Hate Speech

First published Tue Jan 25, 2022

Hate speech is a concept that many people find intuitively easy tograsp, while at the same time many others deny it is even a coherentconcept. A majority of developed, democratic nations have enacted hatespeech legislation—with the contemporary United States being anotable 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 manydifficult questions: What does the ‘hate’ in hate speechrefer to? Can hate speech be directed at dominant groups, or is it bydefinition targeted at oppressed or marginalized communities? Is hatespeech always ‘speech’? What is the harm or harms of hatespeech? And, perhaps most challenging of all, what can or should bedone to counteract hate speech?

In part because of these complexities, hate speech has spawned a vastand interdisciplinary literature. Legal scholars, philosophers,sociologists, anthropologists, political theorists, historians, andother academics have each approached the topic with exceedinginterest. In this current article, however, we cannot hope to coverhow these many disciplines have engaged with the concept of hatespeech. Here, we will focus most explicitly on how hate speech hasbeen taken up within philosophy, with particular emphasis on issuessuch as: how to define hate speech; what are the plausible harms ofhate speech; how an account of hate speech might include both overtexpressions of hate (e.g., the vitriolic use of slurs) as well as morecovert, implicit utterances (e.g., dogwhistles); the relationshipbetween hate speech and silencing; and what might we do to counteracthate speech.

1. What is Hate Speech?

The term ‘hate speech’ is more than a descriptive conceptused to identify a specific class of expressions. It also functions asan evaluative term judging its referent negatively and as a candidatefor censure. Thus, defining this category carries seriousimplications. What is it that designates hate speech as a distinctiveclass of speech? Some claim the term ‘hate speech’ itselfis misleading because it wrongly suggests “virulent dislike of aperson for any reason” as a defining feature (Gelber 2017, 619).That is not, however, the way in which the term is understood amongmost legal theorists and philosophers. Perhaps it would be useful tostart with some examples.

Bhikhu Parekh (2012) lists the following instances as examplesdifferent countries have either punished or sought to punish as hatespeech:

Shouting “[N-words] go home,” making monkey noises,and chanting racist slogans at soccer matches.
“Islam out of Britain. Protect the Britishpeople.”
“Arabs out of France.”
“Serve your country, burn down a mosque.”
“Blacks are inherently inferior, lecherous, predisposed tocriminal activities, and should not be allowed to move intorespectable areas.”
“Jews are conspiratorial, devious, treacherous, sadistic,child killers, and subversive; want to take over the country; andshould 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 “undesiredaliens” from the Netherlands.
A poster of a woman in aburka with text that reads:“Who knows what they have under their sinister and ugly lookingclothes: stolen goods, guns, bombs even?”
Speech that either denies or trivializes the holocaust or othercrimes against humanity.

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

In law, we have to define hate speech carefully to designate the formsof the speech that will receive distinctive legal treatment. This isno easy task. Roughly speaking, we can define hate speech in terms ofthe harms it will cause—physical contingent harms like violenceor discrimination; or we can define hate speech in terms of itsintrinsic properties—the kinds of words it uses; or we candefine hate speech in terms of its connection to principles ofdignity; or we can define hate speech in terms of the ideas itconveys. Each of these definitions has advantages and disadvantages.Each intersects with the first amendment theory in a different way. Inthe end, any definition that we adopt must be justified on the groundthat it will achieve the results we wish to achieve. (Herz and Molnar2012, 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 theways 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’ sodefined in order to determine whether it achieves what we wish toaccomplish and whether the harms of the definition will outweigh itsadvantages” (Herz and Molnar 2012, 31). The upshot is arejection of a univocal definition that captures “theessence” of hate speech as a phenomenon.

It is important to note that many definitions of hate speech will notfall squarely within the categories Post outlines. For instance, theUN’sInternational Convention on the Elimination of AllForms of Racial Discrimination identifies hate speech both interms of its content and its harmful consequences. Most definitionstend to characterize hate speech in multiple ways.

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

Perhaps an immediate reaction to disjunctive definitions of the sortBrison offers is skepticism about the definitiveness of the purportedlist. When we go to test the definition’s application, weinvariably find contestable inclusions and exclusions. Recall theexamples from Parekh at the start of this section. Something like“Arabs out of France” might be included as an instance ofhate speech on Brison’s account on the grounds that it creates ahostile or intimidating environment. Should statements thatcommunicate a similar message in a less abrasive manner also beincluded? Suppose “Only French Nationals should occupyFrance” is roughly equivalent content-wise to “Arabs outof France.” If the former is indeed a less abrasive presentationthough communicating the same content as the latter, what are we tomake of its status? Many will find the statement odious; many willnot. And since it is certainly not a face-to-face vilification or formof group libel, classifying it as hate speech will depend on howlikely it is to create an intimidating or hostile environment.

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

Content-based accounts face the challenge of determining whichcontents meet this standard. If the content that distinguishes hatespeech from other types of speech must express, encourage, or incitehatred towards groups or individuals based on certain features, thenthe proponent of this view will need an account of expression. Is thespeech in view that which signals the presence of a particular mentalstate in the speaker (i.e., hate) or that which is likely to primefeelings of animosity in a specific audience?

Another issue facing content-based approaches concerns distinguishingbetween speech that “respects ‘the decencies ofcontroversy’” and that “which is outrageous andtherefore hate inducing” (Post, 2009, 128). The ability toexpress a wide range of views, even contentious ones, is a cherishedaspect of democratic societies. Failure to observe this distinctionwould broaden the scope of what counts as hate speech perhaps toomuch. In order to make this distinction, one could follow Post intying it to “ambient social norms” that distinguishoutrageous and respectful behavior. One challenge though is indetermining the content of those social norms. For instance, aminority group whose opinions have little impact on the makeup ofnorms are unjustifiably excluded from influencing the shape of theirsociety’s civility norms.

Definitions of hate speech based on intrinsic properties generallyrefer to those that emphasize the type of the speech uttered. What isat issue is the use of speech widely known to instigate offense orinsult among a majority of society. Explicitly derogatory expressionslike slurs are paradigmatic examples of this type of view. In general,the type of speech identified on this account is inherentlyderogatory, discriminatory, or vilifying.

Though attractive at first glance, classifying hate speech along theselines might prove to fall short in two ways. First, defining hatespeech in this way might be too constricting. Some of the examples inour initial list would seem not to count as hate speech since theyarguably lack the intrinsic features. “Arabs out ofFrance,” for example, does not contain explicitly slurringterms. And second, this definition might prove too expansive. In caseswhere slurs are reappropriated by members of the target group or whereartists incorporate them into a creative work, it would appear odd tocount these as instances of hate speech. The concern is tiedspecifically to locating the issue in the terms themselves, as opposedto the use to which the terms are put.

Perhaps a final challenge to intrinsic property views can be derivedfrom the work of Judith Butler (1997). On Butler’s account, hatespeech is a kind of performative that is “always deliveredtwice-removed, that is, through atheory of the speech actthat has its own performative power” (96). More specifically,“[w]hat hate speech does … is to constitute the subjectin a subordinate position” (19). Butler locates the trouble withhate speech in itsperlocutionary effects, a conceptintroduced by J.L. Austin that refers to the effects a speech act canhave on its audience. An example of a perlocutionary effect is feelingamused at a joke or frightened from the telling of a ghost story.Unlike with intrinsic property definitions, Butler shifts focus to thenature of the acts performed rather than the terms in use. (For acritical look at Butler’s account, see Schwartzman (2002).)

Lastly, dignity-based conceptions focus primarily on the role of harmsto the dignity of targets of hate speech. For instance, both StevenHeyman (2008) and Jeremy Waldron (2014) appeal to dignity in theiraccounts. Broadly speaking, hate speech on this kind of conceptionamounts to speech that undermines its target’s “basicsocial standing, the basis of [their] recognition as social equals andas bearers of human rights and constitutional entitlements”(Waldron, 2014, 59). This conception of hate speech will also includecharacterizations in terms of group defamation or group libel. Section130 of Germany’s penal code is an example of legislation thatincorporates a dignity-based conception of hate speech, prohibiting“attacks on human dignity by insulting, maliciously maligning,ordefaming 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 wholeand those about a particular member of a group (Brown, 2017a).Presumably, only the former is consistent with an understanding ofhate speech as a group-based phenomenon. Secondly, an implication ofthe view appears to be that it expands the range of things that wouldcount as hate speech. Any speech that calls into question the basicstanding of certain groups falls under this notion, which may make itmore difficult to distinguish between contentious political speech andhate speech.

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

1.1 The Harms of Hate Speech

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

A distinction between “assaultive hate speech” and“propagandistic hate speech” is helpful when discussingthese harms (Langton 2012; 2018a; see also Gelber and McNamara (2016)who discuss “face-to-face encounters” and“incidences of general circulation”). Hate speech yelledat an individual on the street, or from a passing car, is aface-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 otherhand,propagandistic hate speech is often intra-group speech,spoken by members of one group to fellow ingroup members (e.g., awhite 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 notbe overstated. Summarizing the results of their study which surveyedthe experiences of the victims of hate speech, Katharine Gelber andLuke McNamara conclude that “the distinction betweenface-to-face encounters and general circulation hate speech is notalways clear in the everyday experiences of racism endured bytargets” (2016, 326). Any one instance of hate speech might fallinto both categories. For example, it may occur in its first instanceas an assaultive speech act, and then reports of the event may thentake on a propagandistic aspect, as it is spread among the community.Similarly, even if an instance of hate speech is intended as a pieceof propaganda, it may, when encountered by a member of the communityit disparages, be akin to assaultive speech.

Still, this distinction helps reveal the wide range of thetypesof speech acts that are plausibly harmful, and also offersinsight intohow they harm. For example, Waldron (2014)focuses mainly on hate speech in its propagandistic mode, which heargues undermines the public assurance of equal social standing thatmembers of non-dominant communities are entitled to—in histerms, their assurance of dignity. On this view, public hatespeech—e.g., flyers that read ‘MuslimsOut!’—is “an environmental threat to social peace, asort of slow-acting poison, accumulating here and there, word byword” (2014, 4). Its harm is therefore one that attacks thebroader society, and not just individuals targeted by hate speech.

On the other hand, the essays in the classicWords that Woundtend to focus more on what its authors term “assaultivespeech,” that is, “words that are used as weapons toambush, terrorize, wound, humiliate, and degrade” (Matsuda etal. 1993, 1). This leads them to focus more on hate speech’sability to produce “direct, immediate, and substantialinjury” (Lawrence, 1993, 57), such as “immediate mental oremotional distress” (Delgado, 1993, 93–94). On thisapproach, the most evident harms of hate speech are psychological.These psychological injuries scale up, however, when hate speech isendemic, and so result in the types of community or social harmshighlighted by authors like Waldron. For this reason, the distinctionbetween these approaches may be thought of as more a matter ofemphasis.

This relationship between individual harms and broader social harms isalso evident once we acknowledge the long-term effects of hate speechon victims, in addition to its more immediate impacts (Delgado andStefancic, 2004, 14). Victims of hate speech may first experience“psychological symptoms and emotional distress” likeheightened stress and fear in the immediate aftermath of assaultivehate speech, but they may also experience far-ranging consequences ifthey “modify their behavior and demeanor” to avoidreceiving further hate messages, limiting their ability to participatefully in society (Matsuda, 1993, 24). Gelber and McNamara’sinterview subjects confirm this complex web of effects that hatespeech may cause, highlighting how “harms are often enduring andnot ephemeral” (2016, 336). In this way, hate speech is both animmediate attack on one’s health and dignity, along with athreat to their community’s position in society. The cumulativeeffect of hate speech events, therefore, is a collection of harmslocated both in individuals and communities, which blurs thedistinction between assaultive and propagandistic hate speechevents.

Constitutive and Consequential Harms

Another distinction which is similarly helpful, but also fraught, isthe distinction betweenconstitutive andconsequential harms—that is, harms that occurin the saying of some utterance of hate speech, and thosethat 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 speechfrom 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 correspondto what Austin called theillocutionary act, the actperformedin saying X, while consequential harms correspondtoperlocutionary effects, the results brought aboutbysaying X. Most (though not all) of the harms surveyed abovecomprise consequential harms, as items such as psychological injury,feelings of fear, and societal withdrawal all most naturally fall intothe perlocutionary effects category.

However, philosophers have also drawn attention to how hate speech caninjure in a different way by indirectly affecting the positions of thesocial 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” hatespeech harms in a way not captured in the above account of individualinjuries and their cumulative effects (Maitra and McGowan, 2012, 7).This is animmediate harm that occurs in the saying of thespeech act, which (given appropriate circumstances and uptake)produces a shift in the normative landscape. It is in this way that aninstance of hate speech may not only cause the injuries surveyed abovebut may also, for example,rank Indigenous Peoples asinferior,legitimate discriminatory behavior towards them(perhaps via incitement), or potentiallysilence them. (Wereturn to the notion of silencing as an illocutionary harm of hatespeech in Section 4 below.)

One reason to direct our attention towards the constitutive harms ofhate speech is its potential to productively advance the debate overthe legitimacy of potential restrictions. Mary Kate McGowan (2009) hasmade this case most explicitly. “Rather than focus on what acertain category of speech causes,” she writes, we ought to be“interested in what such speech actually does, in and ofitself” (2009, 389–90). The idea here is that by focusingonly on the harmscaused by hate speech, we are inevitablydrawn into a debate about balancing the costs and benefits ofpermitting or regulating speech, which often leads to an impasse.Alternatively, turning our attention to the acts hate speechconstitutes can reveal features that help us avoid question ofbalancing 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 otheracts of discrimination—like posting a ‘Whites Only’sign up at a hotel—that US law recognizes as illegal. As aspeech act, hate speech can enact discriminatory rules in much thesame way the physical sign does, and so ought to similarly berestricted (McGowan, 2012). This argument proceeds by a development ofAustin’s notion of “exercitives,” which are speechacts that enact rules in a given domain, and is one example of thefruitful use of speech act theory to the philosophy of hatespeech.

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

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

2. Religious Hatred and Anti-Semitism

Religious belief is sometimes the source of putative cases of hatespeech, and sometimes its target. In both cases, assessing theconceptual addition of religion to hate speech is a difficult task.Speech rooted in religious conviction is sometimes subjected toscrutiny to determine whether instances should count as hate speech ornot. For instance, the Westboro Baptist Church’s demonstrationsoften make use of slurs and other explicitly defamatory language. Thisis an extreme case, which can be accommodated by extant hate speechlegislation. Other cases, however, involve religious leaders makingcontentious statements—for instance, questioning the legitimacyor recognition of LGBT+ individuals, while claiming these arestatements of love, not hate. Questions about religious speech of thissort concern whether it is simply contentious speech liberaldemocratic societies must tolerate or speech that runs afoul of deeplyheld norms that ought to be proscribed.

Some wonder whether religious sensibilities should be afforded specialprotection from offense. Amnon Reichman (2009), for instance, notesthat some Israeli scholars have argued that providing specialprotection for religious beliefs is a good idea “so as not topush [religious] believers into having to choose between the authorityof the state and the authority of their religion (namely, theauthority of God)” (338). This relies on an assumption thatreligion is an institutionalized normative regime in competition witha legal regime where clashes over religious beliefs threaten thesocial fabric of society. It is in turn prudent to mitigate suchclashes in order to avoid situations of unrest like the incidentsinvolving comedic cartoons of Mohammed in the Dutch newspaperJyllands-Posten and the French publicationCharlieHebdo.

It is not clear, however, that religious beliefs warrant specialprotection over other forms of belief that may be just as stronglyheld. Clashes over deeply held political beliefs can pose a similarthreat to the social fabric as religious beliefs. Thus, there is noreason to think the same concern should not apply quite broadly.Providing certain types of speech special protection on these groundswould threaten to introduce quite repressive legislation on speech ingeneral.

Holocaust denial, denial of the Armenian Genocide, and the denial ofother crimes against humanity have also been the subject of speciallegislation, especially in Europe. As Michael Whine (2009) notes, 16European states, as well as Israel, have criminalized Holocaust denial(543). In these contexts, at least one rationale for banning speechthat denies or trivializes the Holocaust concerns its role in incitinghatred (Altman, 2012). One possible justification for such legislationrests on claims about what denial speech is. According to MartinImbleau (2011), denial speech poses as an historical endeavor but isreally propaganda. The denier’s aim is to “eradicate theawareness of the truth that prevents the resurgence of past criminalideologies” (2011, 238). But if this is the rationale, itpotentially opens up justifications for much broader application sincesimilar claims might be made of other forms of propaganda. (For ageneral overview on Holocaust denial, see Robert Wistrich (2012) andBehrens et. al (2017).)

3. Slurs, Code Words, and Dogwhistles

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

3.1 Slurs

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

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

(3.1): Crackers make up about 80% of Congress.
(3.2): White people make up about 80% of Congress.

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

Though there seems to be widespread consensus that slurs have or couldhave neutral counterparts, not everyone shares this sentiment. LaurenAshwell (2016), for example, denies that neutral counterparts (whichshe refers to as ‘neutral correlates’) play an essentialrole 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, adefinition need not include reference to neutral counterparts. Infact, making neutral counterparts central to defining slurs rendersone incapable of accounting for terms that function similarly to slursyet lack this purportedly central feature.

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

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

For Ashwell, pragmatic and semantic accounts of slursstructurallyrequire neutral counterparts, and so simply cannot jettison them.One response proponents of these kinds of accounts could give is thatthe gendered insults Ashwell highlights might exhibit properties thatcall their status as slurs into question. It could be open to thesetheorists to suggest that the terms they have identified as a matterof fact do carry neutral counterparts, that this is part of whatdistinguishes them as a class. And while the expressions Ashwellidentifies seem to pattern in some ways like slurs, they also exhibitfeatures that make them dissimilar. Thus, there is no need to wedgeall insulting expressions into one class; there is room to expand ourclassifications in a way that preserves clarity.

Another important issue about slurs is their power to offend. Part ofwhat makes them prime candidates for paradigmatic instances of hatespeech is a widespread belief in their offensive potency. Indeed, muchof the literature on slurs simply assumes they are offensive withoutoffering much (if any) defense of that claim. It is not always clearwhether the reader is supposed to understand offense as theprovocation of a disliked mental state or as the violation ofwidely-accepted public norms.

Renée Bolinger (2017) discusses three ways to understand theclaim 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 theaudience to take offense.

The sense of ‘offense’ in (1) tracks how audiencesactually respond at the moment of utterance. This could not be thesense in which offense is understood for at least two reasons. First,doing so would make the claim ‘slurs are offensive’ toostrong. Since we would be tracking cases of actual offense, we wouldbe focusing on particularuses of slurs, explaining whatmakes those utterances offensive rather than explicating theoffensiveness of a linguistic class. As a result, the most naturalinterpretation of the claim would be that slurring utterances areinvariably offensive, i.e., the use of slurs always provokes dislikedmental states.

This, of course, raises a couple of questions. To begin with, does thestrong claim deny the existence of non-offensive slur uses? Giventhings like linguistic reappropriation, some instances of indirectreports, and even instances of direct reports—especially bymembers of the slur’s targeted group—in which it ispossible to utter slurs without provoking a disliked mental state inthe speaker’s audience, the claim is obviously false. Further,there are also questions about who constitutes theaudience.Is the relevant audience the one intended by the speaker? Everyone whowitnesses the utterance? Only those who are present in the utterancesituation? Because the claim must now be understood to be aboutparticular slur utterances rather than the linguistic type, the claimmust reflect the diversity of reactions provoked by differenttokenings of slurs. A second reason is related to the questions aboutthe audience: does everyone in the audience have to be offended, or isit sufficient if one, some, or a few are? What is the scope of theclaim with respect to offended reactions? The answers to thesequestions will likely render the strong version of the claim untenableand 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 takingoffense. Bolinger identifies three grounds for warranted offense at anutterance: intention, inappropriateness, and associations. A speakermay intend to offend, often doing so with expressions that are taboo,and thus considered inappropriate. Vulgar expletives like‘fuck,’ ‘dick,’ or ‘shithead’ aretypically viewed as inappropriate terms, at least in certain“polite” settings. Some expressions, like slurs, are notonly inappropriate, but also carry associated attitudes and/orpractices that amplify their offense. The swastika and confederateflag, for example, are both deeply associated with oppressive andgenocidal practices towards Jewish people and African Americans,respectively.

This sense of offense still concerns one’s response tosomething, though it is not simply abouthow one reacts butone’s warrant to do so: “An utterance may warrant, butfail to actually generate offense merely because either there is nohearer, or the hearer fails to find the utterance offensive (perhapsbecause she shares the offensive attitude, fails to take it seriously,or misinterprets the utterance)” (Bolinger, 2017, 441). Bolingernotes that the associational offense category in particular is the onethat is often the subject of hate crime legislation (ibid., 442). Suchterms are often backed by formal social institutions,“adequately visible practices,” or a combination ofboth.

In (3), Bolinger uses ‘rational’ or ‘license’to refer to theepistemic justification an audience has intaking offense at a slurring utterance. Here a gap opens up betweenwhat an audience member may be warranted in taking offense at asopposed to when it may be rational to do so. For instance, if anon-native speaker used a slur to refer to someone and we come to findout they were ignorant of the expression’s status as a slur, thetarget would still have been rational to take offense even ifunwarranted. Undoubtedly, any of the three senses discussed may factorinto an explanation of a given slur’s offense. However, theoriesof slurs are more appropriately aimed at capturing warranted andrational offense.

Consider again the following pair of statements:

(3.1): Crackers make up about 80% of Congress.
(3.2): White people make up about 80% of Congress.

The most straightforward explanation of the difference between (3.1)and (3.2) is that ‘cracker’ differs in somesemantic respect from ‘white people’. Two of themost 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 isdetermined by the social institutions that undergird them, whichconsists of two components: anideology and aset ofpractices. Hom defines an ideology as “a set of (usually)negative beliefs about a particular group of people” (431). Asfor the set of practices, these are racist practices that “canrange from impolite social treatment to genocide” (ibid.). Thetwo components combine to produce slurs’ semantic content, whichcontains a normative claim about the way individuals ought to betreated, because of possessing certain characteristics in virtue ofbeing a member of an identifiable social group. (For alternativeaccounts 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 anobservation many will have noticed when considering differentexamples. The slur in (3.1) is typically experienced as less offensivethan ones that target members of marginalized groups. Language usersrecognize variation in offensive potency among slurs, some being moreoffensive than others. Hom refers to this asderogatoryvariation. Difference in the virulence of those backing racistinstitutions 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 muchmore virulent. (One might also wonder if there is any racistinstitution backing slurs for members of dominant groups at all.)

One objection raised against Hom’s view is that the semanticcontent he proposes of slurs is overwrought (Jeshion, 2013b). RobinJeshion argues that Hom’s view “attributes highly specificsets of ideologies and modes of treating the group, yet it is doubtfulthat anything so semantically rich and well defined is semanticallyencoded in the slur” (318). That is, it is doubtful the racistmeans anything this racialized. Jeshion denies that slursexpress anything as robust as Hom claims.

Camp offers an alternative semantic account in which slurs bear aclose relationship to aperspective, which are“open-ended ways of thinking, feeling, and more generallyengaging 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 thetargeted group as a whole” (ibid., 337). The perspective is anegative one that highlights certain characteristics or propertiesspecifically associated with particular groups, ones that are presumedto warrant certain affective and evaluative responses.

What makes slurring perspectives a semantic feature for Camp is thatthey do not “merely signal … allegiance to a certainperspective,” but do so “in an overt and nondefeasibleway, precisely in virtue of employing that expression” (ibid.,340). The use of a slur inserts a willful and noncancelable way ofthinking about the target into a conversation. This is codified in theexpression itself, and not something audiences “figureout” through the use of pragmatic mechanisms. This appears to bebolstered by the fact that one typically cannot erase a slur’sderogation by following up with a statement intending to do so,e.g.,

(3.3): They gave the job I applied for to a s**c,but I don’tmean to suggest there is anything bad about being Hispanic.

The tension of the contrast is one an audience might generally thinkfinds its source in the meaning of the slur itself, rather than fromfeatures that emerge from the way language is used in a particularcontext. Further, as we saw in Jeshion’s objection toHom’s view, the information slurs manage to convey isn’tvery specific. This point is consistent with the open-ended nature ofthe perspectives Camp associates with slurs.

Though Camp’s account represents a marked improvement, criticsstill see shortcomings they believe should give us pause. GeoffNunberg (2018), for instance, argues that Camp’scharacterization of perspectives is too vague to capture the morespecific colorings of slurs for specific groups: “Whateverdistinguishesredskin frominjun orniggerfromcoon, it’s more precise and richer than simply adisposition to think about the referents in certain ways”(Nunberg, 2018, 260–261). According to Nunberg, what is centralfor how slurs work is not the perspective the user employs to thinkabout their target, but the allegiance it signals to a group orcommunity disposed to think negatively of the target.

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

For some theorists, the accounts offered by Hom and Camp leave outwhat they regard as an important aspect of slurring, namely, the roleattitudinal expression plays in their derogatory power. These viewsagree that the difference between slurs and their purportedcounterparts is located in the realm of semantics; the previousaccounts just leave out an important aspect. Jeshion (2013a)identifies three components of slurs’ semantics: (i) atruth-conditional component, (ii) anexpressivistcomponent, and (iii) anidentifying component. Thetruth-conditional component of slurs corresponds to the same groupreferenced by its purported neutral counterpart. The expressivistcomponent captures slurs’ ability to express contempt towardsmembers of socially relevant groups in virtue of their groupmembership. Finally, the identifying component ascribes a property tothe group that is seen as central to its identity. Mark Richard (2010)also proposes a view in which negative attitudes are included in theexplanation of what slurs express. Jeshion and Richard’saccounts are typically referred to asexpressivist views.

One issue expressivist views have been thought to have trouble with isderogatory variation. Derogatory variation refers to thesociolinguistic datum that slurs vary in their offensive potency. Ifwe represent degrees of offense on a scale, slurs like‘n***er’ and ‘k*ke’ are higher up on the scalethan slurs like ‘cracker’ and ‘wop.’Expressivist views have typically attributed one sort of attitude toslurs—contempt—which seems inadequate to capturethe complexity of their offense profiles. For instance, considerco-referring slurs that vary in offense. Expressivist accounts appearto 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 herexpressivist view “is only incompatible with versions ofderogatory variation that stipulate that the variation derives fromthe semantics” (243). Jeshion maintains that focusing onslurring terms rather than particular utterances of those terms causesus to reflect on various factors at play that contribute to theirpower to offend. In effect, such focusing obscures the various factorsbrought to bear on judgments of offensiveness. Thus, Jeshion claims weought to think our intuitive judgments about varying offense supportthe following thesis:

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

Jeshion argues that this thesis is compatible with her account becauseweaponized uses of slurs are offensive for several reasons: semantic,pragmatic, sociocultural, and historical. As a result, there should beno expectation that a semantic view like hers need explain derogatoryvariation semantically.

Inferentialism describes slurs in terms of the kinds ofinferences they license. Proponents of this kind of view includeRobert 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 variousactual and possible sentential roles” (1999, 46). Incharacterizing the meaning of the now-outdated slur‘boche,’ Dummett remarks,

The condition for applying the term to someone is that he is of Germannationality; the consequences of its application are that he isbarbarous and more prone to cruelty than other Europeans. We shouldenvisage the connections in both directions as sufficiently tight asto be involved in the very meaning of the word: neither could besevered 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 beingbarbarous and more prone to cruelty than other Europeans.

Inferentialism also has its challenges. Timothy Williamson (2009), forexample, opposes inferentialism by charging that it has difficultyexplaining how non-bigots, who are not disposed to draw negativeinferences, still understand their use. “We find racist andxenophobic abuse offensive because we understand it, not because wefail to do so” (257). We should note the inferentialist is notwithout resources to respond to Williamson’s charge. Forexample, Brandom’s (1994) inferentialism determinesunderstanding in terms of grasping the broad network of inferentialconnections in which an expression is situated. An importantimplication is thought to be that different speakers will understandthe expression similarly while associating it with differentinferential roles, escaping Williamson’s charge that one must bedisposed to draw slurring inferences to understand the term (seeSteinberger and Murzi, 2017). However, Brandom’s view is itselfcontroversial (For further objections to inferentialism, see Andersonand Lepore (2013b); Hornsby (2001).)

The last view we mention here is a stark alternative to the previousaccounts, opting for a socioculturally-driven explanation. Accordingto Luvell Anderson and Ernie Lepore, slurs are prohibited expressionswhose tokenings provoke offense from those who value and respect theirprohibitions: “What’s clear is that no matter what itshistory, no matter what it means or communicates, no matter whointroduces it, regardless of its past associations,once relevantindividuals declare a word a slur, it becomes one” (2013a,39). The prohibition is meant to apply not only touses butmentions of expressions as well, including direct andindirect reports.

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

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

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

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

3.2 Dogwhistles and Coded Language

In addition to slurs, which are explicitly derogatory, researchershave 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 codedlanguage—dogwhistles—to access existing racialresentment while making surreptitious racial appeals. Saul provides auseful set of distinctions for thinking about dogwhistles: they can beexplicit orimplicit, and further,intentional orunintentional. Saul uses the work oflinguist Kimberly Witten to define anovert intentionaldogwhistle as,

a speech act designed, with intent, to allow two plausibleinterpretations, with one interpretation being a private, codedmessage targeted for a subset of the general audience, and concealedin such a way that this general audience is unaware of the existenceof the second, coded interpretation. (2018, 362)

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

(3.4): Yes there’s power, wonder-working power, in the goodness andidealism and faith of the American people.

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

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

The second message is that Bush identifies with them, that he speakstheir language. Saul thinks both are instances of overt intentionaldogwhistles.

Acovert dogwhistle, according to Saul, is “adogwhistle that people fail to consciously recognize” (2018,365). She is particularly interested in how covert intentionaldogwhistles work in tandem with what psychologists have referred to asracial resentment, a belief system that is measured by thedegree 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 amongwhite Americans even though explicitly racist appeals have come to beviewed as outside the bounds of acceptable political speech. (Atleast, that seemed to be the case up until the 2016 presidentialelection cycle.) White voters, on this model, tend to shy away fromaccepting explicitly racist proposals because they do not want tothink of themselves as racist. The existence of racial resentmentallows for the skilled intentional use of utterances that areunrelated to race on the surface yet access negative racial attitudesin a targeted audience, nudging them towards a particular course ofaction--e.g., voting for a preferred candidate.

An example of a covert intentional dogwhistle is the infamous WillieHorton ad used by the George H. W. Bush campaign in 1988. The adtargeted a prison furlough program in place during MichaelDukakis’s term as governor of Massachusetts. It presented apicture of Horton, an African American man, who while out on furloughraped a white woman and stabbed her husband. Though there was noexplicit mention of race, it was clear to many that the ad drew onracial tropes about Blackness and criminality to stoke fear in whitevoters. In support of the interpretation that this was acovert dogwhistle, Saul notes that once the specter of racewas raised about the ad, its effectiveness started to wane (2018,366). The implication is that while the explicit appeal to racialresentment was a losing strategy, implicit appeal in the form ofcovert dogwhistles could be put to powerful use.

The unintentional dogwhistle is defined as an “unwitting use ofwords and/or images that, used intentionally, constitute anintentional dogwhistle, where this use has the same effect as anintentional dogwhistle” (2018, 368). Dogwhistles of this sortare passed on by unwitting others while achieving similar effects ofthe original intentional one. A special case of unintentionaldogwhistles, what Saul callsamplifier dogwhistles, occurredwhen reporters and TV producers played the Willie Horton adrepeatedly. Presumably, the repeated presentations continued to makethe associations between Blackness and criminality and, thus,continued to stoke fear and racial anxiety in significant portions ofthe white viewing public. For Saul, dogwhistles are therefore bestunderstood functionally, and the difference in speaker-intentionsbetween intentional and unintentional dogwhistles matters only insofaras we define them—their effects, in other words, are oftenidentical.

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

Perhaps there is good reason to think something aboutdogwhistles’ meaning explains their effects. Consider, first, anambiguity thesis that states code words have at least two meanings, aracial and a non-racial meaning. The expression ‘innercity’ in

(3.5): Food stamp programs help many inner-city families,

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

One worry, however, is that terms like ‘inner city’ arenot actually ambiguous. Khoo argues these terms do not behave likegenuinely ambiguous expressions. Compare the following twosentences,

(3.6): Smith is a funny man who is not humorous.
(3.7): Smith is an inner-city pastor who is from, works, and lives in thesuburbs.

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, evencontradictory. If ‘inner-city’ were genuinely ambiguous inthe 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 andnot-at-issue content. At-issue contentis the main point of a speaker’s utterance, the directlyasserted content that is foregrounded whereas not-at-issue content isprojective, meaning it is able to survive embedding underoperators like negation and modals (Tonhauser, 2012). Consider,

(3.8): John stopped smoking.

The at-issue content of (3.8) is represented by (a) and thenot-at-issue by (b):

John does not smoke.
John used to smoke.

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

(3.8′): He did not used to smoke.

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

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

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

Because the racial component of (3.5) is not-at-issue, we have areasonable explanation for why the following pair of sentencesclash,

(3.9): Food stamp programs help many inner-city families.
(3.10): No, poor, high-crime, urban areas are mostly white.

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

(3.11): John stopped smoking. He never smoked at all.
(3.12): Food stamp programs help many inner-city families, most of whomare white.

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

According to a third view, code words are neither ambiguous normultidimensional, but possess only nonracial meaning. What explainsthe phenomenon associated with terms like ‘inner city’ isthe presence of an antecedent belief in the audience member that thenallows them to infer the racial component. For example, an audiencemember 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 membercomes 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 whatKhoo calls theassociation-driven theory of code words. Onthis view, there is “an association between ‘innercity’ (or the concept INNER CITY) and the concept AFRICANAMERICAN (or maybe just RACE) which then primes racist beliefs andprejudices” (2017, 50).

Khoo’s account is simple and compelling, but we may still wonderwhether it is too liberal. For instance, expressions like‘thug,’ ‘illegal alien,’ ‘welfarequeen,’ and ‘terrorist’ seem to behave like theterms Khoo identifies as code words, yet they are generally understoodto be explicitly racial in nature. Patrick O’Donnell (2017)argues that the aforementioned expressions are not code words butracialized terms. O’Donnell characterizes thedifference between racialized terms and code words in the followingway:

Racialized terms involvedirect orpredicativerelations between a term and a racialized group whereas code wordsinvolve indirect inferential or associational relations, and
Racialized terms elicit racial resentment by making salientrace-specific interpretive options, whereas code words function bymaking salient race-neutral interpretive options (2017, 28).

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

Determining whether dogwhistles or coded language count as merelycontentious claims that must be tolerated or as hate speech subject toregulation has implications for broader discussions. The subtlety ofcoded language, for instance, calls its status as hate speech intoquestion. The impact coded language has on an audience lacks the kindof immediacy often attributed to hate speech. Lawrence (1993), forexample, notes that hate speech is often experienced by targets as aslap in the face. On the other hand, Mendelberg’s accountsuggests coded speech can incite racial resentment, and so it may bemore aptly considered similar to propagandistic hate speech, discussedabove. (For more on this point, see Jason Stanley (2015).) This wouldappear to get us closer to how hate speech is purported to function,namely, by inciting racial hatred. Whether it is close enough is ofcourse open for debate.

4. Pornography, Hate Speech, and Silencing

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

Many of the important initial moves in this literature were crafted byfeminist legal scholar Catherine MacKinnon, along with Andrea Dworkin.One of MacKinnon’s most significant claims that has receivedsustained philosophical attention is the idea that (degrading andmisogynist) pornographysilences women. With somemodifications, a similar claim may be applied to hate speech, namely,that hate speechsilences its targets. However, as theliterature 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 thatthere are “words that set conditions” for other speechacts’ successes or failures (1993, 63–68; see also Hornsbyand Langton, 1998, 27). That is, there are some speech acts that fixthe possibility of other speech acts. In other words, they make itpossible for some persons to perform some speech acts, and make itimpossible for others. This is most evident in formal settings, like alegislature, where the formal rules determine who may speak when, andin what manner. Pornography, the argument continues, does just this.It sets rules of behavior that, in effect, inhibit the speech ofwomen. The result of which is that the speech acts ofpornography—performed by those who produce and distributeit—create a climate that undermines women’s capacity toperform 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 sexualrefusal. According to the silencing argument, pornography depictswomen as not (genuinely) refusing sexual advances with utterances of‘no.’ Indeed, according to the myths perpetuated bypornography (among other social influences), a woman’s‘no’ is not a refusal, but rather part of an elaboratesexual script. As a result, when a woman says ‘no’ in anon-pornographic context, intending to refuse a man’s sexualadvances, she may find herself unable to be heard—that is, herwords won’t have the force and effect she intends, and herhearer will not take her to be refusing. She may find herself silencedin this particularly horrendous way, unable to use the standardmethods of refusing another’s sexual advances. The claim is thatthis occurs as a result of pornography silencing women’srefusals in the context of sex. It renders their words powerless.

In making this argument, Langton relies on the distinction betweenlocutionary, illocutionary, and perlocutionary acts, and,correspondingly, locutionary, illocutionary, and perlocutionarysilencing. A couple examples will explain these distinctionsquickly:

WhenX says, ‘Shoot him!’ they are, we can seequite quickly, both saying something: ‘shoot him!’ and atthe same time doing something:ordering the hearer to fire.In Austinian terms, we can say thatX performs thelocutionary act of making an utterance with a certainmeaning, and at the same time is performing anillocutionaryact of ordering the hearer to fire. In addition to these two things,the speaker is also, with their words, bringing about a number ofeffects, which Austin termed the ‘perlocutionaryact.’ 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, andperlocutionary—all occur as part of a single utterance and serveto bring out different aspects of any speech act. Austin (and manyafter him) paid particular attention to the illocutionary act of anutterance, as this, he said, corresponds to theforce of anutterance. That is, what someone is doing with their words.

With this in mind, we can see that there are, in fact, many ways onecould silence someone. You could literally gag or threaten someone toprevent them from speaking at all, which would achieve a type oflocutionary silencing. Alternatively, you could let them say what theywish, recognize what act they are performing, but prevent themachieving their goals, and in doing so achieve a type ofperlocutionary silencing. Finally, a third alternative occurs when onespeaks and is prevented not only from achieving their intendedeffects, but also is prevented from performing the very action theyintend to perform (Langton, 1993, 315). It is this thirdalternative—illocutionary silencing—that is said to occurwhen a man fails to even recognize a woman’s ‘no’ asa refusal, owing at least partially to the influence ofpornography.

The specific mechanics of silencing—along with the underlyingtheory that best explains the phenomena—is subject to muchdispute in the literature, and numerous accounts with differentessential features have been offered (see Langton 1993; Langton andWest 1999; Hornsby 1994; Hornsby and Langton 1998; Maitra2009; McGowan 2004, 2009, 2014; and Mikolla 2011; 2019, amongothers).

Laura Caponetto (2021) distinguishes four different types ofsilencing, demonstrating the breadth of the concept. First, there isessential silencing, which consists in the hearer’sfailure to recognize the illocutionary point of a speech act. Second,there isauthority silencing, where a hearer fails toacknowledge a speaker’s authority in a relevant domain. Third,there issincerity silencing, when the speaker’sutterance is inaccurately taken as insincere. Fourth and finally,there isseriousness silencing, which consists in thehearer’s failure to acknowledge the speaker’s words asappropriately serious. Given these fine-grained ways of understandingsilencing, a broad, comprehensive definition of silencing may be putas follows:

Illocutionary Silencing
A speakerS putting forth a speech actA addressedto a hearerH is illocutionarily silenced iff (i)Hfails to recognize the obtaining of some conditions forA’s success; (ii)S’s attempt atA-ing meets the conditions thatH fails torecognize; (iii) normal input and output conditions are met; (iv) therecognition failure onH’s part is systematic.(Caponetto, 2021)

In nearly all discussions of silencing, one common piece of contentionconcerns the notion of ‘uptake.’ On differentunderstanding of what uptake consists in—ranging from thehearer’s recognition of a speaker’s intent, or the type ofspeech act being performed, up to the material consequences of aspeech act—we are led to different conclusions about whether aspeaker was silenced or not. Disagreement about the conditions ofuptake poses difficulty, therefore, for many accounts of silencing.Drawing on these difficulties, Samia Hesni (2018) has argued that thestandard account of silencing needs significant retooling, in partbecause the necessary distinction between illocutionary silencing andperlocutionary silencing cannot hold, as it relies on aproblematic—and arguably conceptually untenable—notion ofuptake (Hesni, 2018, 957). In an attempt to avoid these difficulties,we might prefer an account of silencing that uses a Gricean, ratherthan Austinian or Searlian framework, bypassing the need to fullydifferentiate the illocutionary from the perlocutionary (see Maitra,2009).

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

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

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

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

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

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

As is the case with pornography and silencing, the details of themechanisms that sustain this type of silencing, along with whatparticular type of silencing racist hate speech results in, aresubject to dispute. But, just like in the pornography debate, theplausibility of the silencing argument lies partly in how it reframesthe overall question surrounding regulation. Rather than simply beinga source of harm that merely infringes on the equality rights of itstargets, if hate speech silences then it also infringes on the speechrights of its targets (West, 2012). As a result, it is not simply aquestion of balancing the speech rights of hate speakers against thewellbeing of their targets, but of competing claims to (substantive,and not just formal) freedom of expression. And given the importancethat most liberal democracies place on freedom of expression, thechallenge presented from hate speech is of central importance. Forthis reason, the silencing question is one of the most disputedaspects of hate speech and has generated great attention.

5. Counteracting Hate Speech

On the presumption that hate speech is harmful—both particularlyharmful for the members of targeted groups, and also generally harmfulto democracy—the natural question that follows is: what shouldwe do about it? This question, however, rests on severalsub-questions—some empirical, some conceptual—thatthemselves admit of rich dispute. For example, depending on how oneconceives of the value and point of free expression—to betterseek the truth, to respect autonomy, to ensure democracy,etc.—different answers to the hate speech question will seemmore worthwhile than others. The same consideration applies toempirical matters as well, which are often difficult to properlyassess in the absence of uncontroversial data. This means thatrelatively straightforward empirical questions—does genocidalspeech pave the way to actual genocidal violence; do governments abusehate speech regulation to punish political rivals and disfavoredminorities; and others—rarely receive unanimous agreement.Despite these challenges, many theorists have addressed the questionof how to counteract hate speech, and what form that response ought totake.

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

5.1 The Case for Bans

First, the case for banning hate speech. While this position may beanathema to many (especially in the United States), it is theconsensus position of most democratic nations around the globe, aswell as the explicit position of the United Nations. In theInternational Covenant on Civil and Political Rights, Article20 requires a ban on hate speech—or, in their words, “anyadvocacy of national, racial or religious hatred that constitutesincitement to discrimination, hostility or violence shall beprohibited by law” (Covenant on Civil and PoliticalRights; see also, Article 4 ofConvention on RacialDiscrimination). It is worth noting the position of the UnitedNations and other democracies on hate speech in part because of thecontrast 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 notgenerally on the basis of (racial) hate. Moreover, these exhortationsto criminalize hate speech from the United Nations sit alongsidecommitments that maintain the importance of freedom of expression. Forinstance, Article 20, quoted above, is immediately preceded by Article19, which affirms right to freedom of expression (Covenant onCivil and Political Rights).

The standard justification offered for restrictions on freedom ofexpression are based on the necessity of (a) respect of the rights orreputations of others; and (b) reasons of national security or ofpublic order. In other words, a ban on hate speech may be thought tofollow from the recognition of the harms it presents, both to thedignity of minority-members of a nation, as well as their physicalsafety. This position maintains, then, that restrictions on hatespeech are a legitimate—and necessary—exception to anotherwise wider understanding of free expression. (For some theorists,it’s worth noting, hate speech is best not understood as thetype of speech that free speech protections are meant toinclude—e.g., it serves no purpose in the pursuit oftruth—and so is not in fact anexception to a freespeech principle, but simply not included in a proper understanding ofthe scope of free speech.) This view naturally follows from theunderstanding that multiple values and rights must be balanced againsteach other. This is true both of countries that explicitly prohibithate speech in order to protect minority rights, as well as in more‘speech-friendly’ nations like the United States, wherespeech that is aimed at and likely to result in “imminentlawless action” may be legitimately restricted (seeBrandenburg v. Ohio).

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

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

On the understanding that the threat of hate speech is not exhaustedby cases that concern “imminent danger,” we might thenground the prohibition of hate speech on the basis that this mayreduce speech that causes harm to its targets, beyond those mostimmediately affected. Of course, there is also an important role to beplayed by non-legal means (e.g., moral and social pressure) in erasingor reducing these harms, so legal bans are best understood as part ofa broader approach to the ills of hate speech. Furthermore, advocatesof bans describe the expressive dimension of these laws as themselvesproviding a reason in favor of legislation (Waldron, 2014). The law,in this sense, serves as a public statement on a community’svalues, and has educational and symbolic importance in itself (Parekh,2012, 46). (For an overview of expressive theories of law, seeAnderson and Pildes, 2000.) A ban on hate speech, therefore, isintended both to reduce harms directly, by decreasing instances ofhate by the threat of law, as well as indirectly, by shaping thecommunity’s moral norms through an expression of value.

5.2 Objections to Bans, and Some Responses

Though many would agree that hate speech can have destructive effects,and that there is a moral imperative on the state to cultivatesomething like respectful relations between its members, objections tohate speech bans abound. In a wide-ranging response to these concerns,Parekh (2012) considers (and rejects) six common objections to theprohibition of hate speech. These six objections are: (1) that theharm of hate speech, while real, is relatively minor and a small priceto pay given the interest of democratic nations; (2) that bans are notthe answer, but rather “better ideas” and “morespeech” are; (3) that a prohibition would have a dangerous“chilling effect” and that hate speech bans are a slipperyslope to all sorts of unwanted restrictions; (4) that bans give thestate too much power to judge the content of speech and decide whatcan or cannot be said, threatening state-neutrality, skewing politicaldebate, and infringing on individual liberty; (5) that bans are anobjectionable form of paternalism or moral authoritarianism, and isincompatible with the assumption that humans are responsible andautonomous individuals and that society is made up of free and equalcitizens; and finally, (6) that bans are ineffective at changingattitudes and removing the hate from the hate speaker’s heart,with the result that bans have the effect of moving extremistsunderground, alienating them from wider society, and in doing sorendering us ignorant of their violent potential and impotent toengage in effective de-radicalization.

Each of these concerns merits more space than can be given to themhere. Still, considering these objections to bans and the responsesavailable, even briefly, is illustrative of the theoretical concernsbans on hate speech bring forth (see the list of references for fullerdevelopment of the relevant theoretical and empirical issues). Againfollowing Parekh (2012, 47–54), we can approach these objectionsas follows.

In response to (1), the objection that the interests of a vibrantdemocracy outweigh the harms imposed by hate speech, it may be arguedthat hate speech does not embody the values of free speech but, infact, undermines them by promoting irrational fears and hatred overreasoned arguments and public scrutiny. How powerful one takes thisresponse to be depends directly on what one takes the value andjustification of a right to free expression to be, which is of coursea 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 othergoods). This is what a ban does, and so may be considered to behelping ensure ‘fair competition’ by countering prevailingprejudices, and encouraging greater participation from the members ofcommunities targeted by hate speech. In other words, bans on hatespeech may promote greater freedom of expression, by preventing thetype 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 theproblems these signal rest on the vague wording and inconsistent orbiased 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 italtogether. Moreover, the appeal to a ‘slippery slope’ maybe inapt, as it implies that once one type of speech is prohibited,society cannot help but prohibit even more types. But we have no clearreason to suppose that this is the case, as existing bans ondefamation have not led to bans on fair critical comment, forexample.

The worries at the core of objection (4) represents a well-foundedfear of the state, and so must be taken seriously. But, to defendersof hate speech bans, its understanding of the threat that hate speechbans pose to state-neutrality is nonetheless flawed. It fails torecognize that the state oftenalready judges the content ofspeech (e.g., in banning commercial fraud, criminal solicitation,public displays of obscenity) and often elides neutrality when itspeaks in favor of certain positions (e.g., the value of humandignity, equality, liberty). While any defense of hate speech bansmust reckon with the possibility of further empowering the state,opponents ought not misrepresent the status quo, exaggerating thereality of state-neutrality.

Objections grounded on the threat of paternalism or moralauthoritarianism, like (5), are similarly serious. However, oneresponse on behalf of bans would be to point out how autonomy isalways exercised under certain conditions and requires variousexternal circumstances for its development and use. When appealing topersonal autonomy, therefore, we should not idealize too greatly sothat its real-world exercise is ignored. Rather, the threats thatracism 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 andmaintain that this is no knock against the law, and indeed is noproblem for hate speech bans. The aim of these bans, in most cases, isnot to prevent hatred but to prevent the harm thatthe publicexpression of hate can cause. The indirect effects of such a law,however, are an empirical matter, and it is unlikely they admit of asingle, general answer, but are highly context-dependent. The subjectof the practicability of hate speech bans deserves special attention,however. Opponents to bans may worry that the suppression of hatespeech is likely to backfire, not only by failing to reduce hatred,but byincreasing the sense of oppression and victimizationthat many bigots thrive on, leading to an escalation of racistviolence (Baker, 2012, 77). Again, as an empirical hypothesis, itcannot be settled simply from the armchair. Still, a further responseavailable on behalf of hate speech bans would be to question thelegitimacy of this objection. If, by hypothesis, bans generated anincrease in violence, it would still be the responsibility of thestate to manage this violence effectively. The role of the state isnot exhausted by implementing a ban, but must be seen alongside itsenforcement.

This, however, leads to a slightly different objection. The opponentof bans may worry that the enforcement of laws against hate speechwould divert the state’s energies away from more effectivemeasures against hated, such as “those directed at changingmaterial conditions in which racism festers, material conditions ofboth the purveyors and targets of hate” (Baker, 2012, 77). Thatis, the energies and resources that would be directed towardsestablishing and enforcing hate speech bans may be better spent onalternative policies. The guiding thought rests on two importantpoints. First, that the intended ends of hate speech bans (e.g.,reduction in the harms of hate speech that fall on those targeted byit, mitigation of the expansion of racist attitudes, lesseningoccurrences of violent hate crimes) may be more effectively achievedvia different means, such as reducing inequality, improving socialsafety nets, political empowerment, and more. Second, though the statecan do more than one thing at once, it is nonetheless working withlimited resources, and efficiency is a value. That these alternativepolicy options may indeed be more effective is an unresolved empiricalmatter. And it remains an open question whether indirect approacheslike this would fail to achieve the expressive ends of hate speechbans, which more directly communicate to those targeted by hate speechthat they are valued members society.

Many of the claims made above, both on behalf of bans and inopposition, raise theoretical and empirical issues whose properexamination spans many articles and books. Suffice to say that thedebate over bans is a highly contested one, and each position rests onan understanding of such issues as the value of free expression, theharm of hate speech, the likely effects a ban might have in aparticular context, and so on. For instance, one who believes thatfree expression is valuable in part because of its role in democraticdecision-making may maintain that specificallypoliticalspeech deserves increased protections, and that some of what othersregard as hate speech might fall into this category, escapingregulation. Alternatively, one may view the immunity for politicalspeech as perhaps a red herring. On the speech act theoretic frameworkoutlined above, some forms of racist hate speech are functionallyidentical to a ‘Whites Only’ sign hanging in a publicrestaurant (McGowan 2012; McGowan and Maitra 2009). The latterexpresses a political opinion in the same way as the former expressiondoes, but it is also regarded as unlawful racially discriminatory. Thesame considerations—legal sanction—might therefore applyto the verbal utterance as the written sign, and the appeal to thepolitical content of the message is irrelevant.

5.3 The Supported Counterspeech Alternative

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

By contrast, the “supported counterspeech” alternativeaims to recognize the specific harms inflicted by hate speech andprovide state support to empower those who are harmed. Gelber, anadvocate for this alternative, places it within the capabilitiesapproach originally developed by Amartya Sen (1992) and MarthaNussbaum (2000; 2003). “If hate-speech acts harm theirtargets’ capacity to develop human capabilities,” Gelbersays, then “this is what needs to be remedied” (2012a,54). The impetus for this approach therefore begins from the idea thatwe must think about remedies to hate speech beyond restrictions andpunishment, as neither of these approaches achieve the goal ofempowering the target of hate speech. (This is especially true of thelatter, punishment, which also carries with it all the negativeconsequences that anti-carceral advocates have noted.) The supportedcounterspeech policy is therefore not focused on hate speakers, butrather the targets of hate speech more directly.

The core of this approach lies in an enlarged conception ofcounterspeech as well as a commitment by the state to provide thematerial conditions necessary for this speech. In practice, this wouldmean that the state is committed to responding to an incident of hatespeech by empowering its targets to engage in more speech, after thefact. The specific forms this support may take will depend on theconditions of different contexts, along with calibration for thespecifics of the incident it is meant as a response to, as well as theneeds of the particular communities. Still, to give a sense of whatthis may entail, examples of the sort of supported counterspeech thatthis position recommends include things such as: assistance in theproduction of a community newsletters, op-eds, radio broadcasts, ortelevision advertisements; the development of antiracism awarenessprograms, or anti-hate-speech workshops; subsidizing community-led artprojects; etc.

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

Bibliography

Anderson, Elizabeth, and Richard Pildes, 2000, “ExpressiveTheories of Law: A General Restatement,”University ofPennsylvania Law Review, 148: 1503–1575.
Anderson Luvell and Lepore, Ernest, 2013a, “SlurringWords,”Noûs, 47(1): 25–48.
–––, 2013b, “What Did You Call Me? Slursas Prohibited Words,”Analytic Philosophy, 54(3):350–363.
Appiah, Kwame Anthony, 2012, “What’s Wrong withDefamation of Religion?” in Michael Herz, and Peter Molnar(eds.),The Content and Context of Hate Speech: RethinkingRegulation and Response, Cambridge; New York: CambridgeUniversity Press, pp. 164–182.
Ashwell, Lauren, 2016, “Gendered Slurs,”SocialTheory 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 andPragmatics of Slurs,” in David Sosa (ed.),Bad Words:Philosophical Perspectives on Slurs, Oxford: Oxford UniversityPress, 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,Holocaustand Genocide Denial: A Contextual Perspective, Abingdon; NewYork: Routledge, Taylor and Francis Group.
Bolinger, Renée Jorgensen, 2017, “The Pragmatics ofSlurs,”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: TheMyth 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 FreeSpeech,”Ethics, 108(2): 312–339.
–––, 1998b, “Speech, Harm, and theMind-Body Problem in First Amendment Jurisprudence,”LegalTheory, 4: 39–61.
Butler, Judith, 1997,Excitable Speech: A Politics of thePerformative, New York: Routledge.
Camp, Elisabeth, 2013, “Slurring Perspectives,”Analytic Philosophy, 54(3): 330–349.
–––, 2018, “Slurs as Dual-ActExpressions,” in David Sosa (ed.),Bad Words, Oxford:Oxford University Press, pp. 29–59.
Caponetto, Laura, 2021, “A Comprehensive Definition ofIllocutionary Silencing,”Topoi, 40:191–202.
Cepollaro, Bianca, 2015, “In Defense of a PresuppositionalAccount of Slurs,”Language Sciences, 52:36–45.
Croom, Adam, 2011, “Slurs,”LanguageSciences, 33: 343–358.
Delgado, Richard, 1993, “Words that Wound: A Tort Action forRacial Insults, Epithets, and Name Calling,” in Mari Matsuda,Charles Lawrence III, Richard Delgado, and Kimberlé Crenshaw(eds.),Words that Wound: Critical Race Theory, Assaultive Speechand the First Amendment, Boulder, CO: Westview Press, pp.89–110.
Delgado, Richard, and Jean Stefancic, 2004,UnderstandingWords 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 Counterspeechin Hate-Speech Policy (with a Focus on Australia),” in MichaelHerz, and Peter Molnar (eds.),The Content and Context of HateSpeech: 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 andAustralia,” 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 theHarms of Hate Speech,”Social Identities, 22 (3):324–341.
Herz, Michael, and Peter Molnar (eds.), 2012,The Content andContext 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, NewHaven: Yale University Press.
Hom, Christopher, 2008, “The Semantics of RacialEpithets,”Journal of Philosophy, 105:416–440.
Hornsby, Jennifer, 1994, “Illocution and itsSignificance,” in Savas L. Tsohatzidis (ed.),Foundations ofSpeech Act Theory: Philosophical and Linguistic Perspectives,London: Routledge, pp. 187–207.
–––, 2001, “Meaning and Uselessness: Howto Think About Derogatory Words,”Midwest Studies inPhilosophy, 25(1): 128 – 141.
Hornsby, Jennifer, and Rae Langton, 1998, “Free Speech andIllocution,”Legal Theory, 4: 21–37.
Imbleau, Martin, 2011, “Denial of the Holocaust, Genocide,and Crimes Against Humanity,” in Ludovic Hennebel and ThomasHochmann (eds.)Genocide Denials and the Law, New York:Oxford University Press, pp. 235–277.
Jeshion, Robin, 2013a, “Expressivism and the Offensivenessof Slurs,”Philosophical Perspectives, 27:231–259.
–––, 2013b, “Slurs and Stereotypes,”Analytic Philosophy, 54 (3): 314–329.
Khoo, Justin, 2017, “Code Words in PoliticalDiscourse,”Philosophical Topics, 45 (2):33–64.
Kirk-Giannini, Cameron Domenico, 2019, “Slurs areDirectives,”Philosopher’s Imprint, 19(48):1–28.
Kukla, Quill [writing as Rebecca], 2014, “PerformativeForce, Convention, and Discursive Injustice,”Hypatia,29 (2): 440–457.
–––, 2018, “Slurs, Interpellation, andIdeology,”The Southern Journal of Philosophy, 56(S1):7–32.
Langton, Rae, 1993, “Speech Acts and UnspeakableActs,”Philosophy & Public Affairs, 22(4):293–330.
–––, 2012, “Beyond Belief: Pragmatics inHate Speech and Pornography,” in Ishani Maitra and Mary KateMcGowan (eds.),Speech & Harm: Controversies Over FreeSpeech, New York: Oxford University Press, pp. 72–93.
–––, 2018a, “The Authority of HateSpeech,” 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 CounterSpeech,” 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 aPornographic Language Game,”Australasian Journal ofPhilosophy, 77(3): 303–319.
Lawrence, Charles III, 1993, “If He Hollers Let Him Go:Regulating Racist Speech on Campus,” in Mari Matsuda, CharlesLawrence III, Richard Delgado, and Kimberlé Crenshaw (eds.),Words that Wound: Critical Race Theory, Assaultive Speech and theFirst Amendment, Boulder, CO: Westview Press, pp.53–88.
Haney Lopez, Ian, 2014,Dog Whistle Politics: How Coded RacialAppeals 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,”CanadianJournal 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 UniversityPress, pp. 94–120.
Maitra, Ishani, and Mary Kate McGowan (eds.), 2012,Speech& Harm: Controversies Over Free Speech, New York: OxfordUniversity Press.
Matsuda, Mari, 1993, “Public Response to Racist Speech:Considering the Victim’s Story,” in Mari Matsuda, CharlesLawrence III, Richard Delgado, and Kimberlé Crenshaw (eds.),Words that Wound: Critical Race Theory, Assaultive Speech and theFirst Amendment, Boulder, CO: Westview Press, pp.17–51.
Matsuda, Mari, Charles Lawrence III, Richard Delgado, andKimberlé Crenshaw (eds.), 1993,Words that Wound: CriticalRace 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 andPhilosophy, 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 RacialDiscrimination,” 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: Speechand the Social World,” in Daniel W. Harris Daniel Fogal, andMatt Moss (eds.),New Works on Speech Acts, Oxford: OxfordUniversity Press, pp. 185–201.
–––, 2019,Just Words: On Speech and HiddenHarm, Oxford: Oxford University Press.
McGowan, Mary Kate, and Ishani Maitra, 2009, “On Racist HateSpeech and the Scope of a Free Speech Principle,”CanadianJournal of Law and Jurisprudence, 23(2): 343–372.
Mendelberg, Tali, 2001,The Race Card: Campaign Strategy,Implicit Messages, and the Norm of Equality, Princeton: PrincetonUniversity Press.
Mikkola, Mari, 2011, “Illocution, Silencing and the Act ofRefusal,”Pacific Philosophy Quarterly, 92:415–437.
–––, 2019,Pornography: a PhilosophicalIntroduction, New York: Oxford University Press.
Miller v. California, 413 U.S. 15 (1973).
Neufeld, Eleonore, 2019, “An Essentialist Theory of theMeaning of Slurs,”Philosopher’s Imprint, 19(35):1–29.
Nunberg, Geoff, 2018, “The Social Life of Slurs,” inDaniel W. Harris, Daniel Fogal, and Matt Moss (eds.),New Works onSpeech Acts, Oxford: Oxford University Press, pp.237–295.
Nussbaum, Martha, 2000,Women and Human Development: TheCapabilities Approach, Cambridge: Cambridge UniversityPress.
–––, 2003, “Capabilities as FundamentalEntitlements: Sen and Social Justice,”FeministEconomics, 9(2/3): 33–59.
O’Donnell, Patrick, 2017, “Generics, Race, and SocialPerspectives,”Inquiry: An Interdisciplinary Journal ofPhilosophy, published online 26 January 2017.doi:10.1080/00207543.2016.1266801
Parekh, Bhikhu, 2012, “Is There a Case for Banning HateSpeech?” in Michael Herz, and Peter Molnar (eds.),TheContent and Context of Hate Speech: Rethinking Regulation andResponse, 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 andJames Weinstein (eds.),Extreme Speech and Democracy, Oxford:Oxford University Press.
Reichman, Ammon, 2009, “Criminalizing Religiously OffensiveSatire: Free Speech, Human Dignity, and Comparative Law,” inIvan Hare and James Weinstein (eds.),Extreme Speech andDemocracy, Oxford: Oxford University Press, pp.331–355.
Richard, Mark, 2010,When Truth Gives Out, Oxford: OxfordUniversity Press.
Saul, Jennifer M., 2018, “Dogwhistles, PoliticalManipulation, and Philosophy of Language,” in Daniel W. HarrisDaniel Fogal, and Matt Moss (eds.),New Works on Speech Acts,Oxford: Oxford University Press, pp. 360–383.
Sbisà, Marina, 2009, “Illocution andSilencing,” in Bruce Fraser, and Ken Turner (eds.),Languagein Life, and a Life in Language: Jacob Mey—a Festschrift,Bradford: Emerald, pp. 351–357.
Schwartzman, Lisa H., 2002, “Hate Speech, Illocution, andSocial Context: A Critique of Judith Butler,”Journal ofSocial 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, AlexanderMiller (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: The2008 Election and the Dream of a Post-Racial America, Chicago:University of Chicago Press.
Tirrell, Lynne, 1999, “Derogatory Terms: Racism, Sexism, andthe Inferential Role Theory of Meaning,” in Christina Hendricks,and Kelly Oliver (eds.),Language and Liberation: Feminism,Philosophy, and Language, Albany: State University of New YorkPress, pp. 41–79.
Tonhauser, Judith, 2012, “Diagnosing (not-)at-issuecontent,”Proceedings of Semantics of UnderrepresentedLanguages in the Americas, 6: 239–254.
UN General Assembly,International Covenant on Civil andPolitical Rights, 16 December 1966.
UN General Assembly,International Convention on theElimination of All Forms of Racial Discrimination, 21 December1965, [available online].
Waldron, Jeremy,2014,The Harm in Hate Speech, Cambridge,MA: Harvard University Press.
West, Caroline, 2012, “Words that Silence? Freedom ofExpression and Racist Hate Speech,” in Ishani Maitra, and MaryKate McGowan (eds.),Speech and Harm: Controversies over FreeSpeech, Oxford: Oxford University Press, pp. 222–248.
Whine, Michael, 2009, “Expanding Holocaust Denial andLegislation Against It,” in Ivan Hare and James Weinstein(eds.),Extreme Speech and Democracy, Oxford: OxfordUniversity Press, pp. 538–556.
White, Ismail, 2007, “When Race Matters and When ItDoesn’t: Racial Group Differences in Response to RacialCues,”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 theSemantics of Pejoratives,” in J. Almog, and P. Leonardi (eds.),The Philosophy of David Kaplan, Oxford: Oxford UniversityPress, pp. 137–158.
Wistrich, Robert, 2012,Holocaust Denial: The Politics ofPerfidy, Berlin: De Gruyter.

Academic Tools

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

Other Internet Resources

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”

Open access to the SEP is made possible by a world-wide funding initiative.
The Encyclopedia Now Needs Your Support
Please Read How You Can Help Keep the Encyclopedia Free

Browse

About

Support SEP

Mirror Sites

View this site from another server:

USA (Main Site)Philosophy, Stanford University

Info about mirror sites

Library of Congress Catalog Data: ISSN 1095-5054

Movatterモバイル変換