tirrell (1999) and Daniel Whiting (2008). Tirrell, for instance, remarks that the “meaning of a word or expression is a matter of its various actual and possible sentential roles” (1999, 46). In characterizing the meaning of the now-outdated slur ‘boche,’ Dummett remarks,

The condition for applying the term to someone is that he is of German nationality; the consequences of its application are that he is barbarous and more prone to cruelty than other Europeans. We should envisage the connections in both directions as sufficiently tight as to be involved in the very meaning of the word: neither could be severed without altering its meaning. (1993, 454)

On Dummett’s account, to know the meaning of ‘boche’ is to make the inference from the referent being German to his being barbarous and more prone to cruelty than other Europeans.

Inferentialism also has its challenges. Timothy Williamson (2009), for example, opposes inferentialism by charging that it has difficulty explaining how non-bigots, who are not disposed to draw negative inferences, still understand their use. “We find racist and xenophobic abuse offensive because we understand it, not because we fail to do so” (257). We should note the inferentialist is not without resources to respond to Williamson’s charge. For example, Brandom’s (1994) inferentialism determines understanding in terms of grasping the broad network of inferential connections in which an expression is situated. An important implication is thought to be that different speakers will understand the expression similarly while associating it with different inferential roles, escaping Williamson’s charge that one must be disposed to draw slurring inferences to understand the term (see Steinberger and Murzi, 2017). However, Brandom’s view is itself controversial (For further objections to inferentialism, see Anderson and Lepore (2013b); Hornsby (2001).)

The last view we mention here is a stark alternative to the previous accounts, opting for a socioculturally-driven explanation. According to Luvell Anderson and Ernie Lepore, slurs are prohibited expressions whose tokenings provoke offense from those who value and respect their prohibitions: “What’s clear is that no matter what its history, no matter what it means or communicates, no matter who introduces it, regardless of its past associations, once relevant individuals declare a word a slur, it becomes one ” (2013a, 39). The prohibition is meant to apply not only to uses but mentions of expressions as well, including direct and indirect reports.

One objection raised against prohibitionism comes from Camp (2018). Camp asserts that though the view is simple and powerful, “it threatens to work too well” by failing to account for some complexities. In particular, Camp claims “slur’s truth-assessibility and projective behavior are more variable than [prohibitionism] predicts” (2018, 33). She believes, for instance, that it is sometimes easy to “quarantine” a slur’s offensiveness within a report like,

John thinks that the s**cs will have taken over the whole neighborhood in another couple years. But of course, I think it’s great that we’re developing such a vibrant Latino community.

The offense of the slur in this statement is judged to be relativized to John rather than the person reporting it.

Which view of slurs one adopts has implications for how one conceives of their harm. For instance, adopting a content-based view of slurs may encourage one to adopt a content-based definition of hate speech, which suggests that the harm produced is in the message being communicated. Adopting an expressivist view, on the other hand, could lead one to lean more towards an intrinsic property account. (For further alternative accounts to the ones mentioned in this section, see Popa-Wyatt & Wyatt (2017), Bach (2018), Croom (2011), Kirk-Giannini (2019), and Neufeld (2019).)

In addition to slurs, which are explicitly derogatory, researchers have also focused on more implicit forms of derogatory communication. Tali Mendelberg (2001), Ian Haney Lopez (2015), Jennifer Saul (2018) and Justin Khoo (2017) detail the use of racially coded language— dogwhistles— to access existing racial resentment while making surreptitious racial appeals. Saul provides a useful set of distinctions for thinking about dogwhistles: they can be explicit or implicit , and further, intentional or unintentional . Saul uses the work of linguist Kimberly Witten to define an overt intentional dogwhistle as,

a speech act designed, with intent, to allow two plausible interpretations, with one interpretation being a private, coded message targeted for a subset of the general audience, and concealed in such a way that this general audience is unaware of the existence of the second, coded interpretation. (2018, 362)

Saul illustrates this kind of dogwhistle with an example from George W. Bush’s 2003 State of the Union speech:

The phrase ‘wonder-working power’ is meant as an overt intentional dogwhistle for Evangelical Christians. According to Saul, there are two possible messages Evangelicals can take away from Bush’s utterance. The first message is simply a translation:

Yes there’s power, the power of Christ, in the goodness and idealism and faith of the American people. (362)

The second message is that Bush identifies with them, that he speaks their language. Saul thinks both are instances of overt intentional dogwhistles.

A covert dogwhistle, according to Saul, is “a dogwhistle that people fail to consciously recognize” (2018, 365). She is particularly interested in how covert intentional dogwhistles work in tandem with what psychologists have referred to as racial resentment , a belief system that is measured by the degree to which participants agree to the following four claims:

• Blacks no longer face much discrimination;
• Their disadvantage mainly reflects their poor work ethic;
• They are demanding too much too fast;
• They have gotten more than they deserve. (2018, 364 quoting Tesler & Sears 2010, 18)

According to Mendelberg, racial resentment remains widespread among white Americans even though explicitly racist appeals have come to be viewed as outside the bounds of acceptable political speech. (At least, that seemed to be the case up until the 2016 presidential election cycle.) White voters, on this model, tend to shy away from accepting explicitly racist proposals because they do not want to think of themselves as racist. The existence of racial resentment allows for the skilled intentional use of utterances that are unrelated to race on the surface yet access negative racial attitudes in a targeted audience, nudging them towards a particular course of action--e.g., voting for a preferred candidate.

An example of a covert intentional dogwhistle is the infamous Willie Horton ad used by the George H. W. Bush campaign in 1988. The ad targeted a prison furlough program in place during Michael Dukakis’s term as governor of Massachusetts. It presented a picture of Horton, an African American man, who while out on furlough raped a white woman and stabbed her husband. Though there was no explicit mention of race, it was clear to many that the ad drew on racial tropes about Blackness and criminality to stoke fear in white voters. In support of the interpretation that this was a covert dogwhistle, Saul notes that once the specter of race was raised about the ad, its effectiveness started to wane (2018, 366). The implication is that while the explicit appeal to racial resentment was a losing strategy, implicit appeal in the form of covert dogwhistles could be put to powerful use.

The unintentional dogwhistle is defined as an “unwitting use of words and/or images that, used intentionally, constitute an intentional dogwhistle, where this use has the same effect as an intentional dogwhistle” (2018, 368). Dogwhistles of this sort are passed on by unwitting others while achieving similar effects of the original intentional one. A special case of unintentional dogwhistles, what Saul calls amplifier dogwhistles , occurred when reporters and TV producers played the Willie Horton ad repeatedly. Presumably, the repeated presentations continued to make the associations between Blackness and criminality and, thus, continued to stoke fear and racial anxiety in significant portions of the white viewing public. For Saul, dogwhistles are therefore best understood functionally, and the difference in speaker-intentions between intentional and unintentional dogwhistles matters only insofar as we define them—their effects, in other words, are often identical.

The use of implicit means like dogwhistling—in both its covert and overt forms—can make the conceptualization and detection of hate speech more difficult. Undoubtedly, this poses a challenge for defining hate speech since dogwhistles are often designed to be innocuous. But what is it that explains the effects often attributed to dogwhistles? That is, how is it possible for language to work in this way?

Perhaps there is good reason to think something about dogwhistles’ meaning explains their effects. Consider, first, an ambiguity thesis that states code words have at least two meanings, a racial and a non-racial meaning. The expression ‘inner city’ in

purportedly expresses two meanings: (i) densely populated, high crime, urban areas, or (ii) poor African American (Khoo, 2017, 40). An ambiguous expression can be used in an utterance to produce a statement that leaves undetermined which interpretation is intended by the speaker.

One worry, however, is that terms like ‘inner city’ are not actually ambiguous. Khoo argues these terms do not behave like genuinely ambiguous expressions. Compare the following two sentences,

A reading of (3.6) is supposed to sound coherent given that ‘funny’ can mean ‘humorous’ or ‘strange’ whereas (3.7) is supposed to sound odd, even contradictory. If ‘inner-city’ were genuinely ambiguous in the way described above, we should be able to use it to mean ‘African American’ and get a coherent reading of (3.7).

A second view posits two dimensions of meaning for code words, at-issue and not-at-issue content. At-issue content is the main point of a speaker’s utterance, the directly asserted content that is foregrounded whereas not-at-issue content is projective , meaning it is able to survive embedding under operators like negation and modals (Tonhauser, 2012). Consider,

The at-issue content of (3.8) is represented by (a) and the not-at-issue by (b):

• John does not smoke.
• John used to smoke.

Note the difficulty in directly denying the not-at-issue content. If one were to follow an utterance of (3.8) with,

you would presumably find this odd and incoherent. A much more elaborate statement is needed to deny the not-at-issue content.

Applying this to view to ‘inner-city’ in (3.5), we end up with:

• At-issue : A poor, densely populated, high-crime, urban area.
• Not-at-issue : Those living in such areas are mostly African American.

Because the racial component of (3.5) is not-at-issue, we have a reasonable explanation for why the following pair of sentences clash,

An objection to this view is that code words do not display non-cancelability the way not-at-issue content typically does; “someone cannot disavow commitment to the not-at-issue content of a sentence S that she utters merely by following up her utterance by asserting the negation of that content” (Khoo, 45). Consider,

The juxtaposition of sentences in (3.11) is supposed to strike the reader as contradictory while those in (3.12) should not.

According to a third view, code words are neither ambiguous nor multidimensional, but possess only nonracial meaning. What explains the phenomenon associated with terms like ‘inner city’ is the presence of an antecedent belief in the audience member that then allows them to infer the racial component. For example, an audience member may already believe

Pre-existing Belief (PB): The inner city is mostly populated by poor African Americans,

so that when hearing a politician proclaim (3.5), the audience member comes to infer

Racial Inference (RI): The food stamp program will primarily benefit poor African Americans.

A contrasting view that draws on the same simple semantics is what Khoo calls the association-driven theory of code words. On this view, there is “an association between ‘inner city’ (or the concept INNER CITY) and the concept AFRICAN AMERICAN (or maybe just RACE) which then primes racist beliefs and prejudices” (2017, 50).

Khoo’s account is simple and compelling, but we may still wonder whether it is too liberal. For instance, expressions like ‘thug,’ ‘illegal alien,’ ‘welfare queen,’ and ‘terrorist’ seem to behave like the terms Khoo identifies as code words, yet they are generally understood to be explicitly racial in nature. Patrick O’Donnell (2017) argues that the aforementioned expressions are not code words but racialized terms . O’Donnell characterizes the difference between racialized terms and code words in the following way:

• Racialized terms involve direct or predicative relations between a term and a racialized group whereas code words involve indirect inferential or associational relations, and
• Racialized terms elicit racial resentment by making salient race-specific interpretive options, whereas code words function by making salient race-neutral interpretive options (2017, 28).

O’Donnell agrees with Khoo that code words are picked out according to their contextual cognitive-pragmatic role, while claiming that this role differs between code words and racialized terms.

Determining whether dogwhistles or coded language count as merely contentious claims that must be tolerated or as hate speech subject to regulation has implications for broader discussions. The subtlety of coded language, for instance, calls its status as hate speech into question. The impact coded language has on an audience lacks the kind of immediacy often attributed to hate speech. Lawrence (1993), for example, notes that hate speech is often experienced by targets as a slap in the face. On the other hand, Mendelberg’s account suggests coded speech can incite racial resentment, and so it may be more aptly considered similar to propagandistic hate speech, discussed above. (For more on this point, see Jason Stanley (2015).) This would appear to get us closer to how hate speech is purported to function, namely, by inciting racial hatred. Whether it is close enough is of course open for debate.

That hate speech and pornography are discussed so frequently together in philosophy might, at first glance, seem surprising. But given the overlap made in the arguments made by anti-porn feminist about pornography and anti-racist theorists about racist hate speech, the two are now intimately linked—for better or for worse. (One important fact that led to this development is, of course, the ruling that pornography is protected by the US first amendment as speech [see Miller v. California (1973)].) According to anti-porn feminists, much of what is said of racist hate speech and the harms that befall its targets also applies, with the appropriate changes, to pornography and women—including, it’s worth emphasizing, women of color.

Many of the important initial moves in this literature were crafted by feminist legal scholar Catherine MacKinnon, along with Andrea Dworkin. One of MacKinnon’s most significant claims that has received sustained philosophical attention is the idea that (degrading and misogynist) pornography silences women. With some modifications, a similar claim may be applied to hate speech, namely, that hate speech silences its targets. However, as the literature has focused on the case of pornography and women, it’s worth examining these arguments in detail first.

This silencing argument begins with MacKinnon’s observation that there are “words that set conditions” for other speech acts’ successes or failures (1993, 63–68; see also Hornsby and Langton, 1998, 27). That is, there are some speech acts that fix the possibility of other speech acts. In other words, they make it possible for some persons to perform some speech acts, and make it impossible for others. This is most evident in formal settings, like a legislature, where the formal rules determine who may speak when, and in what manner. Pornography, the argument continues, does just this. It sets rules of behavior that, in effect, inhibit the speech of women. The result of which is that the speech acts of pornography—performed by those who produce and distribute it—create a climate that undermines women’s capacity to perform certain speech acts of their own. The speech of some (pornographers), therefore, curtails the speech of others (women).

In an influential account of the phenomena of silencing, Langton (1993) deploys speech act theory to examine the case of sexual refusal. According to the silencing argument, pornography depicts women as not (genuinely) refusing sexual advances with utterances of ‘no.’ Indeed, according to the myths perpetuated by pornography (among other social influences), a woman’s ‘no’ is not a refusal, but rather part of an elaborate sexual script. As a result, when a woman says ‘no’ in a non-pornographic context, intending to refuse a man’s sexual advances, she may find herself unable to be heard—that is, her words won’t have the force and effect she intends, and her hearer will not take her to be refusing. She may find herself silenced in this particularly horrendous way, unable to use the standard methods of refusing another’s sexual advances. The claim is that this occurs as a result of pornography silencing women’s refusals in the context of sex. It renders their words powerless.

In making this argument, Langton relies on the distinction between locutionary, illocutionary, and perlocutionary acts, and, correspondingly, locutionary, illocutionary, and perlocutionary silencing. A couple examples will explain these distinctions quickly:

When X says, ‘Shoot him!’ they are, we can see quite quickly, both saying something: ‘shoot him!’ and at the same time doing something: ordering the hearer to fire. In Austinian terms, we can say that X performs the locutionary act of making an utterance with a certain meaning, and at the same time is performing an illocutionary act of ordering the hearer to fire. In addition to these two things, the speaker is also, with their words, bringing about a number of effects, which Austin termed the ‘ perlocutionary act.’ In this case, leading to some unfortunate soul to be shot. (Adapted from Langton, 1993, 295, and Austin, 1962, 101)

To be clear, these three acts—locutionary, illocutionary, and perlocutionary—all occur as part of a single utterance and serve to bring out different aspects of any speech act. Austin (and many after him) paid particular attention to the illocutionary act of an utterance, as this, he said, corresponds to the force of an utterance. That is, what someone is doing with their words.

With this in mind, we can see that there are, in fact, many ways one could silence someone. You could literally gag or threaten someone to prevent them from speaking at all, which would achieve a type of locutionary silencing. Alternatively, you could let them say what they wish, recognize what act they are performing, but prevent them achieving their goals, and in doing so achieve a type of perlocutionary silencing. Finally, a third alternative occurs when one speaks and is prevented not only from achieving their intended effects, but also is prevented from performing the very action they intend to perform (Langton, 1993, 315). It is this third alternative—illocutionary silencing—that is said to occur when a man fails to even recognize a woman’s ‘no’ as a refusal, owing at least partially to the influence of pornography.

The specific mechanics of silencing—along with the underlying theory that best explains the phenomena—is subject to much dispute in the literature, and numerous accounts with different essential features have been offered (see Langton (1993); Langton and West (1999); Hornsby (1994); Hornsby and Langton (1998); Maitra (2009); McGowan (2004, 2009, 2014); Mikolla (2011; 2019), among others).

Laura Caponetto (2021) distinguishes four different types of silencing, demonstrating the breadth of the concept. First, there is essential silencing, which consists in the hearer’s failure to recognize the illocutionary point of a speech act. Second, there is authority silencing, where a hearer fails to acknowledge a speaker’s authority in a relevant domain. Third, there is sincerity silencing, when the speaker’s utterance is inaccurately taken as insincere. Fourth and finally, there is seriousness silencing, which consists in the hearer’s failure to acknowledge the speaker’s words as appropriately serious. Given these fine-grained ways of understanding silencing, a broad, comprehensive definition of silencing may be put as follows:

Illocutionary Silencing A speaker S putting forth a speech act A addressed to a hearer H is illocutionarily silenced iff (i) H fails to recognize the obtaining of some conditions for A ’s success; (ii) S ’s attempt at A -ing meets the conditions that H fails to recognize; (iii) normal input and output conditions are met; (iv) the recognition failure on H ’s part is systematic. (Caponetto, 2021)

In nearly all discussions of silencing, one common piece of contention concerns the notion of ‘uptake.’ On different understanding of what uptake consists in—ranging from the hearer’s recognition of a speaker’s intent, or the type of speech act being performed, up to the material consequences of a speech act—we are led to different conclusions about whether a speaker was silenced or not. Disagreement about the conditions of uptake poses difficulty, therefore, for many accounts of silencing. Drawing on these difficulties, Samia Hesni (2018) has argued that the standard account of silencing needs significant retooling, in part because the necessary distinction between illocutionary silencing and perlocutionary silencing cannot hold, as it relies on a problematic—and arguably conceptually untenable—notion of uptake (Hesni, 2018, 957). In an attempt to avoid these difficulties, we might prefer an account of silencing that uses a Gricean, rather than Austinian or Searlian framework, bypassing the need to fully differentiate the illocutionary from the perlocutionary (see Maitra, 2009).

While much of this literature is explicitly focused on pornography’s potential to silence women in the realm of sexual refusal, the notion that racist hate speech may similarly play a silencing function has also been put forward. For example, in a classic paper on the topic, Lawrence wrote that:

Racist speech … distorts the marketplace of ideas by muting or devaluing the speech of Blacks and other despised minorities. Regardless of intrinsic value, their words and ideas become less saleable in the marketplace of ideas. An idea that would be embraced by large numbers of individuals if it were offered by a white individual will be rejected or given less credence if its author belongs to a group demeaned and stigmatized by racist beliefs. (Lawrence, 1993, 78–79)

Using the above framework, we might therefore say that racist hate speech can itself constitute words that set conditions for the success of other speech acts, and in doing so undermines the speech of its targets—and in some cases, effectively silencing them. That is, racist hate speech may, in cultivating an environment hostile to the voices (and lives) of many, can lead to both locutionary and illocutionary silencing in a way that threatens their freedom of expression. And as is noted above in the section on the harms of hate speech, one long-term consequence of racist hate speech may be the target’s partial withdrawal from certain aspects of public life, including public discourse (West, 2012, 237). One further harmful effect of hate speech, then, is its targets’ silence.

Another way in which racist hate speech might silence is more immediate. Returning to the distinction between propagandistic hate speech on the one hand, and assaultive hate speech on the other, where the latter consists in hate speech uttered directly to its target, we may note that hate speech often serves as a type of attack. So, despite the common refrain of ‘more speech’ offered as advice, conceiving hate speech as a personal attack demonstrates how it, in fact, threatens the speech rights of its targets. As Lawrence puts it: “The visceral emotional response to personal attack precludes speech” (1993, 68). He goes on:

Attack produces an instinctive, defensive psychological reaction. Fear, rage, shock, and flight all interfere with any reasoned response. Words like ‘nigger,’ ‘kike,’ and ‘faggot’ produce physical symptoms that temporarily disable the victim, and perpetrators often use these words with the intention of producing this effect. (ibid.)

So, in both cultivating an environment in which the speech of marginalized groups is systematically devalued, or in serving as an immediate threat, hate speech can be said to silence its targets.

As is the case with pornography and silencing, the details of the mechanisms that sustain this type of silencing, along with what particular type of silencing racist hate speech results in, are subject to dispute. But, just like in the pornography debate, the plausibility of the silencing argument lies partly in how it reframes the overall question surrounding regulation. Rather than simply being a source of harm that merely infringes on the equality rights of its targets, if hate speech silences then it also infringes on the speech rights of its targets (West, 2012). As a result, it is not simply a question of balancing the speech rights of hate speakers against the wellbeing of their targets, but of competing claims to (substantive, and not just formal) freedom of expression. And given the importance that most liberal democracies place on freedom of expression, the challenge presented from hate speech is of central importance. For this reason, the silencing question is one of the most disputed aspects of hate speech and has generated great attention.

5. Counteracting Hate Speech

On the presumption that hate speech is harmful—both particularly harmful for the members of targeted groups, and also generally harmful to democracy—the natural question that follows is: what should we do about it? This question, however, rests on several sub-questions—some empirical, some conceptual—that themselves admit of rich dispute. For example, depending on how one conceives of the value and point of free expression—to better seek the truth, to respect autonomy, to ensure democracy, etc.—different answers to the hate speech question will seem more worthwhile than others. The same consideration applies to empirical matters as well, which are often difficult to properly assess in the absence of uncontroversial data. This means that relatively straightforward empirical questions—does genocidal speech pave the way to actual genocidal violence; do governments abuse hate speech regulation to punish political rivals and disfavored minorities; and others—rarely receive unanimous agreement. Despite these challenges, many theorists have addressed the question of how to counteract hate speech, and what form that response ought to take.

We can divide the most common answers into three broad categories: (1) legally restrict it in some form, as a justified exception to free expression; (2) permit it on the basis of free expression, holding that the harms of censorship outweigh the harms of hate speech; or (3) permit it, but take explicit measures to undo the harm of hate speech.

First, the case for banning hate speech. While this position may be anathema to many (especially in the United States), it is the consensus position of most democratic nations around the globe, as well as the explicit position of the United Nations. In the International Covenant on Civil and Political Rights , Article 20 requires a ban on hate speech—or, in their words, “any advocacy of national, racial or religious hatred that constitutes incitement to discrimination, hostility or violence shall be prohibited by law” ( Covenant on Civil and Political Rights ; see also, Article 4 of Convention on Racial Discrimination ). It is worth noting the position of the United Nations and other democracies on hate speech in part because of the contrast they serve for the dominant position in the United States, which recognizes some exceptions to the right to free expression (e.g., obscenity, libel, child sexual abuse material), but not generally on the basis of (racial) hate. Moreover, these exhortations to criminalize hate speech from the United Nations sit alongside commitments that maintain the importance of freedom of expression. For instance, Article 20, quoted above, is immediately preceded by Article 19, which affirms right to freedom of expression ( Covenant on Civil and Political Rights ).

The standard justification offered for restrictions on freedom of expression are based on the necessity of (a) respect of the rights or reputations of others; and (b) reasons of national security or of public order. In other words, a ban on hate speech may be thought to follow from the recognition of the harms it presents, both to the dignity of minority-members of a nation, as well as their physical safety. This position maintains, then, that restrictions on hate speech are a legitimate—and necessary—exception to an otherwise wider understanding of free expression. (For some theorists, it’s worth noting, hate speech is best not understood as the type of speech that free speech protections are meant to include—e.g., it serves no purpose in the pursuit of truth—and so is not in fact an exception to a free speech principle, but simply not included in a proper understanding of the scope of free speech.) This view naturally follows from the understanding that multiple values and rights must be balanced against each other. This is true both of countries that explicitly prohibit hate speech in order to protect minority rights, as well as in more ‘speech-friendly’ nations like the United States, where speech that is aimed at and likely to result in “imminent lawless action” may be legitimately restricted (see Brandenburg v. Ohio ).

However, most advocates for legally restricting hate speech believe that its proper scope is wider than what US law currently allows. Parekh, for example, rejects the position that hate speech may only be restricted when there is “imminent danger” of violence on the grounds that this understanding is too short-sighted. Moreover, he says,

no action occurs in a historical vacuum, and every action produces consequences not inherently but against a particular background. … Imminent danger occurs against, and is imminent because of, the prevailing social climate, and consistency demands that we concentrate our efforts not only on fighting the immediate source of danger, but also on changing the climate. (2012, 45–46)

On the understanding that the threat of hate speech is not exhausted by cases that concern “imminent danger,” we might then ground the prohibition of hate speech on the basis that this may reduce speech that causes harm to its targets, beyond those most immediately affected. Of course, there is also an important role to be played by non-legal means (e.g., moral and social pressure) in erasing or reducing these harms, so legal bans are best understood as part of a broader approach to the ills of hate speech. Furthermore, advocates of bans describe the expressive dimension of these laws as themselves providing a reason in favor of legislation (Waldron, 2014). The law, in this sense, serves as a public statement on a community’s values, and has educational and symbolic importance in itself (Parekh, 2012, 46). (For an overview of expressive theories of law, see Anderson and Pildes, 2000.) A ban on hate speech, therefore, is intended both to reduce harms directly, by decreasing instances of hate by the threat of law, as well as indirectly, by shaping the community’s moral norms through an expression of value.

Though many would agree that hate speech can have destructive effects, and that there is a moral imperative on the state to cultivate something like respectful relations between its members, objections to hate speech bans abound. In a wide-ranging response to these concerns, Parekh (2012) considers (and rejects) six common objections to the prohibition of hate speech. These six objections are: (1) that the harm of hate speech, while real, is relatively minor and a small price to pay given the interest of democratic nations; (2) that bans are not the answer, but rather “better ideas” and “more speech” are; (3) that a prohibition would have a dangerous “chilling effect” and that hate speech bans are a slippery slope to all sorts of unwanted restrictions; (4) that bans give the state too much power to judge the content of speech and decide what can or cannot be said, threatening state-neutrality, skewing political debate, and infringing on individual liberty; (5) that bans are an objectionable form of paternalism or moral authoritarianism, and is incompatible with the assumption that humans are responsible and autonomous individuals and that society is made up of free and equal citizens; and finally, (6) that bans are ineffective at changing attitudes and removing the hate from the hate speaker’s heart, with the result that bans have the effect of moving extremists underground, alienating them from wider society, and in doing so rendering us ignorant of their violent potential and impotent to engage in effective de-radicalization.

Each of these concerns merits more space than can be given to them here. Still, considering these objections to bans and the responses available, even briefly, is illustrative of the theoretical concerns bans on hate speech bring forth (see the list of references for fuller development of the relevant theoretical and empirical issues). Again following Parekh (2012, 47–54), we can approach these objections as follows.

In response to (1), the objection that the interests of a vibrant democracy outweigh the harms imposed by hate speech, it may be argued that hate speech does not embody the values of free speech but, in fact, undermines them by promoting irrational fears and hatred over reasoned arguments and public scrutiny. How powerful one takes this response to be depends directly on what one takes the value and justification of a right to free expression to be, which is of course a matter of dispute.

One response to (2), the common ‘more speech’ objection, is to note that the “marketplace of ideas” is not neutral, and likely requires some regulation (just like a marketplace of other goods). This is what a ban does, and so may be considered to be helping ensure ‘fair competition’ by countering prevailing prejudices, and encouraging greater participation from the members of communities targeted by hate speech. In other words, bans on hate speech may promote greater freedom of expression, by preventing the type of silencing considered above.

While acknowledging the worries of (3), namely that of a ‘chilling effect’ or a ‘slippery slope,’ represent an important objection, we may respond by noting that the problems these signal rest on the vague wording and inconsistent or biased application of hate speech bans. They are not, therefore, direct objections to hate speech bans as such. The remedy, therefore, lies in correcting these aspects of a ban, rather than abandoning it altogether. Moreover, the appeal to a ‘slippery slope’ may be inapt, as it implies that once one type of speech is prohibited, society cannot help but prohibit even more types. But we have no clear reason to suppose that this is the case, as existing bans on defamation have not led to bans on fair critical comment, for example.

The worries at the core of objection (4) represents a well-founded fear of the state, and so must be taken seriously. But, to defenders of hate speech bans, its understanding of the threat that hate speech bans pose to state-neutrality is nonetheless flawed. It fails to recognize that the state often already judges the content of speech (e.g., in banning commercial fraud, criminal solicitation, public displays of obscenity) and often elides neutrality when it speaks in favor of certain positions (e.g., the value of human dignity, equality, liberty). While any defense of hate speech bans must reckon with the possibility of further empowering the state, opponents ought not misrepresent the status quo, exaggerating the reality of state-neutrality.

Objections grounded on the threat of paternalism or moral authoritarianism, like (5), are similarly serious. However, one response on behalf of bans would be to point out how autonomy is always exercised under certain conditions and requires various external circumstances for its development and use. When appealing to personal autonomy, therefore, we should not idealize too greatly so that its real-world exercise is ignored. Rather, the threats that racism and bigotry pose for autonomy must also be acknowledged, alongside praise for our rational faculties.

One response to (6), that bans are ineffective at changing attitudes, is to admit the law cannot change attitudes (like hatred) directly and maintain that this is no knock against the law, and indeed is no problem for hate speech bans. The aim of these bans, in most cases, is not to prevent hatred but to prevent the harm that the public expression of hate can cause. The indirect effects of such a law, however, are an empirical matter, and it is unlikely they admit of a single, general answer, but are highly context-dependent. The subject of the practicability of hate speech bans deserves special attention, however. Opponents to bans may worry that the suppression of hate speech is likely to backfire, not only by failing to reduce hatred, but by increasing the sense of oppression and victimization that many bigots thrive on, leading to an escalation of racist violence (Baker, 2012, 77). Again, as an empirical hypothesis, it cannot be settled simply from the armchair. Still, a further response available on behalf of hate speech bans would be to question the legitimacy of this objection. If, by hypothesis, bans generated an increase in violence, it would still be the responsibility of the state to manage this violence effectively. The role of the state is not exhausted by implementing a ban, but must be seen alongside its enforcement.

This, however, leads to a slightly different objection. The opponent of bans may worry that the enforcement of laws against hate speech would divert the state’s energies away from more effective measures against hated, such as “those directed at changing material conditions in which racism festers, material conditions of both the purveyors and targets of hate” (Baker, 2012, 77). That is, the energies and resources that would be directed towards establishing and enforcing hate speech bans may be better spent on alternative policies. The guiding thought rests on two important points. First, that the intended ends of hate speech bans (e.g., reduction in the harms of hate speech that fall on those targeted by it, mitigation of the expansion of racist attitudes, lessening occurrences of violent hate crimes) may be more effectively achieved via different means, such as reducing inequality, improving social safety nets, political empowerment, and more. Second, though the state can do more than one thing at once, it is nonetheless working with limited resources, and efficiency is a value. That these alternative policy options may indeed be more effective is an unresolved empirical matter. And it remains an open question whether indirect approaches like this would fail to achieve the expressive ends of hate speech bans, which more directly communicate to those targeted by hate speech that they are valued members society.

Many of the claims made above, both on behalf of bans and in opposition, raise theoretical and empirical issues whose proper examination spans many articles and books. Suffice to say that the debate over bans is a highly contested one, and each position rests on an understanding of such issues as the value of free expression, the harm of hate speech, the likely effects a ban might have in a particular context, and so on. For instance, one who believes that free expression is valuable in part because of its role in democratic decision-making may maintain that specifically political speech deserves increased protections, and that some of what others regard as hate speech might fall into this category, escaping regulation. Alternatively, one may view the immunity for political speech as perhaps a red herring. On the speech act theoretic framework outlined above, some forms of racist hate speech are functionally identical to a ‘Whites Only’ sign hanging in a public restaurant (McGowan 2012; McGowan and Maitra 2009). The latter expresses a political opinion in the same way as the former expression does, but it is also regarded as unlawful racially discriminatory. The same considerations—legal sanction—might therefore apply to the verbal utterance as the written sign, and the appeal to the political content of the message is irrelevant.

The preceding summarizes the two main positions in the debate over hate speech: on the one hand, there are those who defend prohibitions, and on the other, those who maintain hate speech as protected under a wide conception of freedom of expression, and so oppose laws that aim at its prohibition. A third position aims to avoid some of the impasses that haunt this debate. On this view, this impasse is the result of a failure by those who oppose hate speech bans (and, as a result, tend to favor ‘more speech’) to acknowledge the strength of one of the main arguments from those who advocate for bans, namely, that hate speech is a type of assault that often renders one unable to respond. This, along with a failure of those who defend bans from considering non-punitive options for mitigating the harms of hate speech, leads to stalemate. On this understanding, both sides of the debate over bans see the only alternatives as either increased governmental powers to punish, or absent that, ‘unsupported’ counterspeech on the part of those targeted by hate speech (see Gelber, 2012a; 2012b).

By contrast, the “supported counterspeech” alternative aims to recognize the specific harms inflicted by hate speech and provide state support to empower those who are harmed. Gelber, an advocate for this alternative, places it within the capabilities approach originally developed by Amartya Sen (1992) and Martha Nussbaum (2000; 2003). “If hate-speech acts harm their targets’ capacity to develop human capabilities,” Gelber says, then “this is what needs to be remedied” (2012a, 54). The impetus for this approach therefore begins from the idea that we must think about remedies to hate speech beyond restrictions and punishment, as neither of these approaches achieve the goal of empowering the target of hate speech. (This is especially true of the latter, punishment, which also carries with it all the negative consequences that anti-carceral advocates have noted.) The supported counterspeech policy is therefore not focused on hate speakers, but rather the targets of hate speech more directly.

The core of this approach lies in an enlarged conception of counterspeech as well as a commitment by the state to provide the material conditions necessary for this speech. In practice, this would mean that the state is committed to responding to an incident of hate speech by empowering its targets to engage in more speech, after the fact. The specific forms this support may take will depend on the conditions of different contexts, along with calibration for the specifics of the incident it is meant as a response to, as well as the needs of the particular communities. Still, to give a sense of what this may entail, examples of the sort of supported counterspeech that this position recommends include things such as: assistance in the production of a community newsletters, op-eds, radio broadcasts, or television advertisements; the development of antiracism awareness programs, or anti-hate-speech workshops; subsidizing community-led art projects; etc.

In each case, the aim of supported counterspeech is to empower the targets of hate speech, and to increase their capacity for engaging in counterspeech. The goal is thus to undo (as much as one can) the specific harms of hate speech, while avoiding the pitfalls of “private remedies” (as critiqued by Matsuda, 1993). While supported counterspeech could be taken as either an alternative to bans or a supplement to them, it remains an under-explored avenue for considering responses tailored to the particular harms of hate speech.

• Anderson, Elizabeth, and Richard Pildes, 2000, “Expressive Theories of Law: A General Restatement,” University of Pennsylvania Law Review , 148: 1503–1575.
• Anderson Luvell and Lepore, Ernest, 2013a, “Slurring Words,” Noûs , 47(1): 25–48.
• –––, 2013b, “What Did You Call Me? Slurs as Prohibited Words,” Analytic Philosophy , 54(3): 350–363.
• Appiah, Kwame Anthony, 2012, “What’s Wrong with Defamation of Religion?” in Michael Herz, and Peter Molnar (eds.), The Content and Context of Hate Speech: Rethinking Regulation and Response , Cambridge; New York: Cambridge University Press, pp. 164–182.
• Ashwell, Lauren, 2016, “Gendered Slurs,” Social Theory and Practice , 42(2): 228–39.
• Austin. J. L., 1962, How to do Things with Words , Cambridge, MA: Harvard University Press.
• Bach, Kent, 2018, “Loaded Words: On the Semantics and Pragmatics of Slurs,” in David Sosa (ed.), Bad Words: Philosophical Perspectives on Slurs , Oxford: Oxford University Press, pp. 60–76.
• Baker, C. Edwin, 2012, “Hate Speech,” in Michael Herz, and Peter Molnar (eds.), The Content and Context of Hate Speech: Rethinking Regulation and Response , Cambridge, New York: Cambridge University Press, pp. 57–80.
• Bauer, Nancy, 2015, How to Do Things with Pornography , Cambridge, MA: Harvard University Press.
• Beauharnais v. People of the State of Illinois , 343 U.S. 250 (1952).
• Behrens,Paul, Terry, N. and Jensen, O (eds.), 2017, Holocaust and Genocide Denial: A Contextual Perspective , Abingdon; New York: Routledge, Taylor and Francis Group.
• Bolinger, Renée Jorgensen, 2017, “The Pragmatics of Slurs,” Noûs , 51(3): 439–462.
• Brandenburg v. Ohio , 395 U.S. 444 (1969).
• Brandom, Robert, 1994, Making it Explicit , Cambridge, MA: Harvard University Press.
• Brown, Alexander, 2017a, “What Is Hate Speech? Part 1: The Myth of Hate,” Law and Philosophy , 36(4): 419–468.
• –––, 2017b, “What Is Hate Speech? Part 2: Family Resemblances,” Law and Philosophy , 36(5): 561–613.
• Brison, Susan, 1998a, “The Autonomy Defense of Free Speech,” Ethics , 108(2): 312–339.
• –––, 1998b, “Speech, Harm, and the Mind-Body Problem in First Amendment Jurisprudence,” Legal Theory , 4: 39–61.
• Butler, Judith, 1997, Excitable Speech: A Politics of the Performative , New York: Routledge.
• Camp, Elisabeth, 2013, “Slurring Perspectives,” Analytic Philosophy , 54(3): 330–349.
• –––, 2018, “Slurs as Dual-Act Expressions,” in David Sosa (ed.), Bad Words , Oxford: Oxford University Press, pp. 29–59.
• Caponetto, Laura, 2021, “A Comprehensive Definition of Illocutionary Silencing,” Topoi , 40: 191–202.
• Cepollaro, Bianca, 2015, “In Defense of a Presuppositional Account of Slurs,” Language Sciences , 52: 36–45.
• Croom, Adam, 2011, “Slurs,” Language Sciences , 33: 343–358.
• Delgado, Richard, 1993, “Words that Wound: A Tort Action for Racial Insults, Epithets, and Name Calling,” in Mari Matsuda, Charles Lawrence III, Richard Delgado, and Kimberlé Crenshaw (eds.), Words that Wound: Critical Race Theory, Assaultive Speech and the First Amendment , Boulder, CO: Westview Press, pp. 89–110.
• Delgado, Richard, and Jean Stefancic, 2004, Understanding Words that Wound , Boulder, CO: Westview Press.
• Dummett, Michael, 1993, Frege: Philosophy of Language , Cambridge, MA: Harvard University Press.
• Gallie, W. B., 1955, “Essentially Contested Concepts,” Proceedings of the Aristotelian Society , 56: 167–198.
• Gelber, Katharine, 2012a, “Reconceptualizing Counterspeech in Hate-Speech Policy (with a Focus on Australia),” in Michael Herz, and Peter Molnar (eds.), The Content and Context of Hate Speech: Rethinking Regulation and Response , Cambridge, New York: Cambridge University Press, pp. 198–216.
• –––, 2012b, “‘Speaking Back’: The Likely Fate of Hate Speech Policy in the United States and Australia,” in Ishani Maitra and Mary Kate McGowan (eds.), Speech & Harm: Controversies Over Free Speech , New York: Oxford University Press, pp. 50–71.
• –––, 2017, “Hate Speech—Definitions & Empirical Evidence,” Constitutional Commentary , 32(3): 619–629.
• Gelber, Katharine, and Luke McNamara, 2016, “Evidencing the Harms of Hate Speech,” Social Identities , 22 (3): 324–341.
• Herz, Michael, and Peter Molnar (eds.), 2012, The Content and Context of Hate Speech: Rethinking Regulation and Responses , Cambridge, New York: Cambridge University Press.
• Hesni, Samia, 2018, “Illocutionary Frustration,” Mind , 127(508): 947–976.
• Heyman, Steven, 2008, Free Speech and Human Dignity , New Haven: Yale University Press.
• Hom, Christopher, 2008, “The Semantics of Racial Epithets,” Journal of Philosophy , 105: 416–440.
• Hornsby, Jennifer, 1994, “Illocution and its Significance,” in Savas L. Tsohatzidis (ed.), Foundations of Speech Act Theory: Philosophical and Linguistic Perspectives , London: Routledge, pp. 187–207.
• –––, 2001, “Meaning and Uselessness: How to Think About Derogatory Words,” Midwest Studies in Philosophy , 25(1): 128 – 141.
• Hornsby, Jennifer, and Rae Langton, 1998, “Free Speech and Illocution,” Legal Theory , 4: 21–37.
• Imbleau, Martin, 2011, “Denial of the Holocaust, Genocide, and Crimes Against Humanity,” in Ludovic Hennebel and Thomas Hochmann (eds.) Genocide Denials and the Law , New York: Oxford University Press, pp. 235–277.
• Jeshion, Robin, 2013a, “Expressivism and the Offensiveness of Slurs,” Philosophical Perspectives , 27: 231–259.
• –––, 2013b, “Slurs and Stereotypes,” Analytic Philosophy , 54 (3): 314–329.
• Khoo, Justin, 2017, “Code Words in Political Discourse,” Philosophical Topics , 45 (2): 33–64.
• Kirk-Giannini, Cameron Domenico, 2019, “Slurs are Directives,” Philosopher’s Imprint , 19(48): 1–28.
• Kukla, Quill [writing as Rebecca], 2014, “Performative Force, Convention, and Discursive Injustice,” Hypatia , 29 (2): 440–457.
• –––, 2018, “Slurs, Interpellation, and Ideology,” The Southern Journal of Philosophy , 56(S1): 7–32.
• Langton, Rae, 1993, “Speech Acts and Unspeakable Acts,” Philosophy & Public Affairs , 22(4): 293–330.
• –––, 2012, “Beyond Belief: Pragmatics in Hate Speech and Pornography,” in Ishani Maitra and Mary Kate McGowan (eds.), Speech & Harm: Controversies Over Free Speech , New York: Oxford University Press, pp. 72–93.
• –––, 2018a, “The Authority of Hate Speech,” in John Gardner, Leslie Green, and Brian Leiter (eds.), Oxford Studies in Philosophy of Law (Volume 3), New York: Oxford University Press, pp. 123–152.
• –––, 2018b, “Blocking as Counter Speech,” in Daniel W. Harris Daniel Fogal, and Matt Moss (eds.), New Works on Speech Acts , Oxford: Oxford University Press, pp. 144–164.
• Langton Rae, and Caroline West, 1999, “Scorekeeping in a Pornographic Language Game,” Australasian Journal of Philosophy , 77(3): 303–319.
• Lawrence, Charles III, 1993, “If He Hollers Let Him Go: Regulating Racist Speech on Campus,” in Mari Matsuda, Charles Lawrence III, Richard Delgado, and Kimberlé Crenshaw (eds.), Words that Wound: Critical Race Theory, Assaultive Speech and the First Amendment , Boulder, CO: Westview Press, pp. 53–88.
• Haney Lopez, Ian, 2014, Dog Whistle Politics: How Coded Racial Appeals Have Reinvented Racism and Wrecked the Middle Class , Oxford: Oxford University Press.
• MacKinnon, Catharine, 1993, Only Words , Cambridge, MA: Harvard University Press.
• Maitra, Ishani, 2009, “Silencing Speech,” Canadian Journal of Philosophy , 39(2): 309–338.
• –––, 2012, “Subordinating Speech,” in Ishani Maitra and Mary Kate McGowan (eds.), Speech & Harm: Controversies Over Free Speech , New York: Oxford University Press, pp. 94–120.
• Maitra, Ishani, and Mary Kate McGowan (eds.), 2012, Speech & Harm: Controversies Over Free Speech , New York: Oxford University Press.
• Matsuda, Mari, 1993, “Public Response to Racist Speech: Considering the Victim’s Story,” in Mari Matsuda, Charles Lawrence III, Richard Delgado, and Kimberlé Crenshaw (eds.), Words that Wound: Critical Race Theory, Assaultive Speech and the First Amendment , Boulder, CO: Westview Press, pp. 17–51.
• Matsuda, Mari, Charles Lawrence III, Richard Delgado, and Kimberlé Crenshaw (eds.), 1993, Words that Wound: Critical Race Theory, Assaultive Speech and the First Amendment , Boulder, CO: Westview Press.
• McGowan, Mary Kate, 2004, “Conversational Exercitives: Something Else We Do With Our Words,” Linguistics and Philosophy , 27: 93–111.
• –––, 2009, “Oppressive Speech,” Australasian Journal of Philosophy , 87(3): 389–407.
• –––, 2012, “On ‘Whites Only’ Signs and Racist Hate Speech: Verbal Acts of Racial Discrimination,” in Ishani Maitra and Mary Kate McGowan (eds.), Speech & Harm: Controversies Over Free Speech , New York: Oxford University Press, pp. 121–47.
• –––, 2014, “Sincerity Silencing,” Hypatia , 29(2): 458–473.
• –––, 2018, “On Covert Exercitives: Speech and the Social World,” in Daniel W. Harris Daniel Fogal, and Matt Moss (eds.), New Works on Speech Acts , Oxford: Oxford University Press, pp. 185–201.
• –––, 2019, Just Words: On Speech and Hidden Harm , Oxford: Oxford University Press.
• McGowan, Mary Kate, and Ishani Maitra, 2009, “On Racist Hate Speech and the Scope of a Free Speech Principle,” Canadian Journal of Law and Jurisprudence , 23(2): 343–372.
• Mendelberg, Tali, 2001, The Race Card: Campaign Strategy, Implicit Messages, and the Norm of Equality , Princeton: Princeton University Press.
• Mikkola, Mari, 2011, “Illocution, Silencing and the Act of Refusal,” Pacific Philosophy Quarterly , 92: 415–437.
• –––, 2019, Pornography: a Philosophical Introduction , New York: Oxford University Press.
• Miller v. California , 413 U.S. 15 (1973).
• Neufeld, Eleonore, 2019, “An Essentialist Theory of the Meaning of Slurs,” Philosopher’s Imprint , 19(35): 1–29.
• Nunberg, Geoff, 2018, “The Social Life of Slurs,” in Daniel W. Harris, Daniel Fogal, and Matt Moss (eds.), New Works on Speech Acts , Oxford: Oxford University Press, pp. 237–295.
• Nussbaum, Martha, 2000, Women and Human Development: The Capabilities Approach , Cambridge: Cambridge University Press.
• –––, 2003, “Capabilities as Fundamental Entitlements: Sen and Social Justice,” Feminist Economics , 9(2/3): 33–59.
• O’Donnell, Patrick, 2017, “Generics, Race, and Social Perspectives,” Inquiry: An Interdisciplinary Journal of Philosophy , published online 26 January 2017. doi:10.1080/00207543.2016.1266801
• Parekh, Bhikhu, 2012, “Is There a Case for Banning Hate Speech?” in Michael Herz, and Peter Molnar (eds.), The Content and Context of Hate Speech: Rethinking Regulation and Response , Cambridge, New York: Cambridge University Press, pp. 37–56.
• Popa-Wyatt, Mihaela, and Jeremy L. Wyatt, 2017, “Slurs, Roles and Power,” Philosophical Studies , 175(11): 2879–2906.
• Post, Robert, 2009, “Hate Speech,” in Ivan Hare and James Weinstein (eds.), Extreme Speech and Democracy , Oxford: Oxford University Press.
• Reichman, Ammon, 2009, “Criminalizing Religiously Offensive Satire: Free Speech, Human Dignity, and Comparative Law,” in Ivan Hare and James Weinstein (eds.), Extreme Speech and Democracy , Oxford: Oxford University Press, pp. 331–355.
• Richard, Mark, 2010, When Truth Gives Out , Oxford: Oxford University Press.
• Saul, Jennifer M., 2018, “Dogwhistles, Political Manipulation, and Philosophy of Language,” in Daniel W. Harris Daniel Fogal, and Matt Moss (eds.), New Works on Speech Acts , Oxford: Oxford University Press, pp. 360–383.
• Sbisà, Marina, 2009, “Illocution and Silencing,” in Bruce Fraser, and Ken Turner (eds.), Language in Life, and a Life in Language: Jacob Mey—a Festschrift , Bradford: Emerald, pp. 351–357.
• Schwartzman, Lisa H., 2002, “Hate Speech, Illocution, and Social Context: A Critique of Judith Butler,” Journal of Social Philosophy , 33(3): 421–441.
• Sen, Amartya, 1992, Inequality Reexamined , Cambridge, MA: Harvard University Press.
• Stanley, Jason, 2015, How Propaganda Works , Princeton: Princeton University Press.
• Steinberger, Florian, and Julien Murzi, 2017, “Inferentialism,” in Bob Hale, Crispin Wright, Alexander Miller (eds.), A Companion to the Philosophy of Language , second edition, Oxford: Wiley Blackwell, pp. 197–224.
• Tesler, M., and D. O. Sears, 2010, Obama’s Race: The 2008 Election and the Dream of a Post-Racial America , Chicago: University of Chicago Press.
• Tirrell, Lynne, 1999, “Derogatory Terms: Racism, Sexism, and the Inferential Role Theory of Meaning,” in Christina Hendricks, and Kelly Oliver (eds.), Language and Liberation: Feminism, Philosophy, and Language , Albany: State University of New York Press, pp. 41–79.
• Tonhauser, Judith, 2012, “Diagnosing (not-)at-issue content,” Proceedings of Semantics of Underrepresented Languages in the Americas , 6: 239–254.
• UN General Assembly, International Covenant on Civil and Political Rights , 16 December 1966.
• UN General Assembly, International Convention on the Elimination of All Forms of Racial Discrimination , 21 December 1965, [ available online ].
• Waldron, Jeremy,2014, The Harm in Hate Speech , Cambridge, MA: Harvard University Press.
• West, Caroline, 2012, “Words that Silence? Freedom of Expression and Racist Hate Speech,” in Ishani Maitra, and Mary Kate McGowan (eds.), Speech and Harm: Controversies over Free Speech , Oxford: Oxford University Press, pp. 222–248.
• Whine, Michael, 2009, “Expanding Holocaust Denial and Legislation Against It,” in Ivan Hare and James Weinstein (eds.), Extreme Speech and Democracy , Oxford: Oxford University Press, pp. 538–556.
• White, Ismail, 2007, “When Race Matters and When It Doesn’t: Racial Group Differences in Response to Racial Cues,” American Political Science Review , 101(2): 339–54.
• Whiting, Daniel, 2008, “Conservatives and Racists: Inferential Role Semantics and Pejoratives,” Philosophia , 36(3): 375–388.
• Williamson, Timothy, 2009, “Reference, Inference, and the Semantics of Pejoratives,” in J. Almog, and P. Leonardi (eds.), The Philosophy of David Kaplan , Oxford: Oxford University Press, pp. 137–158.
• Wistrich, Robert, 2012, Holocaust Denial: The Politics of Perfidy , Berlin: De Gruyter.

How to cite this entry . Preview the PDF version of this entry at the Friends of the SEP Society . Look up topics and thinkers related to this entry at the Internet Philosophy Ontology Project (InPhO). Enhanced bibliography for this entry at PhilPapers , with links to its database.

• Swanson, Eric, 2015, “ Slurs and Ideology ,” unpublished manuscript
• American Library Association, “ Hate Speech and Hate Crime ”
• European Commission against Hate Speech and Violence
• UN Strategy and Plan of Action on Hate Speech
• UNESCO, “ Countering Online Hate Speech ”
• Rights for Peace, “ Hate Speech ”
• Anti-Defamation Leage, “ Resources for Responding to Hate Speech in the Community ”
• Library of Parliament Research Publication, “ Hate Speech and Freedom of Expression: Legal Boundaries in Canada ”

cognitivism vs. non-cognitivism, moral | freedom: of speech | pornography: and censorship | respect | semantics: proof-theoretic | speech acts

Copyright © 2022 by Luvell Anderson < lander04 @ syr . edu > Michael Barnes < Michael . barnes @ anu . edu . au >

• Accessibility

Support SEP

Mirror sites.

View this site from another server:

• Info about mirror sites

The Stanford Encyclopedia of Philosophy is copyright © 2023 by The Metaphysics Research Lab , Department of Philosophy, Stanford University

Library of Congress Catalog Data: ISSN 1095-5054

Defining Hate Speech

Andy Sellars

There is no shortage of opinions about what should be done about hate speech, but if there is one point of agreement, it is that the topic is ripe for rigorous study. But just what is hate speech, and how will we know it when we see it online? For all of the extensive literature about the causes, harms, and responses to hate speech, few scholars have endeavored to systematically define the term. Where other areas of content analysis have developed rich methodologies to account for influences like context or bias, the present scholarship around hate speech rarely extends beyond identification of particular words or phrases that are likely to cause harm targeted toward immutable characteristics. 

This essay seeks to review some of the various attempts to define hate speech, and pull from them a series of traits that can be used to frame hate speech with a higher degree of confidence. In so doing, it explores the tensions between hate speech and principles of freedom of expression, both in the abstract and as they are captured in existing definitions. It also analyzes historical attempts to define the term in the United States, from the brief period of time when the United States punished hate speech directly. From this analysis, eight traits are surfaced that can be used for the development of a confidence scoring system to help ascertain whether a particular expression should be considered one of hate speech or not.

You might also like

• community GOP reacts to Trump search with threats and comparisons to ‘Gestapo’
• community Why Elon Musk’s Twitter might be (more) lethal
• community Combating Hate Speech Through Counterspeech

Projects & Tools 01

Harmful speech online.

The Berkman Klein Center for Internet & Society is in the third year of a research, policy analysis, and network building effort devoted to the study of harmful speech, in close…

Hate Speech: Its Protection Under the First Amendment and Resisting It With Counterspeech

Introduction.

The First Amendment to the U.S. Constitution states, “Congress shall make no law respecting an establishment of religion, or prohibiting the free exercise thereof; or abridging the freedom of speech, or of the press; or the right of the people peaceably to assemble, and to petition the Government for a redress of grievances.” 2 Each aspect of the First Amendment has its exceptions: people may assemble, but not violently riot; people may practice religion freely, as long as the practice excludes harming others; the press may report, save for printing libel and slander. Regarding the freedom of speech, the Constitution allows for anyone to say anything, unless the words have the potential to incite violence or if the speech includes threatening language. The bounds of free speech, save for these exceptions, are limitless such that extremely offensive language—also known as hate speech—runs rampant, especially in today’s political and social climate. In these cases, using the constitutional freedom of speech to one’s advantage when attempting to put down or prevent hate speech is the most valid option. Attempts to change the law regarding punishment of offensive speech is not feasible, according to previous court cases and current professional opinion.

Case Law Protecting Hate Speech

The law does not look kindly upon cases in which a party attempts to legally limit hate speech. In major hate speech cases, courts have tended to favor the expansion of free speech, not the limitation of it. For example, in the context of Skokie v. National Socialist Party (1977), the National Socalist Party of America (NSPA) made hateful and racist comments about Jewish people and people of color. 3 At the time, the village of Skokie, Illinois had a population of about 5,000 to 7,000 Nazi concentration camp survivors. 4 The Illinois Supreme Court ruled in favor of the NSPA, which sought to hold a public demonstration, and established that “public expression of ideas may not be prohibited merely because the ideas are themselves offensive to some of their hearers.” 5 To reach their decision, the Illinois Supreme Court relied partially on Cohen v. California (1971), a U.S. Supreme Court (“the Court”) case that similarly dealt with free/hate speech that included controversial language and symbolism. In this case, Paul Cohen wore a jacket opposing the Vietnam War, which read “Fuck the Draft. Stop the War.” 6 In Cohen , the Court ruled that government restriction of speech should only occur in the most extreme cases, due to the “premise of individual dignity and choice upon which our political system rests.” 7 Though the language and symbols that the NSPA used were extremely offensive and hateful toward marginalized and persecuted groups of people, the First Amendment right to free speech triumphed in both cases. This Note addresses how the United States can eradicate hateful and offensive expression for the sake of unity without limiting the individual freedoms that the Constitution promises.

Case Law Restricting Hate Speech

In only a select few cases, the Supreme Court has ruled in favor of restricting hate speech. For example, in Virginia v. Black (2003), three defendants had placed burning crosses in the yards of Black families and individuals. 8 They were tried and convicted on this account under a Virginia statute that prohibited cross burning. 9 According to the case law, the Court stated that the First Amendment “affords protection to symbolic or expressive conduct as well as to actual speech.” 10 In this case, the protection would be surrounding the symbolism of a burning cross. However, the Court also considered the group that was being targeted, along with the intention behind the symbol in this case. The burning cross in Virginia was an intimidation, according to Justice Sandra Day O’Connor, and this intimidation fell under the umbrella of a threat “where a speaker directs a threat to a person or group of persons with the intent of placing the victim in fear of bodily harm or death.” 11 This ruling begs the question of how to determine whether a form of speech incites fear within a person or group of people, and whether that incitement constitutes a threat (and is therefore no longer legally protected).

In evaluating Skokie , the Swastika symbol that the NSPA used is similar to the symbolism of a burning cross. Both symbols accentuate hatred and potential threats towards a marginalized group. Burning crosses as a symbol of hatred against people of color originated in 1915 after the debut of the movie The Birth of a Nation , sparking the revival of the Klu Klux Klan. 12 Similarly, the use of the Nazi-appropriated Swastika represents “Aryan identity and German Nationalist pride,” which advances the idea of those of non-Aryan or German descent as inferior and racially impure, leading to physical discrimination against them. 13 In an appellate court review, the NSPA was forbidden to fly the Swastika during their marches. 14 Yet, the use of discriminatory and hateful language was able to persist based on later court rulings, and the organization retained permission to march against Jewish people and people of color. 15 This retainment calls into question the disparate rulings. The courts failed to recognize the incitement of fear in one instance of an organization that runs on racial, ethnic, and religious discrimination and hate speech, and yet, the courts were able to recognize the incitement of fear within another group based on very similar circumstances.

Two Ways to Inhibit Hate Speech

Limiting speech in general, regardless of the personal feelings of the judges rendering the decisions, is an arduous task—though possible. Scholars disagree on what would eradicate it more effectively: providing counterspeech to hate speech or attempting to enact laws to limit it. Jeremy Waldron, a professor at New York University School of Law, and Nadine Strossen, former president of the American Civil Liberties Union and professor at New York Law School, engaged in a public discourse about hate speech and presented differing views on how to fix the problems that hate speech breeds. 16 Before presenting their arguments, both agreed that “a lot of hate speech does satisfy the emergency principle: it constitutes a general threat or targeted harassment or hostile environment harassment or an intentional incitement of imminent violence.” 17 The best way to go about eradicating it, however, was a topic of disagreement: Strossen sides with the counterspeech method, and Waldron favors a legal solution. The biggest takeaway of Strossen’s point of view is the avoidance of censorship laws while still resisting hate speech to the highest extent; Waldron thinks that having legislators come up with ways to legally restrict hate speech is not unattainable, but “reasonable.” 18 While both speakers make valid points, the idea of counterspeech, logically speaking, seems to be the most reasonable if the courts are to adhere to the text of the First Amendment. Waldron’s idea would be the ideal solution, but it is less attainable and has the potential of crossing some constitutional boundaries, while Strossen’s idea stays away from legal impropriety.

The Benefits of Counterspeech

The Supreme Court holds the viewpoint that freedom of speech should be completely protected unless it breaks laws already in place, but it also pushes the idea that “the constitutionally permissible response to speech conveying controversial, disfavored views is ‘counterspeech,’ not censorship—more speech, not silence.” 19 Examples of counterspeech, or speech that directly shuts down discriminatorily-motivated hate speech, would include instances like the dialogue and symbols of the Civil Rights Movement, countless parades and marches for the LGBTQ+ community, and more. According to the American Bar Association, recently, there has been “remarkable and bipartisan outpouring of speech and peaceful dem­onstrations that have denounced hateful ideologies while celebrating our nation’s renewed commitments to equality, inclusivity, and intergroup harmony.” 20 Though instances of counterspeech do not cater to individual hateful biases, it proves that changing the rhetoric and pushing inclusivity by using the freedom of speech for good successfully reduces public forms of hate speech. For example, the outpour of support for the people being discriminated against during the neo-Nazi incident at Charlottesville greatly outnumbered the number of white supremacists supporting the oppressive movement. 21 Not only that, but chief executive officers of many companies offered their own counterspeech arguments in support of marginalized people. Jamie Dimon of JP Morgan Chase stated, “Constructive economic and regulatory policies…will not matter if we do not address the divisions in our country.” 22 He offered support of both unification and of using the freedom of speech to advocate for the oppressed, rather than attempting to make laws to limit speech in general.

The examples of court rulings failing to suppress hate speech and the fact that social movements do more for shutting down discriminatory organizations and parties prove that counterspeech is a favorable alternative to shutting down hateful discourse. Making laws that contravene the First Amendment just to theoretically rid the United States of hate speech does not necessarily help to eradicate wrongful prejudice itself. However, through counterspeech and the social movements that typically go along with it, hateful discourse is reduced to an extent, and provides an educational experience on putting aside bias, stereotypes, and prejudice against marginalized communities at the same time.

• *B.A. Candidate for Philosophy and History, Fordham College at Rose Hill, Class of 2025. Thank you to the Fordham Undergraduate Law Review team, my friends, and my family for supporting me in writing this note.
• U.S. Const. amend. I.
• Skokie v. National Socialist Party, 69 Ill. 2d 605, 610 (Ill. 1978); see Chris Demaske, Village of Skokie v. National Socialist Party of America (Ill) (1978) , The First Amendment Encyclopedia, https://www.mtsu.edu/first-amendment/article/728/village-of-skokie-v-national-socialist-party-of-america-ill.
• Skokie , 69 Ill. 2d at 610.
• Id . The U.S. Supreme Court heard National Socialist Party of America v. Skokie on appeal in 1977 and reversed and remanded the Illinois Supreme Court’s denial to lift the lower court’s injunction. The Court wrote that the state must allow the stay because the injunction will deprive the petitioners of their First Amendment rights “during the period of appellate review which, in the normal course, may take a year or more to complete.” National Socialist Party v. Skokie, 432 U.S. 43 (1977). However, this case still shows a reverence of the protection of free speech over its limitation since it represents a move in the Supreme Court to limit the application of restrictions on free speech. See Nadine Strossen, Hate: Why We Should Resist It with Free Speech, Not Censorship (2018).
• Cohen v. Cal., 403 U.S. 15, 16 (1971).
• Id . at 25.
• Virginia v. Black, 538 U.S. 343, 348 (2003).
• See id . at 344.
• Id . at 358.
• Id . at 360.
• The Various Shady Lives of the Ku Klux Klan , Time (Apr. 9, 1965), http://content.time.com/time/subscriber/article/0,33009,898581,00.html.
• United States Holocaust Memorial Museum, Introduction to the Holocaust , Holocaust Encyclopedia (Nov. 5, 2021), https://encyclopedia.ushmm.org/content/en/article/introduction-to-the-holocaust.
• Skokie v. National Socialist Party, 51 Ill. App. 3d 279, 293 (Ill. App. Ct. 1977). The court said that “[t]he swastika is a personal affront to every member of the Jewish faith, in remembering the nearly consummated genocide of their people committed with memory by those who used the swastika as their symbol.” This appellate court decision was affirmed in part and reversed in part by the Illinois Supreme Court in Skokie v. National Socialist Party .
• See High Court Judge Upholds Ban on Swastikas in Skokie March , The Chicago Tribune (Aug. 27, 1977) (The article refers to the appellate case occurring after the original court case, in which the NSA was allowed to march but not hold offensive and hateful symbols).
• See Susan Kelley, Legal Experts Discuss Hate Speech and How to Limit It , Cornell Chronicle (Apr. 11, 2018), https://news.cornell.edu/stories/2018/04/legal-experts-discuss-hate-speech-and-how-limit-it.
• Id (Waldron finds that presenting hate speech instances in front of legislators and persuading them to make laws protecting against it is feasible).
• See Nadine Strossen, Counterspeech in Response to Changing Notions of Free Speech , American Bar Association (2018), https://www.americanbar.org/groups/crsj/publications/human_rights_magazine_home/the-ongoing-challenge-to-define-free-speech/counterspeech-in-response-to-free-speech/.

• Search Menu
• Advance articles
• Author Guidelines
• Book Reviews
• Open Access
• Submission System
• Why Submit?
• About Journal of International Criminal Justice
• Editorial Board
• Advertising and Corporate Services
• Journals Career Network
• Self-Archiving Policy
• Dispatch Dates
• Journals on Oxford Academic
• Books on Oxford Academic

Article Contents

1. introduction, 2. the most egregious form of speech — direct and public incitement to commit genocide, 3. the early jurisprudence on hate speech, 4. the case against vojislav šešelj — an overview, 5. hate speech as persecution — the hrtkovci speech, 6. instigating crimes through hate speech, 7. conclusion.

• < Previous

Defining Hate Speech: A Seemingly Elusive Task

• Article contents
• Figures & tables
• Supplementary Data

Audrey Fino, Defining Hate Speech: A Seemingly Elusive Task, Journal of International Criminal Justice , Volume 18, Issue 1, March 2020, Pages 31–57, https://doi.org/10.1093/jicj/mqaa023

• Permissions Icon Permissions

This article looks at the international criminal law on hate speech that falls short of direct and public incitement to commit genocide. Using the most egregious form of hate speech that has been prosecuted as an international crime — that of direct and public incitement to genocide — as a baseline, the author analyses the legal parameters of hate speech as persecution (a crime against humanity) and hate speech as instigation (a mode of liability). In so doing, the author critically reviews the International Residual Mechanism for the International Criminal Tribunals’ (IRMCT) appeal judgment in the Šešelj case ( Šešelj Appeal Judgment) in the light of prior case law of the International Military Tribunal of Nuremberg (IMT) and the International Criminal Tribunals for Rwanda and the former Yugoslavia (ICTR and ICTY respectively). The author submits that a plain reading of the Šešelj Appeal Judgment supports the view that it is only the more extreme form of incitement to violence, incitement to commit crimes, followed by actual violent acts, that may constitute hate speech amounting to the crime of persecution: incitement to discrimination or incitement to hatred as such do not qualify. Whether ‘incitement to violence’ absent the commission of crimes could qualify as persecution (a crime against humanity) remains an unsettled point. With regard to hate speech as instigation, the Šešelj Appeal Judgment’s restatement and application of the law causes less controversy: the substantial causal connection required for instigation was found to be direct in the circumstances of that case — even though directness is not a legal requirement for instigation. The author concludes that both these interpretations of hate speech are consistent with the earlier ad hoc tribunals’ jurisprudence and, more generally, with international human rights law which, with some controversial exceptions, allows criminalization only of the most extreme forms of incitement to violence.

With the rise in incidents of hate speech and hate crime around the world, exacerbated by the rise of nationalist political discourse against ‘others’, belonging to, for instance, a different colour, religion, nationality or ethnic origin or descent, there is a corresponding interest in what hate speech actually is. 1 Hate speech and hate crime are indicative of escalating internal strife in society and can possibly constitute an early warning sign of mass violations of human rights, crimes against humanity or even genocide. 2 Criminalizing hate speech is one tool in an array of measures that States can take to deter crime in all situations, including during armed conflict.

However, there is a paucity of jurisprudence on hate speech before international criminal tribunals. Most of the case law has focused on direct and public incitement to genocide, a crime first clearly spelled out in the 1948 Convention on the Prevention and Punishment of the Crime of Genocide (Genocide Convention), and replicated into the statutes of the International Criminal Tribunal for Rwanda (ICTR), the International Criminal Tribunal for the former Yugoslavia (ICTY) and the International Criminal Court (ICC). The grey area, and thus admittedly more interesting legal issues, however, arise when hate speech is not so clear-cut, especially because it is neither direct nor public. More recently, the International Residual Mechanism for the International Criminal Tribunals (IRMCT), the successor of both the ICTR and ICTY, grabbing the media’s and scholars’ attention, rendered the Šešelj appeal judgment ( Šešelj Appeal Judgment) on such hate speech. 3

This article attempts to address the question of what category of hate speech, short of direct and public incitement to genocide, is criminal. Neither the Šešelj Trial Judgment 4 nor the Šešelj Appeal Judgment set out a clear definition of hate speech. The only positive definition of hate speech thus remains the one given by the Nahimana et al. Appeal Judgment, which distinguished between direct and public incitement to commit genocide on the one hand, and ‘hate speech in general (or inciting discrimination or violence)’ on the other. 5 This failure by international tribunals to clearly address some of the basic questions that practitioners grapple with when prosecuting or defending hate speech cases, has given rise to various interpretations of what types of hate speech may be considered criminalized under international law. Some scholars argue that in certain circumstances, hate speech is criminal even if the words used do not call for violence. 6 Others, on the other hand, acknowledging that hate speech falling short of calls to violence is not criminalized, opine that States should adopt a unified liability treaty on ‘atrocity speech offences’. 7 In their view, such a treaty would codify the new crimes of both ‘incitement to commit war crimes’ and ‘incitement to crimes against humanity’, which would include criminalizing hate speech which does not amount to calls for violence but is nonetheless part of a widespread or systematic attack against a civilian population. 8

In this context of legal and judicial uncertainty, this article proposes to fill the gap and thus addresses the — seemingly basic — questions that have yet to be answered. First: is hate speech amounting to calls to hatred or discrimination, but short of calls to violence, criminal — or does it have to call to violence to constitute a crime? What is a ‘call to violence’? Does it include calls to commit crimes or is a call to commit a crime the same as a call to violence? In other words, has the jurisprudence used the terms ‘call to violence’ interchangeably with ‘call to commit crimes’? If a call to violence is a call to commit a crime, does the latter have to be a certain category of crime, such as crimes against the person? 9 Secondly: does violence actually have to ensue for hate speech to be criminalized? Put differently, is there a causal nexus required between the hate speech and subsequent acts of violence, hatred, or discrimination? Can hate speech amount to an inchoate crime, whereby prosecutors would have no need to show causation between the hate speech and any subsequent acts of violence, hatred, or discrimination?

To assist in this inquiry, the article will first briefly review the crime of direct and public incitement to genocide, thus considering what hate speech is not . This will serve as a baseline against which to analyse hate speech. By identifying the legal test required for a conviction for the most egregious cases of direct and public incitement to genocide, the author will be able to effectively discuss the relatively less obvious cases of hate speech that are also prohibited under international criminal law. The article does so through a practical assessment of the International Military Tribunal of Nuremberg (IMT) and ICTR jurisprudence in the light of the more recent Šešelj Appeal Judgment, rather than from a theoretical perspective of whether hate speech is an inchoate crime or not.

The discussion will then turn to a critical analysis of the Šešelj Trial Judgment and the Šešelj Appeal Judgment. Without delving into the various controversies surrounding it, eloquently articulated by Judge Lattanzi in her partially dissenting opinion, 10 the Šešelj Trial Judgment presents us with a unique opportunity to investigate what constitutes hate speech that is criminal, including: (i) whether calls to violent action are required; (ii) what this entails, e.g. calls to commit crimes; and (iii) the causal nexus required. This will be done by outlining the legal elements of the crimes and modes of liability for which Šešelj was convicted by the Appeals Chamber. The analysis of these issues is glaringly missing in the Šešelj Trial Judgment and, while the Šešelj Appeal Judgment sheds some light on certain aspects of this, the discussion is still by no means comprehensive — something inherent in an appeal procedure, which does not — and should not — entail a de novo review of the facts. 11

Since the Šešelj Appeal Judgment primarily focused on hate speech as the underlying act of persecution (a crime against humanity) and as a form of instigation (a mode of liability) the discussion will revolve around the legal test required for hate speech to qualify as persecution and as instigation of crimes, such as persecution, deportation and other inhumane acts (forcible transfer) as crimes against humanity, as well as murder, torture, cruel treatment and plunder of public or private property as violations of the laws or customs of war. 12 The case will be made that a contextual reading of the jurisprudence shows that only hate speech that incites violence, in the form of inciting the commission of crimes, is criminal under international criminal law, while incitement to discrimination or to hatred do not qualify. Furthermore, in the context of hate speech as a crime of persecution, a crime against humanity, and as a mode of liability of instigation of crimes, a causal connection between such speech and the subsequent act of violence needs to be proven. The article concludes that the criminalization of hate speech should not be expanded nor broadly interpreted as this would go against both a holistic reading of the jurisprudence and international human rights law. 13 A look at human rights law is instructive not only because the application and interpretation of international criminal law must be consistent with it, but also because it is also facing similar challenges in defining hate speech, as will be noted below. 14

As a starting point, we know what hate speech is not — it is not direct and public incitement to commit genocide. 15 It is therefore useful to delve into what direct and public incitement to commit genocide is, so that we can exclude it — by a process of elimination — when trying to define the contours of hate speech.

As a statutory crime first set forth in the Genocide Convention, the elements of the crime of direct and public incitement to commit genocide are easier to ascertain, even in the light of the paucity of jurisprudence. This is because the text of the Genocide Convention itself, and an assessment of its travaux préparatoires , already explain the main purpose and object of the drafters. This is also why it is helpful to use this crime as a baseline to discuss the legal parameters of hate speech both as persecution (a crime against humanity) and instigation (a mode of liability).

Looking at the plain language of the Genocide Convention, it is the most extreme form of incitement that is criminalized — public and direct. 16 The Genocide Convention also clearly excludes from its ambit hate speech ‘only’ inciting to racial discrimination or hatred, short of incitement to violence which is not genocide. 17 For instance, showing support for other persons’ speeches which incited genocide is not enough to constitute direct and public incitement to genocide. 18

ICTR jurisprudence on the elements of this crime is also pertinent in understanding its contours: the actus reus of direct and public incitement to commit genocide requires that the accused directly and publicly incites the commission of genocide; 19 the mens rea required is the intent to directly prompt or provoke (an)other individual(s) to commit genocide. 20

With regard to what can constitute the direct element of incitement, the Nahimana et al. Appeal Judgment, upholding the trial chamber’s position, 21 stressed that what is relevant is the meaning of the words used in the specific context. 22 If the discourse remains ambiguous, even when considered in its context, it cannot constitute direct incitement to commit genocide. 23 The ICTR Appeals Chamber also required specifically urging another individual to take immediate criminal action , rather than merely making a vague or indirect suggestion. 24 Being an inchoate offence, the crime of direct and public incitement to commit genocide is punishable even where the incitement fails to produce the result expected by the perpetrator, that is, even if no genocide actually follows. 25 Against this backdrop, the author will now turn to the early jurisprudence on hate speech which does not amount to direct and public incitement to commit genocide.

Any inquiry into hate speech should commence with the IMT media cases against Julius Streicher, editor-in-chief of the anti-Semitic newspaper, Der Stürmer , and Hans Fritzsche, Head of the Radio Division Nazi’s Germany’s Ministry of Propaganda. In 1946, the IMT convicted Streicher of persecution based on his publications from 1938 through 1945, 26 while it acquitted Fritzsche because, though his speeches showed anti-Semitism, they did not urge persecution or extermination of Jews. 27

Some scholars argue that the two findings by the IMT support opposing views: on the one hand, that criminal prosecution can target a wide range of speeches amounting to persecution while, on the other hand, that such charges should only be directed at a narrow and extremely serious category of hate speech which amounts to direct calls to violence. 28 One line of interpretation goes even further and holds that Streicher was only convicted on the basis of unambiguous calls for extermination of Jews — a call to commit violent crimes — and not because of his earlier pre-World War II publications. 29

Forty years later, the ICTR was faced with the same issues when it was dealing with hate speech as persecution, a crime against humanity in, amongst others, the famous Media case against Ferdinand Nahimana, Jean-Bosco Barayagwiza and Hassan Ngeze ( Nahimana et al. ). Nahimana and Barayagwiza were found guilty of persecution on the basis of the fact that they had been, inter alia, advocating ethnic hatred or inciting violence against the Tutsi population for RTLM radio broadcasts in 1994; Ngeze was convicted in part for the same crime in relation to the content of Kangura newspaper publications, of which he was editor-in-chief. 30

The Nahimana et al. Appeal Judgment reversed those convictions for persecution based on hate speech without a call to violence against the Tutsis, while upholding those based on hate speech that was accompanied by direct calls to violence against Tutsis. 31 It held that ‘hate speeches and calls for violence’, considered as a whole and in the context of a widespread campaign of persecution against the Tutsis, constituted acts of persecution. 32 Being crimes against humanity, such acts of persecution would have been committed as part of a widespread or systematic attack against the civilian population. Thus, the holding in the Nahimana et al. Trial Judgment that there need not be a ‘call to action in communications that constitutes persecution’ nor a ‘link between persecution and acts of violence’ was reversed on appeal. 33 Yet, the Appeal Chamber did not explicitly state which was the determinative factor for its finding of persecution: the hate speech, as part of a widespread and systematic attack against the Tutsi population, or the calls to violent acts, which in this case involved calls to commit crimes against the person, including murder, torture, ill-treatment and rape. 34

These early cases on hate speech do not therefore fully answer the question of whether hate speech short of calls to violence may be sufficient for a finding of persecution as a crime against humanity. From this case-law, it is also not clear if violence has to actually ensue and be caused by such calls to violence — though of course, factually, violence in the form of murder, deportation, torture, ill-treatment, and rapes did take place during and after Streicher’s and Nahimana’s respective writings or speeches. 35

Due to this lack of clarity, divergent viewpoints continue to plague the commentaries on the jurisprudence on hate speech. This is also why the Appeals Chamber’s judgment in the Šešelj case raised so many expectations, and why it likely disappointed so many when it was issued, since it still did not fully address the point of whether incitement to discrimination or hatred, short of calls to violence alone can amount to persecution, and if so, in which circumstances. To put the analysis in perspective, a short review of the case against Šešelj is in order.

Vojislav Šešelj, a nationalist politician, President of the Serbian Radical Party and member of the Serbian Parliament, was charged by the ICTY Office of the Prosecutor, inter alia, with having instigated and committed crimes against humanity and war crimes, including deportation and persecution, during the conflicts that ravaged the former Socialist Federal Republic of Yugoslavia during the 1990s. 36 The prosecution specifically charged Šešelj with persecution as a crime against humanity by direct and public denigration through hate speech against the non-Serb population of Vukovar in Croatia, Zvornik in Bosnia and Herzegovina and Hrtkovci in Vojvodina, Serbia. 37 The prosecution further alleged that Šešelj recruited and organized Serb volunteer units, known as ‘Šešeljevci’, who committed crimes against non-Serb civilians. 38

The case against Šešelj was that he had instigated the direct perpetrators of the alleged crimes by, inter alia, using inflammatory and denigrating propaganda against non-Serbs in his speeches and publications. 39 It further held that Šešelj was aware of the power of his propaganda and his influence with Serb volunteers, in particular the Šešeljevci. 40 Šešelj, who represented himself, claimed that his speeches were intended to galvanize Serb troops and to articulate his own political vision. 41 He acknowledged that he advocated his ideology, but claimed that this did not amount to persecution. 42 The Trial Chamber in Šešelj , by majority, acquitted Šešelj of all charges even though it found that three of his speeches — one given in Hrtkovici (Vojvodina, Serbia) on 6 May 1992 and two to the Serbian Parliament on 1 and 7 April 1992 — were calls for the expulsion and forcible transfer of Croats. 43

On appeal, the prosecution averred that the Trial Chamber had erred in not finding Šešelj responsible for crimes on the basis of his speeches. 44 It submitted in this regard that the Trial Chamber had failed to engage with its core argument that his ‘relentless propaganda campaign’ instigated the commission of crimes against non-Serbs. 45 In the end, the Appeals Chamber reversed certain acquittals by the Trial Chamber, and entered convictions for instigating persecution (forcible displacement), deportation, and other inhumane acts (forcible transfer) as crimes against humanity and for committing persecution (violation of the right to security) as a crime against humanity in Hrtkovci, Vojvodina. 46 With this in mind, a brief overview of one of Šešelj’s speeches, the Hrtkovici speech, which was the basis of his convictions on appeal, is called for. This is followed below by a discussion on various issues arising out of both the Šešelj Trial Judgment and Šešelj Appeal Judgment.

On 6 May 1992, Šešelj spoke in the village of Hrtkovci in Vojvodina, Serbia and claimed ‘there was no room for Croats in Hrtkovci … that the Croats who had not yet left of their own accord would be escorted to the border by bus; … that he firmly believed that the Serbs from Hrtkovci and the surrounding villages … would promptly get rid of the remaining Croats in [their] village and the surrounding villages.’ 47 The crowd then chanted ‘ Ustashas out’, ‘Croats, go to Croatia’ and ‘This is Serbia’. 48

The Trial Chamber found by majority that the speech given by Šešelj clearly constituted a call for the expulsion of Croats from the village. 49 However, it deemed, by a different majority, that the prosecution had failed to prove that this speech was the reason for the departure of the Croats or for the campaign of persecution carried out in the village following the speech. 50

The Appeals Chamber, recalling the Trial Chamber’s own finding that this speech constituted a ‘clear appeal’ for the expulsion of Croats in Hrtkovci, instead held that no reasonable trier of fact could have found that it did not ‘incite violence that denigrated and violated the right to security of members of the Croatian population’. 51 According to the Appeals Chamber, Šešelj's speech was grave enough to amount to persecution as a crime against humanity. 52 With his speech, Šešelj ended the ‘relative peace in Hrtkovci’ and the ‘sense of safety by infecting the village with hatred and violence’ and led to the departure of Croats in the ensuing months. 53 It held that Šešelj’s speech denigrated Croats on the basis of their ethnicity, in violation of their right to respect for dignity as human beings. 54 The Appeals Chamber thus found Šešelj had committed persecution, a crime against humanity, based on a violation of the right to security. 55

A. Terminology Used

Looking first at the prosecution’s framing of the charge of persecution, the ‘direct and public denigration’ of non-Serbs echoes the terminology of the crime of direct and public incitement to genocide. This epitomizes the way the jurisprudence on incitement to genocide has shaped prosecutorial strategy and influenced the discourse on hate speech in judicial settings.

The Trial Chamber noted that there was no proof that the speech was the reason for the Croats’ departure from Hrtkovci and the persecutory campaign against them, but did not explain why such proof would have been required. 56 It did not lay down the law on persecution and on hate speech as the actus reus of such crime, so we are left guessing why it reached its conclusions. Clarity on appeal was, therefore, even more necessary. The Appeals Chamber restated that persecution as a crime against humanity under Article 5(h) of the ICTY Statute was an act or omission which

discriminates in fact and which denies or infringes upon a fundamental right laid down in international customary or treaty law ( actus reus ); and (ii) was carried out deliberately with the intention to discriminate on one of the listed grounds, specifically race, religion, or politics ( mens rea ). 57

Unlike the crime of direct and public incitement to genocide, the crime of denigrating hate speech as persecution requires that the persecution in fact takes place against the targeted person or group in violation of a fundamental right — that there be discrimination in fact . For this reason, it was important for the prosecution to prove the impact of the speech. For ease of discussion, the remainder of this section will follow the elements of persecution set forth above.

B.   Actus reus of Persecution: Hate Speech as Incitement, including the Language of the Speaker, the Targeted Audience, and the Influence over such Audience

At the outset, it should be noted that the terminology and methodology used by the Appeals Chamber in the Šešelj case is at times less than ideal. The Appeals Chamber first recalled that the ICTR Appeals Chamber in Nahimana et al. had held that ‘speech inciting to violence against a population on the basis of ethnicity, or any other discriminatory ground, violates the right to security of the members of the targeted group and therefore constitutes “actual discrimination”’, and that the context is important. 58 It went on to establish that Šešelj, by ‘instigating the forcible expulsion of Croatians from Hrtkovici’, ‘incited violence against them, in violation of their right to security’, adding that he had also denigrated the Croatians of Hrtkovci on the basis of their ethnicity, in violation of their right to respect for dignity as human beings. 59 The Appeals Chamber here relied on its previous factual finding on instigation of forcible displacement and forcible transfer (amounting to persecution) to make a finding on commission of persecution, considering the violations of fundamental human rights it entailed (right to security). 60 Put differently, in this case the factual findings made to find instigation were used as the ‘vehicle’ for the finding of commission of persecution.

Recalling the Trial Chamber’s own finding that Šešelj’s speech was a ‘clear appeal’ for the expulsion of Croats from Hrtkovci, the language used by šešelj — ‘there is no room for Croats in Hrtkovci’, 61 his direct address to the Croats, the targeted persons — ‘no, you have nowhere to return to’, 62 his influence over the Serb crowd and the similarities between his words and the acts (‘repeated mistreatment, threats, and violence resulting in a large percentage of them leaving Hrtkovic’) that subsequently took place, the Appeals Chamber found that this speech amounted to incitement as the actus reus of persecution. 63 However, it did not explicitly define ‘incitement’. More specifically: is incitement a call to violent action or to crime?

The Appeals Chamber’s description of what it considers incitement, as actus reus of the crime of persecution, is rooted in contextual factors. 64 This has in turn led some scholars to argue that hate speech occurring in the context of a widespread or systematic attack against a civilian population can constitute persecution if it violates a fundamental right and if, considered ‘cumulatively’ with other persecutory acts, it meets the gravity threshold required for persecution. 65 This is so, in their opinion, even if the words in question do not call explicitly for violence. 66 In any case, they also point out that the Appeals Chamber emphasized the fact that Šešelj ‘incited violence (particularly in the form of expulsion of the Croatian population in Hrtkovci)’. 67 Other scholars however suggest that, in this case, there were actually no explicit calls to violence and ‘no need for an immediate context of violence and crimes’. 68

A plain reading of the Šešelj Appeal Judgment confirms that the Appeals Chamber considered incitement to commit crimes, which could also reasonably be expected to involve or lead to violence, as an act of inciting violence. 69 It held that ‘no reasonable trier of fact could have found that Šešelj’s speech did not incite violence ’ and repeated that he had in fact ‘ incited violence’. 70 The Appeals Chamber clearly equates incitement to violence with incitement to commit crimes, and more specifically, forcible transfer and deportation. There is no indication in the Šešelj Appeal Judgment that the appellate judges considered the related issues of whether ‘incitement to violence’ could be broader than inciting crimes. This therefore remains an unsettled issue.

It should also be noted that, in addition, the Appeals Chamber did not address the possibility that Šešelj, by inciting forcible transfer and deportation on the basis of ethnicity, was inciting discrimination. Incitement to discrimination has been described as ‘beseeching listeners or readers to oppress the victim group in certain non-violent ways.’ 71 Thus, it remains unclear from this judgment and the prior jurisprudence, whether a case of incitement to violence, absent any incitement to commit a crime, or incitement to discrimination only (without other accompanying discriminatory acts) could amount to persecution under certain circumstances.

A sideways glance at human rights law is informative on this point. Incitement has been defined as statements about national, racial, or religious groups creating an imminent risk of discrimination, hostility or violence against persons belonging to them — where imminence is related to directness. 72 With regard to racist hate speech, incitement has been defined as influencing others to engage in certain forms of conduct, including the commission of crimes, through advocacy or threats. 73 Advocacy is generally understood to be ‘explicit, intentional, public and active support and promotion of hatred’ towards a target group. 74 Meanwhile, violence is ‘the use of physical force or power against another person, or against a group or community, which either results in, or has a high likelihood of resulting in, injury, death, psychological harm, maldevelopment or deprivation’. 75

While there are elements of convergence between some aspects of international human rights and international criminal law, such as the inclusion of incitement to commit crimes as an act of incitement and the concepts of ‘imminence’ and ‘publicity’, human rights law does not provide clear guidance as to whether incitement to discrimination, hostility, or violence should be criminalized or otherwise sanctioned. 76 The only emerging consensus among states seems to be, vis-à-vis the most extreme forms of incitement to violence, that of criminalizing incitement to imminent violence, and then only on the ground of religion or belief. 77 In light of this, the author is of the view that the only correct legal conclusion is that international criminal jurisprudence, similar to the consensus emerging in human rights law, may only be interpreted as criminalizing forms of incitement to violence, which are incitement to commit crimes where ensuing violence occurs. It does not criminalize incitement to ‘simple’ discrimination, hatred, or hostility when there is no incitement to actual violence.

In short, the Appeals Chamber in the Šešelj case chose to pursue, as the ICTY and ICTR had often done in the past, a factual, case-by-case assessment of whether the conduct in question constituted incitement, rather than laying down a general definition in abstracto . This may be judicially expedient, of course, but it does make identifying the legal standard of what constitutes incitement amounting to persecution somewhat harder, thus leaving more room to conflicting interpretations.

C. Actus reus of Persecution: Causality or Impact of Speech and Discrimination in Fact

The Appeals Chamber also noted that following Šešelj’s speech, Croats were increasingly harassed, threatened and became victims of violence, which led to a large number of them leaving Hrtkovci. 78 The Appeals Chamber’s reasoning is actually consistently peppered with references to violence against the Croats following Šešelj’s speech. For instance, in both its discussions on commission of persecution and instigation, the climate of ‘repeated mistreatment, threats and violence’, 79 Šešelj’s ‘infecting the village with hatred and violence’, 80 and ‘the context of coercion, harassment, and intimidation, which was met with inaction by the local authorities’ 81 are regularly referred to. It seems that, equally important to the language used by Šešelj, are the resulting crimes and violent acts committed against the Croats.

The Appeals Chamber’s discussion of incitement and the impact of such incitement (i.e. the threats and violence which indeed followed it) were intertwined to such an extent that the impact can be described as featuring as a decisive factor for the finding of persecution. Indeed, the impact of the incitement is usually relevant in the context of persecution to show that discrimination in fact occurred — it is a legal requirement, as mentioned above. 82 Ideally, therefore, it should have been dealt with separately, as a discrete legal element of the crime of persecution, not as part of the finding of the incitement discussion. In this case, however, the Appeals Chamber’s finding of what constituted the actus reus of incitement for persecution was also riddled with references to the violence which followed Šešelj’s speech. The centrality given to the impact of the speech (i.e., the ensuing threats and violence) cannot be underestimated. This may explain why Presiding Judge Meron, who had partially dissented on hate speech as persecution in the Nahimana et al. Appeal Judgment, agreed to this finding. 83

Thus, although some commentators argue that a link between the speeches and the subsequent acts that constitute persecution as a crime against humanity is not required, and that it is not necessary for a particular result to be proven as a direct consequence of the speech, the Šešelj Appeal Judgment, read as a whole, offers scant support for this viewpoint. 84 For a finding of persecution, the Šešelj Appeal Judgment — which is the most recent in the developing line of jurisprudence of ICTR, ICTY and IRMCT — does actually require incitement to violence being inciting crimes and violent acts that occur as a consequence of such incitement.

D. Actus reus of Persecution: In Violation of a Fundamental Right

As briefly mentioned above, the Šešelj Appeal Judgment goes on to hold that Šešelj’s speech denigrated the Croats on the basis of their ethnicity, in violation of their ‘right to respect for dignity as human beings’. 85 It further endorsed the ICTR Appeals Judgment in the Nahimana et al. case that ‘speech inciting to violence against a population on the basis of ethnicity, or any other discriminatory ground, violates the right to security of the members of the targeted group and therefore constitutes “actual discrimination”’. 86

The use of the term ‘respect for dignity’ can be traced back to the Nahimana et al. Appeal Judgment. 87 Some have argued that it is unclear how Šešelj’s words were degrading or dehumanizing 88 and that ‘respect for dignity as human beings’ is not a right in and of itself. 89 However, reading this part of the Šešelj Appeal Judgment in context, it is the right to security of the Croats and other non-Serbs that is effectively violated by the persecution. 90 It would indeed have been preferable had the Appeals Chamber not used this terminology, ‘their right to respect for dignity as human beings’, but instead referred to their right of equality and freedom from discrimination — a cornerstone of the human rights system 91 — which is what it was in fact arguably referring to, in context. In any event, it seems relevant that the Appeals Chamber only finds that Šešelj’s words reach the requisite level of gravity amounting to the actus reus of persecution as a crime against humanity only after considering both aspects (violation of the right to security and violation of the right to human dignity/freedom from discrimination).

E. Mens rea of Persecution

The mens rea of persecution is the ‘specific intent to cause injury to a human being because he belongs to a particular community or group’. 92 This specific intent must be proven over and above the general intent to commit the crime of persecution. 93 The Appeals Chamber has held that, while the requisite discriminatory intent may not be inferred directly from the general discriminatory nature of an attack characterized as a crime against humanity, the ‘discriminatory intent may be inferred from such a context as long as, in view of the facts of the case, circumstances surrounding the commission of the alleged acts substantiate the existence of such intent’. 94

In the case of Hrtkovci, based on the same facts from which it concluded that there was incitement amounting to persecution, and in line with its jurisprudence, the Appeals Chamber also inferred Šešelj’s discriminatory intent. 95 It was thus able to exclude inferences that the speech was given as propaganda, in support of the war effort, or to strengthen the morale of the Serb troops (as Šešelj had invariably argued throughout the proceedings). 96 A more detailed explanation of why it excluded these other inferences (which, in the minds of the appellate judges, would have had to be unreasonable given the circumstances) would have helped shed some light on how triers of fact ought to weigh evidence and would have provided insight into the legal standard applicable with regard to the mens rea of persecution carried out through hate speech.

The Šešelj Appeals Chamber also addressed how hate speech may constitute instigation, not as a crime but rather a mode of liability. In the case of instigation, the accused may be convicted if the instigation to a crime under the jurisdiction of the court or tribunal was a factor substantially contributing to the conduct of another person committing a crime. 97 After an overview of the relevant speeches, this section will discuss them in the light of the elements of instigation as a mode of liability, that is, prompting another to commit a crime, and the awareness of the substantial likelihood that a crime will be committed in the execution of that instigation. 98

A. Overview of Relevant Speeches

With regard to the speech in Hrtkovci, the Appeals Chamber found that Šešelj had instigated deportation, persecution (forcible displacement) and other inhumane acts (forcible transfer) as crimes against humanity against non-Serbs. 99 In this context, the Appeals Chamber noted that Šešelj had influence over the members of his political party, that he was even seen by some ‘as if he were a god’, and that his speeches had a significant impact on the audience. 100 The Appeals Chamber further noted that, after Šešelj’s speech, many Croats left for Croatia either out of fear, or by fraudulent housing exchanges with Serb refugees amid an atmosphere of coercion, harassment, and intimidation which was met by inaction on the part of law-enforcement and other officials. 101 In fact, the Appeals Chamber explained, Serbs, including the then Hrtkovci mayor himself, who had listened to Šešelj's speech, regularly threatened non-Serbs who remained in the town. 102 The Appeals Chamber thus considered that in the light of Šešelj's influence over the crowd, the striking parallels between his words and the acts subsequently perpetrated by others including, inter alia , members of his audience, no reasonable trier of fact could have found that he had not substantially contributed to the conduct of the perpetrators. 103

Other speeches considered in the context of instigation were the ones given by Šešelj in the Serb Parliament. In the speech of 1 April 1992, Šešelj had stated: ‘[w]e are going to expel the Croats … We are simply going to pack you into trucks and trains and let you manage in Zagreb.’ 104 In his second relevant speech, on 7 April 1992, he stated: ‘[P]erhaps the best solution … would be simply putting [Croats] … on buses and trucks and taking them to Zagreb.’ 105

The Trial Chamber found that the Serb Parliament speeches clearly constituted calls for the expulsion and forcible transfer of Croats; by another majority, the Trial Chamber however reached the conclusion that they were an ‘expression of an alternative political programme that was never implemented’. 106 It also held that, given the lack of measurable impact and the harsh criticism Šešelj had received for his speeches, it could not find that they amounted to incitement to war crimes. 107 Furthermore, the Trial Chamber found that the prosecution had not shown a causal link between Šešelj’s speeches to the Serb Parliament on 1 and 7 April 1992 and the crimes committed in April 1992 in Mostar, Zvornik, and Greater Sarajevo, or that the crimes committed between May 1992 and September 1993 could be attributed to him. 108 The Trial Chamber therefore found Šešelj not guilty of instigating such crimes. 109

The Appeals Chamber left this finding unchanged because, in its view, the prosecution had not demonstrated at trial the extent of the dissemination of Šešelj’s speeches and the specific impact they had had on the commission of crimes in Mostar, Zvornik, and Greater Sarajevo. 110 The Appeals Chamber also found that the temporal link between the speeches and the subsequent crimes was tenuous. 111

In March 1992, Šešelj had further given a speech in Mali Zvornik in which he had called on his Serb ‘brother Chetniks, especially you across the Drina’ … to clear up Bosnia from the pagans and show them the road to the east where they belong’. 112 Šešelj himself confirmed that he had ‘attacked fundamentalist Muslims and pan-Islamists who wanted Bosnia to separate from Yugoslavia, and called them “ pogani ”’, (which according to him meant ‘waste’ or ‘faeces’) and ‘ balijas ’ (a derogatory term for Muslims). 113 The Trial Chamber, by majority, did not find that, by calling on the Serbs to ‘clear up’ Bosnia of the ‘ pogani ’ and the ‘ balijas ’, Šešelj was calling for the ethnic cleansing of the non-Serbs of Bosnia. 114 The Trial Chamber found that, given the context, this could have been a call by Šešelj to galvanize the Serbs in support of the war effort. 115 Again, the Trial Chamber noted that there was no proof of the impact of this speech. 116

With regard to this speech, the Appeals Chamber instead considered that (based on the political context of a possible declaration of independence by Bosnia and on evidence before the Trial Chamber) no reasonable trial chamber could have found that Šešelj’s speech did not call for ethnic cleansing. 117 It held that the inflammatory language of Šešelj’s speech could have prompted other persons to commit crimes against non-Serb civilians. 118 The Appeals Chamber further assessed whether the Trial Chamber’s finding that the speech (as well as other statements made by Šešelj) ‘had an impact on’, or ‘causal link’ to the commission of crimes against non-Serbs. 119 It then however went on to find that, because the prosecution’s argument on appeal was the temporal link between Šešelj’s speeches and the contemporaneous or subsequent commission of crimes in various locations, given the time span of ‘nearly 3 weeks’ from the time of his speech, a reasonable trier of fact could have found such a link tenuous. 120 It therefore upheld the Trial Chamber’s overall finding. 121

On 7 November 1991, on his way to Vukovar (Croatia), Šešelj had held a press conference in the town Šid in Serbia, close to the border with Croatia (at the time Croatia had declared its independence but had not been recognized as a separate state), stating that ‘this entire area will soon be cleared of the Ustashas’. 122 Then, on 12 and 13 November 1991, he made several additional speeches in Vukovar, saying to Serb forces and to the Šešeljevci that ‘no Ustashas must leave Vukovar alive’ and that they should ‘show [them] no mercy’. 123 Going around the town in a vehicle with a loudspeaker, Šešelj also allegedly called on Croat soldiers to surrender, and according to some of the evidence, he told the ‘Ustashas’ that if they did not, they would die. 124

The Trial Chamber found that, with respect to Šešelj’s speeches in November 1991, the content of what he actually said was equivocal — noting that there was a reasonable possibility that the speeches were rather made to support the morale of the Serb troops. 125

On appeal, the prosecution argued that, as the war in Croatia escalated and after months of ‘building a reservoir of hate’, Šešelj ‘triggered’ the crimes committed in Vukovar. 126 It referred to evidence that these statements were understood by the Šešeljevci to mean that Croat detainees should be executed, as well as to evidence allegedly showing that Šešelj deliberately equated the broader Croat population with ‘ Ustashas ’. 127 The Appeals Chamber held that the prosecution had not addressed the Trial Chamber's main reservation — the content of Šešelj's statements and thus left this finding undisturbed. 128

B. Actus reus of Instigation: Direct Causal Connection

Turning to the Appeals Chamber’s assessment of instigation, the discussion on the Hrtkovci speech is the most instructive. Addressing the prosecution’s argument that Šešelj’s speech was given to a large audience and was quickly disseminated, the Appeals Chamber noted that the Hrtkovci speech was given to ‘some 700 Serb Radical Party sympathizers and citizens[,] 60% of whom were Serbian refugees from Croatia’. 129 While at first glance the size of the audience and the quick dissemination of Šešelj’s speech may appear to be irrelevant — they often go to prove the public element of the crime of direct and public incitement to commit genocide — 130 they also show the causal connection required between the instigation and the crimes committed. Hence, the Appeals Chamber focused carefully on the presence of some perpetrators of crimes, including the future mayor of the town himself, in Šešelj’s audience. 131 The thread running through this assessment is the direct impact the speech had on some members of his audience who subsequently went on to commit crimes, such as forcible transfer. 132 This appears to be so despite the fact that the Appeals Chamber did not explicitly use the term ‘direct’. In this regard, it appears reasonable to conclude that directness is not required to prove a substantial contribution to the conduct of the person committing the instigated crime, although it would probably be easier to meet the burden of proof if there is proof instigation was indeed direct. One wonders if this is again the lingering influence of the jurisprudence on direct and public incitement to genocide. 133

Regarding the speeches in the Serb Parliament, the Appeals Chamber recalled that the prosecution had not demonstrated at trial the breadth of their dissemination and the ‘specific impact’ that they had had on the commission of crimes. 134 These are both puzzling statements. With regard to the former, this is hardly a fair comment on the part of the Appeals Chamber because, by its nature, parliament is a public place, and its impact is, by definition, public. In fact, parliament is so obviously a public place that the prosecution would have met the ‘publicity’ threshold, if this were legally required. But instigation does not require publicity, according to previous statements of the law by ICTY and ICTR. Indeed, wide dissemination of a speech is also not necessary to prove instigation. Why the Appeals Chamber mentioned this as though it were a legal requirement remains a mystery.

As to the ‘specific impact’ comment, it is unclear what this means as well. There is no legal explicit requirement amounting to ‘specific impact’ for a finding of instigation: there need only be a substantial contribution through this speech of the crimes that ensued. Is this reference to specific impact possibly related to a temporal link? 135 In the context of instigation, the ‘causal connection’ 136 between the instigation and the actus reus of the crime is necessary to show the instigation was substantial; maybe the Appeals Chamber was trying to detail what is needed, especially in the context of ‘political’ speeches, for the link (i.e. the contribution) between the words of the instigator and the crimes committed by others, to be substantial enough. The Appeals Chamber’s use of the term ‘specific’, possibly meaning direct, unfortunately causes more confusion in an area where jurisprudence is scant.

C. Causal Connection: Temporal

As to the Mali Zvornik and the Serb Parliament speeches, the Appeals Chamber found that the time elapsed between them and the subsequent crimes could lead a reasonable trier of fact to conclusions other than a finding of causality, although this argument is definitively under-developed. 137 For instance, the time span of ‘nearly 3 weeks’ from the time of Šešelj’s Mali Zvornik speech and the subsequent crimes committed in Zvornik was considered a ‘tenuous’ link. 138 With regard to the speeches in the Serb Parliament, the Appeals Chamber merely held that, on appeal, the Prosecution had failed to show that no reasonable trier of fact could have reached the impugned conclusion. 139

In this regard, it is noteworthy that the Nahimana et al. Appeal Judgment upheld a finding of instigation following an RTLM broadcast which urged Tutsis to return to their homes, when many of those who returned were immediately killed — on the same day of their return and the broadcast. 140 The Nahimana et al. Appeals Chamber also left undisturbed a finding of instigation in the killing of three priests who had been named in a RTLM broadcast of 20 May 1994 and were subsequently killed. One of them, Father Ngoga, who had earlier managed to escape, was killed 11 days after the broadcast. 141 Thus, even when a broadcast did not include a call for the killing of Tutsis, a direct temporal link was still found between the broadcast and the commission of the subsequent crimes. In this case, scholars have argued that the context of the RTLM other broadcasts (and the overall situation in the area) was also relevant to the finding of instigation. 142

Looking at other earlier cases, such as Streicher and Ruggiu , the time elapsed between the speech or broadcast, on the one hand, and the commission of crimes for which they were held responsible, on the other, varied between weeks and months (in the case of Nazi Germany, arguably even years). 143 Granted, those cases referred to persecution as a crime against humanity (and direct and public incitement to genocide) and the temporal link was considered in the context of discrimination in fact, an aspect of the actus reus of persecution: these findings were not taking into account the temporal link as relevant to the causal connection with regard to instigation. However, when assessing the impact of speech, courts tend to look at a series of events that culminate in the commission of crime(s), regardless of the rubric under which the charges are being considered. 144 Causation is a tool to gauge the mental effect a speech had on the perpetrators of the crime(s) in question. 145 As such, the test is the same whether the tool is used in the context of examining the crime of persecution or the mode of liability of instigation.

In light of this, it appears somewhat arbitrary for the Šešelj Appeals Chamber to implicitly make a finding that, in the context of instigating a crime, a temporal link of approximately three weeks is tenuous. The Appeals Chamber appears to be considering that a three-week gap between the speech and the subsequent (instigated) crimes does not translate into a substantial contribution as other factors may have also contributed to the crimes in question, thus negating the substantial nature of the contribution of the original speech. Indeed, the longer the time elapsed between the speech and the crimes that took place, the greater the possibility that other events could have intervened and (at least partially) contributed to the actual commission of such crimes. 146 Moreover, being an appellate review, the Appeals Chamber was only looking at whether a reasonable trier of fact could have reached such findings, even if maybe some of the appellate judges themselves would have reached a different conclusion on the same facts. Nonetheless, even if other factors had intervened to partially contribute to the crimes committed, a trier of fact should be careful to keep in mind that it is possible to have multiple factors causing a crime (each contributing), and still find instigation 147 : what is required is that the speech was the cause that made a substantial contribution to the crime, not the sole or essential contribution.

What if the prosecution had been able to prove a link other than the temporal link to show that Šešelj’s speech amounted to a substantial contribution and therefore instigation? For instance, if it could have shown that some Serb followers of Šešelj, or the Šešeljevci, over whom he had at least de facto moral authority, committed some of these subsequent crimes? What if some Šešeljevci were proven to have been present in the audience when he gave this speech, as was the case with the Hrtkovci speech? What if evidence could have been found, from insiders in the Serb groups, of how the speech had prompted them to commit crimes? In such instances, a trier of fact, applying the law correctly, would have had to establish instigation, and reach a verdict very different from the one entered by the Trial Chamber and upheld by the Appeals Chamber.

Understandably, on the evidence discussed by the Trial Chamber, the Appeals Chamber instead seemed reluctant to assess evidence on the causal connection itself, since the relationship in those instances of speeches was somewhat more indirect than what it had in the case of the Hrtkovci speech. Yet, we would have benefited from a more thorough analysis by the Appeals Chamber on this point. 148

D. Actus reus of Instigation: Content and Purpose of the Speech

It is interesting to note that with regard to the Vukovar speeches, the Appeals Chamber framed the main issue as being that of the content of the message. What a speaker actually says is the starting point of any inquiry as to whether a speech can be considered as criminal. The term ‘ Ustasha ’ according to Šešelj himself, referred to fascist Croats who had sided with the Nazis in World War II, and who had killed Serbs. 149 In the 1990s, according to him, this term meant the Croatian extremists, supported by Franjo Tuđman, then President of Croatia, who persecuted Serbs. 150 The Trial Chamber did not explain the historical and cultural meaning of this word further, nor why it chose to believe Šešelj’s own explanation of who he was referring to when he used this word. The Trial Chamber also did not explain why it did not give weight to the testimony of those who explained what the words meant to them. Such an approach was clearly inconsistent with that taken in the Nahimana et al. and Bikindi Trial Judgments, which relied on testimony to decipher the meaning of historical references made by the speakers, and how these references had changed over time. 151 Such analysis is unfortunately lacking in the Šešelj Trial Judgment.

A closer look at what the Šešelj Trial Chamber actually held shows that the emphasis in its analysis was on another issue ― the purpose of the speech. Besides the unclear content of the speech, the Trial Chamber found that:

more decisively , [it][ ]…does not deem that the reported speeches, even if we assume they have been proven, can be considered as acts of instigating a crime. Their context rather suggests that these were speeches aimed at reinforcing the Accused’s political party. 152

Assuming the speech was actually given (something the Trial Chamber should have probably established beyond the reference to it being ‘reported’), a reasonable trier of fact should have assigned more weight to it, especially in the atmosphere of witness intimidation surrounding this trial, together with the out-of-court statements of recanted witnesses. 153 In light of these circumstances and if Šešelj had actually uttered the words alleged (including that ‘no Ustashas must leave Vukovar alive’, that Serbs should ‘show [them] no mercy’, and the other speeches against Croats), it would have been reasonable for a trier of fact to reach the conclusion that these speeches had prompted crimes: at the end of the day, whether a speech has a substantial effect on the commission of a crime is a fact-based inquiry. 154 In any event, and even considering the role of appellate instances, a more in-depth analysis on the part of the Appeals Chamber on the content and purpose of the speech given would still have been useful for further clarity of the law.

E. Mens rea of Instigation

The ICTY Appeals Chamber had previously held that a person who instigates an act or omission with the awareness of the substantial likelihood that a crime will be committed in the execution of that instigation, has the requisite mens rea for establishing criminal liability. 155 This means that an individual who instigates an act with the awareness of the substantial likelihood that persecution as a crime against humanity will be committed in the execution of such instigation may be held liable for the crime of persecution. 156

The Šešelj Appeal Judgment held that instigation implies ‘prompting another person to commit an offence’, confirming the above-mentioned standard. Although proof of a causal connection between the instigation and the actus reus of the crime is required, the prosecution need not prove that the crime would not have been perpetrated without the accused’s prompting: in other words, the contribution does not need to be essential.

Applying this standard, the Šešelj Appeals Chamber found that Šešelj intended to prompt the commission of crimes or, at the very least, was aware of the substantial likelihood that they would be committed as a result of his instigation through his speech of 6 May 1992 in Hrtkovci. 157 Again, the mens rea discussion is factual: the Appeals Chamber explicitly referred to the ‘content’ of his speech and inferred his mens rea from the ‘inflammatory words’ used. 158 This does highlight the importance of the actual words used, which should in fact be the starting-point of any analysis of hate speech cases.

The Šešelj Appeal Judgment was an opportunity to clarify the earlier case-law on what kind of hate speech can be regarded as criminal. The IRMCT Appeals Chamber however failed to explicitly and unequivocally address the issue. Based on a plain reading of the judgment, it is only the more extreme form of incitement to violence, that is inciting to commit crimes and ensuing violent acts that constitutes persecution as a crime against humanity, not ‘mere’ incitement to discrimination or hatred. This requirement for incitement to violence to be incitement to commit crimes with violent acts as a result of such incitement is not only supported by a holistic reading of the more recent ad hoc tribunals’ appeals chambers’ jurisprudence, but is also reflective of the current state of international human rights law, with its emerging consensus that allows criminalization of hate speech, but only when it amounts to incitement to imminent violence. 159 International human rights law, as it stands today, does not extend to the criminalization of other forms of incitement. 160

With regard to the second question this article posed, i.e. whether a causal nexus is required between incitement to violence — incitement to commit crimes, and any subsequent acts of violence, the answer is in the affirmative. Not only is actual discrimination an element of the crime of persecution: in Šešelj , the actual violent acts that ensued were actually decisive to the finding of persecution committed through incitement to violence. Because the appellate judges did not discuss the lapse of time between the Hrtkovci speech and the subsequent violent acts, it is hard to ascertain if imminence is also required, and what this might entail in various factual circumstances. 161

Turning to hate speech as a form of instigation of crimes (a mode of liability, rather than a crime itself), the Šešelj Appeals Chamber’s restatement of the law was less controversial. The substantial causal connection required for instigation of a crime must be direct, whether temporal or otherwise — although the Appeals Chamber did not explicitly use the term ‘direct’, and directness is not a legal requirement for instigation.

In any event, in the Šešelj case, the difference between, on the one hand, hate speech as the actus reus of persecution, which requires incitement to violence, that is, incitement to commit crimes and subsequent acts of violence and, on the other, instigation, which requires a substantial contribution to the ensuing crime, remained somewhat blurred. This is in part owing to the reliance by the Appeals Chamber on the same underlying factual findings, which formed the basis for the prosecution’s arguments on both persecution and instigation. But it is also because the Appeals Chamber did not clearly define whether criminal hate speech amounts to the commission of crimes. Defining hate speech in international criminal law thus continues to be elusive.

She is also a PhD researcher on hate speech in international criminal and human rights law. She has worked as Legal Officer with REDRESS, at the International Criminal Tribunal for the former Yugoslavia, and at the UN Office of the High Commissioner for Human Rights. The views expressed are those of the author alone.

For the purposes of this article, the term ‘hate speech’ is used to refer to incitement to violence as incitement to commit crimes, falling short of direct and public incitement to commit genocide. It does not cover ‘mere’ incitement to hatred or discrimination. On the issue, see e.g. Combating Racist Hate Speech (CERD Recommendation No. 35), UN Doc. CERD/C/GC/35, 26 September 2013, at 46; ECRI General Policy Recommendation No. 15 on Combating Hate Speech (ECRI Recommendation No. 15), Council of Europe, European Commission against Racism and Intolerance, CR (2016)15, 8 December 2015, § 4. See also UN Secretary-General's remarks at the launch of the United Nations Strategy and Plan of Action on Hate Speech, 18 June 2019, available online at https://www.un.org/sg/en/content/sg/statement/2019-06-18/secretary-generals-remarks-the-launch-of-the-united-nations-strategy-and-plan-of-action-hate-speech-delivered (visited 17 March 2020) ; Opening Statement by UN High Commissioner for Human Rights, 41st session of the Human Rights Council, 24 June 2019, available online at https://www.ohchr.org/EN/NewsEvents/Pages/DisplayNews.aspx?NewsID=24724&LangID=E (visited 17 March 2020).

See Framework of Analysis for Atrocity Crimes, A Tool for Prevention , July 2014, at 18–24; Decision on Follow-Up to the Declaration on Prevention of Genocide: Indicators of Patterns of Systematic and Massive Racial Discrimination , CERD/C/67/1, 14 October 2005, §§ 8–9. See also Report of the independent international fact-finding mission on Myanmar , A/HRC/39/64, 12 September 2018, §§ 73, 81, 85.

Judgment, Šešelj (MICT-16-99-A), Appeals Chamber, 11 April 2018 ( Šešelj Appeal Judgment).

Judgment, Šešelj (ICTY-03-67-T), Trial Chamber, 31 March 2016 ( Šešelj Trial Judgment).

Judgment, Nahimana, Barayagwiza, Ngeze (ICTR-99-52-A), Appeals Chamber, 28 November 2007 ( Nahimana et al. Appeal Judgment), § 692.

See e.g. W. Timmermann, ‘International Speech Crimes Following the Šešelj Judgment’, in P. Dojčinović (ed.), Propaganda and International Criminal Law: From Cognition to Criminality (Routledge, 2019), chap 4, at 115–116, 118; R. Wilson and M. Gillett, The Hartford Guidelines on Speech Crimes in International Criminal Law (Peace and Justice Initiative, 2018), at 144.

See G. Gordon, Atrocity Speech Law: Foundation, Fragmentation, Fruition (Oxford University Press, 2017), at 19–24, 316–321, 346–347, 349–356; 373–382, 403–404; G. Gordon, Reply by Gregory S. Gordon: On the General Part, the New Media and the Responsibility to Protect , 14 July 2017, available online at http://opiniojuris.org/2017/07/14/reply-by-gregory-s-gordon-on-the-general-part-the-new-media-and-the-responsibility-to-protect/ (visited 17 March 2020), at 1. See also W. Timmerman, ‘Inciting Speech in the former Yugoslavia: The Šešelj Trial Chamber Judgment’, 15 Journal of International Criminal Justice (2017) 133–155, at 154–155; Timmerman, supra note 6, at 119.

The jurisprudence is silent on this point. A discussion on what constitutes ‘calls to violence’ is found below in the context of the discussion on incitement as the actus reus of persecution (a crime against humanity).

See Partially Dissenting Opinion of Judge Flavia Lattanzi – Amended Version, Šešelj Trial Judgment (Dissenting Opinion), Vol. 3.

See Šešelj Appeal Judgment, §§ 12, 14.

These were the crimes for which the prosecution requested the Appeals Chamber to convict Šešelj. See Third Amended Indictment, Šešelj (Indictment) (IT-03-67), 7 December 2007, §§ 18, 20–22, 24–30, 34. The author will not address the chapeau requirements of crimes against humanity but will limit the analysis to the underlying act of hate speech and when and how this may amount to persecution.

References are made to human rights law throughout but an in-depth analysis of hate speech under international human rights law is beyond the scope of this article.

See Art. 21(3) ICCSt.

Nahimana et al. Appeal Judgment, § 692.

Art. III(c) Genocide Convention.

See also D.F. Orentlicher, ‘Criminalizing Hate Speech in the Crucible of Trial: Prosecutor v. Nahimana ’, 21 American University International Law Review (2006) 557–596, at 561, 563; W. Schabas, ‘Hate Speech in Rwanda: The Road to Genocide’, 46 McGill Law Journal (2000) 141–171.

Judgment, Nyiramasuhuko, Ntahobali, Nsabimana, Nteziryayo, Kanyabashi, Ndayambaje (ICTR-98-42-A), Appeals Chamber, 14 December 2015 ( Nyiramasuhuko et al. Appeal Judgment), § 3340.

Nahimana et al. Appeal Judgment, § 677.

Ibid. ; Judgment, Ngirabatware (MICT-12-29-A), Appeals Chamber, 18 December 2014 ( Ngirabatware Appeal Judgment), § 58.

Judgment and Sentence, Nahimana, Barayagwiza, Ngeze (ICTR: 99-52-T), Trial Chamber I ( Nahimana et al. Trial Judgment), 3 December 2003.

Nahimana et al. Appeal Judgment, §§ 701, 703, 711, 715.

Ibid. , §§ 701, 711.

Nyiramasuhuko et al. Appeal Judgment, § 3338.

Nahimana et al. Appeal Judgment, §§ 678, 720, 766. Some scholars argue that the jurisprudence of the ICTR is ambiguous on this point as many trial judgments also refer to the direct causal link between speeches and subsequent commission of crimes. This possibly reflects the position taken by the International Law Commission in its 1996 Draft Code of Crimes Against the Peace and Security of Mankind , in which it held that direct and public incitement attracts individual criminal liability when an act of incitement ‘in fact occurs’). See R. Wilson, Incitement on Trial (Cambridge University Press, 2017), at 25, 32 (arguing that the Nahimana et al. Appeal Judgment did not completely clear up this causation issue), 42–43. See also Report of the International Law Commission, Draft Code of Crimes Against the Peace and Security of Mankind with Commentaries , UN Doc. A/C/CN.4/L.532, 17 July 1996, available online at http://legal.un.org/ilc/texts/instruments/english/commentaries/7_4_1996.pdf (visited 17 March 2020), §§ 18, 22. The author is inclined to consider the Appeals Chamber’s views as a mere factual assessment, i.e. that genocide followed the incitement in the Rwanda case, rather than a reflection of an actual legal standard.

See Trial of the Major War Criminals before the International Military Tribunal , Vol. 1 (IMT, Nuremberg, 1947), Julius Streiche r ( Streicher Finding), 302–304.

Ibid. , Hans Fritzsche ( Fritzsche Finding), 336–338. See also Trials of War Criminals before the Nuremberg Military Tribunals under Control Council Law No. 10 , Vol. XIV (Nuremberg: October 1946–April 1949), US v. Ernst Weizsacher et al. (Ministries Case) , Case 11, Otto Dietrich ( Dietrich Finding), 565–576.

Orentlicher, supra note 17, at 582–585; Wilson and Gillett, supra note 6, at 122–123.

Orentlicher, supra note 17, at 585–586; M. Kearney, The Prohibition of Propaganda for War in International Law (Oxford University Press, 2007), at 40 and note 120.

Nahimana et al. Trial Judgment, §§ 7, 1081–1082, 1084.

See Nahimana et al. Appeal Judgment, §§ 988, 993, 995–997, 1001–1002, 1013–1014, 1016; Nahimana et al. Trial Judgment, §§ 1073, 1078–1079, 1081–1084. The Ruggiu Trial Judgment had also considered incitement amounting to advocacy of ethnic hatred short of violence as persecution. The judgment was however not appealed so it is unsure if an appeals chamber would have upheld these findings. See Ruggiu Trial Judgment, § 44(v)–(xii). Cf. Judgment, Kordić and Čerkez (IT-95-14/2-T), Trial Chamber, 26 February 2001, § 209 and note 272. ICTR and ICTY trial judgments are not binding upon each other. Judgment, Karemera and Ngirumpatse (ICTR-98-44A), Appeals Chamber, 29 September 2014, § 52; Judgment, Lukić and Lukić (IT-98-32/1-A), Appeals Chamber, 4 December 2012 § 260.

Nahimana et al. Appeal Judgment, § 988 (emphasis added).

Nahimana et al. Trial Judgment, § 1073.

See Nahimana et al. Appeal Judgment, §§ 987-988.

See Streicher Finding, 302-304; Nahimana et al. Appeal Judgment, § 988; Ruggiu Trial Judgment, § 44 (viii)-(ix), (xi).

Indictment, §§ 2–5, 10(b)(c)(d), 15–17, 31–33.

Ibid ., §§ 17(k), 20, 22, 33.

Ibid. , §§ 10(a), (g),16, 20–22, 24, 26–27, 29, 32–33.

Šešelj Trial Judgment, §§ 286–287.

Ibid. , § 288.

Ibid. , §§ 9, 11, 289, 291, 329.

Ibid. , §§ 291, 297.

Ibid. , §§. 333, 338, 343 and pp. 109–110. The Trial Chamber further found that war crimes had been committed by Serb forces including the Šešeljevci but that there was no hierarchical link between Šešelj and his volunteers once they were integrated with Serb forces ( Ibid. , §§ 116, 205–220, 249), and that there had been no JCE involving him ( Ibid. , § 281. See also ibid. , §§ 220-280).

Šešelj Appeal Judgment, § 120.

Ibid. , § 123.

Ibid. , §§ 155, 165-166, 181.

Ibid. , § 331.

Ibid. , § 332.

Ibid. , § 333.

Ibid. , §§ 161, 163.

Ibid. , § 163. The Appeals Chamber was satisfied that Šešelj’s speech amounted to discrimination in fact and that it was delivered with discriminatory intent, that his conduct formed part of the widespread or systematic attack against the civilian population encompassing also parts of Croatia and Bosnia and Herzegovina, and that he was aware that such conduct formed part of the attack. Ibid. , § 164.

Ibid. , § 163.

Ibid. , §§ 165–166.

See supra text before note 50.

Šešelj Appeal Judgment, § 159 (internal references omitted).

Ibid ., § 163 and note 578 (referring to Šešelj Appeal Judgment, §§154–155).

See infra discussion in 6. Instigating Crimes through Hate Speech.

Šešelj Appeal Judgment, § 161.

Ibid ., § 163.

Ibid. , §§ 162, 164. See also ibid. , § 154.

The Appeals Chamber had also not defined it in the Nahimana et al. Appeal Judgment. See Nahimana et al. Appeal Judgment, § 987.

See Wilson and Gillett, supra note 6, at 144. See also Judgment, Bikindi (ICTR-01-72-T), Trial Chamber III, 2 December 2008 ( Bikindi Trial Judgment), § 394 (arguing that the same facts that could lead a trial chamber to find the existence of a widespread or systematic attack against a civilian population may also support a finding of other underlying acts of persecution as both have to meet the ‘discriminatory grounds’ threshold). Cf. Nahimana et al. Appeal Judgment, §§ 987–988; Art. 7(1)(h) ICCSt. Under the ICCSt., persecution as a crime against humanity must be directed against ‘any identifiable group or collectivity on political, racial, national, ethnic, cultural, religious, gender’ or ‘other grounds that are universally recognized as impermissible under international law.’ Unlike the ad hoc tribunals’ jurisprudence, however, this is generally understood to mean that persecution must occur in connection with another crime listed in Art. 7 of the ICC Statute. See Wilson and Gillett, supra note 6, at 119.

Wilson and Gillett, supra note 6, at 144.

Ibid ., at 138.

Timmermann, supra note 6, at 115–116.

Šešelj Appeal Judgment, § 163.

Ibid. (emphasis added).

Gordon, at 313. Discrimination has been described as ‘any distinction, exclusion or restriction made on the basis of race, colour, descent, national or ethnic origin, nationality, gender, sexual orientation, language, religion, political or other opinion, age, economic position, property, marital status, disability, or any other status that has the effect or purpose of impairing or nullifying the recognition, enjoyment or exercise, on an equal footing, of all human rights and fundamental freedoms in the political, economic, social, cultural, civil or any other field of public life.’ Report of the Special Rapporteur on the promotion and protection of the right to freedom of opinion and expression (SR Report), UN Doc. A/67/357, 7 September 2012, § 44 (d). See also ECRI Recommendation No. 15, § 7 (h).

See SR Report, § 44 (c); ECRI Recommendation No. 15, § 7 (q). See also Rabat Plan of Action on the prohibition of advocacy of national, racial or religious hatred that constitutes incitement to discrimination, hostility or violence, Conclusions and recommendations emanating from the four regional expert workshops organised by OHCHR, in 2011, and adopted by experts in Rabat, Morocco on 5 October 2012 (Rabat Plan of Action), UN Doc. A/HRC/22/17 Add. 4, 11 January 2013, § 22 (likelihood, including imminence, implies that there should be some causation which is ‘rather direct’).

See CERD Recommendation No. 35, § 16; 16/18 Combating intolerance, negative stereotyping and stigmatization of, and discrimination, incitement to violence and violence against, persons based on religion or belief (HRC 16/18 Resolution), Resolution adopted by the Human Rights Council, A/HRC/RES/16/18, 12 April 2011 § 5(e)–(f).

See SR Report, § 44(b). See also ECRI Recommendation No. 15, § 7 (a).

See SR Report, § 44(f). See also ECRI Recommendation No. 15, § 7 (ff).

See Reservations to UN treaties available online at https://treaties.un.org/Pages/ViewDetails.aspx?src=TREATY&mtdsg_no=IV-4&chapter=4&clang=_en (visited 17 March 2020); https://treaties.un.org/Pages/ViewDetails.aspx?src=IND&mtdsg_no=IV-2&chapter=4&clang=_en (visited 17 March 2020 ) . There are no states objecting to the reservations made by the reserving states. See also CERD Recommendation No. 35, § 12 (referring to General Comment No. 34 , Article 19: Freedoms of opinion and expression , UN Doc. CCPR/C/GC/34, 12 September 2011, §§ 22–35, 33–35)-13. Cf. CERD had earlier taken a strict interpretation on the mental element required to support criminal liability for incitement. Positive measures designed to eradicate all incitement to, or acts of, racial discrimination: implementation of the International Convention on the Elimination of All Forms of Racial Discrimination, Article 4 , UN Doc. A/CONF.119/10, 1986, §§ 83, 96, 235.

See e.g. HRC 16/18 Resolution, §5(f) (adopted without a vote). See also Rabat Plan of Action, at 4–7 (Recommendations).

Šešelj Appeal Judgment, § 164 and note 580 (referring to ibid. , §§149–150). See also ibid. , § 154

Ibid. , § 164.

Ibid. , § 163. This is a term borrowed from the Streicher Finding and Ruggiu Trial Judgment. See Streicher Finding, 302; Judgment and Sentence, Ruggiu (ICTR-97-32-I), Trial Chamber 1, 1 June 2000 ( Ruggiu Trial Judgment), § 19.

Šešelj Appeal Judgment, § 154.

See supra text before note 57.

Cf. Nahimana et al. Appeal Judgment, § 988; Nahimana et al. Appeal Judgment, Partly Dissenting Opinion of Judge Meron, §§ 3–4, 12–13.

See Timmermann, supra note 7, at 144 and note 93 (commenting on Nahimana et al. Trial Judgment, § 1073 and the Šešelj Trial Judgment); See also Wilson and Gillett, supra note 6, at 144.

Ibid. , § 159

See Nahimana et al. Appeal Judgment, §§ 986–987.

See Timmermann, supra note 6, at 116–118.

Wilson and Gillett, supra note 6, § 140 and note 157. The opposite view is also prevalent. Some states consider that the right to dignity should be respected and protected in and of itself. See e.g. Constitution of South Africa, Art. 10 available online at https://www.gov.za/documents/constitution/chapter-2-bill-rights#10 (visited 17 March 2020); Basic Law of Germany, Art. 1(1) available online at https://www.bundesregierung.de/breg-en/chancellor/basic-law-470510 (visited 17 March 2020). See also generally D. Kretzmer and E. Klein, The Concept of Human Dignity in Human Rights Discourse (Brill, 2002).

See also Šešelj Appeal Judgment, §§ 165–166.

See e.g. General Comment No. 34 , Article 19: Freedoms of opinion and expression (CCPR Comment No. 34), UN Doc. CCPR/C/GC/34, 12 September 2011, § 2; HRC 16/18 Resolution, § 1.

Judgment, Kordić and Čerkež (IT-95-14/2-A), Appeals Chamber, 17 December 2004 ( Kordić & Čerkež Appeal Judgment), § 111.

Ibid. , § 111.

Judgment, Šainović, Pavković, Lazarević, Lukić (IT-05-87-A), Appeals Chamber, 23 January 2014, § 579.

Šešelj Appeal Judgment, § 164.

See supra text before notes 41 and 42.

Kordić & Čerkež Appeal Judgment § 27. See Šešelj Trial Judgment, §§ 294-296; Šešelj Appeal Judgment, § 124.

See infra discussions at 6.B. Actus reus of Instigation: Direct Causal Connection – 6.E. Mens rea of Instigation

Šešelj Appeal Judgment, § 150, 154–155.

Ibid. , § 147, referring to Šešelj Trial Judgment, § 341. See also Dissenting Opinion, § 12.

Šešelj Appeal Judgment, §§ 150, 154.

Ibid. , § 154.

Šešelj Trial Judgment, § 336.

Ibid. , § 337.

Ibid. , §§ 335, 338.

Ibid. , §§ 338–339.

Ibid. , § 343.

Ibid. , § 350.

Šešelj Appeal Judgment, § 132. The Appeals Chamber however found that Šešelj had the mens rea for persecution including discriminatory intent. Ibid. , §§ 150, 155 The Appeals Chamber reversed the Trial Chamber’s finding that there was no widespread or systematic attack against the civilian population in Croatia and Bosnia and Herzegovina. Ibid. , §§ 71, 76–78. Moreover, given the content of Šešelj' s speech and the contemporaneous events in Croatia and Bosnia and Herzegovina, the Appeals Chamber found that the perpetrators were also aware that their acts formed part of the attack. Ibid. , § 150.

Ibid. See infra discussion on 6.C . Causal Connection: Temporal .

Šešelj Trial Judgment, §§ 324, 327. See also ibid. , §§ 321–322, 325–327.

Ibid. , § 325.

Ibid. , § 328. The Trial Chamber failed to consider the pejorative connotation of these words in the Bosnian Serb language.

Ibid. , § 328.

Ibid. See infra discussion at 6.C. Causal Connection: Temporal .

Šešelj Appeal Judgment, § 130.

Ibid. , § 131.

Ibid. , § 132.

Ibid. See infra discussion at 6.C. Causal Connection: Temporal.

Šešelj Trial Judgment, § 306. The word ‘ Ustashas ’, from the name of a nationalist Croat organization active in the 1930s and 1940s, is linked to fascist ideology and practices, including persecution of Serbs, Jews, Romani and other minorities, as well as to brutal violence.

Ibid. , §§ 309–310.

Ibid. , §§ 310, 318.

Ibid. , §§ 304–318. See infra discussion at 6.D. Actus reus of Instigation: Content and Purpose of the Speech.

Šešelj Appeal Judgment, § 136.

Šešelj Appeal Judgment, § 136 . See infra discussion at 6.D. Actus reus of Instigation: Content and Purpose of the Speech.

Ibid. § 137. See infra discussion at 6.D. Actus reus of Instigation: Content and Purpose of the Speech.

Šešelj Appeal Judgment, § 147.

See supra text before notes 19–20.

Ibid. , §§ 147, 149, 154.

Some ICTR and ICTY trial chambers have indeed held that for instigation, the causal connection must have ‘directly and substantially contributed’ to the subsequent commission of crimes or must have been a ‘clear contributing factor’. Judgment and Sentence, Ndindabahizi (ICTR-2001-71-I), Trial Chamber 1, 15 July 2004, § 456; Judgment , Kvočka, Kos, Radić, Žigić, Prcać (IT-98-30/1-T), Trial Chamber, 2 November 2001, § 252. Cf. Decision Pursuant to Art. 61(7)(a) and (b) of the Rome Statute on the Charges of the Prosecutor Against Bosco Ntaganda, Ntaganda (ICC-01/04-02/06-309 09-06-2014 1/98 EC PT), Pre-Trial Chamber II, 9 June 2014, § 153 (holding that for inducement, there has to be, inter alia, ‘a direct effect on the commission or attempted commission of the crime’); Judgment pursuant to Art. 74 of the Statute, Bemba Gombo, Kilolo Musamba, Mangenda Kabongo Babala Wandu, Arido (ICC-01/05-01/13-1989-Red 19-10-2016 1/458 NM T), Trial Chamber VII, 19 October 2016, § 73 (holding that the modes of liability of ‘soliciting and inducing’ in Art. 25(3)(b) of the ICCSt. ‘fall into the broader category of “instigating”’).

See e.g. Nahimana et al. Appeal Judgment, §§ 677, 701, 711; Ngirabatware Appeal Judgment, § 58; Nyiramasuhuko et al. Appeal Judgment, § 3338. It is interesting to note Judge Lattanzi’s Dissenting Opinion on instigation. She referred to the content of the speech, means of dissemination and impact, seemingly implying that these are legal requirements for instigation on the basis of speech. See Dissenting Opinion, §§ 95–123. Rather, impact is likely relevant only to the extent that it shows the substantial contribution of the instigator on the perpetrators of crimes. On the other hand, the means of dissemination of the speech are not directly related to a legal requirement for instigation: while it is relevant on an evidentiary level, the emphasis on dissemination of the speech (and its modalities) once again seems to lead back to the elements for the crime of direct and public incitement to commit genocide, not incitement as a mode of liability, and thus may lead to judicial confusion. In fact, it is the purpose of the speech and the context , discussed in Judge Lattanzi’s dissent together with the impact of the speech, which are more pertinent. See Dissenting Opinion, §§ 117–119.

See supra text before note 110.

See infra discussion at 6.C. Causal Connection: Temporal.

See supra discussion at 6.B. Actus reus of Instigation: Causal Connection.

Šešelj Appeal Judgment, § 132.

Ibid. Meanwhile, in its discussion on the Hrtkovci speech, the Appeals Chamber noted that the Croats left Hrtkovci in the four months after his speech. However, the temporal link was not the main argument of the prosecution in that case, so the Appeals Chamber did not comment on it. It also appears not to have been determinative of its finding of persecution. See ibid. , § 163.

Ibid. , § 132. See also supra text before note 111.

Nahimana et al. Appeal Judgment, § 515; Nahimana et al. Trial Judgment, §§ 449, 482.

Ibid. , § 515; Nahimana et al. Trial Judgment, §§ 411, 482.

Timmermann, supra note 7, at 149.

Streicher Finding, 301–304; Ruggiu Trial Judgment, §§ 24, 43, at 18–19.

See Wilson, supra note 25, at 166–168.

See Nahimana et al. Appeal Judgment, § 513.

See Kordić & Čerkež Appeal Judgment, § 27.

This case also highlights the dilemma prosecutors face with causation and why some scholars argue that at least in the context of public and direct incitement to genocide , ‘reasonable probability’ or ‘likelihood’ should replace ‘causation’ as a test (see S. Benesch, ‘The Ghost of Causation in International Speech Crime Cases’, in P. Dojčinović (ed.), Propaganda, War Crimes Trials and International Law (Routledge, 2012) 254–267, at 256, 262–264; H. van der Wilt, ‘Between Hate Speech and Mass Murder: How to Recognize Incitement to Genocide’, in H.G. van der Wilt, J. Vervliet et al. (eds), The Genocide Convention: The Legacy of 60 Years (Brill, 2012), at 41–50. However, in relation to instigation, causation remains a legal requirement so this suggestion is less relevant in this specific context.

Šešelj Trial Judgment, § 317. See also ibid. , § 14.

Ibid. , § 317 & note 371.

See e.g. Nahimana et al. Trial Judgment, §§ 394, 456-458, 473, 481, 652-656, 666; Bikindi Trial Judgment, §§ 247–252. See also Ruggiu Trial Judgment, §§ 44(iii)–(iv).

Šešelj Trial Judgment, § 307 (emphasis added).

See Dissenting Opinion, § 22. In an extraordinary line of procedural obstructions and contempt proceedings related to the main criminal case against him, Šešelj was inter alia convicted for revealing the identities of protected witnesses on several occasions, while persons closely collaborating with him were convicted for witness intimidation. See e.g. Judgment, Šešelj (IT-03-67-R77.2), Trial Chamber, 24 July 2009 (public redacted version); Judgment, Šešelj (IT-03-67-R77.2-A)‚ Appeals Chamber, 19 May 2010; Judgment, Šešelj (IT-03-67-R77.3), Trial Chamber, 31 October 2011 (public redacted version); Judgment, Šešelj (IT-03-67-R77.3-A)‚ Appeals Chamber, 28 November 2012; Judgment, Šešelj (IT-03-67-R77.4)‚ Trial Chamber II, 28 June 2012 (public redacted version); Judgment, Šešelj (IT-03-67-R77.4-A)‚ Appeals Chamber, 30 May 2013 (public redacted version). See also Judgment, Petković (IT-03-67-R77.1), Trial Chamber III, 11 September 2008 (public redacted version); Order Lifting Confidentiality of Order in Lieu of Indictment and Arrest Warrants, Jojić et al. (IT-03-67-R77.5), Trial Chamber 1, 1 December 2015; Order of Transfer to the International Residual Mechanism for Criminal Tribunals, Jojić et al. (IT-03-67-R77.5), President’s Office, 29 November 2017.

Cf. Judgment, Popović, Beara, Nikolić, Miletić, Pandurević (IT-05-88-A), Appeals Chamber, 30 January 2015, § 1741 (on aiding and abetting); Judgment, Nzabonimana (ICTR-98-44D-A), Appeals Chamber, 29 September 2014, § 489 (on aiding and abetting).

Kordić & Čerkež Appeal Judgment, § 112; Judgment, Blaškić (IT-95-14-A), Appeals Chamber, 29 July 2004, § 166.

See supra text before note 77.

See supra text before note 76.

It is of note that imminence has been identified as an aspect of incitement by the Appeals Chamber in the context of direct and public incitement to genocide. See supra text before note 24.

Email alerts

Citing articles via.

• Recommend to your Library

Affiliations

• Online ISSN 1478-1395
• Print ISSN 1478-1387
• Copyright © 2024 Oxford University Press
• About Oxford Academic
• Publish journals with us
• University press partners
• What we publish
• New features  
• Open access
• Institutional account management
• Rights and permissions
• Get help with access
• Accessibility
• Advertising
• Media enquiries
• Oxford University Press
• Oxford Languages
• University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

• Copyright © 2024 Oxford University Press
• Cookie settings
• Cookie policy
• Privacy policy
• Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

Tackling Hate Speech: In Conversation With Caitlin Ring Carlson

Hate speech is ubiquitous and can happen anywhere — in Charlottesville, Virginia, where young men in khakis shouted, “Jews will not replace us”; in Myanmar, where the military used Facebook to target the Muslim Rohingya, calling them dogs and worse; in Capetown, South Africa, where a pastor called on ISIS to rid South Africa of the “homosexual curse.”

Yet defining hate speech is an elusive task, and scholars haven’t come to a clear and shared definition of the concept. Broadly, it’s an expression that seeks to malign an individual for their immutable characteristics, says Caitlin Ring Carlson, an expert in communication and mass media and the author of “ Hate Speech ,” a recent addition to the MIT Press Essential Knowledge series . “Hate speech represents a structural phenomenon,” Carlson explains in the book’s introduction, “in which those in power use verbal assaults and offensive imagery to maintain their preferred position in the existing social order.” Although legal action and social stigmatization seek to minimize the spread and impact of hate speech, the issue persists both in person and online.

Carlson’s book doesn’t parse specific words and phrases. It instead investigates legal approaches and controversies around the world in order to provide suggestions for limiting its spread. These solutions may be applied from the highest point of order, by governments and media organizations, to the individual and personal level. In short, Carlson writes, “logging off is not the answer.” Widespread change, specifically on social media platforms, requires vigilance from users, regulators, and institutions.

Zoë Kopp-Weber : You open your book highlighting the ubiquity of hate speech, yet despite the problems it causes, the term is expansive and often contested. What about the term makes it so hard for scholars to agree upon a clear definition?

Caitlin Ring Carlson : What makes hate speech so difficult to define is the subjective nature of hate speech itself. Phrases, images, and terms that I may see as maligning an individual based on their fixed characteristics, such as race, gender identity, or sexual orientation, may not be seen the same way by others. Intent plays a role as well, and intent is difficult to determine. When slurs are used by members of the group they were originally meant to harm, it is rightly considered a reclamation or reappropriation of the term and, thus, not hate speech since the intent is not to malign someone.

There has not been an incident of genocide in recorded history that was not accompanied by discourse seeking to dehumanize the targeted groups.

ZKW : How do you demonstrate that hate speech is a driving force for issues like bias-motivated violence and genocide?

CRC : I don’t think it’s possible to empirically demonstrate that hate speech causes bias-motivated violence. However, a historical analysis of genocide and bias-motivated violence clearly illuminates the relationship between hate speech and these atrocities. There has not been an incident of genocide in recorded history that was not accompanied by discourse seeking to dehumanize the targeted groups. Thus, hate speech creates the ideological conditions for people to act out against members of another ethnic or religious group, for example.

ZKW : How has the U.S.’s position on freedom of expression informed legal responses to hate speech, and how does this compare to the way other countries approach the issue?

CRC : In the United States, we tend to place the right to free expression above other rights. We consider the harm caused by hate speech to be less costly to society than the harm associated with restrictions on our right to free expression, particularly as it relates to political dissent. Only when hate speech crosses the line and becomes a true threat or incitement to violence can it be punished. Interestingly, we have other categories of speech that are exempt from First Amendment protection, such as obscene speech or speech that is injurious to another’s reputation. Hate speech is just not one of those categories.

This approach is vastly different from most other Western democracies, which prohibit hate speech and punish it with fines or jail time. From Canada to the European Union, several countries have laws against expression that incites hatred based on a person’s race, gender, ethnicity, religion, etc. Citizens of these countries tend to place the right to human dignity over the right to free expression.

ZKW : Given its history, Germany’s stringent laws restricting hate speech are not surprising, even being the international leader in shaping standards for online communication and content. What of their efforts, if any, would be the greatest takeaway for countries like the U.S. in their individual regulation of hate speech?

CRC : While the German law NetzDG, which requires social media platforms to remove illegal hate speech quickly or risk substantial fines, is not perfect, there are several lessons we can take from this approach. First is transparency. Part of this law requires large social media companies to create and disseminate reports regarding which content and accounts were removed and why. In addition, this approach serves as a reminder that regulation can and perhaps should be used to motivate social media and other computer services to act not only in the best interest of their shareholders but also in the interest of the public.

ZKW : In addition to being hurtful and a foundation for greater, more physical threats, how else does hate speech create personal, even political barriers?

CRC : Several scholars, including Danielle Citron and Helen Norton, have argued that the proliferation of hate speech, particularly online, makes it difficult for those targeted to engage in the political process. For example, let’s say there’s a discussion happening on a neighborhood Facebook page about a new City Council ordinance to reduce police funding. It’s easy to imagine how, after posting her opinion, a Muslim woman might be met with a barrage of hate speech calling her names and encouraging her to “go back to her country.” To protect herself from this abuse, the woman leaves the discussion. A week later, when a spokesperson from the neighborhood is invited to speak at a City Council meeting, the Muslim woman’s perspective on the issue is not included or represented in the testimony because she was driven from the page by the vitriolic hate speech she encountered. In the future, the Muslim woman may be far less likely to engage in any online civic discourse for fear of similar attacks.

In terms of personal barriers, Mari Matsuda has, for decades, warned us about what she sees as the most significant potential harm caused by hate speech, which is that those targeted will come to believe in their own inferiority. If children are raised in a world where the public discourse tells them they’re subhuman because they are Black, transgender, or Jewish, they may come to believe that they are less worthy of dignity than other people.

ZKW : Historically, bias-motivated violence and political dissent have both flourished on college campuses. What complicates higher education institutions’ ability to address hate speech?

CRC : It is difficult for colleges and universities to address hate speech because of the tension between their dual goals of being places where new ideas are considered and places where people live and work. For centuries, students at universities have been asked to wrestle with concepts they disagree with in order to form their own opinions and, eventually, their broader worldview. Professors have been given academic freedom and tenure to explore alternative perspectives, test hypotheses, and speak out on critical public issues without interference from administrators. In so many ways, free expression is integral to higher education.

However, problems arise when that expression, whether from faculty or outside speakers, threatens the physical and emotional safety of students who in many instances are a captive audience that cannot simply “look away” when a professor or speaker uses an offensive slur or claims one race or gender is inferior to another. Therefore, colleges and universities must engage in the hard work of finding a balance between exposure to new ideas and creating a community where people feel safe and supported enough to engage with those ideas.

ZKW : Greg Lukianoff and Jonathan Haidt, the authors of “The Coddling of the American Mind,” argue that higher education has incorrectly taught students that they are fragile, emotional beings, creating a culture of extreme safety that leads to intimidation and violence. What do such claims miss about the existence of trigger warnings and safe spaces?

CRC : What’s missing from the argument in “The Coddling of the American Mind” is the students’ perspective, particularly students with historically marginalized identities. In the book, I include a great quote from Mary Blair, a Black woman who was a student at the University of Chicago. CBS News interviewed her and several of her fellow students. In response to another student’s comment about the real world not being a safe space, she said, “I can assure you, all people of color who have existed in a white space know that the real world is not a safe space.”

What’s missing from the argument in “The Coddling of the American Mind” is the students’ perspective, particularly students with historically marginalized identities.

In my experience, students are not “fragile, emotional beings” and instead are mindful, empathetic people who want to be able to engage with controversial ideas in a meaningful and productive way. Along those lines, there is a fundamental misunderstanding regarding the term “safe spaces.” These are not intellectual safe spaces, but rather an environment where everyone feels comfortable in expressing themselves and participating fully without fear of attack, ridicule, or denial of experience. No one is suggesting that students in the classroom avoid or ignore ideas they disagree with. Instead, these tools allow students to engage with these concepts in a respectful and constructive way.

Finally, content or trigger warnings are simply tools that some instructors use to let students know that a sensitive topic or issue is about to be discussed so that students are not caught off guard. Rather than avoiding certain topics, such as sexual assault or bias-motivated violence, altogether, content warnings allow professors to communicate with students about the nature of the upcoming material.

ZKW : Facebook has been a significant subject in conversations regarding the proliferation of hate speech through social media. What responsibility does social media legally have to address these issues; how could we see these responsibilities change in the coming years?

CRC : Legally, social media has no responsibility to address these issues. As private virtual spaces, social media platforms are free to create whatever community standards they want for their platforms. As users, we agree to these rules when we sign the terms of service that allow us to access the site.

From an ethical perspective, social media have an essential role to play in decreasing hate speech in public discourse. However, as publicly traded companies, social media organizations’ first responsibility is often to their shareholders. It seems unrealistic to think that they will take any action detrimental to their bottom line. If, as I suspect it does, hate speech and other offensive content leads to greater engagement on the platform, it is unlikely that these companies will act differently unless users or advertisers demand it or the government steps in to regulate it.

ZKW : As social media organizations seek to more aggressively remove hate speech from their platforms, what ways can content moderation be improved algorithmically and logistically?

CRC : The algorithms and artificial intelligence used by social media companies to remove hate speech from their platforms have improved a great deal in recent years. Natural language processing allows companies to identify and remove all instances of particular words. However, the algorithms still struggle to identify hate speech when the meaning of a comment or post depends on its context. For example, the phrase “go home b*tch” would not be considered hate speech if posted as a comment on a news story about a football team beating the visiting team. However, it would be considered hate speech if posted to a story about Representative Ilhan Omar’s most recent bill in Congress. In terms of the logistics of content moderation, moving the process in-house, rather than outsourcing it to firms that offer low wages and problematic working conditions would improve human content moderators’ efficacy. Dedicating more resources to identifying and removing hate speech (along with disinformation, harassing speech, and nonconsensual pornography) should be a top priority for social media organizations.

ZKW : What don’t we understand about the phenomenon of hate speech? What should future research focus on?

CRC : Right now, we don’t fully understand the various impacts, big and small, that hate speech has on individuals and on society as a whole. I would love to see future research that unpacks the psychological, emotional, and physiological impacts hate speech has on individuals. For example, how are people influenced by hate speech that is about a group they are a member of compared to hate speech directed at them personally? From a structural perspective, we should investigate the role hate speech plays in establishing and maintaining racial and other forms of discrimination and inequality. Future research should also examine the relationship between hate speech and extremism, particularly online. We need to know how, specifically, hateful rhetoric translates into offline violence and whether there are interventions that have been or could be successful.

Zoë Kopp-Weber is a publicist at the MIT Press.

Caitlin Ring Carlson is Associate Professor in the Communication Department at Seattle University and the author of “ Hate Speech .”

What you need to know about hate speech

What is hate speech.

The UN Strategy and Plan of Action on Hate Speech refers to the working definition as: "any kind of communication in speech, writing or behaviour that attacks or uses pejorative or discriminatory language with reference to a person or a group on the basis of who they are, in other words, based on their religion, ethnicity, nationality, race, colour, descent, gender or other identity factor."

Forms of hate speech can include scapegoating, stereotyping, stigmatization and the use of derogatory language. It is often employed in the promulgation of conspiracy theories, disinformation and denial and distortion of historical events such as genocide. States are required to prohibit most severe forms of hate speech that constitute incitement to violence, hostility or discrimination, or incitement to genocide or other violations of international law in line with Article 20 of the International Covenant on Civil and Political Rights   

Some other forms of hate speech, such as individual threats, may be restricted too. But international law protects forms of speech that may be offensive and raise concerns over prejudice and intolerance, which can be the object of other measures to prevent harm. The Rabat Plan of Action provides guidance for defining restrictions on freedom of expression and incitement to hatred and for the application of Article 20. 

What are the effects of hate speech?

Hate speech not only causes harm at the personal level and can incite violence, it is an attack on inclusion, diversity and human rights. It undermines social cohesion and erodes shared values, setting back peace, stability, sustainable development and the fulfillment of human rights for all. 

How does UNESCO work to counter hate speech?

UNESCO cooperates with many diverse partners to address and counter hate speech within the framework of the United Nations Strategy and Plan of Action on Hate Speech , which was launched by the UN Secretary-General in 2019. At all times, UNESCO emphasizes the importance of a human rights-based approach to tackling hate speech, including through safeguarding freedom of expression. 

Education is a powerful tool to combat disinformation, misinformation and hate speech. UNESCO supports countries to support regulators and judicial operators, develop education responses, and policy and legislation that promote, protect and uphold international human rights.  

How does UNESCO tackle online hate speech?

Hate speech spreads with unprecedented speed and reach through digital tools, notably social media platforms. UNESCO works to address online hate speech by equipping learners with digital citizenship skills, so people of all ages learn to navigate the internet safely and responsibly.  

Media and Information Literacy can strengthen the resilience of learners to hate speech and build their capacity to recognize and counter mis- and disinformation, violent extremist narratives and conspiracy theories.  

UNESCO advocates for increased transparency and accountability for digital platforms to counter online disinformation and speech that incite hatred and discrimination. This includes calling on social media companies to report on hate speech, how their algorithms may affect its spread, and the policies they apply to counter it. UNESCO has issued a set of 26 high-level principles to increase transparency among internet platform companies. 

UNESCO also supports the enabling of a free, diverse, pluralistic media sector, including professional media self-regulation, as well as disseminating good practices against hate speech and training judicial actors and law enforcement on international standards on freedom of expression.   

UNESCO guide for policy-makers to counter hate speech

UNESCO and the United Nations Office on Genocide Prevention and the Responsibility to Protect (UNOSAPG)  have jointly developed  the first  guide for policy-makers and teachers to  explore educational responses to this phenomenon and give practical recommendations for strengthening education systems. The guide is part of the implementation of the UN Strategy and Plan of Action on Hate Speech.

Related items

• Information and communication
• Civic education
• Antisemitism
• Nonviolence
• See more add

Essay on Hate Speech

A well-established democratic society should allow and protect human rights, and freedom of speech is amongst these rights. Individuals should be allowed the right to express themselves without censorship, restrain, and without legal consequences. Nevertheless, when the right to express oneself is substituted for threats, abuse, and prejudice, this right and freedom should be denied. More significantly, when a person’s speech or writing threatens the nation’s unity, he or she should be held accountable. Hate speech undermines the very concept of democracy, and therefore it should not be protected by the First Amendment.

It is important to note that freedom of speech is a concept protected by the Constitution, and it is viewed as a critical element of the United States democracy. On the other hand, hate speech is not directly regulated because of the extensive and robust right to free speech established in the Constitution. However, the U.S. Supreme Court has, over the years, defined freedom of speech to include certain aspects that many viewed as underlying elements of hate speech. For instance, in the 1971  Cohen v. California case, the  Supreme Court ruled that freedom of speech may include the right to include certain offensive words and phrases to convey a message.

But to critically understand the general concept surrounding the freedom of speech and why the Constitution should not protect hate speech, it is vital to first look to the past. The First Amendment was adopted in 1791 and is meant to protect people’s religions, freedom of speech, and, more importantly, protect the press and people’s right to protest (Coates, 223). In 1787, a group of legislators like James Madison and Alexander Hamilton met and drafted a new U.S. Constitution. Anti-federalists, including the first governor of Virginia, Patrick Henry argued that it (the Constitution) gave the federal government too much control at the detriment of the states; they rejected this very Constitution. They also felt that the Constitution lacked protection for people’s individual rights (Coates, 223). James Madison drafted the Bills of Rights, which was presented to Congress in 1789. The Bill contained ten amendments to the United States. The First Amendment guaranteed free speech, and citizens could express themselves without thinking about punishment for the first time in the history of the United States; speech was an essential aspect of freedom of expression. People were allowed the right to say whatever they liked, whenever they liked, and in the way they saw fit.

However, as time went by, this freedom was excused for the right to abuse others; in fact, many believed that as long as the First Amendment remained unchanged, they were allowed to attack and use pejorative and discriminatory language with reference to other people and groups. Yes, the First Amendment guarantees this. But as long as the ideas and information expressed in this case threaten the fabric that unites society, this freedom should perhaps be tamed. This idea was established in the 1919  Schenck v. the United States  U.S. Supreme Court Case. In  Schenck v. the United States , the U.S. Supreme Court ruled that if the words spoken or printed posed a “clear and present danger,” the freedom of expression rights guaranteed in the Constitution could be limited. In other words, this meant that people could be held accountable for what they said, either publicly or on social media. An excellent example that fits this case is encouraging other people to commit hate crimes, a prevalent instance with white supremacists.

In addition to harming and presenting danger, hate speech limits the democratic process; it threatens the communitarian view of a healthy democracy and societal pursuit of common purpose. In other words, hate speech drives people to pursue different adverse objectives at the expense of reason and democracy. It limits the very concept of freedom, as humans are only free when their actions are regulated by reason. This idea is prevalent in Immanuel Kant’s Critique of Practical Reason (1788). Immanuel Kant (1724-1804) believed that a society governed by practical reason decides the best way to act (Hill 67). In most cases, people with practical reasons guide their conduct and act according to the conception of the law.

Hate speech, even though it is a concept protected by the Constitution, practicality defames the practicality and concept of freedom. Practical reason, in this case, should guide individuals to understand the difference between expression and abuse (Neblo 925). Hate speech may serve to undermine the political status of some people. Equal political standing is a necessary condition of democratic policy-making, and while this may be true, the legitimacy of this process may be undermined by hate speech. This concept is motivated by the assumption that the Legitimacy Claim underscores both that democratic legitimacy presupposes a degree of relationship equality in the political sphere between people and that hateful acts of speech will erode those relationships in some circumstances.

The First Amendment is of critical importance in the democratic United States. Here, racism, hate speech, and offensive speeches are protected and viewed as individual rights. The First Amendment was established to encourage the free exchange of ideas and to create a form of redress of people against their government. Likewise, this amendment was structured to protect unpopular forms of speech. However, when these speeches threaten the nature and codes that govern society, they should be deterred. Protecting hate speech in the excuse of providing citizens with freedom undermines the very concept of freedom. It is ironic when the Constitution guarantees freedom and similarly limits it in the pretense of allowing free speech. More importantly, if an idea or a piece of writings threatens other people’s survival, then the idea or the writing should not be protected as free speech. American unity should always come first as dangers tend to occur when a nation forgets that it is one country. Therefore, hate speech should not be protected. More so, there ought to be deterrent measures that discourage individuals from emitting utterances that threaten American unity.

Works Cited

Coates, IV, John C. “Corporate Speech and The First Amendment: History, Data, And Implications.”  SSRN Electronic Journal , 2015, p. 223.

Cohen v. California , 403 U.S. 15, 91 S. Ct. 1780, 29 L. Ed. 2d 284 (1971).

Hill, Thomas E. “Dignity and Practical Reason in Kant’s Moral Theory.” 1992, p. 67.

Neblo, Michael A. “Impassioned Democracy: The Roles of Emotion in Deliberative Theory.”  American Political Science Review , vol. 114, no. 3, 2020, pp. 923-927.

Schenck v. United States , 249 U.S. 47, 39 S. Ct. 247, 63 L. Ed. 470 (1919).

Cite this page

Similar essay samples.

• Journal article.
• Essay on the Native American Flute
• Essay on Masculinity
• Essay on International Business
• Project Management Issues in the Construction Industry
• Essay on Prenatal Screening Tests

Home — Essay Samples — Social Issues — Freedom of Speech — Social Media and Freedom of Speech: Combating Misinformation and Hate Speech

Social Media and Freedom of Speech: Combating Misinformation and Hate Speech

• Categories: Freedom of Speech

About this sample

Words: 833 |

Published: Jan 29, 2024

Words: 833 | Pages: 2 | 5 min read

Table of contents

The importance of freedom of speech, the challenges of freedom of speech in social media, balancing freedom of speech with responsibility, implications of limiting freedom of speech, case studies of social media platforms and their approach to freedom of speech, references:.

• Balkin, J. M., & Zittrain, J. (2010). A grand compromise to preserve a free and open Internet. Harvard Law Review, 124(6), 1333-1396.
• Cohen-Almagor, R. (2019). Freedom of expression, hate speech, and disinformation: The challenges of our time. Journal of Social and Political Psychology, 7(1), 364-380.
• Forelle, M., Helderman, R., & Timberg, C. (2020). Zuckerberg defends Facebook as bastion of free speech against lawmakers' skepticism. Retrieved from https://www.washingtonpost.com/technology/2020/07/29/zuckerberg-antitrust-testimony/
• Royakkers, L., & Timmer, J. (2020). Combating misinformation on social media: A review of the effectiveness of current solutions and recommendations for the future. Journal of Contingencies and Crisis Management, 28(4), 332-344.
• Tandoc Jr, E. C., Zheng, N., & Ling, R. (2019). Shaping and being shaped: The reciprocal relationship between media and society. Journalism & Mass Communication Quarterly, 96(2), 341-358.

Cite this Essay

Let us write you an essay from scratch

• 450+ experts on 30 subjects ready to help
• Custom essay delivered in as few as 3 hours

Get high-quality help

Verified writer

• Expert in: Social Issues

+ 120 experts online

By clicking “Check Writers’ Offers”, you agree to our terms of service and privacy policy . We’ll occasionally send you promo and account related email

No need to pay just yet!

Related Essays

2 pages / 1064 words

2 pages / 953 words

2 pages / 849 words

2 pages / 1011 words

Remember! This is just a sample.

You can get your custom paper by one of our expert writers.

121 writers online

Still can’t find what you need?

Browse our vast selection of original essay samples, each expertly formatted and styled

Related Essays on Freedom of Speech

Freedom of speech is a foundational pillar of democratic societies and a fundamental human right. It serves as the bedrock of open and inclusive societies, allowing individuals to express their thoughts, opinions, and ideas [...]

The Turner v Driver case has been an important and highly controversial legal battle that has sparked heated debates regarding freedom of speech, privacy rights, and the responsibilities of social media platforms. This case [...]

Freedom of speech is a fundamental right that has been the subject of much debate and controversy in recent years. From historical origins to modern-day implications, the concept of freedom of speech has far-reaching [...]

Eric Foner, a prominent historian and author, explores the concept of American freedom in his book, "The Story of American Freedom." In this work, Foner delves into the complexities of freedom in America, examining its evolution [...]

In music, censorship can be defined as ‘the suppression or prohibition’, of any parts of music ‘that are considered obscene or politically unacceptable’ (Oxford University Press, 2019). This includes the editing of musical and [...]

The First Amendment of the United States Constitution gives us the right to freedom of speech and freedom of press. In my opinion, these rights have been abused in such a manner that the freedom of press has been pushed beyond [...]

Related Topics

By clicking “Send”, you agree to our Terms of service and Privacy statement . We will occasionally send you account related emails.

Where do you want us to send this sample?

By clicking “Continue”, you agree to our terms of service and privacy policy.

Be careful. This essay is not unique

This essay was donated by a student and is likely to have been used and submitted before

Download this Sample

Free samples may contain mistakes and not unique parts

Sorry, we could not paraphrase this essay. Our professional writers can rewrite it and get you a unique paper.

Please check your inbox.

We can write you a custom essay that will follow your exact instructions and meet the deadlines. Let's fix your grades together!

Get Your Personalized Essay in 3 Hours or Less!

We use cookies to personalyze your web-site experience. By continuing we’ll assume you board with our cookie policy .

• Instructions Followed To The Letter
• Deadlines Met At Every Stage
• Unique And Plagiarism Free

Search the United Nations

• What is hate speech?
• Hate speech vs freedom of speech
• Hate speech and real harm
• Why tackle hate speech?
• A pandemic of hate
• The many challenges of tracking hate
• Targets of hate
• The preventive role of education
• UN Strategy and Plan of Action
• The role of the United Nations
• International human rights law
• Further UN initiatives
• Teach and learn
• Test yourself
• International Day
• Key human rights instruments
• Thematic publications
• Fact sheets [PDF]
• UN offices, bodies, agencies, programmes

Understanding hate speech

Hate speech versus freedom of speech

The need to preserve freedom of expression from censorship by States or private corporations’ is often invoked to counter efforts to regulate hateful expression, in particular online.

Freedom of opinion and expression are, indeed, cornerstones of human rights and pillars of free and democratic societies. These freedoms support other fundamental rights, such as to peaceful assembly, to participate in public affairs, and to freedom of religion. It is undeniable that digital media, including social media, have bolstered the right to seek, receive and impart information and ideas. Therefore, legislative efforts to regulate free expression unsurprisingly raise concerns that attempts to curb hate speech may silence dissent and opposition.

To counter hate speech, the United Nations supports more positive speech and upholds respect for freedom of expression as the norm. Therefore, any restrictions must be an exception and seek to prevent harm and ensure equality or the public participation of all. Alongside the relevant international human rights law provisions, the UN Rabat Plan of Action provides key guidance to States on the difference between freedom of expression and “incitement” (to discrimination, hostility and violence), which is prohibited under criminal law. Determining when the potential of harm is high enough to justify prohibiting speech is still the subject of much debate. But States can also use alternative tools – such as education and promoting counter-messages – to address the whole spectrum of hateful expression, both on and offline.

“Addressing hate speech does not mean limiting or prohibiting freedom of speech. It means keeping hate speech from escalating into something more dangerous, particularly incitement to discrimination, hostility and violence, which is prohibited under international law.”

— United Nations Secretary-General António Guterres, May 2019

Services on Demand

Related links, potchefstroom electronic law journal (pelj), on-line version  issn 1727-3781, per vol.23 n.1 potchefstroom  2020, http://dx.doi.org/10.17159/1727-3781/2020/v23i0a7520 .

Demystifying Hate Speech under the PEPUDA

J Geldenhuys * ; M Kelly-Louw **

University of South Africa. Email: [email protected] ; [email protected]

The factual matrix that is considered in each hate speech case differs from that in the next. However, certain factors always remain key in the process of balancing the different constitutional rights at play: who the victim is, who the perpetrator is and the nature of the expression. Additional factors to be considered in deciding whether an expression constitutes hate speech include: historical associations; who the utterer is as against the victim(s); the audience that is addressed and where the utterance is made; and the prevailing social conditions. How South African courts and the South African Human Rights Commission factor in these specific issues in assessing whether an utterance constitutes hate speech is examined in this contribution. Applicable international law principles and comparable foreign law reveal certain areas of the South African hate speech protection requiring refinement.

Keywords : Equality Court; freedom of speech; hate speech; factors to assess hate speech; South African Human Rights Commission.

1 Introduction

In 2012 the Legal Resource Centre of South Africa (the LRC) cited as reason for taking guidance from other jurisdictions in interpreting and applying the hate speech protection in the Promotion of Equality and Prevention of Unfair Discrimination Act 4 of 2000 (the PEPUDA) that South Africa's jurisprudence was still "in its infancy". 1 Little appears to have changed since then. In March 2019, almost two decades after the enactment of the PEPUDA, the South African Human Rights Commission (the SAHRC) in an official report 2 commented that divergent views exist in the various Equality Courts as to what would constitute hate speech. The SAHRC proceeded to explain that no consensus exists on how section 10 of the PEPUDA ought to be interpreted 3 and that this area of the law is still in a developmental phase. 4 The SAHRC emphasised that even the legislature is still grappling with the issue and that the Department of Justice has repeatedly advised that it is in the process of reviewing and amending the PEPUDA in order to provide clarity. 5 The purpose of this contribution is to peel back some of the layers of complexity surrounding the application of the hate speech protection in the PEPUDA in its current state.

The surrounding circumstances or factual matrix in which racial utterances are made is important in the assessment of whether an utterance is judiciable and sanctionable as hate speech. But what significance is attached to the different factors? And how does the hate speech protection offered in terms of the PEPUDA as applied by the SAHRC and the Equality Court measure up to international standards? This contribution, which is made against the backdrop of decided cases and recent factual scenarios that have been reported on in the media, 6 in particular in newspapers and via social media, answers these questions.

2 The factors that are considered

2.1 Introduction

The contextual factors to be considered in a hate speech case depend on the facts of a particular matter. Consequently, it is impossible to provide a closed list of factors that must be assessed in each hate speech case.

The PEPUDA was enacted to ensure that South Africa complies with its international obligations. 7 The PEPUDA must be interpreted in accordance with the Constitution, in a manner that abides by international law 8 and where appropriate having regard to foreign law. 9 Moreover, the legislation must be interpreted keeping in mind the context of the particular case and the purpose of the PEPUDA. 10 Section 233 of the Constitution requires that the court "must prefer any reasonable interpretation that is consistent with international law" over a possible construct of the provision that contradicts international law principles. 11 The interpretation afforded to the PEPUDA's hate speech protection, therefore, should not be out of kilter with the construct of comparable hate speech provisions in foreign jurisdictions, particularly those having shared international law obligations. 12 The hate speech protection in section 10(1) of the PEPUDA must be interpreted in a manner that balances the right to dignity, to equality and to freedom of expression. 13 These rights contained in the Bill of Rights 14 are implicated in the application of the hate speech provision in the PEPUDA. Therefore, it is apt to consider foreign law when interpreting section 10(1) of the PEPUDA in relation to its scope of application and in determining how it should be applied. 15

In the Memorandum 2012 the LRC had examined and compared hate speech provisions and jurisprudence in ten different jurisdictions. 16 The inquiry therefore covered the principles of international law and foreign law. 17 However, the LRC noted that a direct comparison between South Africa's hate speech provision and the hate speech protection in any other jurisdiction is problematic. 18 First, different ideologies, and historical and social circumstances inform the adoption of different hate speech provisions. 19 Secondly, "hate speech" is not a universally defined concept. 20

This means that countries regulate hate speech in different ways: some recognise hate speech as a crime, 21 while others, including South Africa, consider it to be a civil offence. 22 In some jurisdictions the intention of the utterer is considered key to determining liability, 23 whereas in South Africa it is considered to be of no significance. 24 In most jurisdictions the hate speech prohibitions require proof that the utterance is likely to cause harm. However, what the harm is that must ensue differs from jurisdiction to jurisdiction. Some jurisdictions require proof of incitement to violence, or of hatred, the causing of insult or hurt and/or of humiliation. 25

The LRC noted that much as in South Africa, the European Court of Human Rights (the ECtHR) does not have a framework for how the contextual factors must be weighed up in the assessment of hate speech. 26 Instead, a "common-sense" approach is followed. 27 However, three contextual facts are usually considered to be incremental in the weighing up of the right to freedom of expression and the rights to equality and dignity: the nature of the utterance; who the victims of the hate speech are; 28 and the identity of the alleged perpetrator. 29 The LRC further identifies factors that play a pivotal role in the assessment of whether an utterance constitutes hate speech, but stresses that the list is not exhaustive. The factors are: historical associations and in relation thereto who the utterer is as against the victims; where and to whom the utterance is made; and the socio-political circumstances at the time of making of the utterance. 30 The Committee on the Elimination of Racial Discrimination (hereafter the CERD) General Recommendation 35: Combating Racist Hate Speech (2013) 31 (hereafter the Recommendation), which the SAHRC is required to uphold, 32 indicates similar contextual factors that must be considered in deciding whether or not an utterance is sanctionable as hate speech. 33

2.2 The nature of the utterance

Section 10(1) of the PEPUDA, which regulated hate speech, read:

Subject to the proviso in section 12, 34 no person may publish, propagate, advocate or communicate words based on one or more of the prohibited grounds, 35 against any person, that could reasonably be construed to demonstrate a clear intention to -

(a) be hurtful;

(b) be harmful or to incite harm;

(c) promote or propagate hatred.

However, on 29 November 2019 the Supreme Court of Appeal in the appeal to the Qwelane Equality Court case 36 handed down a pivotal judgment declaring section 10(1) of the PEPUDA unconstitutional and invalid. 37 In the court's view the provision as it stood contradicted the right to freedom of speech as envisaged in section 16 of the Constitution. 38

The result of the judgment is that Parliament must within eighteen months of the judgment amend the wording of section 10(1) of the PEPUDA in order to remedy the defect. 39 In the interim the Supreme Court of Appeal ordered that courts must apply the following amended wording of section 10(1):

No person may advocate hatred that is based on race, ethnicity, gender, religion or sexual orientation and that constitutes incitement to cause harm. 40

In order to succeed in hate speech claims the victims must on a balance of probabilities prove that the offending statement qualifies as hate speech. 41

South Africa's hate speech protection, at least as it was applied before Qwelane SCA, was peculiar in two ways. First, there was no need for the victim to prove that actual or likely harm might follow from the making of the utterance. Secondly, the requirement that a "reasonable person should understand" the utterance to demonstrate a "clear intention" to harm was not used in other jurisdictions. 42

The test that was used to assess whether an utterance could be interpreted as conveying a clear intention to do harm as envisaged in section 10(1) of the PEPUDA, in its original form, was an objective test, and the speaker's subjective intention was irrelevant. 43 The test was whether the utterance "could be reasonably construed to demonstrate a clear intention to 'incite harm'". 44 The standard of the reasonable person as it was applied was: "whether a reasonable person could conclude (not inevitably should conclude) that the words mean the author had a clear intention to bring about the prohibited consequences." 45 This test has not in our view been affected by the amendment pursuant to the Qwelane SCA judgment. To succeed in a claim under the common law actio iniuriarum, it must be proven that the conduct was objectively and subjectively demeaning. 46 The reason for requiring objective insult is that not doing so would result in the court's being inundated with referrals from "hypersensitive persons". 47 This is also the reason why the test for hate speech must even under the amended interim definition in our view be an objective one.

The impact that the utterance has on the person/s to whom it is addressed, and the fact that he or she felt offended remains relevant. However, the victim cannot also be the reasonable audience in the assessment of whether the utterance has the effect of advocating hatred and inciting harm to the victim group. 48 A clear distinction should be drawn between the assessments to determine whether an utterance is hurtful or harmful, and whether it is likely to advocate (incite) hatred. 49

In the Mandela Foundation case 50 it was held that displaying the old South African land flag constituted hate speech. 51 The lobby group AfriForum had argued that different people have different intentions in displaying the flag, and that the intention is not always to cause harm. 52 The court dismissed this argument on the basis that the "clear intention" in section 10(1) of the PEPUDA does not depend on the subjective intention of the person who is displaying the flag. Whether there is malice in the expression that is under scrutiny, therefore, in the view of the Equality Court, cannot be determined on a case-by-case basis. 53

2.2.1 Words or expressions of hate speech

Section 10(1) of the PEPUDA, in its original format, referred to hate speech in the form of "words". An utterance which can potentially qualify as hate speech can take on several forms: it can consist of a verbal expression; 54 be reduced to writing; 55 or be contained in the lyrics of songs. 56 Hate speech can even consist of symbols or take the form of other types of expression, such as waiving, burning or saluting a flag. 57 This is illustrated by the recent hate speech case concerning the old South African flag. Mojapelo DJP in the Mandela Foundation case considered section 10(1) in its original format and held that "words" must be interpreted generously 58 to extend beyond verbal representations. 59 The Equality Court accepted that the old flag is a demeaning symbol, the embodiment of the oppression of black South Africans by whites and of racial segregation during apartheid. 60 The court noted further that the old flag is used internationally as a symbol of white supremacy. 61 In the result the Equality Court held that gratuitously 62 displaying the old South African flag constitutes hate speech, 63 harassment 64 and unfair discrimination 65 as intended in the PEPUDA.

The fact that so many different types of acts can qualify as an utterance that can potentially be hate speech is a consequence of the wide construction afforded to "words" as used in the initial version of section 10(1) of the PEPUDA. Botha and Kok contend that in order for that version of section 10 to pass constitutional muster, non-verbal communication should not be considered as falling under the hate speech prohibition. 66 However, the wide construction that is preferred is aligned with international and foreign law regulating hate speech that bans not only words but also all expressions that expose target groups to hatred. 67 Therefore, Botha and Govindjee are correct that "words" in section 10(1) of the PEPUDA ought to be replaced by or interpreted as "expressions". 68 The interim amended version of section 10(1) omits the term "words" and merely provides "[n]o person may advocate". This is more open-ended, potentially embracing all qualifying expressions. This amendment is better aligned with international and foreign law.

2.2.2 The value attached to the expression

From a foreign law perspective, not all expressions are considered equal. In each case the court must assess whether the expression (a) advances democratic discourse, and/or (b) has been expressed in the process of truth-finding and/or (c) is made as a means of self-fulfilment. 69 The more of these purposes an expression is found to advance, the more value should be attached to it. 70 A "valuable form" of expression should carry more weight in the balancing of the constitutional rights at play. 71

However, in De Reuck v Director of Public Prosecutions, Witwatersrand Local Division 72 the Constitutional Court indicated that the right to freedom of expression in the Constitution is subject to the rationality test in section 36 of the Constitution only, and not also to a threshold test considering the traditional value attached to the expression, as mentioned above, that is viewed as forming the basis of the right. 73 The PEPUDA forbids "categories of expression" that fall beyond the constitutional protection in section 16. 74 Consequently, to prove that an expression constitutes hate speech as envisaged in section 10(1) of the PEPUDA, it must be proven that the expression oversteps the bounds of the right to freedom of expression. 75

In the Qwelane Equality Court case 76 the court concluded that the contents of Qwelane's article in the Sunday Sun entitled "Call me names - but gay is not okay'' constituted hate speech in terms of section 10(1) of the PEPUDA in its original formulation. In the statement concerning homosexuals, 77 Qwelane had compared their sexual conduct to bestiality. 78 The court held that inviting homophobia had no constitutional value. 79 The court based this finding on the fact that the utterance had not on the face of it, or on the evidence presented to the court, been made to spark a debate on lesbians and gays. 80 Rather, in the view of the Equality Court Qwelane had made the utterance to persuade the readers to support his homophobic views. The Equality Court found that the utterances against homosexuals were hurtful, incited harm and propagated hatred and accordingly amounted to hate speech. Therefore, it could not be argued that the utterance fell within the realm of the protection of freedom of speech. 81 On appeal, the Supreme Court of Appeal 82 declared section 10 of the PEPUDA unconstitutional and invalid. 83 In terms of the interim section 10 Qwelane's utterances did not constitute hate speech. Accordingly, the judgment of the Equality Court was overturned. 84

Generally, it can be said that if an utterance is made to spark a debate, it would be protected. Moreover, in terms of the Recommendation 85 if the purpose for the making of the utterance is to protect or to defend human rights, the conduct should not be sanctionable as hate speech.

2.2.3 Political statements

In most of the jurisdictions that were evaluated in Memorandum 2012, political speech is viewed as carrying more value in the process of balancing free speech and the rights to equality and dignity. 86 Particular value is also attached by the ECtHR to political speech which is viewed as an important vehicle for democratic discourse. 87 However, whether these types of utterances will be absolved from scrutiny in relation to constituting possible hate speech depends on whether they add value to truth-finding, or whether they in fact bear no truth. 88

The SAHRC found that the "white slaughter" remark of the leader of the Economic Freedom Front (the EFF), Julius Malema, 89 did not qualify as hate speech and reasoned that the statement had been made by Malema in the context of the land debate. 90 The SAHRC's ruling in Malema's favour rested at least in part on the fact that there was a measure of truth to the premise that white South Africans still occupy land belonging to black South Africans, and that whites enjoy substantial economic power. 91 The other reason for finding the statement to be permissible 92 was that it did not according to the SAHRC contain a threat of imminent violence. 93

However, in the Masuku case the Equality Court rejected the defence against the hate speech charge that the utterance was true, fair comment or in the public interest. 94 Truth and public interest are defences accepted against charges of defamation, not to avoid liability against hate speech. 95 We agree with the finding in the Masuku case that it is inappropriate to excuse hate speech on the basis that an utterance holds a measure of truth. The purpose of the protection offered by section 10(1) of the PEPUDA differs significantly from the protection against defamation, which is to address the injury to the plaintiff's reputation. If the defendant in a defamation case made a statement which is true and if it is in the public interest that he made the statement, the utterance can be viewed as merited and deserved. 96 The same cannot be said in a hate speech case. 97 The SAHRC appears to have introduced a defence that is out of place in the context of hate speech, which is not recognised in the same way by the Equality Court.

The fact that the "white slaughter" remark made by Malema did not convey a threat of imminent physical harm would also not exclude liability based on hate speech. The "harm" envisaged in the PEPUDA is not restricted to physical violence only. Other legislation caters for threats of physical harm, besides the PEPUDA. 98 Harm can also be emotional harm, 99 but its impact would need to be "more than merely 'hurtful' in the dictionary sense". 100 As was held in the Khumalo case, the rehabilitative objective of the PEPUDA suggests that harm could also relate to the reaction to the utterances, 101 i.e. the negative effect it has on the nation-building project in that it does not promote non-racialism. 102 The Labour Appeal Court in Crown Chickens (Pty) Ltd t/a Rockland Poultry v Kapp 103 made it clear that the courts must conspicuously safeguard the fragile relationships between the different population groups in South Africa. 104 This duty extends to the SAHRC too. The amended wording of section 10(1) of the PEPUDA as formulated in Qwelane SCA still supports the premise in the Qwelane Equality Court case that to qualify as hate speech an utterance must "be hurtful and harmful and have the potential of inciting harm and plainly propagating hatred." 105 In our opinion the "white slaughter" remark ticks all the boxes to qualify as hate speech under either version of section 10(1).

2.2.4 Historical value

In South Africa, as in other jurisdictions, if the alleged perpetrator is a member of a previously disadvantaged group which was subject to historical oppression, offensive speech is tolerated more readily, for reasons explained below. Offensive utterances that can be viewed as venting frustration with historically dominant groups is considered as a "form of self-fulfilment". 106 The LRC also indicated that struggle songs, like Dubula ibhunu (Xhosa: "Shoot the Boer") are arguably political expression bearing a historical value. Singing them remains relevant to vulnerable groups. They are used to invoke solidarity and courage to face current challenges too. 107

In the Duncanmec case, 108 black workers, members of a trade union, during an unprotected strike sang an isiZulu struggle song. The lyrics translated into English are: "Climb on top of the roof and tell them that my mother is rejoicing when we hit the boer." 109 Following a disciplinary enquiry, they were dismissed for misconduct.

In the Bargaining Council where the workers challenged their dismissal, 110 the arbitrator concluded that although the singing of the song was inappropriate, it did not constitute racism. 111 Having looked at a video of the singing and dancing workers, she remarked that the singing was not violent, but rather "peaceful and short-lived". 112 She found the employees' dismissal substantively unfair and ordered the employer to reinstate them. 113

On review, the trade union acting on behalf of its dismissed members explained to the Labour Court that the song was not hate speech against whites, but a struggle song that black workers had sung during apartheid. 114 The Labour Court accepted this explanation, and the factual finding which had been made by the bargaining council arbitrator that singing a struggle song is different from making a crude racist remark to someone. 115 The workers conceded that the singing of this song was more appropriate in the currency of apartheid, and that the Constitution currently affords workers rights that they were denied under the old order. Notwithstanding, they argued that the effects of apartheid continue to affect them in the workplace. They explained that the economic structure had not changed, and that in many instances whites fulfil management functions whereas black workers are employed in the lower ranks. 116 They contended that the motivation for singing struggle songs was to achieve "solidarity and defiance of the authority of the employer" and not racial hatred. 117

The court agreed, noting that a distinction ought to be drawn between the singing of struggle songs and making other racially loaded utterances because of the history attached to struggle songs. The court found unconvincing the reasoning that the singing of struggle songs should be sanctionable as hate speech because it infringes on the victim/s right to dignity. 118

In the final round of litigation, the Constitutional Court pointed out as to the lyrics that the only word in the song that referred to race was "boer", which could mean "farmer" or a "white person", neither of which is racially offensive. 119 The employer argued that it was the context in which the word "boer" was used that could hurt or offend those who heard it. 120 As the employer and the trade union had both accepted the arbitrator's factual finding that the song did not contain racist words, but that singing it at work was inappropriate, the Constitutional Court accepted the arbitrator's finding that the employee's conduct was "racially offensive" but not racism justifying dismissal. 121

In the AfriForum case Lamont J found that Malema, then president of the African National Congress Youth League, had acted in contravention of the hate speech provision in the PEPUDA for singing "Shoot the Boer" at political rallies. The court held that if he were to sing the song in future, he would face criminal charges and a potential prison spell. After the judgment Malema simply replaced the word "shoot" with the word "kiss" and sang "Kiss the Boer". 122

The SAHRC has also received several other complaints of alleged hate speech made by Malema and other EFF members. On 27 March 2019 the SAHRC ruled on four comments made by Malema and one by the EFF's general secretary at the time, Godrich Gardee. The SAHRC found that, although the statements were "quite offensive", they did not qualify as hate speech. 123 Besides the contentious "white slaughter" comment 124 and the singing of "Kiss the Boer", Malema had made the following statement concerning Indians and Coloureds: 125

We were not all oppressed the same. Indians had all sorts of resources Africans didn't have, Coloureds as well... The majority of Indians hate Africans. The majority of Indians are racist. I'm not saying all, I'm saying majority.

Gardee on Twitter had referred to the former leader of the DA, Mmusi Maimane, as "a garden boy". 126

As to the statement concerning Indians, the SAHRC again perceived some truth in Malema's utterance. It found that on an "objective and contextual assessment" the statements indicate that although both of the population groups mentioned were disadvantaged during apartheid, the African group was and remains more vulnerable than Indians. Moreover, the SAHRC remarked that under apartheid Indians had enjoyed certain political and economic privileges from which black Africans were excluded. 127 The SAHRC continued that "although a minority, the white population group is socio-economically powerful" 128 and that Malema belongs to the vulnerable black population group. 129 The SAHRC found that the statements made by Malema 130 may have been offensive and disturbing, but that they bore constitutional value for dealing with land reform and race relations. Likewise, the SAHRC found Gardee's statement on twitter "offensive and demeaning", but not constituting hate speech. The SAHRC ruled that there would be "no political or constitutional value" in affording a remedy for this statement. 131

As to Malema's statement that the majority of Indians are racist, the SAHRC pointed out that he did not say that all Indians are racist. This finding is problematic. The SAHRC failed to indicate whether it would have been hate speech if Malema stated that all Indians were racists. If so, we agree with Naidoo. The SAHRC's ruling "sets a problematic precedent by adopting a hierarchical approach to the racial status of the alleged offender and the target of the speech." 132 Naidoo asks if this means that if the song "Kill the Boer" was sung by someone other than a black African it could be hate speech, or whether if a white person had made the statement concerning Indians it would constitute hate speech. 133 The SAHRC's chairperson explained that freedom of speech would be endangered if the SAHRC were quick to deem robust and offensive speech to be hate speech. Finding that Malema had contravened the prohibition against hate speech in the PEPUDA based solely on the hurt caused would threaten the constitutional right. 134

A senior researcher of the SAHRC commented that although Malema's "white slaughter" utterance could be interpreted as hurtful by a white audience, the reasonable listener would have realised that the utterance concerned land reform, and that it was not intended to harm white people. 135 Her view is that 136

[t]he historical context in which the speech is made is one of unjust land dispossessions by both colonialists and the Apartheid government. Reference to slaughtering is made within this context. The statement calls for the peaceful invasion of land. Malema explicitly stated that he is not calling for a slaughter of white people.

While Malema's utterance was again viewed by the SAHRC to be a permissible "political statement", 137 the posting of Helen Zille, who was formerly the Western Cape Premier (now the chairperson of the federal executive) of the official political opposition party, the Democratic Alliance (the DA), was not for the same reason absolved from further scrutiny. 138 She had posted on Twitter:

For those claiming that the legacy of colonialism was only negative, think of our independent judiciary, transport infrastructure, piped water etc.

Here the approach accepted in international law was preferred, i.e. to be stricter against politicians and individuals holding status positions in so far as statements that they make are racially divisive. 139

This was not the last from Malema. Following the death of the previous president of Zimbabwe, Robert Mugabe, he posted on Twitter: "the only white man you can trust is a dead white man". The SAHRC has indicated its intention to refer the matter to the Equality Court as a possible instance of hate speech. 140

The fact that the SAHRC is lethargic in acting against these types of serious and repeated utterances by certain politicians could have devastating consequences. In the Masuku case 141 the SAHRC led the evidence of a doctor, Dr Stanton, who has done extensive research on the topic of the prevention of genocide. In his research Stanton analysed the processes that lead to genocide. He has discovered that there is a pattern. The first stage of genocide starts with words having consequences. It is repetition of these utterances that incites genocide. The fact that Malema has been allowed to make repeated negative, racial utterances revolving around the same theme and that they have gone unpunished could in our view be considered the allowance of this first step of genocide against the race groups that he targets.

The identity and status of the perpetrator can potentially increase the likelihood, and the extent of the harm suffered. 142 How well known and influential the perpetrator is also impacts on the size of the audience who will take note of the expression and the value placed on it. 143 For these reasons, the ECtHR places a higher premium on utterances made by politicians. 144 The European Council's case law suggests that the court is more inclined to find politicians guilty of hate speech. 145 This principle does not appear to be consistently applied, particularly by the SAHRC. The "defences" that appear to be accepted by the SAHRC have syphoned through to the Equality Courts, where they are raised as defences against charges of having uttered hate speech.

Black First Land First (the BLF), a political party which has since been deregistered, often used slogans emanating from the apartheid era, including "kill the farmer, kill the boer", "one settler, one bullet" and "land or death". 146 Andile Mngxitama, the president of the BLF, answering to charges of alleged hate speech, observed that the slogans were responses to the historic and current land dispossession. 147 The Equality Court disagreed and declared the slogan "land or death" to be hate speech. 148

On 8 December 2018 at a rally in Potchefstroom Mngxitama stated: 149

You kill one of us (black Africans), we will kill five of you (whites). We will kill their women, we will kill their children, we will kill their dogs, we will kill their cats, we kill anything that comes for us.

The BLF denied that this was hate speech, arguing that supporters were only being instructed to defend themselves should they be attacked and killed by whites. 150 Even if this is so, no provision is made in the PEPUDA for any defences besides the threshold test for the application of section 10(1) of the PEPUDA.

In 2019 a fatal tragedy occurred at the high school Driehoek in Vanderbijlpark. Four white pupils died after an overhead walkway caved in. Various other pupils, mostly white, were seriously injured. Following the tragedy, the BLF spokesperson posted on social media that the death of the young children should be celebrated. 151 Someone responded: "minus three land criminals - great news" and "[d]on't have heart to feel pain for white kids. Minus 3 future problems." The BLF spokesperson responded, "God is responding, why should we frown on the ancestors' petitions to punish the land thieves including their offspring." 152 The SAHRC received several complaints and indicated that it would refer the matter to the Equality Court.

Mngxitama responded that the country should rather have an open dialogue about why people hold such views. 153

In the Johannesburg Equality Court, facing charges of hate speech along with other members of the BLF, Mngxitama, argued that the utterances should not be viewed as hate speech. Instead, he argued, they were a spontaneous reaction emanating from a place of pain. He contended that the history of the country should be considered and that the statement should be viewed from the perspective that if a black child falls into a pit toilet there would not be as much publicity as in the case of these white children who died. He argued that the utterances were posted out of historical anger against whites. 154 Initially, Mokgoatlheng J found that the members' statements constituted hate speech, but this finding was nullified by Qwelane SCA in which section 10(1) of the PEPUDA was amended in the interim. 155

It is not correct to apply the legislation differently to different population groups. It is also the incorrect position from which to start the assessment of whether or not an utterance constitutes hate speech. The PEPUDA expressly determines that in hate speech cases considerations of fairness, including the personal context of the victim and the perpetrator, 156 do not play a role. 157 However, the PEPUDA requires that the context in which an utterance is made should be considered. Factors that are considered part of the context are the social and historical context and whether on the facts the utterance was directed at a group or individual that is recognised as being vulnerable. 158

In Canada (Human Rights Comm) v Taylor 159 the Canadian court approached the prohibition against hate messages from the point of view of those affected. 160 This orientation, that the court should view the equality guarantee from the view of the victim, appears to be correct. It makes sense when considering the context in which the utterance is made to look at the act of alleged hate speech from the victim's point of view, taking into consideration the surrounding circumstances applicable to him or her at the time that the utterance is made. This approach conforms to the Critical Legal Theory which determines that in the context of equality one must deal with individuals in accordance with their true conditions of disadvantage. 161 Further, the PEPUDA explicitly requires taking into consideration the disadvantage and the context of the complainant. 162

The appeal court in Herselman v Geleba 163 also preferred the view that "the perceptions of the receiver or listener" are determinative. 164 The court reasoned that the expressed purpose of the PEPUDA to protect victims of hate speech dictates this approach. 165 Similarly, in the Masuku case the court iterated that what is important is how the statement was perceived. 166

2.3 The identity of the victim(s)

2.3.1 Individual v group protection

Section 10(1) of the PEPUDA, as it stood before Qwelane SCA, prohibited the communication of words "against any person" on a prohibited ground. These words are no longer contained in the interim amended provision. Nevertheless, no effort was made in the interim to clarify that protection against hate speech is not applicable to individuals as opposed to vulnerable groups. South Africa's legislation is not the only hate speech regulation that can be interpreted as providing individual protection. 167

However, to use it as a mechanism for the protection of individuals does not appear to accord with the purpose of the PEPUDA as a remedial human rights statute with the transformation of the South African society in mind. 168

The purpose of the PEPUDA is to promote democracy, to reconcile South African society and to uphold the constitutional values. 169 Moreover, to interpret section 10(1) as a mechanism to protect individual interests as opposed to group rights would be out of alignment with the hate speech regulation in other jurisdictions. 170 The Canadian Supreme Court declared unconstitutional section 14(1)(b) of the Saskatchewan Human Rights Code, 171 a provision which regulated offensive speech, but which was not aimed at the protection of a vulnerable group. 172 The Canadian Supreme Court in Saskatchewan (Human Rights Commission) v Whatcott 173 held:

[H]ate speech must rise to a level beyond merely impugning individuals: it must seek to marginalize the group by affecting its social status and acceptance in the eyes of the majority.

In Islamic Unity Convention v Independent Broadcasting Authority,' 174 the Constitutional Court confirmed that the purpose of regulating hate speech is to ban any expression which "reinforces and perpetuates patterns of discrimination and inequality" and undermines unity, tolerance and reconciliation. 175 In the Qwelane Equality Court case it was also acknowledged that section 10(1) of the PEPUDA is supposed to protect vulnerable groups, to give effect to their rights to equality, and to prevent unfair discrimination against them. 176 Although it is not irrelevant that individuals belonging to a group are offended and hurt, "[u]ltimately, it is the need to protect the societal standing of vulnerable groups that is the objective of legislation restricting hate speech". 177

Lindsay Maasdorp, spokesperson of the BLF, posted on Facebook: "I really dislike Max Price! I want to deal with him. Please cadres, tell that clown I'm coming for him, and he should expect me!" On a different day he wrote:

"When will we kill them?" The next day he posted on Twitter and on Facebook: "I have aspirations to kill white people, and this must be achieved!" About two weeks later he posted: "Make sure the struggle implicates whiteness (white-power/white people)" and shortly thereafter: "Max Price must be dealt with personally!" and "The Price on your life has been set to Max!" 178

The threats made against Max Price on social media by Maasdorp would in terms of foreign law and section 10(1) of the PEPUDA not qualify as hate speech. They were directed at Price as an individual and did not relate to one of the prohibited grounds enumerated in the PEPUDA. 179 However, the references to whites could qualify as hate speech. The statements are directed at a target group and relate to race, one of the prohibited grounds.

2.3.2 The vulnerability of the target group

It is an accepted principle in foreign jurisdictions that the more vulnerable the target group is to which the racial utterance is directed, the more likely it is that the group will be harmed by and because of the hate speech. 180 The potential harm that can be inflicted by a racial utterance also increases if a power disparity exists between the perpetrator and the victims. 181 However, the likely effect of the utterance is less important in South Africa, as section 10(1) of the PEPUDA does not require of the victims to prove actual or potential harm.

In recognition of the obligation to uphold international obligations and to have regard to foreign law when interpreting and applying the legislation adopted under the Constitution, the interests of all of the population groups in South Africa should be taken into account.

It is trite in international law and on a national level that minorities are often defenceless against racial discrimination. The Equality Court recently iterated that South African equality courts, in the fulfilment of the obligations under the Constitution and the PEPUDA, cannot allow hate speech against minority groups. 182 Therefore, it is incremental that the court must act in protection of minority groups, particularly those which historically have fallen victim to hate crimes. 183

During March 2019 the Equality Court heard a hate speech case concerning postings on Facebook. This followed a Carte Blanche episode concerning the slaughtering and skinning of donkeys for use in Chinese traditional medicine. A spate of negative individual comments directed at Chinese people followed. For instance, "[t]hey (Chinese people) are the most despicable things on the planet! Hate the Chings"; "[th]ere are no more disgusting humans than the Chinese people. I wish they all just die!"; "[ca]n we stop these slant-eyed freaks from coming into the country"; and also "we should start killing their children for a cure of the common babalaas." 184

The Chinese Association lodged an application to declare the statements of eleven people to be hate speech. Historically, the Chinese as a minority population group in South Africa, like black South Africans, were subjected to legislation that was unfairly discriminatory based on race. 185 The Chinese were also required to fight a legal battle in order to secure recognition as falling within the "designated groups of employees" that are entitled to the benefits of affirmative action under the Employment Equity Act 55 of 1998 (the EEA). 186 In 2008 the High Court of South Africa declared Chinese South Africans who became citizens of the Republic before 1994 and their descendants as being eligible for the benefits of affirmative action. 187 The outcome of this pending hate speech case should provide meaningful insight as to the value that the Equality Court ascribes to the factor of the vulnerability and historical discrimination of a minority population group in South Africa.

2.4 The identity of the perpetrator

Who the perpetrator is plays a role in different ways. The more powerful the utterer is relative to the target group, the greater the threat of harm. How much value is attached to the expression is also linked to the perpetrator's identity. The ECtHR is stricter when it comes to hate speech perpetrated by politicians. 188 Office bearers, governmental employees and politicians, because of their status in society and their obligation to the country's tax payers, ought to be subjected to closer scrutiny and to more severe punishment if they are found liable for hate speech. They are under a duty to act in the best interest of all citizens, regardless of their race. 189 The Recommendation also determines that office bearers and opinion makers have more of a responsibility to foster harmonious race relations than members of the general public and that they ought to be dealt with more strictly by institutions responsible for the regulation of hate speech. However, this principle appears not to be applied consistently in relation to hate speech in South Africa.

In Dagane v SSBC 190 (hereafter the Dagane case) it was noted that Dagane had been dismissed for posting racist statements 191 on Malema's Facebook page. The Labour Court held that dismissing him for doing so was fair because Dagane had made the racist utterances in his capacity as a police officer, and police officers are responsible for the safety of all citizens. 192

On the other hand, the SAHRC apparently absolved the utterances made by Zindzi Mandela-Hlongwane, the daughter of the late Winnie Madakizela Mandela and Nelson Mandela, from scrutiny. While serving as the ambassador of South Africa to Denmark Mandela-Hlongwane had, among other things, on her Twitter account referred to white South Africans as "trembling white cowards", "thieving rapist descendants of Van Riebeck (sic)" and "shivering land thieves". The spokesperson of the EFF, Mbuyiseni Ndlozi, stated that the EFF supported Mandela-Hlongwane's tweets and her views. 193 AfriForum requested the Minister of International Relations and Cooperation, Dr Naledi Pandor, to dismiss Mandela-Hlongwane, arguing that as an ambassador of South Africa there rested a duty on her to act in the interest of all the country's residents without prejudice. 194 She was not recalled or dismissed. Rather, her contract was extended for a further six months. AfriForum lodged a complaint with the SAHRC. 195 In an official response the SAHRC declared that "the history of Zindzi Mandela and her family needs to be taken into account during any investigation into her tweets on land reform." 196

2.5 Historical associations

The words that are used may bear cultural or historical associations that qualify them as hate speech. Some expressions are viewed as being manifestly heinous by the SAHRC and the Equality Court alike, and they are inclined to be viewed as hate speech. Examples include calling a black South African a "baboon" or likening him to a monkey, 197 calling the members of a population group "cockroaches" 198 and using the k-word. 199 Referring to a black person using the k-word is always hate speech, even if the perpetrator is black. 200 It was recently confirmed that the k-word, when used by one black individual to address another, is not considered by the court to be part of the culture of black South Africans. 201 "Boer" has in some cases been recognised as pejorative, but this is not always the case. 202

In assigning meaning to words, South African courts generally use dictionary definitions and refer to meanings assigned to words in previous court cases. 203 However, in some cases even if the wording does not state what is meant, the true meaning of the utterance has been deduced. For instance, in Dyonashe v Siyaya Skills Institute (Pty) Ltd, 204 even though the expression did not expressly refer to Whites but rather to "Boers", which could be viewed as a neutral race descriptor, 205 the commissioner was satisfied that objectively viewed the reasonable person would read "Kill the Boer" to mean kill white people. In the AfriForum case, 206 the court had held that the struggle song "Kill the Boer" is understood by the reasonable person to mean kill white people. 207 In the Hotz case the court declared that a T-shirt with the inscription "Kill all whites" was racist even though from less than a meter away a tiny "s" is visible. In other words, the caption read "skill all whites". 208

Expressions can also be ambiguous, bearing more than one possible meaning. While saying one thing you may be implying something else. 209 Speech acts can be divided into three parts: "locutionary (what is said), illocutionary (what is meant) and perlocutionary (the effect)". 210 According to the principles of pragmatics, it is often not what you say, but how you say something that matters in determining the level of politeness of a verbal exchange. 211 Carney posits convincingly that when a court in a hate speech case assesses whether an utterance is hurtful or harmful, it would assist to employ principles of pragmatic linguistics focussing in particular on "speech acts" and principles of "politeness". 212 Utterances may for instance have different meanings to different population groups and cultures. 213 The skills of communication are often culture-bound, a fact which may result in misinterpretations of expressions by individuals not belonging to the utterer's cultural group. 214 For instance, the lyrics of songs and what the intention is in singing them often do not correlate with the literal meanings. 215 In the Memorandum 2012 the LRC posited that interpreting what a song means and why a song was composed and ultimately sung in an assessment of whether the singing thereof constitutes incitement to harm or of hatred would require "insider" knowledge. 216

This principle is well illustrated by the case concerning the old land flag. As to the meaning of the expression, the Federasie van Afrikaanse Kultuurvereniginge in the Mandela Foundation case argued that the old land flag is a cultural symbol of reconciliation between the boers and the English. 217 The dictum in the Mandela Foundation case contrasts with the findings of the ECtHR. In Vajnai v Hungary 218 the ECtHR acknowledged that whereas the red star was a symbol that signified Soviet totalitarianism for some, in other sections of the Hungarian society it was a sign of solidarity and social justice. The court emphasised the importance of not imposing unduly narrow restrictions on expressions that may have multiple meanings. 219 Making representations of history is also acknowledged by the ECtHR as intrinsic to permissible democratic discourse. 220 Moreover, a narrow limitation in an instance where an expression has different meanings as in this case is contrary to article 1 of the Declaration of the Rights of Persons belonging to National, Ethnic, Religious and Language Minorities, which was adopted by the United Nations General Assembly on 18 December 1992. It reads: 221

States shall protect the existence and the national or ethnic, cultural, religious and linguistic identity of minorities within their respective territories and shall encourage conditions for the promotion of that identity.

2.6 The audience to whom the utterance is made and where it is made

Where and to whom the utterance is made may determine whether it constitutes hate speech or only rude or distasteful speech. 222

Whether it makes a difference if an utterance that qualifies as potential hate speech occurred in the workplace or outside of the workplace is an issue in the PEPUDA that requires clarification through legislative amendment. Section 5(3) of the PEPUDA apparently excludes "employees" from the protection offered by this piece of legislation, ostensibly leaving them remediless.

Whether witnesses were present when a verbal utterance was made is an important consideration in the assessment of whether an expression is derogatory. 223 The testimony of witnesses is important to confirm that the words were indeed uttered, and to convey to the court the manner in which the utterance was made. For instance, recently, after a black licensing department official allegedly made aggressive racist utterances to a young white male client, the witness who had reported the incident described the manner in which the official had treated the young man as "berating, insulting and humiliating". 224

The potential harm is also dependent on the way the hearers interpret the speech. 225 However, whether all the witnesses considered the utterance to be racist is not the test to determine whether the expression qualifies as hate speech. The correct question to ask, according to the Constitutional Court, 226 is whether objectively the words were reasonably capable of conveying to the reasonable hearer a pejorative meaning. 227

In the Dagane case the fact that Dagane had posted the racially loaded utterance on Malema's Facebook page is significant. As a member of the EFF and by posting the statement where many likeminded individuals were likely to read it, Dagane showed his intention to incite hatred and violence. 228 Moreover, the likelihood that the audience would share his ideology increased the likelihood that his utterance would incite hatred or harm. 229

2.6.1 Private or public utterances

In terms of foreign jurisprudence, it is evident that hate speech regulation is not intended to censor ideas. Consequently, it is required that the utterance must be made in public to be sanctionable. 230 To regulate hate speech which occurs in public sets a benchmark of what is acceptable behaviour and may assist in changing the mind-set of individuals too. 231 Unfortunately, the PEPUDA is not clear that "publicity" of the expression is required for it to qualify as hate speech, as is the case in other jurisdictions. 232

In terms of the Recommendation, if a racial utterance is disseminated via the mainstream media, or via the internet, or if it is repeated, it must be viewed as the fruit of a deliberate plan to instil hostility. It appears that the Equality Court considers how widely the utterance is publicised as a factor counting against the perpetrator. For instance, in the Dagane case the Labour Court considered the fact that the police officer had posted the racially loaded statements on a quasi-public forum [which is] accessible to potentially thousands of Facebook users.

However, the amount of publicity that hate speech attracts does not always depend on who the perpetrator is. Neither is it necessarily connected to his relative social standing and importance.

There is a public interest dimension to hate speech cases. 233 The PEPUDA requires that these cases must be heard in open court. 234 In Khumalo the Roodepoort proceedings which had been instituted by the ANC were conducted in camera to evade publicity. 235 Moreover, no press statement was made after these proceedings. 236 The Johannesburg Equality Court lamented the fact that contrary to the prescripts of the PEPUDA the Roodepoort proceedings were not conducted in public. 237 In the light of the lack of publicity, the Johannesburg Equality Court was not convinced that the public interest had been served. 238

The research conducted by Brink and Mulder in 2017 239 suggests that utterances made by whites against blacks enjoy more media coverage than those made by blacks against whites, even if what is said by the black perpetrators appears to be much more malignant. 240 The black case studies referred to in the survey enjoyed markedly less media coverage than those of white transgressors. In fact, the largest number of media reports for a black transgressor (Malema, with 163) was reported on nearly 100 times fewer than the lowest number of reports concerning a white transgressor. 241

Although one would expect that utterances made by individuals in influential positions would receive more media attention, this is not always the case. Angelo Agrizzi had made the utterance in which he referred to two Bosasa directors by the k-word, in the privacy of his home with only a few members of his family and a colleague present. The SAHRC stated in a newspaper report that this fact made no difference. A secret recording which was played at the 2019 Zondo Commission of Enquiry into State Capture was publicised widely. 242 Penny Sparrow had posted her utterance on her private Facebook account, for the attention of a few of her personal friends and family only. However, someone got a screenshot of it and disseminated the utterance further, and hers became the most thoroughly publicised case of hate speech in South Africa to date. 243 Likewise, Adam Catzavelos' cases enjoyed a lot of publicity. His video went viral. 244 Whereas Agrizzi was an executive businessman, Sparrow and Catzavelos were relatively unknown, working as an estate agent and in a family business respectively. 245

The skewed numbers of media reports along racial lines could possibly emanate from the fact that it is not widely accepted that blacks can be guilty of racism against whites. 246 It could also be attributable to the fact that the vast majority of South Africans are black and that reports concerning racism against blacks could increase newspaper sales. 247 Whatever the reason, it is contrary to the Recommendation that the media coverage on hate speech should be so one-sided. It requires "[i]nformed, ethical and objective media" which "does not refer to race in a manner that may promote intolerance." 248 The "Concluding observations on the combined fourth to eighth periodic reports of South Africa" 249 indicate that the CERD is concerned about the governing party's racist pronouncements and that the media chooses to ignore them, which contributes to the general racial polarisation in South Africa.

2.7 The social conditions at the time of making the utterance

The Recommendation indicates that the socio-economic and political circumstances at the time of the making of the utterance must be considered by South African institutions in assessing whether an utterance constitutes hate speech. The assessment into whether the intention of the utterer of alleged hate speech is to be hurtful, harmful and to propagate hatred requires a scrutiny of the content of the offending utterance in its social context. 250 The meaning of an utterance and whether it constitutes hate speech may be affected by the social conditions at the time that the utterance is made. 251

The current unemployment rate in South Africa stands at a staggering 29 per cent. 252 The StatsSA "Quarterly Labour Force Survey for Quarter 2 of 2019" indicates that black South Africans still account for the biggest segment of the unemployed group at 32,7 per cent, followed by coloureds (22,5 per cent), Indian/Asians at 11,2 per cent and whites at 7,4 per cent.

Although many measures have been introduced since 1994 to promote participation in the economy by those who were during apartheid deprived of the opportunity, 253 vast inequity remains. 254 These measures do not necessarily address the interests of the vast majority of indigent black South Africans 255 and where there are successes, change is considered to be too slow. 256 The country is currently caught up in a land debate threatening the property rights on which the economy rests, while holding promise for black South Africans that they will be afforded land that they believe to belong to them. 257 This has raised racial tensions even further.

In Modikwa Mining Personnel Services v CCMA 258 the employee had been dismissed for saying: "we need to get rid of the whites" at a meeting at which both white and black employees were present. The court considered the utterance to be overtly racist. 259 However, having considered the social, political and historical context, Gaibie AJ noted that the utterer as a black employee may have felt that he was being discriminated against unfairly in the workplace. Notwithstanding, the court held that the proper grievance channels ought to be followed instead of resorting to making racist statements. 260 The SAHRC appears to take into consideration the social conditions of different population groups in different ways. 261

3 Conclusion

Given the uncertainties surrounding the proper interpretation and application of section 10(1) of the PEPUDA, in its initial as well as interim amended format, it is appropriate to consider international and foreign law for guidance.

In the hate speech provision as it was before the judgment in Qwelane SCA, the PEPUDA did not make it clear that the utterance must incite hatred or be likely to cause harm to qualify as hate speech. The Equality Court recently confirmed that the "clear intention" requirement must be separated from the subjective intention of the utterer. 262 The objective test that has been laid down to establish hate speech should remain unaffected by the amended wording. 263

A wide construction was afforded to "words" in the initial version of section 10(1) of the PEPUDA so that the term covered any expression, although some commentators have disagreed with this interpretation. The amended version of section 10(1) by Qwelane SCA completely omits the term "words" and simply provides "[n]o person may advocate", thereby removing any uncertainty that its meaning is limited to words only. The version is broad enough also to embrace any qualifying expressions. This broad interpretation accords with international and foreign law.

Whereas in foreign law additional criteria are set for measuring the value of expressions, this is not done in South Africa. The protection in section 10(1) of the PEPUDA is subject only to section 16 of the Constitution and the limitation clause in section 36. 264 The Equality Court 265 and the SAHRC 266 appear to interpret this as meaning that if the utterance has constitutional value or is made in the protection of a constitutional right, it would not constitute hate speech.

In foreign jurisdictions value is placed on political speech as part of democratic discourse. The SAHRC also applies this principle, but inconsistently along racial lines. In the process, it appears as if the SAHRC has created certain defences that are not provided for in the legislation.

South Africa, like other jurisdictions, is more lenient to perpetrators belonging to groups which have suffered previous disadvantage. However, it appears that in South Africa this leniency is abused and even raised as a defence against hate speech charges. Moreover, the SAHRC appears to have created an untenable hierarchy of disadvantage in applying the hate speech protection.

In South Africa protection against charges of hate speech is available in instances where inter-personal speech directed at individuals is offensive, whereas in foreign and international law the purpose is to protect victim groups.

South Africa's population consists of various population groups. It is incumbent on the SAHRC and the Equality Courts in terms of the principles of international law to pay specific regard to the protection of the interests of minority groups.

In relation to hate speech perpetrated by politicians, public officers and other highly placed individuals, the ECtHR is stricter in applying the hate speech provisions. This principle has not been implemented consistently in South Africa, a fact which has contributed to the repetition of hate speech by certain political figures.

Certain words are recognised in South Africa as hate speech by historical association. However, the meaning assigned to expressions may not necessarily be the same for all cultures. Whereas the ECtHR has accepted that where an expression has more than one meaning it should be scrutinised further to assess its possible constitutional value, this is apparently not done in South Africa.

Where the utterance is made can on a literal interpretation of the PEPUDA potentially exclude the possibility of its qualifying as hate speech. It also affects how many people will be exposed to the utterance, and the likely effect thereof. In terms of foreign jurisprudence only public utterances are accepted as hate speech. However, section 10(1) of the PEPUDA appears to cover utterances made in private also.

International law requires fair and equitable reporting of hate speech incidents. However, in South Africa the tendency is to report instances widely in the media where the perpetrator of the alleged hate speech is white, but not if he or she is black. The CERD has recognised this ostensible condonation of hate speech against whites by blacks as a concern which has the effect of entrenching the extant racial polarisation.

The socio-economic circumstances in South Africa dictate that the SAHRC and the Equality Courts must take a productive stance. It is important that they address the issue of hate speech, particularly when it is perpetrated by individuals holding status positions, in an equitable and consistent manner. Failure to do so may have dire consequences for the country and its inhabitants.

Bibliography

Eatock v Bolt 2011 FCA 1130

Hagan v Trustees of the Toowoomba Sports Grounds Trust 2000 FCA 1615

Canada (Human Rights Comm) v Taylor 1990 13 CHRR D/435 (SCC)

Grant v Torstar Corporation 2009 3 SCR 640

R v Ahenakew 2006 SKQB 17-18

R v Bahr 2006 ABPC 360

Saskatchewan (Human Rights Commission) v Whatcott 2013 1 SCR 467

European Court of Human Rights

Fratanolo v Hungary App No 29459/10 (ECtHR 3 November 2011)

Garaudy v France App No 65831/01 (ECtHR 3 July 2003)

Vajnai v Hungary App No 33629/06 (ECtHR 8 July 2008)

South Africa

AfriForum v Malema 2011 6 SA 240 (EqC)

Borgin v De Villiers 1980 3 SA 556 (A)

Chinese Association of South Africa v Minister of Labour 2008 ZAGPHC 174 (18 June 2008)

Crown Chickens (Pty) Ltd t/a Rocklands Poultry v Kapp 2002 6 BLLR 493 (LAC)

Dagane v SSSBC 2018 7 BLLR 669 (LC)

Delange v Costa 1989 2 SA 857 (A)

De Reuck v Director of Public Prosecutions, Witwatersrand Local Division 2004 1 SA 406 (CC)

Duncanmec (Pty) Limited v Gaylard 2018 6 SA 335 (CC)

Dyonashe v Siyaya Skills Institute (Pty) Ltd 2018 3 BALR 280 (CCMA)

Freedom Front v South African Human Rights Commission 2003 11 BCLR 1283 (SAHRC)

Herselman v Geleba 2011 ZAEQC 1 (1 September 2011)

Hotz v University of Cape Town 2017 2 SA 485 (SCA)

Human Rights Commission of South Africa v SABC 2003 1 BCLR 92 (BCCSA)

Islamic Unity Convention v Independent Broadcasting Authority 2002 4 SA 294 (CC)

Lebowa Platinum Mines Ltd v Hill 1998 19 ILJ 1112 (LAC)

Le Roux v Dey 2011 3 SA 274 (CC)

Makhanya v St Gobain 2019 7 BALR 720 (NBCCI)

Masuku v South African Human Rights Commission obo South African Jewish Board of Deputies 2019 2 SA 194 (SCA)

Minister of Home Affairs v Fourie 2006 1 SA 524 (CC)

Modikwa Mining Personnel Services v CCMA 2012 ZALCJHB 61 (29 June 2012)

Mohammed v Jassiem 1996 1 SA 673 (SCA)

Nelson Mandela Foundation Trust v AfriForum 2019 4 All SA 237 (EqC)

Phillips v DPP, Witwatersrand Local Division 2003 3 SA 345 (CC)

Qwelane v South African Human Rights Commission 2020 2 SA 124 (SCA)

Rustenburg Platinum Mine and SAEWA obo Meyer Bester 2018 8 BCLR 951 (CC)

SAEWU v Rustenburg Platinum Mines (LAC) (unreported) case number JA 45/2016 of 3 May 2017

South African Breweries (Pty) Ltd v Hansen (unreported) case number CA6/2016 of 25 May 2017

South African Human Rights Commission obo South African Jewish Board of Deputies v Masuku 2018 3 SA 291 (GJ)

South African Human Rights Commission v Khumalo 2019 1 All SA 254 (GJ)

South African Human Rights Commission v Qwelane 2018 2 SA 149 (GJ)

South African Revenue Service v Commission for Conciliation, Mediation and Arbitration 2017 38 ILJ 97 (CC)

Strydom v Black First Land First 2019 ZAEQC 1 (6 May 2019)

The Citizen 1978 (Pty) Ltd v McBride 2011 4 SA 191 (CC)

Thembani v Swanepoel 2017 3 SA 70 (ECM)

Vodacom (Pty) Ltd v Byrne 2012 33 ILJ 2705 (LC)

United States

West Virginia State Board of Education v Barnette 319 US 624 1943

Legislation

Racial Discrimination Act 52 of 1975

Canadian Criminal Code RSC 1985, c C-46

Saskatchewan Human Rights Code SS 1979, c S-24.1

Strafgesetzbuch, 1998

Broad-Based Black Economic Empowerment Act 53 of 2003

Cape Chinese Exclusion Act, 1904

Constitution of the Republic of South Africa, 1996

Employment Equity Act 55 of 1998

Prevention and Combating of Hate Crimes and Hate Speech Bill, 2016 (B9-2018)

Promotion of Equality and Prevention of Unfair Discrimination Act 4 of 2000

South African Human Rights Commission Act 40 of 2013

Transvaal Immigration Restriction Act, 1902

United Kingdom

Public Order Act, 1986

Government publications

GN 167 in GG 41543 of 29 March 2018 (Explanatory Summary of the Prevention and Combating of Hate Crimes and Hate Speech Bill, 2018)

International instruments

Declaration of the Rights of Persons belonging to National, Ethnic, Religious and Language Minorities (1992)

International Convention on the Elimination of All Forms of Racial Discrimination (1969)

Internet sources

List of Abbreviations

BLF Black First Land First

CCMA Commission for Conciliation, Mediation and Arbitration

CERD Committee on the Elimination of Racial Discrimination

DA Democratic Alliance

EEA Employment Equity Act

ECtHR European Court of Human Rights

EFF Economic Freedom Front

Harv L Rev Harvard Law Review

LRC Legal Resource Centre of South Africa

OHCHR Office of the United Nations High Commissioner for Human Rights

PEPUDA Prevention of Unfair Discrimination Act 4 of 2000

PELJ Potchefstroom Electronic Law Journal

PSLR Pretoria Student Law Review

SAHRC South African Human Rights Commission

SAJHR South African Journal of Human Rights

SAPL Southern African Public Law

SA Merc LJ South African Mercantile Law Journal

SCA Supreme Court of Appeal

StatsSA Statistics South Africa

THRHR Tydskrif vir die Hedendaagse Romeins-Hollandse Reg

UK United Kingdom

Date Submission: 25 October 2020 Date Revised: 27 April 2020 Date Accepted: 4 May 2020 Date published: 4 June 2020

Editor Prof O Fuo * Judith Geldenhuys. LLB LLM (UP) LLD (Unisa). Associate Professor Department of Mercantile Law, Unisa, South Africa. Email: [email protected] . We are grateful for the valuable comments of the reviewers. We remain responsible for any errors. ORCID ID https://orcid.org/0000-0001-9956-7071 . ** Michelle Kelly-Louw. B IURIS LLB LLM LLD (Unisa), Dip Insolvency Law and Practice (SARIPA) (UJ). Professor Department of Mercantile Law, Unisa, South Africa. Email: [email protected] . ORCID ID https://orcid.org/0000-0003-0145-3119 . 1 Oxford Pro Bono Publico 2012 https://www.law.ox.ac.uk/sites/files/oxlaw/L_comparative_hate_speech_-_lrc.pdf (the LRC Memorandum) para 5. 2 SAHRC 2019 https://www.sahrc.org.za/home/21/files/SAHRC%20Finding%20Julius%20Malema%20&%20Other%20March%202019.pdf (hereafter SAHRC Findings). 3 Different interpretations are afforded to s 10(1) of the Promotion of Equality and Prevention of Unfair Discrimination Act 4 of 2000 (the PEPUDA) by different authors. For the latest commentary see Marais and Pretorius 2019 PELJ. The relationship between ss 10(1), 7(a) and 12 of the PEPUDA has been discussed sufficiently. See Botha and Govindjee 2017 PELJ 5; Botha and Govindjee 2016 SAJHR 304; Kok and Botha 2014 Litnet Akademies 208-209. 4 SAHRC Findings para 13.3. 5 SAHRC Findings para 13.3. Also see South African Human Rights Commission v Khumalo 2019 1 SA 289 (GJ) (hereafter the Khumalo case) paras 115-116. Sutherland J expresses his concerns regarding the fact that the amendments are set to be completed only in about 5 years. 6 Although the contents of media reports and posts on social media may at times be one-sided, exaggerated or even misleading, we refer thereto in instances where the hate speech cases referred to have not been reported. 7 Section 3 of the PEPUDA and s 233 of the Constitution of the Republic of South Africa, 1996 (hereafter the Constitution). 8 Section 39(1)(b) of the Constitution; Nelson Mandela Foundation Trust v AfriForum 2019 4 All SA 237 (EqC) (hereafter Mandela Foundation case) paras 108-109. 9 Sections 3(1) and 3(2) of the PEPUDA. S 39(1)(c) of the Constitution determines that the court "may consider foreign law" when interpreting the rights contained in the Bill of Rights. Mandela Foundation case paras 116, 118-120. 10 Sections 3(3) of the PEPUDA; Mandela Foundation case para 121. 11 Mandela Foundation case paras 110-113. 12 The importance of aligning the interpretation afforded with international and foreign law is illustrated in the Mandela Foundation case para 19. 13 Islamic Unity Convention v Independent Broadcasting Authority 2002 4 SA 294 (CC) para 45; also see LRC Memorandum para 8. 14 Chapter 2 of the Constitution. 15 This was expressly stated in the Mandela Foundation case para 117. 16 See the LRC Memorandum para 3. Australia, Canada, the European Court of Human Rights (ECtHR), Germany, India, Ireland, Slovenia, South Africa, the United Kingdom (UK) and the United States. 17 LRC Memorandum para 3. 18 LRC Memorandum para 5. 19 LRC Memorandum para 5; also see Currie and De Waal Bill of Rights Handbook 356-357; Khumalo case para 99. 20 LRC Memorandum para 5. 21 For instance, in Canada and Slovenia. LRC Memorandum fn 6. 22 For instance, in Ireland and the UK. LRC Memorandum fn 6. Notably, the South African Parliament has considered adopting an instrument to criminalise hate speech, namely the Prevention and Combating of Hate Crimes and Hate Speech Bill, 2016. The second draft was published in March 2018, but it lapsed before being adopted. See Bill B9 of 2018 and the Explanatory Summary in Gen N 167 in GG 41543 of 29 March 2018. 23 This is usually the case in countries where hate speech is regulated as a crime. Examples include Canada and Slovenia. 24 LRC Memorandum para 5. The UK prohibits utterances made with the intention to or which are likely to "stir up" hatred. See s 18 of the Public Order Act, 1986. In Germany hate speech must be likely to cause public disorder (volksverhetzung). See s 130 of the German Strafgesetzbuch, 1998. In Australia the speech must be "reasonably likely in all the circumstances to offend, insult, humiliate or intimidate." Refer to s 18C of the Racial Discrimination Act 52 of 1975. 25 LRC Memorandum para 5. 26 LRC Memorandum para 26. 27 LRC Memorandum para 26. 28 The courts appear to be more inclined to assist groups who are historically or currently oppressed. Also see LRC Memorandum para 10. 29 The court and the SAHRC are more lenient if the perpetrator falls into a group that is or has been oppressed or marginalised. LRC Memorandum para 10. 30 LRC Memorandum para 29. These factors overlap largely with those put forward by Botha and Govindjee 2017 PELJ 17. 31 OHCHR 2013 https://www.ohchr.org/EN/Issues/Education/Training/Compilation/Pages/d)GeneralrecommendationNo35Combatingracisthatespeech(2013).aspx (the Recommendation). 32 Section 13(b)(vi) and (vii) of the South African Human Rights Commission Act 40 of 2013. 33 Paragraph 35 of the Recommendation lists: the content and form of speech; the economic, social and political climate; the position or status of the speaker in society and the audience at which the utterance is directed; and how widely the utterance is disseminated, but then adds also the objectives of the speech. 34 The proviso in s 12 of the PEPUDA excludes "bona fide engagement in artistic creativity, academic and scientific inquiry, fair and accurate reporting in the public interest or publication of any information, advertisement or notice in accordance with s 16 of the Constitution" from the ambit of s 10(1). The proviso is a defence available to an utterer whose speech would otherwise fall foul of s 10(1). If the respondent can prove that the utterance falls within the ambit of s 12, the utterance must be absolved from scrutiny despite meeting the s 10(1) threshold test. South African Human Rights Commission v Qwelane 2018 2 SA 149 (GJ) (hereafter the Qwelane Equality Court case) para 65. 35 Section 1 of the PEPUDA enumerates: "(a) race, gender, sex, pregnancy, marital status, ethnic or social origin, colour, sexual orientation, age, disability, religion, conscience, belief, culture, language, birth and HIV/AIDS status; or (b) any other ground where discrimination based on that other ground (i) causes or perpetuates systemic disadvantage; (ii) undermines human dignity; or (iii) adversely affects the equal enjoyment of a person's rights and freedoms in a serious manner that is comparable to discrimination on a ground set out in (a)." 36 Qwelane v South African Human Rights Commission 2020 2 SA 124 (SCA) (hereafter Qwelane SCA). 37 Qwelane SCA paras 1, 36. 38 Qwelane SCA para 96. 39 Qwelane SCA para 96. 40 Section 10(1) of the PEPUDA as amended in the interim. 41 Qwelane Equality Court case para 53. 42 LRC Memorandum para 6. 43 Khumalo case para 100. 43 Khumalo case para 98; also see South African Human Rights Commission obo South African Jewish Board of Deputies v Masuku 2018 3 SA 291 (GJ) (hereafter the Masuku case) para 7. 44 Khumalo case para 88. 45 Khumalo case paras 88, 90. 46 Delange v Costa 1989 2 SA 857 (A) 860-861; Le Roux v Dey 2011 3 SA 274 (CC) para 143. 47 Currie and De Waal Bill of Rights Handbook 256. 48 LRC Memorandum para 47. 49 LRC Memorandum para 48. 50 Notably, this case was decided before Qwelane SCA. 51 Mandela Foundation case para 1. For a discussion of the case, see Herd 2019 PSLR. 52 Mandela Foundation case para 167. 53 Mandela Foundation case paras 167-168. 54 In the Masuku case the equality court found that Anti-Zionist statements made by Masuku while he was delivering a speech at the University of the Witwatersrand qualified as hate speech. However, on appeal the Supreme Court of Appeal held that Masuku's right to freedom of expression ensured in s 16 of the Constitution trumped the statements made. Even though his statements were hurtful of people's feelings, or wounding, offensive, politically inflammatory or blatantly offensive, they did not constitute hate speech given the circumstances that prevailed at the time and the place where they were made. Therefore, he could not be deprived of his constitutional protection. Masuku v South African Human Rights Commission obo South African Jewish Board of Deputies 2019 2 SA 194 (SCA) paras 27 and 31; also see the discussion in fn 263. 55 For instance, in the Qwelane case the offending utterances were made in an article in the Sunday Sun newspaper. 56 See for instance Duncanmec (Pty) Limited v Gaylard 2018 6 SA 335 (CC) (hereafter the Duncanmec case); Human Rights Commission of South Africa v SABC 2003 1 BCLR 92 (BCCSA); AfriForum v Malema 2011 6 SA 240 (EqC) (hereafter the Afriforum case). 57 West Virginia State Board of Education v Barnette 319 US 624 1943 632. 58 Mandela Foundation case para 128. 59 Mandela Foundation case para 141. 60 Mandela Foundation case paras 84-89. 61 Mandela Foundation case para 90. 62 The old flag may still be displayed in instances where it would serve a journalistic, academic or artistic purpose in the public interest. Mandela Foundation case para 56. 63 In terms of s 10(1) of the PEPUDA. 64 Section 11 of the PEPUDA. 65 As envisaged in s 7 of the PEPUDA. 66 Botha and Kok 2019 SAPL 33-34. 67 Article 4 of the International Convention on the Elimination of All Forms of Racial Discrimination (1969); the Recommendation para 7. 68 Botha and Govindjee 2017 PELJ 28. 69 These criteria were laid down by the Canadian Supreme Court in Grant v Torstar Corporation 2009 3 SCR 640 para 47; also see The Citizen 1978 (Pty) Ltd v McBride 2011 4 SA 191 (CC) fn 120, where they were cited with approval. 70 LRC Memorandum paras 11-12. 71 LRC Memorandum para 17. 72 De Reuck v Director of Public Prosecutions, Witwatersrand Local Division 2004 1 SA 406 (CC). 73 De Reuck v Director of Public Prosecutions, Witwatersrand Local Division 2004 1 SA 406 (CC) para 48. The court cites: "truth-seeking, free political activity and self-fulfilment" (sic). 74 Section 16(1) of the Constitution determines: "Everyone has the right to freedom of expression, which includes - (a) freedom of the press and other media; (b) freedom to receive or impart information or ideas; (c) freedom of artistic creativity; and (d) academic freedom and freedom of scientific research." 75 LRC Memorandum para 11. 76 South African Human Rights Commission v Qwelane 2018 2 SA 149 (GJ) (Qwelane Equality Court case). 77 Qwelane Equality Court case para 9. 78 Qwelane Equality Court case para 10. 79 Qwelane Equality Court case para 52. 80 The potential that an utterance can spark meaningful and necessary discourse is cited as an important consideration in determining whether it should enjoy constitutional protection. See Herd 2019 PSLR 131. 81 Qwelane Equality Court case para 53. 82 Qwelane SCA. 83 See the discussion in 2.2. 84 Qwelane SCA para 96. 85 OHCHR 2013 https://www.ohchr.org/EN/Issues/Education/Training/Compilation/Pages/d)GeneralrecommendationNo35Combatingracisthatespeech(2013).aspx . 86 LRC Memorandum para 51. 87 LRC Memorandum para 13. 88 So, for instance, denying the holocaust is not viewed as a protectable historical representation. See Garaudy v France App No 65831/01 (ECtHR 3 July 2003); also see LRC Memorandum para 16. 89 He had stated: "We [the EFF] are not calling for the slaughter of white people, at least for now ... The rightful owners of the land are black people. No white person is a rightful owner of the land here in SA and the whole of the African continent." 90 SAHRC Findings para 16; also see the discussion under 2.2.4. 91 SAHRC Findings para 16. 92 SAHRC Findings para 16.2; Hotz v University of Cape Town 2017 2 SA 485 (SCA) (hereafter the Hotz case) para 67. 93 SAHRC Findings para 17. 94 Masuku case para 45; Botha and Govindjee 2017 PELJ 28. 95 Masuku case para 45; Botha and Govindjee 2017 PELJ 28 for similar reasons do not support importing a "truth" defence for use in hate speech cases. 96 Borgin v De Villiers 1980 3 SA 556 (A); also see Iyer 2018 Speculum Juris 126, 130. 97 For an opposing view to the effect that truth seeking and self-fulfilment are incremental considerations in assessing an utterance, see Herd 2019 PSLR 134135. 98 Khumalo case para 93. 99 Freedom Front v South African Human Rights Commission 2003 11 BCLR 1283 (SAHRC). 100 Qwelane SCA para 70. 101 Khumalo case para 94. Also see Qwelane SCA para 70, where the court agreed that the "harm" envisaged s 10(1) of the PEPUDA (as also is the case in s 16(2)(c) of the Constitution) need not necessarily be physical but may include psychological harm. 102 Khumalo case para 102. 103 Crown Chickens (Pty) Ltd t/a Rocklands Poultry v Kapp 2002 6 BLLR 493 (LAC) para 35. 104 Also see the Duncanmec case para 7. 105 Qwelane Equality Court case para 53. However, the proposal in the Qwelane Equality Court case to read the requirements as "hurtful and harmful and promote hatred" has not been applied in this manner by the SAHRC. The SAHRC in 2019 noted that no precedent exists as to whether a conjunctive reading or a disjunctive reading is preferable. See SAHRC Findings 7-15. 106 LRC Memorandum paras 22-24. 107 LRC Memorandum para 17. 108 Even though this case was not a hate speech case under the PEPUDA and other considerations may apply in the employment context, it provides valuable insight. The PEPUDA expressly excludes considerations of fairness from the test for hate speech (see ss 14 and 15 of the PEPUDA). However, the fact that the expression in this case was not hate speech based on race was argued to be a reason why the dismissal of the workers for singing the struggle song was unfair. Duncanmec case para 24. 109 Duncanmec case paras 1, 10. 110 Duncanmec case para 16. 111 Duncanmec case para 17. 112 Duncanmec case para 18. This serves as an example of how the manner in which the expression is made is considered in assigning meaning. See the discussion under 2.5. 113 Duncanmec case paras 18-19. 114 Duncanmec case para 21. 115 Duncanmec case para 17. 116 Duncanmec case para 24. 117 Duncanmec case para 25. 118 Duncanmec case para 27. 119 Duncanmec case para 37. 120 Duncanmec case paras 17, 37. 121 Duncanmec case paras 38-39. 122 Laing 2011 https://www.telegraph.co.uk/news/worldnews/africaandindianocean/southafrica/8757630/Julius-Malema-found-guilty-of-hate-speech-for-singing-Shoot-the-Boer.html . 123 SAHRC Findings para 13.2; see also Naidoo 2019 https://city-press.news24.com/Voices/hate-speech-human-rights-commission-did-not-consider-malemas-influence-20190409 . 124 Discussed under 2.2.3. 125 SAHRC Findings para 10.1. 126 SAHRC Findings para 5.7; Naidoo 2019 https://city-press.news24.com/Voices/hate-speech-human-rights-commission-did-not-consider-malemas-influence-20190409 . 127 SAHRC Findings para 8.1.2. 128 SAHRC Findings para 9.4.2. 129 Naidoo 2019 https://city-press.news24.com/Voices/hate-speech-human-rights-commission-did-not-consider-malemas-influence-20190409 . 130 SAHRC Findings para 12.1. 131 SAHRC Findings para 12.2. 132 Naidoo 2019 https://city-press.news24.com/Voices/hate-speech-human-rights-commission-did-not-consider-malemas-influence-20190409 . 133 Naidoo 2019 https://city-press.news24.com/Voices/hate-speech-human-rights-commission-did-not-consider-malemas-influence-20190409 . 134 Feketha 2019 https://www.iol.co.za/news/politics/malema-hate-speech-ruling-furore-20131591 . 135 Feketha 2019 https://www.iol.co.za/news/politics/malema-hate-speech-ruling-furore-20131591 . 136 Feketha 2019 https://www.iol.co.za/news/politics/malema-hate-speech-ruling-furore-20131591 . 137 Jana 2019 https://www.politicsweb.co.za/documents/malemas-white-slaughter-remarks-the-sahrcs-finding . 138 The Public Protector is currently investigating this matter further, even though Zille no longer holds office. Maughan 2019 https://www.businesslive.co.za/bd/politics/2019-08-07-helen-zille-and-busisiwe-mkhwebane-to-square-off-over-colonialism-tweets/ . 139 Maughan 2019 https://www.businesslive.co.za/bd/politics/2019-08-07-helen-zille-and-busisiwe-mkhwebane-to-square-off-over-colonialism-tweets/ . 140 Eyewitness News 2019 https://www.msn.com/en-za/news/politics/sahrc-to-take-malema-to-court-for-hate-speech/ar-AAHt7Fp?li=BBqfP3n . 141 Masuku case para 11. 142 LRC Memorandum para 41. 143 LRC Memorandum paras 20-22. 144 LRC Memorandum paras 20-22. 145 Weber Manual on Hate Speech 37; LRC Memorandum paras 20-22. 146 Strydom v Black First Land First 2019 ZAEQC 1 (6 May 2019). 147 Mabuza 2019 https://www.sowetanlive.co.za/news/south-africa/2019-01-29-blf-hate-speech-accusations-to-be-heard-in-equality-court/ . 148 SAHRC 2019 https://sahrc.org.za/index.php/sahrc-media/news-2/item/1907-media-statement-the-south-african-human-rights-commission-welcomes-the-equality-court-s-decision-on-the-blf-and-hate-speech ; Shange 2018 https://www.heraldlive.co.za/news/2018-07-19-blf-in-court-for-hate-speech-against-whites/ ; Lindeque 2019 https://ewn.co.za/2019/05/06/blf-ordered-to-remove-slogan-after-equality-court-hate-speech-ruling . 149 Pijoos 2019 https://www.timeslive.co.za/news/south-africa/2018-12-10-watch--you-kill-one-black-person-we-kill-five-white-people-blf-president/ . 150 Pijoos 2019 https://www.timeslive.co.za/news/south-africa/2018-12-10-watch--you-kill-one-black-person-we-kill-five-white-people-blf-president/ . 151 ANA Reporter 2019 https://www.iol.co.za/news/politics/driehoekhighschool-da-youth-reports-blf-to-sahrc-over-most-vile-remarks-19095305 . 152 Pijoos 2019 https://www.timeslive.co.za/politics/2019-02-03-blf-to-be-reported-to-human-rights-commission-over-racist-horskool-driehoek-remarks/ . 153 Penny and Nqola 2019 https://ewn.co.za/2019/02/05/sahrc-appalled-by-racist-comments-on-hoerskool-driehoek-tragedy . 154 Beukes 2019 https://www.pressreader.com/search?query=BLF%20vier%20die&languages=en&groupBy=Language&hideSimilar=0&type=1&state=1 . 155 Unfortunately, the Equality Court case has not been reported. Consequently, we are reliant on news reports for information regarding the outcome of the case. See Mitchley 2019 https://www.news24.com/SouthAfrica/News/nullified-equality-court-judgment-against-blf-leaders-paves-way-for-criminal-case-solidarity-20191204 . 156 Section 14 of the PEPUDA. 157 Section 15 of the PEPUDA. 158 SAHRC Findings 15; also see Grootboom 2019 PSLR 101. 159 Canada (Human Rights Comm) v Taylor 1990 13 CHRR D/435 (SCC). 160 Compare Eatock v Bolt 2011 FCA 1130 (Eatock) paras 243-252. The court follows a similar approach. 161 Bohler-Muller and Tait 2000 Obiter 406, 410. 162 Preamble and ss 3(1)(a), 4(2) and 14(2)(a) of the PEPUDA; also see Kok 2008 SAJHR 446 fn 26. 163 Herselman v Geleba 2011 ZAEQC 1 (1 September 2011). 164 See too Eatock paras 243-252, 273; LRC Memorandum para 39. 165 Carney 2014 Language Matters 330. 166 Masuku case para 3. Although the Supreme Court of Appeal in Masuku v South African Human Rights Commission obo South African Jewish Board of Deputies 2019 2 SA 194 (SCA) did not expressly deal with this aspect, it appears from its judgment that this line of argument may not hold up. 167 Section 18C of the Australian Racial Discrimination Act 52 of 1975 declares it unlawful "to offend, insult, humiliate or intimidate another person or a group of people" based on race. 168 For a discussion of the drafting history, see Gutto Equality and Non-discrimination 17-95. 169 Section 2 of the PEPUDA. 170 Botha and Govindjee identify as a shortcoming of the hate speech protection in the PEPUDA that to "communicate" hatred is insufficient to establish liability. They suggest replacing "communicate" with "advocate" and that instead of banning the "communication" of "words" that are "hurtful" directed at "individuals" the hate speech protection should protect vulnerable groups against utterances that promote hatred against them, and which are likely to cause harm. Botha and Govindjee 2017 PELJ 27. 171 Saskatchewan Human Rights Code SS 1979, c S-24.1. 172 Saskatchewan (Human Rights Commission) v Whatcott 2013 1 SCR 467 para 92 et seq. 173 Saskatchewan (Human Rights Commission) v Whatcott 2013 1 SCR 467 para 80. 174 Islamic Unity Convention v Independent Broadcasting Authority 2002 4 SA 294 (CC). 175 Islamic Unity Convention v Independent Broadcasting Authority 2002 4 SA 294 (CC) paras 29-30, 33, 46. 176 Islamic Unity Convention v Independent Broadcasting Authority 2002 4 SA 294 (CC) para 53. 177 Saskatchewan (Human Rights Commission) v Whatcott 2013 1 SCR 467 para 82. 178 GroundUp Staff 2016 https://www.groundup.org.za/article/max-price-Maasdorp-assault-claims/ . 179 Section 1 of the PEPUDA. 180 Waldron 2010 Harv L Rev 1596, 1626; LRC Memorandum paras 18, 27. 181 LRC Memorandum para 7. 182 Masuku para 54. 183 Jana 2019 https://www.politicsweb.co.za/documents/malemas-white-slaughter-remarks-the-sahrcs-finding para 10. 184 Ho 2019 https://www.dailymaverick.co.za/article/2019-03-14-hate-speech-case-a-message-about-racism-discrimination/ . 185 Refer to the Cape Chinese Exclusion Act, 1904 and the Transvaal Immigration Restriction Act, 1902. 186 See the definition is s 1 of the Employment Equity Act 55 of 1998 (the EEA). 187 Chinese Association of South Africa v Minister of Labour 2008 ZAGPHC 174 (18 June 2008). 188 LRC Memorandum para 21. 189 LRC Memorandum para 20. 190 Dagane v SSSBC 2018 7 BLLR 669 (LC) (the Dagane case). 191 His utterances included: "F*** this white racist s**t! We must introduce Black apartheid. Whites have no ROOM in our heart and mind. Viva MALEMA" and "When the Black Messiah (NM) [Nelson Mandela] dies, we'll teach whites some lesson, We'll commit a genocide on them. I hate whites". 192 Dagane case para 49; also see Botha 2018 THRHR 671, 673. 193 Maroela Redaksie 2019 https://maroelamedia.co.za/nuus/sa-nuus/eff-beaam-zindzi-mandela-se-uitlatings/ . 194 AfriForum 2019 https://allafrica.com/stories/201906170365.html . 195 Cornelissen 2019 https://maroelamedia.co.za/nuus/sa-nuus/klagte-teen-zindzi-mandela-by-menseregtekommissie-ingedien/ . 196 Jacaranda FM 2019 https://sahrc.org.za/index.php/sahrc-media/news/item/2030-sahrc-will-consider-probe-into-zindzi-mandela-tweets . 197 Sparrow on Facebook had compared black holiday-makers on the beach to "monkeys". eNCA 2016 https://www.enca.com/south-africa/penny-sparrow-feels-twitter-wrath . Also see Lebowa Platinum Mines Ltd v Hill 1998 19 ILJ 1112 (LAC) paras 12, 41, 58; and Kok 2009 SAPL 655. 198 This is considered an epithet that led to the Tutsi genocide in Rwanda. O'Grady 2015 https://foreignpolicy.com/2016/04/15/rwandan-who-called-tutsis-cockroaches-in-1992-gets-life-sentence/ . Morota posted on his Facebook account: "I hate white people and must go back wherever they come from or alternatively to hell (sic)." He also referred to white people as white cockroaches. Burger 2016 https://nuus.info/dosent-na-wit-kokkerotte-nog-ongestraf ; Burger 2016 https://www.netwerk24.com/Nuus/Politiek/ernstig-gewaarsku-oor-wit-kakkerlakke-20161022 ; Natasha Stop White Genocide 2016 https://stopwhitegenocideinsareports.blogspot.com/2016/06/black-unisa-lecturer-who-called-white.html . 199 South African Breweries (Pty) Ltd v Hansen (unreported) case number CA6/2016 of 25 May 2017; SAEWU v Rustenburg Platinum Mines LAC (unreported) case number JA 45/2016 of 3 May 2017; South African Revenue Service v Commission for Conciliation, Mediation and Arbitration 2017 38 ILJ 97 (CC) paras 4, 7-9, 53; Thembani v Swanepoel 2017 3 SA 70 (ECM); Crown Chickens (Pty) Ltd t/a Rocklands Poultry v Kapp 2002 6 BLLR 493 (LAC) para 35. 200 Khumalo agrees that uttering the k-word is "inherently racist irrespective of the context." Khumalo 2018 SA Merc LJ 392. 201 Khumalo 2018 SA Merc LJ 392. Jordaan 2019 https://select.timeslive.co.za/news/2019-08-23-final-word-on-the-k-word-its-not-ok-whether-youre-black-or-white/ ; Maphanga 2019 https://www.netwerk24.com/Nuus/Hof/man-wat-skuldig-is-na-k-woord-wil-appelleer-20190822 . Compare Hagan v Trustees of the Toowoomba Sports Grounds Trust 2000 FCA 1615 para 7. 202 Freedom Front v South African Human Rights Commission 2003 11 BCLR 1283 (SAHRC) 1290; Makhanya v St Gobain 2019 7 BALR 720 (NBCCI). However, in contrast, see the Duncanmec case para 37. 203 Carney 2014 Language Matters 328. 204 Dyonashe v Siyaya Skills Institute (Pty) Ltd 2018 3 BALR 280 (CCMA). 205 The Constitutional Court in the Duncanmec case para 37 held that "boer" is not racially offensive. However, compare Freedom Front v South African Human Rights Commission 2003 11 BCLR 1283 (SAHRC) 1290, in which the court held that "boer" is derogative. 206 AfriForum case para 108. 207 AfriForum case para 108. 208 Hotz case para 55. 209 Saeed Semantics 242. 210 Saeed Semantics 242. 211 Goffman Interaction Ritual 5; Carney 2014 Language Matters 334. 212 Carney 2014 Language Matters 327, 329. 213 LRC Memorandum para 33. 214 Janney and Arndt "Intracultural Tact versus Intercultural Tact" 25. 215 LRC Memorandum para 53. 216 LRC Memorandum para 60. 217 Mandela Foundation case paras 62, 79. 218 Vajnai v Hungary App No 33629/06 (ECtHR 8 July 2008) (hereafter the Vajnai case) paras 52-53. 219 Vajnai case paras 51, 54, 57; also see Fratanolo v Hungary App No 29459/10 (ECtHR 3 November 2011) para 25. LRC Memorandum paras 33-34. In these cases the court banned the display of the red start only in so far as it was used to propagate totalitarian ideology. 220 LRC Memorandum paras 14-16. 221 Declaration of the Rights of Persons belonging to National, Ethnic, Religious and Language Minorities (1992) 5. Also see Brink and Mulder 2017 https://solidariteit.co.za/wp-content/uploads/2017/10/Racism-hate-speech-and-double-standards-by-no-means-a-mere-matter-of-bla.pdf . 222 LRC Memorandum para 43. 223 Rustenburg Platinum Mine and SAEWA obo Meyer Bester 2018 8 BLLR 735 (CC) (hereafter the Bester case) para 50. 224 Njilo 2019 https://www.timeslive.co.za/news/south-africa/2019-06-04-racist-licensing-official-in-viral-video-has-been-dismissed/?utm_campaign=recommen-dation&utm_medium=banner&utm_source=newsroom . 225 LRC Memorandum para 46. 226 Bester case para 50. 227 Bester case para 50; Mohammed v Jassiem 1996 1 SA 673 (SCA) 711. 228 LRC Memorandum paras 44-46. 229 LRC Memorandum para 47. 230 R v Ahenakew 2006 SKQB 17-18; R v Bahr 2006 ABPC 360 para 30. 231 Minister of Home Affairs v Fourie 2006 1 SA 524 (CC) para 138. 232 Compare s 319(1) of the Canadian Criminal Code RSC 1985, c C-46, which sets as a requirement for hate speech that the statements must be communicated "in a public place". In s 319(7) "public place" is defined as somewhere that the public has the right to access or where someone is in attendance by invitation, whether the invitation is expressed or implied. Section 18C of the Australian Racial Discrimination Act 52 of 1975 makes it an offense "otherwise than in private" to make offensive, insulting or humiliating statements. 233 Khumalo case paras 20, 69. 234 Section 19 of the PEPUDA. 235 As mentioned above, this is a contravention of s 19(2) of the PEPUDA as highlighted under 2.3. Khumalo case para 20. 236 Khumalo case para 69. 237 Khumalo case para 24.2. 238 Khumalo case para 69. 239 Brink and Mulder 2017 https://solidariteit.co.za/wp-content/uploads/2017/10/Racism-hate-speech-and-double-standards-by-no-means-a-mere-matter-of-bla.pdf . 240 Brink and Mulder 2017 https://solidariteit.co.za/wp-content/uploads/2017/10/Racism-hate-speech-and-double-standards-by-no-means-a-mere-matter-of-bla.pdf . 241 There were 251 reports concerning the utterance by Justin van Vuuren. Brink and Mulder 2017 https://solidariteit.co.za/wp-content/uploads/2017/10/Racism-hate-speech-and-double-standards-by-no-means-a-mere-matter-of-bla.pdf . 242 Daniels 2019 https://www.iol.co.za/news/south-africa/gauteng/agrizzi-sahrc-in-settlement-talks-over-k-word-slur-24348287 . 243 An online search for "Penny Sparrow" and "monkeys" generated 27500 results. Sparrow's matter featured in 4501 media reports. 244 Catzavelos had stated on a video while on holiday in Greece: "Not a f***en [k-word] in sight. Heaven on earth". Lindeque 2018 https://ewn.co.za/2018/08/22/adam-catzavelos-fired-from-family-business-after-racist-video . 245 In the 2017 survey the persons who had made the offensive utterances in the black case studies were mostly highly placed individuals. Among them was the previous President of the Republic of South Africa, a leader of a political party, the rector of the University of the Free State, and the spokesperson of the Ngwathe Municipality. In the cases in which the alleged transgressor was white, media coverage was relatively extensive, even though besides Mabel Jansen, a judge, and Chris Hart, a senior economist, the utterers in the white category of the survey were not particularly influential. Brink and Mulder 2017 https://solidariteit.co.za/wp-content/uploads/2017/10/Racism-hate-speech-and-double-standards-by-no-means-a-mere-matter-of-bla.pdf . 246 In Vodacom (Pty) Ltd v Byrne 2012 33 ILJ 2705 (LC) para 15 Van Niekerk J noted that it is believed that a black person cannot be racist. Also see Dlula 2014 https://m.news24.com/MyNews24/Black-people-cant-be-racist-20141124 ; Mzwakali 2015 https://www.pambazuka.org/governance/black-people-can%E2%80%99t-be-racist . 247 Black South Africans form an overwhelming majority. StatsSA 2019 http://www.statssa.gov.za/publications/P0302/MYPE%202019%20Presentation_final_for%20SG%2026_07%20static%20Pop_1.pdf . 248 The Recommendation paras 39-40. 249 The Recommendation para 12. 250 Qwelane Equality Court case para 53. 251 LRC Memorandum para 49. 252 Trading Economics 2019 https://tradingeconomics.com/south-africa/unemployment-rate . 253 Including the EEA and the Broad-Based Black Economic Empowerment Act 53 of 2003. 254 Pike, Purchert and Chinyamurindi 2018 Acta Commercii 3. 255 Pike, Purchert and Chinyamurindi 2018 Acta Commercii 3; also see De Lange 2019 Rapport 7. 256 Munnik date unknown https://labourguide.co.za/recent-articles/50-new/most-recent-publications/2660-mere-compliance-with-the-employment-equity-act-results-in-no-or-insignificant-ee-transformation ; also see Department of Labour 2019 https://www.labourguide.co.za/workshop/1692-19th-cee-annual-report/file 7,10, 2529. See also Merten 2019 https://www.dailymaverick.co.za/article/2019-08-28-employment-equity-20-years-down-the-line-a-marginal-movement-to-diversity/ . 257 Friedman 2018 https://mg.co.za/article/2018-03-05-land-debate-in-south-africa-is-about-dignity-and-equality-not-the-constitution . 258 Modikwa Mining Personnel Services v CCMA 2012 ZALCJHB 61 (29 June 2012). 259 Modikwa Mining Personnel Services v CCMA 2012 ZALCJHB 61 (29 June 2012) para 35. 260 Modikwa Mining Personnel Services v CCMA 2012 ZALCJHB 61 (29 June 2012) para 33; cf Khumalo 2018 SA Merc LJ 388. 261 See the discussion of the hierarchic approach it followed under 2.2.4 and in dealing with the postings of Mandela-Hlongwane under 2.4. 262 Mandela Foundation case paras 167-168. 263 See the discussion under 2.2. 264 Section 10 of the PEPUDA is a limitation of the right to freedom of speech as enunciated in s 16 of the Constitution. In order to strike a balance between the different constitutional rights at play, s 16(2) of the Constitution expressly excludes harmful expressions from constitutional protection. Notably, s 16(2)(c) of the Constitution excludes the "advocacy of hatred" from the constitutional protection of freedom of speech. The boundaries of the protection of the right to freedom of expression are demarcated in s 16(2). The fact that a specific "expression may be hurtful of people's feelings, or wounding, distasteful, politically inflammatory or downright offensive, does not exclude it from protection." See Masuku v South African Human Rights Commission obo South African Jewish Board of Deputies 2019 2 SA 194 (SCA) paras 14-15, 19 and 31. Also see Grootboom 2019 PSLR 100101. Any limitation of the right to freedom of expression not contemplated in s 16(2) must meet the requirements of the limitation clause in s 36 of the Constitution. See Phillips v DPP, Witwatersrand Local Division 2003 3 SA 345 (CC) para 17. 265 Qwelane Equality Court case para 53. 266 SAHRC Findings para 12.2.

Hate speech on social media

Checked : Olivia S. , Vallary O.

Latest Update 21 Jan, 2024

Table of content

Hardly anybody would march up to a crowd in person and announce their hate for all members of a certain ethnic group, race or religion just out of the blue. However, this is sometimes an everyday occurrence online, where the anonymity of an online profile affords users the confidence to make bold statements on social media.

One would think that in this day and age, we live in relative harmony at least in our real lives. While people may still have their differences, they have mostly learned to keep it civil in the community and at the workplace. Most countries have laws protecting minority groups from discrimination, and there are numerous organizations that champion freedom, tolerance and diversity. However, with the prevalence of hate speech in the online world, one may question if people really are living in harmony these days or if they have simply taken their quarrels elsewhere.

Social media   is a golden platform for connecting people and allowing them to network, share news and air their personal opinions. Unfortunately, it is often also filled with negativity, one of which is hate speech. To share one’s personal views on social media is to open oneself up to a barrage of online attacks, sometimes fighting fire with fire. Many arguments over hate speech escalate into “flame wars”, where most commenters may not even remember what they were originally debating about, but rather resort to personal attacks and sometimes even threats. If words could kill, the participants would probably be dead many times over. Whether hate speech targets a single person or a group of people, it is undeniably destructive and violent.

What constitutes hate speech? There is no definite set of rules that can reliably identify hate speech for what it is. At its core, hate speech tends to attack people for having certain characteristics, such as their race, skin color, ethnic group, religion, gender or sexual orientation – essentially harassing, intimidating or calling for violence against people for who they are. Some hate speech comments may be general – for example, “I hate all members of a certain race”. Others may be targeted at a single person or generalizing all members of a group that the person belongs to. Hate speech can occur in varying severity, from voicing one’s opinion about a certain group or person to threatening them with violence, including statements about wanting to kill them.

There are also other forms of hate speech, such as the posting of violent images or images with such a connotation. For instance, some anti-Semitic groups post images of the Holocaust, including gas chambers, Nazi symbols or Adolf Hitler. Anti-black groups may post gory images from the Jim Crow era of African American victims. Even if no “speech” is technically used in such posts, the message is quite clear and is commonly regarded as hate speech.

The wealth of information on the Internet may have played its part in inciting hate speech. Many individuals still harbor prejudiced feelings towards members of certain demographic groups. For instance, racism, misogyny or homophobia very likely still lurk in the hearts of many, even if it is not visible in their daily lives. The Internet is home to critics, toxicity and “triggers” that can goad the inclined to violence. Additionally, word spreads quickly on social media, allowing people the ability to spread their ideas, threats and violent actions quickly – most of which tend to go viral once the online population fans the flames.

Some sociologists believe that the way social media platforms work is not helping to curb the spread of hate speech either. Social media platforms make money off advertising. These depend on maximizing the user’s browsing time so that more advertisements will be fed to them. Since it is in the social media platforms’ best interests to keep users on the page for as long as possible, algorithms are used to direct users to pages and media they would likely want to browse. This makes it easy for a user to chance upon violent or extremist views, which can then lead them to more similar posts.

Unfortunately, it often turns out that the more time people spend around these kinds of negative sentiments on social media, the more inclined they may be to air their honest views. In fact, the open policy of the Internet encourages people to post and respond freely, especially on social media pages and comments sections which are not moderated. Rarely is hate speech incited on its own. Usually,   media   coverage of an ongoing hot topic tends to provoke responses which are critical and condescending, which then provoke more responses and culminate in a chain reaction of hate speech. The media may also tend to paint some easy targets in a more negative light, resulting in stronger feelings from the public and thus stronger responses.

Under the policy of freedom of speech, hate speech is not illegal. In fact, hate speech is constitutionally protected in the United States under the First Amendment, as long as it does not contain a threat of violence. Most social media platforms and online forums are privately owned, allowing the owners to restrict the content allowed on their platform as per their own terms of service. However, while forums may be more strictly policed, social media tends to be looser in terms of what it defines as hate speech, since it deals with much higher volumes of content and is essentially each user’s own personal “space”.

Even platforms with the strictest policies on hate speech usually have to moderate it manually due to the many forms hate speech can take on. Systems can be trained to recognize word patterns, but it is more difficult to discern the context or intentions of those words. With the vast number of posts on popular social media platforms, it is usually up to individual users to report offensive posts and optionally block the offending user.

We Will Write an Essay for You Quickly

Some may say that hate speech is harmless as long as it remains a verbal assault on the Internet. After all, the Internet is not meant for those with thin skin. Since hate speech is technically protected under freedom of speech, why should people not be allowed to air their personal opinions publicly?

It appears that the prevalence of hate speech can easily provoke people to violence. In recent years, there have been a number of hate crimes that have come about as a result of hate speech on social media. In June 2015, Dylann Roof shot and killed nine African American worshippers at a church in Charleston, thereafter describing his act in vivid detail and without remorse in a video which he posted on social media. It was believed that Roof had “self-radicalized” online through material containing violent white supremacist beliefs on the Internet, leading him to think that an appropriate course of action would be to resort to violence.

The Charleston church shooting is only one of many cases where perpetrators consumed online media that led them to develop extreme beliefs, leading to violent courses of action. While people are certainly entitled to their personal opinions, we should not underestimate the potential of social media in dividing mankind.

Looking for a Skilled Essay Writer?

• Mississippi College Bachelor of Arts (BA)

No reviews yet, be the first to write your comment

Write your review

Thanks for review.

It will be published after moderation

Latest News

What happens in the brain when learning?

10 min read

20 Jan, 2024

How Relativism Promotes Pluralism and Tolerance

Everything you need to know about short-term memory

Researchers leverage AI to fight online hate speech

Any frequent denizen of cyberspace can confirm that online hate speech is a widespread issue. With our daily lives becoming increasingly virtual, and would-be perpetrators emboldened by the anonymity of the digital space, online hate and harassment have risen to unprecedented heights.

A recent survey on online hate and harassment by the Anti-Defamation League shows that over half of American adults report being harassed online at some point in their lives; over a quarter have experienced online harassment just within the last year. “Overall, reports of each type of hate and harassment increased by nearly every measure and within almost every demographic group,” the survey finds.

Efforts to address online hate speech have faced significant hurdles, however. Online content moderation is often the subject of controversy, straddling a fine line between protecting free speech and safeguarding internet users from harm. Emerging innovations are tasked with fostering digital environments free from toxicity and discrimination while also avoiding censorship of inaccurately flagged language.

Rising to this challenge is an innovative solution recently developed by researchers at the University of Michigan and Microsoft, which combines cutting-edge deep learning models with traditional rule-based approaches to better identify hateful speech online. This approach is reported in their paper titled Rule By Example: Harnessing Logical Rules for Explainable Hate Speech Detection , presented at the Annual Meeting of the Association for Computational Linguistics (ACL).

“More and more tech companies and online platforms are developing automated tools to detect and moderate harmful content,” said Christopher Clarke, doctoral student in computer science and engineering at U-M and lead author of the study. “But the methods we’ve seen so far have a lot of room for improvement.” 

Traditional methods for hate speech detection are rule-based, meaning they operate based on set guidelines concerning what constitutes harmful speech, often taking the form of block lists or keyword flagging. Although attractive due to their transparency and customizability, these methods have proven largely insufficient; they are difficult to scale and the rules they rely on to dictate what is flagged do not adequately capture the context and nuances of online content. 

“Detecting hate speech and toxicity is a subjective task, and language is very ambiguous,” said Clarke. “It is relatively easy for users to switch around tokens and bypass a rule-based system, so these methods are quite fragile.”

Data-driven deep learning methods have emerged as a promising alternative for online content moderation. Such models are trained on large amounts of data and leverage deep neural networks to learn richer, more accurate representations and generalize these representations to new data. Despite their improvements in accuracy compared to one-size-fits-all rule-based approaches, the application of deep learning techniques is not without challenges.

“With out-of-the-box, pre-trained models, the user inputs some text and the model generates a prediction and a probability score as to whether the content is hateful or not,” said Clarke. “The main issue with these models is a lack of transparency—users aren’t able to see or understand what the model is learning and how it could be improved.”

The failure of deep learning models to give users any explanation or guidance as to the reasoning behind their choices has hindered their widespread adoption and has added to growing distrust among consumers.

Seeking to harness the enhanced performance of deep learning models while preserving the transparency and customizability of rule-based methods, Clarke and his coauthors developed Rule By Example (RBE), an exemplar-based approach that uses deep learning to compare text inputs with examples of hateful content. 

“RBE is a contrastive learning framework that pairs rules with what we call exemplars, examples of text that defines a given rule,” said Clarke. “The framework pairs logical rules that are very explainable with these exemplars and then encodes and learns them.”

To accomplish this, RBE relies on two neural networks: a rule encoder as well as a text encoder. Using these tools, the model learns robust, accurate embeddings of hateful content and the rules behind them, enabling it to accurately predict and classify online hate speech.

This groundbreaking two-part framework also solves the persistent issue of transparency that other deep learning models demonstrate. It gives users a clear picture of how its predictions are formed as well as the option to revise the rules being used.

“RBE displays what we call rule grounding, allowing users to trace back model predictions directly to the rules that govern them as well as the examples tied to those rules,” said Clarke. 

Instead of only receiving a prediction and probability score, as with other deep learning methods, RBE allows users to see the factors that influence the model’s precision, building in transparency without sacrificing performance.

RBE also boasts exceptional customizability. Its unique grounding feature means that customers can edit the rules and examples influencing predictions in real-time without having to retrain the model.

RBE’s unique two-step approach combines the best of both worlds, preserving the advantages of deep learning models and rule-based techniques, ultimately yielding a hate speech detection system that is robust, accurate, and transparent.

“Our approach not only ensures greater transparency, but also enhances performance,” said Clarke. “RBE outperforms several benchmarks and actually shows better performance than existing hate speech classifiers.”

In fact, compared to the closest competing classifier, RBE showed a 2% increase in accuracy, a substantial difference considering the massive amounts of online data these models process. Together with its built-in transparency mechanism, this performance boost demonstrates RBE’s potential to significantly improve online content moderation and make digital spaces safer for everyone.

The goal is to integrate RBE into Microsoft Cloud, and plans are underway to patent this new technology. In the future, Clarke and his collaborators are also hoping to push RBE’s capabilities even further, testing its ability to accommodate more complex rules and extending it to other classification tasks, such as predicting the intent behind a given piece of text. 

In all, the hope is that RBE and its future iterations will benefit end users by broadening protections against online hate speech and harassment while preserving openness and transparency.

Clarke’s coauthors on the above paper are Prof. Jason Mars of U-M as well as Matthew Hall, Gaurav Mittal, Ye Yu, Sandra Sajeev, and Mei Chen of Microsoft.

Harman Hate Speech And Matthew Desmond's Analysis

Both Jeremy Waldron's book, the Harman Hate Speech, and Matthew Desmond's analysis of the housing crisis in America talk about the harm in social hierarchies due to external factors such as poverty and bias. While Waldron's argument about assurance versus dignity in terms of preserving one's right to free speech emphasizes Americans' constitutional rights, Desmond's eviction is an in-depth portrayal of actual real-life American stories that the justice system has betrayed. Both Harms in Hate Speech and Evicted highlight the importance of injustice and the humanitarian need for survival when faced with the theme of just relationships and individuality. During part three of Desmond's eviction, the readers follow the story of Arlene and her two …show more content…

For Arleen, relationships with others besides her children became a means of survival instead of something that would be a positive force in her life. In Jeremy Waldron's final chapter of The Harm in Hate Speech, he brings up the argument of tolerance and calumny. Toleration is crucial when observing relationships since it allows individuals to gain knowledge of social customs and interactions. While observing the case of Osborne VR, Waldron states, "Is a tolerant society just a free society free from religious persecution, or is it a society in which people cohabit and deal with one another despite their religious differences in an atmosphere of civility and respect, an atmosphere that is not disfigured by grotesque defamations of the sort that we saw in the case of our V Osborne" (Waldron 206). The feeling of tolerance when faced with navigating a struggle in a relationship, especially as Americans, is a testament to how social hierarchies and social expectations play a role in the ideologies of a harmonious

More about Harman Hate Speech And Matthew Desmond's Analysis

The Volokh Conspiracy

Mostly law professors | Sometimes contrarian | Often libertarian | Always independent

• Editorial Independence
• Volokh Daily Email

Free Speech

Call to "Update Free Speech Policies" to Address Supposed Hate Speech at Public University

I'm against it, whomever it's coming from..

Eugene Volokh | 3.29.2024 8:27 AM

I've seen plenty before, but this one is from Texas Gov. Greg Abbott, and the particular target is "antisemitic speech." An excerpt from Wednesday's Executive Order :

WHEREAS, Section 51.9315(f) of the Texas Education Code requires all higher education institutions to adopt policies detailing students' responsibilities regarding free expression on campus; and WHEREAS, Section 51.9315(c)(2) of the Texas Education Code provides that students should not participate in, and higher education institutions should not allow, expression that is unlawful or disrupts the operations of the institution; and WHEREAS, antisemitism and the harassment of Jewish students have no place on Texas university campuses and will not be tolerated by my administration; NOW, THEREFORE, I … hereby direct all Texas higher education institutions to do the following: 1. Review and update free speech policies to address the sharp rise in antisemitic speech and acts on university campuses and establish appropriate punishments, including expulsion from the institution. 2. Ensure that these policies are being enforced on campuses and that groups such as the Palestine Solidarity Committee and Students for Justice in Palestine are disciplined for violating these policies. 3. Include the definition of antisemitism, adopted by the State of Texas in Section 448.001 of the Texas Government Code, in university free speech policies to guide university personnel and students on what constitutes antisemitic speech.

Texas Government Code Section 448.001 incorporates " the International Holocaust Remembrance Alliance's 'Working Definition of Antisemitism,' " which includes (among other things),

Denying the Jewish people their right to self-determination, e.g., by claiming that the existence of a State of Israel is a racist endeavor. Applying double standards by requiring of it a behavior not expected or demanded of any other democratic nation. Using the symbols and images associated with classic antisemitism (e.g., claims of Jews killing Jesus or blood libel) to characterize Israel or Israelis. Drawing comparisons of contemporary Israeli policy to that of the Nazis.

As readers might gather, I oppose those on the Right trying to ban supposed "hate speech" at universities, just as I oppose those on the Left trying to do the same. Of course students have to be free to argue that Jews don't have a right to "self-determination" in the sense of having their own country, just as they have to be free to argue that Palestinians don't have such a right, or Basques, Kurds, Catalonians, North Cypriots, South Ossetians, etc. don't have such a right. I support Israel's right to exist, but that is a subject that has to be up for free debate just like any other subject.

Of course students' free speech rights can't be limited by judgments about when the students are supposedly applying " double standards ." What would we think, for instance, of criticisms of Communist China being punishable if some university administrator concludes that the critic isn't holding some other country to the same standards? Of course they have to be free to analogize the Israeli government or Hamas or the Chinese government or the Russian government or for that matter Trump or Biden to the Nazis, whatever we think of the soundness of such analogies.

Students also have the right to express anti-Semitic sentiments (or anti-Palestinian or anti-trans or anti-gay or anti-black or anti-white or anti-male or anti-female or anti-Catholic or anti-Protestant or anti-Muslim sentiments), just as they can any other sentiments. There is no "hate speech" exception to the First Amendment, and no anti-Semitic speech exception in particular.

The rules are also the same at public universities. Papish v. Board of Curators  (1973) expressly rejected the notion that the university's "legitimate authority to enforce reasonable regulations as to the time, place, and manner of speech and its dissemination" extends to "disapproved  content … rather than the time, place, or manner of its distribution." And  Papish expressly held that "the First Amendment leaves no room for the operation of a dual standard in the academic community with respect to the content of speech," which is to say a different standard than the First Amendment applies to the government as regulator. Or, to quote  Healy v. James   (1972),

[T]he precedents of this Court leave no room for the view that, because of the acknowledged need for order, First Amendment protections should apply with less force on college campuses than in the community at large. Quite to the contrary, "[t]he vigilant protection of constitutional freedoms is nowhere more vital than in the community of American schools."

As I've said before, universities certainly have the power to punish trespassing, blocking entrances, excessive noise, vandalism, and a wide range of other behavior, so long as they do it apart from content. They likewise have the power to punish true threats of illegal conduct, solicitation of illegal conduct, and the like, so long as they do it evenhandedly.  They should generally use this power, and at times they haven't used it enough.

But they can't target "antisemitic speech" based on its antisemitic sentiments, and punish it with "expulsion" or for that matter with lesser punishment. And they certainly can't define criticism of Israel, however misguided or even ill-motivated, as punishable "antisemitic speech."

Be Skeptical of That 60 Minutes Report on Havana Syndrome

Robby Soave | 4.4.2024 11:46 AM

Licenses and Dead Bodies

Liz Wolfe | 4.4.2024 9:30 AM

With Trump and Biden on the Ballot, Many Voters Prefer 'None of the Above'

Katherine Mangu-Ward | From the May 2024 issue

Brickbat: Walking the Dog(s)

Charles Oliver | 4.4.2024 4:00 AM

Politicians Are Showering Manufacturing Companies With Crony Subsidies for 'Job Creation.' It Won't Work.

Veronique de Rugy | 4.4.2024 12:01 AM

Recommended

Accessibility links

• Skip to main content
• Skip to main Navigation
• Skip to Search

Scotland’s contentious new hate crime law may impact free speech

• By Associated Press

A new law against hate speech came into force in Scotland on Monday, praised by some but criticized by others who say its sweeping provisions could criminalize religious views or tasteless jokes.

The Hate Crime and Public Order (Scotland) Act makes it an offense to stir up hatred with threatening or abusive behavior based on characteristics including age, disability, religion, sexual orientation and transgender identity. Racial hatred was already banned under a law dating from 1986.

The maximum sentence is seven years in prison.

The legislation does not specifically ban hatred against women. The Scottish government says that will be tackled by a separate forthcoming law against misogyny.

Scottish Minister for Victims and Community Safety Siobhian Brown said the new law would help build “safer communities that live free from hatred and prejudice.”

“We know that the impact on those on the receiving end of physical, verbal or online attacks can be traumatic and life-changing,” she said. “This legislation is an essential element of our wider approach to tackling that harm.”

Critics argue that the law will have a chilling effect on free speech, making people afraid to express their views. The legislation was passed by the Scottish Parliament almost three years ago but has been delayed by wrangling over its implementation.

Veteran human rights activist Peter Tatchell said the law was well-intended but vague, relying on “subjective interpretation” of what constitutes abuse and allowing people to report alleged offenses anonymously.

The Scottish National Party-led government in Edinburgh says the legislation includes free speech protections, including a specific guarantee that people can still “ridicule or insult” religion.

“The threshold of criminality in terms of the new offenses is very, very high indeed,” First Minister Humza Yousaf said. “Your behavior has to be threatening or abusive and intended to stir up hatred.”

“Harry Potter” author J.K. Rowling, who has called the law “ludicrous,” is among critics who say it could be used to silence what are known as “gender-critical” feminists, who argue that rights for trans women should not come at the expense of those who are born biologically female.

In a series of posts on X, formerly Twitter, Rowling referred to several prominent trans women as men. Misgendering could be an offense under the new law in some circumstances.

“I’m currently out of the country, but if what I’ve written here qualifies as an offence under the terms of the new act, I look forward to being arrested when I return to the birthplace of the Scottish Enlightenment,” Rowling wrote.

Scottish National Party lawmaker Joanna Cherry, another critic of the law, said that “if you are a woman, you have every right to be concerned.”

“Biological sex is not included as a protected characteristic in the act, despite women being one of the most abused cohorts in our society,” she wrote in The National newspaper.

Meanwhile, police organizations are concerned the law will trigger a flood of reports over online abuse.

David Kennedy, general secretary of the Scottish Police Federation, said the law could “cause havoc with trust in police.” And the Association of Scottish Police Superintendents wrote to lawmakers to express worry that the law could be “weaponized” by an “activist fringe.”

The law is the latest case of Scotland's semi-autonomous government, which is led by the pro-independence SNP, diverging from the Conservative U.K. administration in London. In 2022, the Scottish Parliament passed a law allowing people to change their legally recognized gender through self-declaration, without the need for medical certification.

The gender-recognition legislation was vetoed by the British government, which said it conflicted with U.K.-wide equalities legislation that, among other things, guarantees women and girls access to single-sex spaces such as changing rooms and shelters.

More Europe News

Spain's decision to recognize Palestinian state marks potential turning point for Europe

Biden heralds 75th anniversary of NATO’s founding

Macron believes france, allies 'could have stopped' 1994 rwanda genocide, russia arrests 3 more over moscow concert hall attack , ukrainian officials say deadly russian drone attack hit kharkiv, the day in photos.

April 4, 2024

• Share full article

Advertisement

Supported by

Guest Essay

A.I.-Generated Garbage Is Polluting Our Culture

By Erik Hoel

Mr. Hoel is a neuroscientist and novelist and the author of The Intrinsic Perspective newsletter.

Increasingly, mounds of synthetic A.I.-generated outputs drift across our feeds and our searches. The stakes go far beyond what’s on our screens. The entire culture is becoming affected by A.I.’s runoff, an insidious creep into our most important institutions.

Consider science. Right after the blockbuster release of GPT-4, the latest artificial intelligence model from OpenAI and one of the most advanced in existence, the language of scientific research began to mutate. Especially within the field of A.I. itself.

Adjectives associated with A.I.-generated text have increased in peer reviews of scientific papers about A.I.

Frequency of adjectives per one million words

Commendable

A study published this month examined scientists’ peer reviews — researchers’ official pronouncements on others’ work that form the bedrock of scientific progress — across a number of high-profile and prestigious scientific conferences studying A.I. At one such conference, those peer reviews used the word “meticulous” more than 34 times as often as reviews did the previous year. Use of “commendable” was around 10 times as frequent, and “intricate,” 11 times. Other major conferences showed similar patterns.

Such phrasings are, of course, some of the favorite buzzwords of modern large language models like ChatGPT. In other words, significant numbers of researchers at A.I. conferences were caught handing their peer review of others’ work over to A.I. — or, at minimum, writing them with lots of A.I. assistance. And the closer to the deadline the submitted reviews were received, the more A.I. usage was found in them.

If this makes you uncomfortable — especially given A.I.’s current unreliability — or if you think that maybe it shouldn’t be A.I.s reviewing science but the scientists themselves, those feelings highlight the paradox at the core of this technology: It’s unclear what the ethical line is between scam and regular usage. Some A.I.-generated scams are easy to identify, like the medical journal paper featuring a cartoon rat sporting enormous genitalia. Many others are more insidious, like the mislabeled and hallucinated regulatory pathway described in that same paper — a paper that was peer reviewed as well (perhaps, one might speculate, by another A.I.?).

What about when A.I. is used in one of its intended ways — to assist with writing? Recently, there was an uproar when it became obvious that simple searches of scientific databases returned phrases like “As an A.I. language model” in places where authors relying on A.I. had forgotten to cover their tracks. If the same authors had simply deleted those accidental watermarks, would their use of A.I. to write their papers have been fine?

What’s going on in science is a microcosm of a much bigger problem. Post on social media? Any viral post on X now almost certainly includes A.I.-generated replies, from summaries of the original post to reactions written in ChatGPT’s bland Wikipedia-voice, all to farm for follows. Instagram is filling up with A.I.-generated models, Spotify with A.I.-generated songs. Publish a book? Soon after, on Amazon there will often appear A.I.-generated “workbooks” for sale that supposedly accompany your book (which are incorrect in their content; I know because this happened to me). Top Google search results are now often A.I.-generated images or articles. Major media outlets like Sports Illustrated have been creating A.I.-generated articles attributed to equally fake author profiles. Marketers who sell search engine optimization methods openly brag about using A.I. to create thousands of spammed articles to steal traffic from competitors.

Then there is the growing use of generative A.I. to scale the creation of cheap synthetic videos for children on YouTube. Some example outputs are Lovecraftian horrors, like music videos about parrots in which the birds have eyes within eyes, beaks within beaks, morphing unfathomably while singing in an artificial voice, “The parrot in the tree says hello, hello!” The narratives make no sense, characters appear and disappear randomly, and basic facts like the names of shapes are wrong. After I identified a number of such suspicious channels on my newsletter, The Intrinsic Perspective, Wired found evidence of generative A.I. use in the production pipelines of some accounts with hundreds of thousands or even millions of subscribers.

As a neuroscientist, this worries me. Isn’t it possible that human culture contains within it cognitive micronutrients — things like cohesive sentences, narrations and character continuity — that developing brains need? Einstein supposedly said : “If you want your children to be intelligent, read them fairy tales. If you want them to be very intelligent, read them more fairy tales.” But what happens when a toddler is consuming mostly A.I.-generated dream-slop? We find ourselves in the midst of a vast developmental experiment.

There’s so much synthetic garbage on the internet now that A.I. companies and researchers are themselves worried, not about the health of the culture, but about what’s going to happen with their models. As A.I. capabilities ramped up in 2022, I wrote on the risk of culture’s becoming so inundated with A.I. creations that when future A.I.s are trained, the previous A.I. output will leak into the training set, leading to a future of copies of copies of copies, as content became ever more stereotyped and predictable. In 2023 researchers introduced a technical term for how this risk affected A.I. training: model collapse . In a way, we and these companies are in the same boat, paddling through the same sludge streaming into our cultural ocean.

With that unpleasant analogy in mind, it’s worth looking to what is arguably the clearest historical analogy for our current situation: the environmental movement and climate change. For just as companies and individuals were driven to pollute by the inexorable economics of it, so, too, is A.I.’s cultural pollution driven by a rational decision to fill the internet’s voracious appetite for content as cheaply as possible. While environmental problems are nowhere near solved, there has been undeniable progress that has kept our cities mostly free of smog and our lakes mostly free of sewage. How?

Before any specific policy solution was the acknowledgment that environmental pollution was a problem in need of outside legislation. Influential to this view was a perspective developed in 1968 by Garrett Hardin, a biologist and ecologist. Dr. Hardin emphasized that the problem of pollution was driven by people acting in their own interest, and that therefore “we are locked into a system of ‘fouling our own nest,’ so long as we behave only as independent, rational, free-enterprisers.” He summed up the problem as a “tragedy of the commons.” This framing was instrumental for the environmental movement, which would come to rely on government regulation to do what companies alone could or would not.

Once again we find ourselves enacting a tragedy of the commons: short-term economic self-interest encourages using cheap A.I. content to maximize clicks and views, which in turn pollutes our culture and even weakens our grasp on reality. And so far, major A.I. companies are refusing to pursue advanced ways to identify A.I.’s handiwork — which they could do by adding subtle statistical patterns hidden in word use or in the pixels of images.

A common justification for inaction is that human editors can always fiddle around with whatever patterns are used if they know enough. Yet many of the issues we’re experiencing are not caused by motivated and technically skilled malicious actors; they’re caused mostly by regular users’ not adhering to a line of ethical use so fine as to be nigh nonexistent. Most would be uninterested in advanced countermeasures to statistical patterns enforced into outputs that should, ideally, mark them as A.I.-generated.

That’s why the independent researchers were able to detect A.I. outputs in the peer review system with surprisingly high accuracy: They actually tried. Similarly, right now teachers across the nation have created home-brewed output-side detection methods , like adding hidden requests for patterns of word use to essay prompts that appear only when copied and pasted.

In particular, A.I. companies appear opposed to any patterns baked into their output that can improve A.I.-detection efforts to reasonable levels, perhaps because they fear that enforcing such patterns might interfere with the model’s performance by constraining its outputs too much — although there is no current evidence this is a risk. Despite public pledges to develop more advanced watermarking, it’s increasingly clear that the companies are dragging their feet because it goes against the A.I. industry’s bottom line to have detectable products.

To deal with this corporate refusal to act we need the equivalent of a Clean Air Act: a Clean Internet Act. Perhaps the simplest solution would be to legislatively force advanced watermarking intrinsic to generated outputs, like patterns not easily removable. Just as the 20th century required extensive interventions to protect the shared environment, the 21st century is going to require extensive interventions to protect a different, but equally critical, common resource, one we haven’t noticed up until now since it was never under threat: our shared human culture.

Erik Hoel is a neuroscientist, a novelist and the author of The Intrinsic Perspective newsletter.

The Times is committed to publishing a diversity of letters to the editor. We’d like to hear what you think about this or any of our articles. Here are some tips . And here’s our email: [email protected] .

Follow the New York Times Opinion section on Facebook , Instagram , TikTok , WhatsApp , X and Threads .

• Newsletters
• Account Activating this button will toggle the display of additional content Account Sign out

Conservatives Are Getting Comfortable Talking Openly About a National Abortion Ban

After this week’s oral argument, few court watchers believe the Supreme Court is now ready to limit the Food and Drug Administration’s authority to approve mifepristone , a drug used in more than half of all abortions , as opponents of abortion sought. At oral argument in FDA v. Alliance for Hippocratic Medicine , it did not appear that the plaintiff doctors persuaded the court that the law inflicted injuries that would give them standing to sue. The reason for the justices’ skepticism is not hard to find. The doctors built their case on a mountain of remote possibilities. Patients might suffer complications from mifepristone—a drug with an impressively low complication rate—and might seek treatment at emergency rooms, where the plaintiffs may happen to practice, when the plaintiffs might not be able to find another physician willing to intervene. And all of that might mean that the plaintiffs would have to act in violation of their conscience. But then again, it might not. That’s why this case seems dead on arrival: The justices seemed unwilling to engage in the sort of rank speculation the plaintiffs have in mind. If this chain of hypotheticals is enough, anyone can bring a constitutional challenge to any drug approval or any law.

But the case was also a vehicle for advancing ever more expansive conscience-based arguments that have become common currency among Christian conservatives—claims of the kind we have seen in well-known cases like the 2014 Hobby Lobby decision recognizing conscience objections to the contraceptive mandate of the Affordable Care Act or even last year’s ruling in 303 Creative v. Elenis that allowed a conservative Christian graphic designer to refuse to make custom websites for same-sex weddings.

Today, those with conscience-based objections seek more than to pray or dress in conformity with religious belief. They object to laws providing Americans access to health care or freedom from discrimination. Compliance with these laws, they claim, would make the objector complicit in the assertedly sinful conduct of others.

Objectors bringing this new generation of complicity-based conscience claims invite courts to deny other Americans the protections of the law. In the FDA case, the plaintiffs do not even seek an exemption from the law; through an expansive standing claim, the doctors claim the only way the court could protect their conscience is to strike down FDA approvals providing all Americans access to medication abortion. Simply having mifepristone on the market, they argue, risks making them complicit in abortion.

Justice Ketanji Brown Jackson zeroed in on the problems with this argument. She observed that Erin Hawley, the attorney for the Alliance for Hippocratic Medicine, had identified a “broad” and “narrow” idea of conscience. The “narrow” reading was straightforward: “participating in a procedure.” This reading had problems of its own: In fact, no doctor was obliged to prescribe mifepristone, and in any event, federal law provides doctors conscience protections.

Yet Hawley didn’t think complicity ended there. Jackson seemed confused. Did Hawley mean that a handful of other doctors who participated in post-abortion procedures, such as the removal of tissue, were also complicit? Or was Hawley asking the court to recognize the complicity claims of someone who worked in an emergency room where abortions took place, or handed an abortion provider a water bottle?

Jackson spotlighted a defining feature of “conscience-war” claims that one of us (Reva Siegel), writing with Douglas NeJaime, has identified : Conservatives assert ever-expanding complicity-based conscience claims, urging the government to accommodate their claims without making any provision for other Americans who would lose the protection of law. Appealing to the value of conscience obscures the material and dignitary harm that accommodating the objection inflicts on others.

Solicitor General Elizabeth Prelogar stressed this point: While the plaintiffs could not say their conscience had been or would be harmed, their claim to conscience obscured harm done to a variety of other parties. That includes the FDA, which had its own scientific judgments displaced, and the pharmaceutical industry, which relies on the FDA approval process to ensure some sort of uniform industry standards. First and foremost, it includes, as Prelogar noted, “women who need access to medication abortion .”

Conscience claims have been alluring to conservatives because, like colorblindness, they allow conservatives to speak as a “minority,” and to assert traditional family values as individual freedom claims. But there is a telling shift. When groups like Alliance Defending Freedom asserted complicity-based conscience claims at the time of Hobby Lobby , they worried about losing in a Supreme Court that was far less conservative—and about alienating a Republican Party that still prioritized electability rather than ideological purity.

By contrast, in Alliance for Hippocratic Medicine, ADF talked not only about protecting women or safeguarding conscience; it made claims around the Comstock Act, a symbol of Victorian sexual morality focused not on protecting fetal life but on discouraging illicit sex , that ADF seeks to reinvent as a de facto national abortion ban. ADF argued that FDA could not have had the authority to approve telehealth abortions in 2021 because the Comstock Act bars the mailing of abortion pills—and indeed, any abortion-related item. Justices Clarence Thomas and Samuel Alito both seemed interested in transforming the 1873 Comstock Act into an abortion ban that American voters would never choose to enact. Alito seemed shy about mentioning Comstock by name, instead referring to the hard-to-recognize number in the U.S. Code. Thomas was not so reserved, all but telling attorneys for Danco, the maker of the name-brand mifepristone, that the Comstock Act barred the mailing of the drug.

The very fact that ADF wants to talk about the Comstock Act is remarkable. It seems unwise to hitch the anti-abortion movement’s star to a 19 th -century anti-vice movement known for “Comstockery”: censoring political speech, undermining democratic norms, and condemning any form of sex not intended for procreation .

Voters have already rejected state abortion bans. Just imagine what most Americans would make of it if an already unpopular Supreme Court interpreted a law from 1873 as a sweeping, punitive zombie abortion ban. But worrying about the public’s reaction assumes the movement is seeking to persuade voters rather than simply looking for ways to use power to enforce traditional family values and punish those who become pregnant or might provide them medical care. Anti-abortion groups are planning to revive enforcement of the Comstock Act if Donald Trump wins the presidency , claiming they would not need Congress to act .

The argument in the mifepristone case was a potent reminder of why conservatives have gravitated to conscience claims—and demonstrated the hidden harms that these claims can inflict on other Americans. But the conversation at the Supreme Court this week also suggested that conservatives are preparing to express the values underlying complicity-based conscience claims more openly. As it gains power, members of the anti-abortion movement seem increasingly ready to take off the mask.

Activist investor Nelson Peltz defeated in Disney board battle

Entertainment giant wins the support of enough investors to see off proxy battle

Billionaire activist investor Nelson Peltz has failed in his attempt to secure a seat on Disney’s board after shareholders backed the Hollywood giant instead.

At its annual general meeting on Wednesday, Disney confirmed it had won the support of enough investors to see off a proxy battle mounted by Mr Peltz, who holds a stake in Disney worth $3.5bn (£2.8bn).

The result marks a victory for Bob Iger, the Disney chief executive, who will be re-appointed to the company’s board alongside 11 other existing directors.

Mr Peltz, who holds a stake in Disney through his vehicle Trian Fund Management, waged a fierce campaign against the media giant, arguing its performance was lagging behind rivals and taking aim at its bungled succession planning.

He also lashed out at the diverse casts of Disney’s recent films . He told the Financial Times: “People go to watch a movie or a show to be entertained. They don’t go to get a message.”

Disney last week said Mr Peltz was “ill-equipped to serve on Disney’s board” and that he would “harm Disney and jeopardise our strategic transformation”.

The activist, who is the father-in-law of Brooklyn Beckham, was pushing for board seats for himself and former Disney finance chief Jay Rasulo.

He won the support of influential advisory group Institutional Shareholder Services, which recommended that investors should vote to add the billionaire to Disney’s board.

However, rival advisor Glass Lewis sided with Disney and Mr Iger in the dispute, alongside high-profile figures including JP Morgan boss Jamie Dimon and Star Wars creator George Lucas.

Vanguard Group and BlackRock, Disney’s two largest shareholders, were among the investors to vote in support of the company.

In a statement, Trian said: “While we are disappointed with the outcome of this proxy contest, Trian greatly appreciates all of the support and dialogue we have had with Disney stakeholders. We are proud of the impact we have had in refocusing this company on value creation and good governance. 

“We will be watching the Company’s performance and be focusing on its continued success.”

Mark Parker, chairman of Disney, said: “We are immensely grateful to our shareholders for their investment in Disney and their belief in its future, particularly during this period of great change in the broader entertainment industry.

“We are fortunate to have a highly qualified board of directors who possess a profound commitment to the enduring strength of this company and an enormous amount of experience and expertise, including succession planning.”

The vote also marks a defeat for Blackwells Capital, another activist investor that was seeking three board seats.

Earlier, Elon Musk vowed to take a stake in Disney if Mr Peltz won his battle to win a board seat.

Mr Musk posted on his social media site X that the hedge fund manager “should definitely be on the Disney board”. He added: “While I don’t own any Disney shares today, I would definitely buy their shares if Nelson were elected to the board.”

The Tesla boss has been a vocal critic of Disney and Mr Iger’s leadership over its decision to halt advertising on Twitter, now rebranded as X.

The world’s third richest man has said Mr Iger should be “fired immediately”, called Disney’s recent films “unwatchable”, and claimed that “Walt Disney would despise Bob Iger”.

Wednesday’s result represents a boost to Mr Iger, a long-serving Disney executive who was parachuted back into the top job at the end of 2022.

Mr Iger has slashed around 7,000 jobs as part of a $5.5bn cost-cutting strategy as he aims to steer Disney through the streaming age and turn its fortunes around after a string of disappointing box office results.

• Walt Disney,
• Facebook Icon
• WhatsApp Icon