Abstract
Applying speech act theory to the phenomenon of hate speech, some philosophers seek to explain how even ordinary people can obtain the capacity, power, or authority to oppress, subordinate, or marginalise the targets of their verbal attacks. Such explanations are answers to what is called the authority problem. However, hitherto these philosophers have focused exclusively on standard examples of racist speech in which members of historically oppressor groups verbally attack members of oppressed groups. In this paper, I address the (or an) authority problem in relation to examples of reverse hate speech, where members of historically oppressed groups verbally attack members of oppressor groups. I critically examine two conventional answers to the authority problem: the first is that even ordinary people can successfully perform prototypical hate speech acts by virtue of being participants in norm-governed social practices of racism; the second is that such people can do so by virtue of the silence of bystanders granting them licensed authority. I argue these conventional answers have lesser explanatory heft as applied to examples of reverse hate speech. I then introduce and defend a third answer to the authority problem, which I believe can adequately account for examples of both standard and reverse hate speech, namely the possession of informal social authority.
I Introduction
Concepts of hate speech are often explained, although not exclusively, in terms of canonical hate speech acts such as stereotyping, essentialising, Othering, discriminating, oppressing, ranking as inferior, subordinating, marginalising, excluding, threatening, inciting, and so on. It is widely supposed that some, but not all, such acts require the speaker to possess some form of power, capacity, or authority for successful performance. The authority problem consists in plausibly accounting for how different kinds of speakers, including ordinary people (i.e. people who lack formal authority), could acquire the authority to perform certain hate speech acts.
Various answers to the authority problem have been canvassed in the philosophical literature on hate speech (Barnes 2016; Bianchi 2019; Brown and Sinclair 2023, 59–61; Cousens 2023; Gelber 2021; Langton 1993, 302–3; Langton 2012, 2018a; Levin 2010; Maitra 2012; Matsuda 1989; McGowan 2009, 2012, 2019; Sadurski 1999, ch. 4). 1 However, the standard examples discussed in the literature all involve attacks by members of groups or classes of people who are, generally speaking, powerful, dominant, advantaged, protected, in good standing, and/or historical oppressors, against members of groups who are relatively powerless, subordinate, disadvantaged, vulnerable, low status, and/or historically oppressed. By comparison, a lot less has been said about the authority problem in relation to attacks by members of oppressed groups against members of oppressor groups (see Richardson-Self 2018; Brown and Sinclair 2023, 130–65; Lepoutre 2023). I call such attacks reverse hate speech.
At first glance, reverse hate speech would seem to intensify the authority problem: if it is a challenge to explain how ordinary people (i.e. people who lack formal authority) obtain the power, capacity, or authority to perform hate speech acts, it seems an even tougher challenge to explain how ordinary people can do so when they are themselves members of oppressed groups and are verbally attacking members of oppressor groups. This paper seeks to address this gap in our understanding of both the (or an) authority problem and reverse hate speech.
Now it might be objected at this early stage that the phrase ‘the authority problem’ was originally coined in relation to standard cases involving speakers who are members of oppressor groups and targets who are members of historically oppressed groups. However, as I see it, at its heart, the (or an) authority problem is about explaining how even ordinary people could acquire the authority to perform recognisable hate speech acts, including but not limited to acts such as subordinating or marginalising, for instance. And so, examples of reverse hate speech can trigger an authority problem which is also valid and important, even if it is not ‘the authority problem’ as originally cast in the literature.
I look at three answers to the authority problem: first, that sometimes speakers are able to successfully perform certain speech acts by virtue of being involved in norm-governed social practices of racism (Gelber 2021; McGowan 2009, 2012, 2019, 2019); second, that sometimes speakers can obtain authority by virtue of being licensed by the silence of audiences (Maitra 2012); and third, that sometimes speakers can obtain informal authority due to their social position (Brown and Sinclair 2023; Langton 2018a, 59–61). This short list of answers is not exhaustive nor intended to be so. Other ways might be relevant in some situations. 2 Nonetheless, I light upon these three answers because of the prominence given to the first two in the existing literature on the authority problem and the relative lack of attention given to the third answer. Moreover, I look at these three answers because on initial appearances they all seem transposable to examples of reverse hate speech: first, even persons involved in reverse hate speech could potentially be involved in norm-governed social practices of racism, depending on the scope of those practices; second, audience silence might also seem to be a potential source of authority for reverse hate speakers; and third, it is conceivable that even a reverse hate speaker could possess informal social authority, under the right circumstances. The aim of the paper is to assess whether initial appearances do in fact translate into equal explanatory heft after more careful and sustained analytical inspection, and, if not, therefore, whether a pluralistic approach is needed to explain all cases.
In a highly abstract sense, all answers to the (or an) authority problem posit some kind of basic capability, power, or authority, namely something of the order required to successfully perform certain sorts of wrongful or harmful speech acts involving hate speech. The main substantive differences are the varieties and the sources or grounds of the relevant capability, power, or authority, and the relevant speech acts being theorised.
I shall argue that, in reality, the first two answers (norm-governed social practices and licenced authority, respectively) are harder to transpose to examples of reverse hate speech than the third answer (informal social authority), and are inferior because of that. By way of clarification, however, when I say the first two answers are inferior, I do not mean to claim they have zero explanatory heft and that the third answer can explain every situation. Rather, I mean that explanatory heft is comparative and that the first two answers have less explanatory heft than they promise and less explanatory heft than the third answer, specifically in relation to cases of reverse hate speech. Moreover, the lesser degree of explanatory heft is significant as opposed to trivial: the first two answers are unable, whereas the third answer is able, to explain one basic and arguably representative example of reverse hate speech (i.e. test case).
All of that being said, none of these three answers is likely to be capable by itself of explaining every single example of standard and reverse hate speech – nor are they intended to do so by those who propose them. 3 To assume that explanatory adequacy means the ability of a theory (qua answer to the authority problem) to explain every example would set the bar too high. In fact, in the conclusion I will defend a pluralist approach as being needed precisely because no single theory (or answer) is capable of explaining every kind of situation.
Before looking at these three answers, however, I first need to run through some necessary conceptual preliminaries relating to hate speech, speech acts, and authority. I shall also introduce the relevant test case.
II Concepts of hate speech
Whilst concepts of hate speech are often partly understood in terms of speech acts (see Altman 1993; Brown 2017b; Gelber 2021; Langton 2012; Langton et al., 2012; McGowan 2009, 2012, 2019), they do not begin and end with speech acts (Brown 2017a). Concepts of hate speech are also closely related to the target, style, message, act, and effect of hate speech, for example (Brown 2015; Brown and Sinclair 2023; Delgado and Stefancic 2004; Matsuda 1989; Sadurski 1999; Sumner 2004; Tsesis 2002). Nevertheless, for the purposes of this paper I shall focus exclusively on hate speech acts. I do so because it provides an especially acute and interesting lens through which to see the difference between standard and reverse hate speech.
Of course, reverse hate speech is not the only controversial grey area or hard case when it comes to understanding concepts of hate speech. Other grey areas include misgendering, blackface, and righteous hate speech, for example (Brown and Sinclair 2023, chs. 3 and 4); and also microaggressions (Elder 2021; Brown and Sinclair 2023, 320–1, 493). However, this raises the spectre that so-called reverse hate speech is a proxy for righteous hate speech. I shall address this possibility in the conclusion, but for now I shall simply treat reverse hate speech as a sui generis form of hate speech.
Now if the phrase ‘authority problem’ is interpreted to mean purely and simply the problem of explaining how certain speech can oppress or subordinate others, and if oppression and subordination are interpreted to mean something like being put into and/or kept in a low social, political, and economic position within society as whole, then at first sight it is unclear that reverse hate speech could trigger an authority problem. For, some people might argue that it is impossible for targets of reverse hate speech to be victims of oppression and subordination in the aforementioned sense. However, there is more to hate speech than illocutionary oppression and subordination. Hate speech is a heterogeneous and broad concept that is associated with a bewildering array of illocutionary acts (see Brown and Sinclair 2023, 58–59). And so, there remains the pressing problem of explaining how ordinary speakers can gain the power or authority to successfully perform all those other recognisable hate speech acts, including in cases of reverse hate speech. This is a socially and politically pressing matter partly because some members of oppressor groups harbour the grievance that they are also sometimes verbally attacked on the basis of their protected characteristics, and it is epistemically unjust to precipitously dismess (i.e. without careful consideration) the claim they are also victims of hate speech (ibid, 130–65).
Nevertheless, some people might still object to the very notion of reverse hate speech on the grounds that it is a confusion of the term ‘hate speech’ in its original meaning. To explain, in her seminal article ‘Public Response to Racist Speech: Considering the Victim’s Story’ published in 1989, Mari Matsuda coined the term ‘hate speech’ to refer to the sort of racist speech that ought to be legally punished. Moreover, Matsuda argued the term should only apply to speech by oppressors against the oppressed, which is to say that hate speech as she conceived it is based on certain power dynamics and is unidirectional. On Matsuda’s definition of the legal concept hate speech, to speak of reverse hate speech would be an oxymoron.
However, to assume that hate speech remains an essentially unidirectional concept is to fundamentally misunderstand both the legal and ordinary concepts of hate speech as they are today, even if this is not how Matsuda originally conceived them. For one thing, current legal usage of the term ‘hate speech’ and associated hate speech laws clearly part company with Matsuda on the question of unidirectionality. Nearly all current legal usage treats illegal hate speech as multidirectional, meaning hate speech laws apply as written equally to the speech of members of oppressed groups against members of oppressor groups and to the speech of members of oppressor groups against members of oppressed groups (see Brown 2015, ch. 2). Indeed, one of the common objections to hate speech laws has been the fear that although as written they are multidirectional, as applied they will be used disproportionately against the speech of members of oppressed groups. 4 This objection concedes the point that speech by oppressed groups against oppressor groups can be illegal hate speech; the objection is that such speech should not be made illegal disproportionately. Importantly, the facial multidrectionality of hate speech laws stems from the fact that they are typically defined in terms of general protected characteristics like race or ethnicity as opposed to subcharacteristics like being Black or subgroups like Black people (Brown 2016, 276–81). 5 Given this standard legal usage, the qualified terms ‘reverse hate speech’ or ‘anti-White hate speech’ can be meaningful insofar as they deliberately pick out only a certain subset of illegal hate speech, akin to how the phrases ‘standard hate speech’ or ‘anti-Black hate speech’ might pick out only a certain subset of illegal hate speech.
A further point worth emphasising here is that not only has conventional legal usage of the term ‘hate speech’ parted company from Matsuda’s definition but the decades since 1989 has seen the evolution of two concepts of hate speech: the legal concept and the ordinary concept (Brown 2017a; Brown and Sinclair 2023, chs. 1 and 5). For example, the ordinary concept covers speech such as racial slurs, for instance, that used in isolation would be unlikely to be unlawful based on the most common varieties of hate speech laws such as laws banning incitement to hatred, discrimination, and violence. Social media companies have played a significant role in defining and popularising the ordinary concept of hate speech. But the crucial point here is that, like the legal concept, the ordinary concept is also multidirectional. Consider Meta’s (Facebook, Instagram) policy on hate speech. 6 Even though the public-facing ‘community standard’ on hate speech does not explicitly state that the policy is multidirectional and none of its illustrative examples are forms of reverse hate speech, there is clear evidence from current patterns of enforcement and other public statements and off the record comments from company staff that the policy is multidirectional (Brown and Sinclair 2023, 134–7).
In order to put some flesh on the bones of the idea of reverse hate speech, and to provide a test case for the above-mentioned answers to the authority problem (norm-governed social practices, licensed authority, informal social authority), I shall now provide a basic hypothetical example or vignette of reverse hate speech. Suppose an African American woman brings her White boyfriend to a family barbeque at which he is the only White person present. Now suppose her uncle pulls the boyfriend aside and, with grim faced sincerity, drops the following bombshell. (1) Just so you know whitey, we don’t want your kind here.
I take it as read that (1) could be hate speech, semantically, and that, as described, the vignette might also involve the uncle’s successful performance of a wrongful or harmful hate speech act – for example, the act of marginalising someone on the grounds of their race. The three answers to the authority problem are to be tested against this vignette as a sort of benchmark measure of their relative explanatory heft.
As a final clarification, the broad conceptual framework I shall employ in thinking about the family barbeque vignette owes an intellectual debt to the philosopher of language J. L. Austin. In his 1962 book How to do Things With Words Austin distinguished between the ‘locutionary’, ‘illocutionary’, and ‘perlocutionary’ force of words. In this paper I focus on illocutionary force, which is essentially a matter of the acts people might perform or not perform with words. As Austin (28–29, 160) pointed out, some illocutionary acts require for their successful performance that the speaker have ‘authority’. In the words of Rae Langton (1993, 305), ‘authoritative illocutions [are] actions whose felicity conditions require that the speaker occupy a position of authority in a relevant domain’. Which illocutions are also authoritatives? Austin (1962, 154) proffered the English verbs ‘to demote’, ‘to excommunicate’, and ‘to degrade’ as examples pertaining to a particular family of illocutionary acts he labelled ‘exercitives’. ‘[T]he exercitive is an assertion of influence or exercising of power’ (162), including ‘the giving of a decision in favour of or against a certain course of action, or advocacy of it’ (154). In other words, ‘[i]t is a decision that something is to be so, as distinct from a judgement that it is so’ (ibid.). When the boss of a company, Smyth, utters the words ‘Jeffers, you’re demoted’ to an employee, for example, the boss is making a decision that Jeffers is to be demoted or is exercising his power to demote Jeffers. It is not hard to see why this act would be an authoritative illocution. It is only by virtue of the fact that the boss occupies the relevant position of authority, namely that the boss possesses the right or power to demote, that the boss successfully performs the exercitive act of demoting Jeffers.
What about prototpypical hate speech acts such as discriminating, ranking, marginalising, excluding, inciting, and so on? Returning to the family barbeque vignette, one way to interpret the uncle’s behaviour is as an act of marginalising the White boyfriend, that is, giving a decision that the boyfriend should not be made welcome at the barbeque. The decision implies that the boyfriend should be treated not as a valued guest but as an unwelcome intruder, such as given the cold shoulder, ignored, spoken to harshly, given dirty looks, subject to teeth sucking, not offered food and drink, and so on. This act also looks like an exercitive (‘the giving of a decision in favour of or against a certain course of action, or advocacy of it’).
Several further things need to be made clear at this stage. For one thing, like most examples, it may be possible to interpret the family barbeque vignette differently, including as involving more or less serious hate speech acts. 7 However, for the sake of simplicity, I shall focus solely on the issue of how the uncle has performed the hate speech act of marginalising the White boyfriend, specifically the act of deciding for others and/or calling on others to treat him as an unwelcome visitor such as by giving him the cold shoulder. Arguably, the act of marginalising the White boyfriend is not quite as serious as the act of giving a decision that the boyfriend should be excluded entirely from the family barbeque, that is, a decision that the boyfriend ought to be physically ejected from the property. There are countless ways a person could be made unwelcome without actually being removed.
A second point is that to qualify as an exercitive and an act of marginalising the White boyfriend, arguably it is not enough that the uncle merely makes a personal decision for himself about the boyfriend, such as the decision that the boyfriend will not be made welcome by the uncle (‘As far as I am concerned, you are not going to be made welcomed by me’). That sort of act might be more accurately classified as an ‘expressive’, ‘behabitive’, or ‘commisive’ within standard taxonomies of illocutionary acts (Austin 1962, 83, 154–62; Searle 1975, 10, 12). Rather, there must be a sense in which the uncle gives a decision on behalf of everyone present that the boyfriend should not be made welcome by people at the barbeque. Furthermore, if the uncle speaks authoritatively, and other conditions are met, then the course of action he speaks in favour of will come to pass, and the boyfriend will be made unwelcome by everyone at the barbeque, meaning they come to see and treat him as unwelcome. Nevertheless, I do not mean to say that if the uncle’s merely publicly asserted a personal decision that the boyfriend will not be made welcome by him, then this could not count as hate speech. But such an act would not be an exercitive nor an authoritative; instead it might be a behabitive or even a commissive, for example. In this paper, I want to interrogate examples of reverse hate speech through the lens of the authority problem.
A third point is that prototypical hate speech acts come in many different varieties and are certainly not limited to speech acts of marginalising others (Brown and Sinclair 2023, 58–59). Many philosophers of hate speech focus on speech acts of subordinating others, including ranking others as inferior (Gelber 2021; Langton 2012; Maitra 2012). If a particular speech act were purely the rendering of a judgement that certain groups of people are socially inferior and thereby count as lesser in the eyes of others, then it might not qualify as an exercitive under Austin’s classification given the distinction he himself draws between exercitives and another family of illocutionary acts, namely ‘verdictives’. On the one hand, an exercitive ‘is a decision that something is to be so, as distinct from a judgement that it is so’ (Austin 1962, 154). On the other hand, Austin cites the English verbs ‘to rate’, ‘to grade’, and ‘to rank’ as examples pertaining to verdictives: ‘the delivering of a finding, official or unofficial, upon evidence or reasons as to value or fact so far as these are distinguishable’ (ibid., 152). Then again, there might be instances in which certain hate speech acts involve both verdictives and exercitives. Consider someone who uses hateful language in such a way as to combine the judgement that certain groups of people are socially inferior and count as lesser in the eyes of others with the giving of a decision that, on this basis, henceforth they are also to be treated as inferior. Indeed, Austin makes clear that sometimes exercitives can be based on verdictives (155). Consider a judge handing down a prison sentence, as in, sentencing the defendant (exercitive), on the basis of an earlier judgement that the defendant is guilty (verdictive). At any rate, what is relatively clear is that both exercitives and verdictives are authoritatives (Brown 2019b, 210; Langton 1993, 304), and so both types of speech act trigger the authority problem.
Finally, it is important to clarify at the outset that the uncle, as far as we know, does not possess any formal authority, at least no formal authority on a par with the formal authority possessed by the boss in the demotion example mentioned earlier or a judge in a court case, for insance.
Having cleared the ground, I shall now assess the three explanations as to how even a person such as the uncle, who is lacking a position of formal authority in the discursive situation, might nevertheless obtain the requisite power, capacity, or authority to successfully perform a recognisable hate speech act against his target.
III Norm-governed social practices
As Mary Kate McGowan has theorised, one way ordinary people might obtain the capacity to successfully perform hate speech acts is by being participants in rule- or norm-governed social practices of racism (McGowan 2009, 132–8; 2012; 2019; 2022; see also Brown 2015, 86–90; Gelber 2021, 402–3; Lepoutre 2023, 545–6). According to this broad type of explanation, such social practices are governed by secondary rules or general norms, G-norms, of the sort that effectively give speakers the power or capacity to enact primary rules or specific norms, S-norms, including S-norms concerning the targeting of individuals. Strictly speaking, this answer to the authority problem says that the power or capacity to enact the S-norms – enactments that could constitute acts of subordination or marginalisation, for example – is located not in the speaker as such but in the G-norms upon which they rely.
Nevertheless, the exact nature of the G- and S-norms, and the sorts of wrongful or harmful acts involved, depends on the situation and requires analyses of a sociological as well as philosophical-linguistic kind. As McGowan herself puts it: The social world is a messy place; g-norms are both complex and contested. We need social science to settle some of the relevant issues. What the g-norms are, what the harms are, how speech brings those harms about, and what the remedies ought to be are all questions that require multiple types of expertise to satisfactorily settle. As a philosopher of language, my main contribution is the identification of this hidden mechanism of norm enactment as well as establishing its relevance to these issues. In short, I am a firm believer in the division of intellectual labor. (McGowan 2022, 1072n.38)
I concur with McGowan that one of the insights that a sociological analysis of speech pragmatics can provide is the additional distinction between norms that have a wide scope qua social conventions applicable to many speech situations or contexts, or even applicable across all of society, what I shall call wide-scope norms, and norms that have a more narrow or limited scope or are more ‘localized’ in the sense that they are applicable only to particular speech situations or highly contextualised settings, what I call narrow-scope norms (see also McGowan 2019, 85).
To explain this distinction, I shall now set forth a pair of descriptions of wide scope and narrow scope norms, respectively. At this stage, however, I may be parting company from McGowan since it is unclear whether she would recognise these as instances of what she calls ‘norms’ (see McGowan 2009, 396–7, 2019, 84–86). Wide-Scope Norms 1 (WSNs1): There exists a general racist social practice across society which is governed, among other things, by the following secondary rule or general norm G: If a White person addresses the phrase-type ‘Hey nigger, we don’t want your kind here’ to a Black person, then, other things remaining equal, the White person is giving a decision for other people at the relevant place or event that a primary rule or specific norm S should hold which says the Black person should not be made welcome. Narrow-Scope Norms 1 (NSNs1): At a majority-White golf club called Forrest Pines, there is a particular racist social practice governed, among other things, by the following secondary rule or general norm G: If a White club member walks up to a Black person and says, ‘Hey nigger, we don’t want your kind here’, then, other things remaining equal, the White club member is giving a decision for other club members that a primary rule or specific norm S should hold which says the Black person should not be made welcome at the golf club.
The WSNs1 are norms to the effect that a White person can make a decision for others that a Black person should not be made welcome at a place or an event. The NSNs1, by contrast, are norms to the effect that a White member of the Forrest Pines golf club can make a decision for other club members that a Black person should not be made welcome at that golf club. However, both involve highly complex G-norms that are conditional and social in nature. 8 They govern what happens/should happen when a speaker uses specific words towards certain categories of people in given circumstances.
It is important to ask what kind of relationship exists between WSNs and NSNs. Are NSNs derivative norms or grounded in WSNs meaning that the former hold only in virtue of the latter? Or are the two kinds of norms independent in the sense that, for example, NSNs1 could hold even if WSNs1 did not hold? I believe that it is conceivable for these different kinds of norms to hold independently of each other and that this independence opens up at least the possibility of speakers, based on NSNs, being able to use reverse hate speech with the power to perform certain sorts of wrongful or harmful acts, even in the absence of corresponding WSNs. However, I also believe that the two kinds of norms, WSNs and NSNs, are mutually influencing in the sense that the absence of certain WSNs can mean it is less likely for certain NSNs to occur in practice. This, then, makes the presence of powerful reverse hate speech if not impossible, then at least less likely to occur, under the norm-governed social practices theory, compared to the occurrence powerful standard hate speech.
Before making this argument, I need to clarify two things. The first is that identifying relevant norms in cases of reverse hate speech does not solely depend on finding a semantically equivalent slur to ‘nigger’, even if that were possible (Brown and Sinclair 2023, 130–65). Rather, what matters is finding cases in which people of colour deploy a string of words against White people, perhaps in a social context below the level of the entire society, in a manner capable of rising to the level of recognisable hate speech acts like subordinating, marginalising, or inciting hatred against them (ibid., 152, 162–3).
It is also worth clarifying that, according to McGowan herself (2012, 143–4), ‘plenty of uses of racial epithets (and perhaps especially those by people of colour) do not constitute discriminatory acts’. Nevertheless, this comment falls short of the stronger, blanket claim that no reverse attacks of any form could ever constitute recognisable hate speech acts such as oppressing, subordinating, marginalising, or inciting, under any circumstances. McGowan (as far as I am aware) does not make that stronger, blanket claim; and I believe she is right not to do so.
The political scientist Katharine Gelber relies on McGowan’s account to construct a definition of hate speech as ‘discriminatory speech’. In places, however, Gelber does appear to rule out the possibility of reverse attacks constituting recognisable hate speech acts. For example, Gelber writes: Some group identities constitute straightforward cases in which the speech would not constitute hate speech, since the slurs are not representative of, and a move in perpetuating, systemic discrimination. Examples of these kinds of utterances include claims of racism against White people in Western democratic, liberal countries like Australia and the United States, in which although some people might express views of dislike, even intense dislike, towards White people, systemic discrimination against, and disadvantage of, whites as a group has not been evidenced. Therefore speech-acts cannot be moves in a rule-governed activity that systemically marginalizes and discriminates against White people. (Gelber 2021, 404) I am aware of, but unconvinced by, arguments that groups such as men or White people do face systemic discrimination. (410n.13)
However, in other places Gelber concedes there is more than one way for speakers to obtain power to perform hate speech acts. Being participants in rule- or norm-governed activities of racism and discrimination is one way but not the only way (403, 407). I shall consider two other ways in Sections III and IV, respectively. Before doing so, however, I shall focus on Gelber’s scepticism.
At first glance, Gelber appears to overlook the possibility of speakers operating within ‘domains of influence’ or ‘communicative jurisdictions’ (Langton 1998, 264) that exist at a level below that of the entire society (sub-societal). I have in mind particular speech situations, or what I shall call discursive microenvironments, in which speakers can perform hate speech acts qua participants in limited or highly contextualised social practices of racism and discrimination, including those in which standard society-wide racism and systemic discrimination does not apply, at least not fully, and is even reversed. Put simply, could not there exist within a particular Black subcommunity or even within a small-scale speech situation or discursive microenvironment like a family barbeque, a norm-governed practice of anti-White prejudice of the sort that could give ordinary Black speakers the power to perform recognisably wrongful or harmful hate speech acts against White people?
Consider the following description of some narrow-scope norms (NSNs) that are applicable only to a particular discursive microenvironment, namely the family barbeque introduced earlier. Narrow-Scope Norms 2 (NSNs2): At a particular family barbeque, there is a particular racist social practice governed, among other things, by the following secondary rule or general norm G*: If a Black family member walks up to a White person and says, ‘Hey whitey, we don’t want your kind here’, then, other things remaining equal, the Black family member is giving a decision for others at the barbeque that a primary rule or specific norm S* should hold which says the White person should not be made welcome at the barbeque.
If the uncle at the family barbeque says to the White boyfriend, ‘Hey whitey, we don’t want your kind here’, and assuming the general norm G* holds, then, other things remaining equal, would it not also be true that the uncle has enacted a specific norm S* which says the boyfriend should not be made welcome at the barbeque? And would it not be the case that the uncle has thereby successfully performed the recognisable hate speech act of marginalising the White boyfriend on grounds of race (i.e. enacting an S* rule for others which says a White person should not be made welcome at the family barbeque)? 9 This would seem to hold even in the absence of a set of corresponding WSNs that make the use of ‘Hey whitey, we don’t want your kind here’ conventionally marginalising across society as a whole or in many situations.
However, in cases of reverse hate speech things can be much less straightforward than they seem. Compared to the example of standard hate speech at the majority-White golf club and the general norm G, arguably it is much less likely that the general norm G* will emerge (i.e. come to exist), even at the particular level of the family barbeque. Perhaps it is more difficult in a majority-White society for a racist social practice of the sort governed by G* to take hold. This might be due in part to the absence of a history of widespread, sustained, and systemic anti-White racism and discrimination across society as a whole (e.g. attitudes, activities, practices, campaigns, movements, structures). 10 It may also reflect the fact that many Black people are focused on the day-to-day challenges of living in a majority-White society with its systemic anti-Black racism, and so they unconsciously close their minds to or consciously reject the ideology and performance of anti-White racism for reasons of self-preservation (being anti-White in a White majority society would be imprudent) or principle (they reject racism in all its forms because they know at first-hand how racism feels). This might make it far less likely that G* emerges. The absence (or impotence) of G* might also be due to the ability of White speakers to block certain sorts of speech acts, such as by saying things like ‘You have absolutely no right to tell me I’m not welcome here’.
Putting this the other way around, in the case of standard hate speech there are good reasons to think the secondary rule or general norm G would emerge at a macro societal level. One is that it is likely to have been instantiated on many occasions previously. A key point is that the performance of a hate speech act is not a singular isolated act but the repetition of similar acts that make up the history of the racist activity and together these acts are instantiations of a secondary rule or general norm. In the words of Judith Butler, ‘hate speech is an act that recalls prior acts’ (1997, 20). ‘If a performative provisionally succeeds […], then it is not because an intention successfully governs the action of speech, but only because that action echoes prior actions, and accumulates the force of authority through the repetition or citation of a prior, authoritative set of practices’ (51). ‘The speaker who utters the racial slur is thus citing that slur, making linguistic community with a history of speakers’ (52). By contrast, if phrase-types like ‘Hey whitey, we don’t want your kind here’ are not associated with a history of speakers, then the general norm G* may not come to exist across society as a whole and might be unlikely to emerge even in particular settings or small-scale social environments (discursive microenvironments) such as a family barbeque. So, if G* does not exist in that particular setting or discursive microenvironment, and partly because it does not exist across society as a whole, then a Black speaker who utters the phrase ‘Hey whitey, we don’t want your kind here’ in that particular setting may also fail to enact a specific norm S* that White people should not be made welcome.
Another reason to suppose the general norm G could emerge at a macro societal level is if a substantial number of people living in the society are not merely aware of G but also come to expect that other people will live up to G and are responsive to them doing so. 11 Some people may even internalise G, meaning their recognition of it becomes second nature. These contributing factors are themselves likely to reflect the widespread presence of anti-Black racism across society, including a range of interconnected and mutually reinforcing prejudiced attitudes and discriminatory practices. In the absence of these contributing factors in the case of reverse hate speech, partly due to the absence of a large supporting cast of anti-White prejudiced attitudes and discriminatory practices, it seems less likely the general norm G* would come to exist at the society-wide level and, consequently, less likely, although not impossible, that G* will emerge in particular speech situations or highly contextualised settings.
In other words, I believe successful performance of wrongful or harmful reverse hate speech acts is still likely to be more uncertain than the successful performance of standard hate speech acts even when the relevant general norm G* is a NSN, as in, has narrow scope covering only a particular or highly contextual discursive microenvironment. Suppose the uncle repeats the phrase ‘Hey whitey, we don’t want your kind here’ multiple times during the course of the family barbeque but is the only person doing so. Arguably, even though this would constitute a history of previous usage, it may not be sufficient to establish the secondary rule or general norm G*, such as if the phrase is not given any credence by the rest of the people at the barbeque. Even if the uncle succeeds in performing the brute act of repeating the phrase multiple times without interruption, this in itself does not demonstrate the presence of a racist social practice governed by the secondary rule or general norm G*. After all, the use of racist language is often social or collective in nature and this means that other people at the barbeque and not solely the speaker get to determine the meaning and significance given to the offending phrase and whether it is part of a norm-governed racist social practice. If other people at the barbeque do not understand the meaning and significance of the phrase to be such that its usage constitutes a racist social practice governed by the secondary rule or general norm G*, then G* is unlikely to emerge. As a result, the uncle would be unable to rely on G* to enact a primary rule or specific norm S* and as such unable to perform the act of marginalising the boyfriend, according to the current account.
Nevertheless, what if not just the uncle but several other people at some point during the family barbeque said to the boyfriend ‘Hey whitey, we don’t want your kind here’? Once again, however, the mere repetition may not be decisive in establishing the existence of the general norm G*, that is, to make it come to exist. After all, these other people might be using the phrase with a very different meaning and intention to that of the uncle. For example, they might be using it to break the ice or to make the boyfriend feel part of the family through playful ribbing (in their eyes). Arguably, in order for iterations of the offending phrase to make the general norm G* emerge, a significant number of the speakers (qua putative participants in the norm-governed social practice) would also need to share the aim of marginalising the boyfriend and this would need to be the dominant aim if there were more than one aim in play. Yet insofar as the family barbeque is representative of the wider Black community, it is questionable whether these conditions would be met in the real world.
Of course, someone might argue, as McGowan (2012, 132) does, that speaker intentions and awareness are not needed for speakers to enact S-norms – where this enactment may constitute wrongful hate speech acts of subordination or marginalisation, for example – precisely because the G-norms are already in play. However, the pressing question is what grounds the general norms, as in, what makes G-norms come to exist. I would argue that intentions and awareness are needed for the creation of some general norms (for some general norms it is hard to see how those norms could emerge without intentions and awareness). General norm G* sets in motion normative expectations about what happens/should happen when speakers use the phrase ‘Hey whitey, we don’t want your kind here’, and this would seem to require intentions and awareness of at least some kind.
Furthermore, to claim that a social practice is norm-governed is not merely to say that within the social practice some behaviours are rendered appropriate or inappropriate by the governing norms (e.g. welcoming or not welcoming certain people); normativity in human activity also means that the governing norms exert a special ‘pull’ on participants in the activity. In other words, the norms give participants in the activity a reason to behave or not behave in certain ways simply because the relevant behaviour is rendered appropriate or inappropriate by those governing norms. If the general norm G* is active or present but only exerts a pull on the uncle, then one could argue that strictly speaking he is the sole participant, or the only full participant in the aforementioned sense. And if the uncle is the sole participant, then it seems implausible to describe him as capable of making a decision on behalf of other people at the barbeque that the boyfriend should not be made welcome (i.e. capable of enacting a specific norm S* that governs the behaviour of everyone who is present towards the White boyfriend).
Similarly, even if the general norm G* did come to exist, that does not mean the uncle could automatically rely on G* to enact a specific norm S* and thereby successfully perform the act of marginalising the White boyfriend. After all, another member of the family could block the uncle’s hate speech act by challenging the validity of the general norm G* and/or by invoking a competing rule, such as by saying the following. (2) Nothing you said makes the slightest bit of difference to whether her boyfriend should be made welcome here because he is our guest and the most important thing is that we do right by our guests and make them welcome; you can’t give a decision for everyone that he shouldn’t be made welcome here simply by saying the crap you said.
The key point here is that, like all social events, there is likely to be more than one norm-governed social practice occurring simultaneously at the barbeque, meaning other family members could invoke alternative rules or norms (e.g. norms of hospitality, comity, and anti-racism) to block the uncle’s attempt to rely on G* to enact S*, including through their public utterance of (2), for example.
Of course, if the family in question were all avowed anti-White racists, then several speakers at the barbeque might use the offending phrase to attack the White boyfriend. What is more, members of the family would probably attach a shared meaning and significance to the offending phrase such that G* is active or present and they would be unlikely to block the uncle’s act of marginalising the White boyfriend. However, if the family were of this sort, then it is highly unlikely the woman would have a White boyfriend or, assuming she is the sole non-racist, that she would choose to bring him to the barbeque knowing the hostile reception he would receive.
There is an additional concern about applying the norm-governed social practice theory to examples of reverse hate speech. When thinking about racist social practices governed by general norms like G*, what qualifies speakers as participants in those social practices such that the relevant norms apply to them and give them the capacity to enact specific norms such as S*? The need to answer this further question means that, for McGowan and other defenders of the theory, the authority problem is supplemented with what I shall call the participant problem.
For example, in order for the uncle to qualify as a participant in a racist social practice governed by the general norm G*, must he first prove his ‘standing’ or ‘credentials’ as an anti-White bigot or does he qualify simply due to the brute fact of using the phrase ‘Hey whitey, we don’t want your kind here’? 12 If the latter, does this not beg the question? After all, surely it is possible for people to use the phrase yet not be participants in the social practice, which is to say they could use the offending phrase but fail to enact the specific norm S* because the general norm G* does not apply to them (because they are non-participants, meaning persons whose actions fall outside of the scope of the social practice and its norms). To insist on the impossibility of such an outcome merely by stipulating that every user of the offending phrase is by definition a participant in the social practice would be to simply assume the general norm G* is present and universally applies rather than to demonstrate that it does.
Similarly, must the uncle have some understanding or awareness of the general norm G* to qualify as a participant in the racist social practice governed by it? McGowan (2012, 132) thinks a speaker having ‘awareness’ of the norms that enable them to subordinate others is not a necessary condition of being a participant in the relevant norm-governed racist social practice. But this seems implausible (see also Brown 2015, 88–89). Suppose the woman’s three-year-old brother repeats the uncle’s phrase ‘Hey whitey, we don’t want your kind here’ without having the slightest idea of its meaning and significance, and does so merely because he likes repeating what adults say or because he mistakenly thinks it is how adults are welcoming the boyfriend. Would that qualify the little brother as a participant in the racist social practice governed by G*? Does the general norm G* apply to the little brother such that merely by using the offending phrase he has also successfully performed the hate speech act of marginalising the boyfriend? Surely not.
Putting the participant problem slightly differently, perhaps all White people could be seen as potential participants in the racist social practice governed by the secondary rule or general norm G – that is, as having G apply to them or having standing under G – simply by virtue of being White – that is, by virtue of being members of a group (White people) that has been involved in the history of anti-Black racism and has benefited from racial hierarchies between Black and White people (Cousens 2023). By contrast, if some sort of general norm G* were to be present in relation to a social practice involving reverse hate speech, it is unclear that it would take the same unqualified form as the general norm G, which holds for any White person. In fact, if something like G* were to hold, it is doubtful it would hold for any given (i.e. potentially every) Black person. It is doubtful precisely because of the position of Black people as victims of racial oppression and the effect this oppression may have not merely on their mindset towards White people but also their participation in social practices involving such things as giving decisions that White people should not be made welcome. This is not to deny that some Black people are anti-White racists and have the power to perform acts of marginalising White people. Rather, it is to say many Black people are not realistically viewed as participants in anti-White racist social practices. And this surely makes a difference to whether, if something like G* were to hold, it would hold for any given Black person who happened to utter a particular string of words.
All of this would seem to suggest the uncle needs some additional authority to successfully perform the hate speech act of marginalising the White boyfriend (to make the decision on behalf of the people at the barbeque that the boyfriend should not be made welcome). In Sections III and IV, I shall explore two further answers to the authority problem that invoke different kinds of additional authority.
IV Licensed authority
A second way that ordinary people might obtain the authority to successfully perform prototypical hate speech acts is by means of bystander silence granting them the authority (Maitra 2012, 107–17). For example, ordinary people might be able to use certain forms of language to judge or rank others as lesser or socially inferior and thereby make it the case that they have a lower social status in the eyes of others (subordinate) or give a decision that others should be made socially unwelcome and thereby make it the case that they are to be made socially unwelcome (marginalise), if the silence of bystanders licences them to authoritatively perform those acts. Of course, in some situations it might be difficult tell whether all bystanders are part of the same conversation as the speaker (115–6) and whether their silence actually means the granting of authority (Brown 2015, 78–80; Brown 2019b, 211–20). 13 But so long as it is possible to tell at least some of the time, then the answer has some explanatory force.
Maitra originally developed this account in relation to the following standard example of hate speech. An Arab woman is on a subway car crowded with people. An older White man walks up to her, and says, “F***in’ terrorist, go home. We don’t need your kind here.” He continues speaking in this manner to the woman, who doesn’t respond. He speaks loudly enough that everyone in the car hears his words clearly. All other conversations cease. Many of the passengers turn to look at the speaker, but no one interferes. (Maitra 2012, 100–1)
Maitra analyses this as an act of ‘illocutionary subordination’ in the sense of not merely ranking someone as inferior but also thereby making them inferior, as in, making it the case that in the eyes of others their social statutes is lowered. Then again, arguably, the very same scenario might just as plausibly be interpreted as involving an act of stereotyping, Othering, marginalising, or even inciting discrimination.
Two clarifications are needed about this second answer to the authority problem. One is that the relevant authority is grounded in bystanders lending their authority to the speaker. As Maitra (2012, 107) puts it, ‘what authority the speakers come to have is granted to them by their audiences’. ‘Licensing, then, can be regarded as a kind of […] authorizing, of a speaker by relevant others’. In one sense, this authorizing is a matter of bystanders delegating the speaker to act on their behalf, as in, bystanders implicitly giving their permission for their own verdict, decision, or judgement towards certain groups of people to be substituted with that of the speaker.
A second clarification is that one putative benefit of this answer to the authority problem is that it appears to explain how an ordinary person could obtain the authority to successfully perform a prototypical hate speech act even if there is no norm-governed racist social practice connected with the language they are using and even if the norm-governed social practice does exist but the speaker is not a qualifying participant and the relevant norm does not apply to them for some reason.
Now, at first glance, it does not seem hard to imagine situations in which ordinary people could successfully perform reverse hate speech acts based on licensed authority (see also Lepoutre 2023, 543–5). Suppose a Caucasian man enters a grocery store on a street in a predominantly Arab quarter of a Western city. All the other customers and the storekeeper are Arab. One of the customers turns and says the following to the Caucasian man, after which everyone remains silent. (3) Fucking infidel, go use a different grocery store. Your kind do not belong on this street.
In this scenario, it seems plausible to think the bystander silence has a meaning and significance consistent with the bystanders licensing the speaker to perform one or other reverse hate speech act. However, to underscore the importance of situational context, it is worth briefly exploring what might happen if the example is slightly altered so that although the speaker and the target continue to display the formal properties of a reverse verbal attack, all the other customers in the store just so happen to be Caucasian at that particular moment, and the Arab storekeeper did not hear the comment because he was chatting on his mobile phone. Arguably, in these circumstances it is much less plausible to think the speaker obtains an authority to perform the intended hate speech act via bystander silence. This is because bystander silence can no longer be safely assumed to be an instance of licensing. This breakdown in preconditions for licensing might happen in several ways. For example, the Caucasian audience might take the speaker to have performed a different speech act to that which was intended. For example, the Arab speaker might have intended to perform the act of ranking as inferior (verdictive) or marginalising (exercitive) the Caucasian man, but instead the audience of fellow Caucasians might have, because they are Caucasian, interpreted the meaning and significance of the words as merely a diatribe, that is, an act of forcefully and bitterly expressing how the speaker feels about Caucasians (expressive). 14 In this eventuality, arguably, the bystanders’ silence would not have licensed the speaker to rank as inferior or marginalise the Caucasian man.
I believe a somewhat similar failure of licensing might occur in the family barbeque scenario, even if everyone at the barbeque is African American save for the boyfriend. After all, silence might reflect the audience receiving the uncle’s utterance of the phrase ‘Hey whitey, we don’t want your kind here’ merely as him expressing his feelings about White people, as opposed to giving a decision about how the whole family should treat the boyfriend. If the family thinks the uncle is just angry and letting off some steam and does not really mean that everyone should not welcome the boyfriend, then at most the silence would licence the uncle to perform the act of venting (expressive); it would not licence him to perform an act of ranking others as inferior (verdictive) or marginalising (exercitive).
Furthermore, although the licensed authority is grounded in bystander silence (depending on the meaning and significance of the silence), the silence itself is a contingent fact that is causally linked to other contributing factors. And typically, these factors might diverge between scenarios of standard and reverse hate speech. To explain, in scenarios involving standard hate speech, like Maitra’s subway example, silence is causally linked to other factors associated with systemic racism and discrimination. For one thing, silence is often the result of fear of violence. In racist societies there is often a history of temporal conjunction between the use of racist language and violence. Unsurprisingly, people riding the subway carriage might fear violence against the target but also against themselves – or at least aggressive abuse – if they speak up. 15 In addition, in a racist society it is more likely that a proportion of bystanders will remain silent precisely because of their own racist attitudes: which is to say, silence might be a function of the speaker mirroring or giving voice to some of the bystanders’ prejudices against members of the relevant group (Barnes 2016). Moreover, it might be common for bystanders to remain silent in such situations, even though they are not actively racist themselves, because they are experiencing cognitive dissonance caused by their ambivalence. If they have conflicting beliefs about the target group or lack strong feelings of concern, they might experience internal conflict when confronted with overt racism, feeling both disgusted by it but also not compelled to step in despite thinking of themselves as good people – the upshot being they concentrate on or get locked into their own unpleasant emotionals or confusing thoughts and remain silent at the crucial moment.
By contrast, in scenarios involving reverse hate speech, these contributing factors might be less likely to occur and less prominent even when they do occur. For one thing, there might not be the same perception of a threat of imminent violence because historically there is a less frequent connection between anti-White rhetoric and violence. Furthermore, it might be less common for a speaker to be mirroring or voicing what bystanders already think in scenarios of reverse hate speech compared to standard hate speech, such as if in the relevant society anti-White prejudice is less widespread than anti-Black prejudice. In the words of Thomas C. Grey: The denigrating epithets [e.g. ‘nigger’] draw their capacity to impose the characteristic civil-rights injury to “hearts and minds” from the fact that they turn the whole socially and historically inculcated weight of these prejudices upon their victim. Each hatemonger who invokes one of these terms summons a vicious chorus in his support. It is because, given our cultural history, no such general prejudices strike against the dominant groups that there exist no comparable terms of universally understood hatred and contempt applicable to whites, males, and heterosexuals as such. (Grey 1992, 504)
In addition to this, in scenarios of reverse hate speech, it is perhaps less likely that the speaker will find an audience of people frozen by their own internal conflict or ambivalence. Being more likely to have experienced racial discrimination themselves and being strongly motivated to be unlike the racists that have tormented them, at least some people at the family barbeque may be quicker to achieve moral clarity in the face of anti-White racism – spared from crippling internal mental struggles they are free to speak up and not remain silent at the crucial moment.
If the absence of these contributing factors predicts that at least one bystander will have the mindset and courage to voice dissent against a reverse hate speaker, then licensed authority based on bystander silence would seem to have less explanatory heft in solving the authority problem in scenarios of reverse hate speech compared to standard hate speech. This is distinct from the previous point about variations in the meaning and significance of silence between scenarios of reverse and standard hate speech. This is a point about the likelihood that there will be silence in the first place.
Sticking with the family barbeque vignette, it seems eminently likely that at least one other person – if not the African American woman who brought her White boyfriend, then someone else – would counter the uncle’s words. (4) For Christ’s sake, don’t speak to him like that!
If the voicing of (4) means there is to be no silence after the uncle’s utterance of the phrase ‘Hey whitey, we don’t want your kind here’, then the current answer to the authority problem no longer applies. At this point, it might be tempting to assume that under the altered scenario in which someone counters the uncle’s words it would be simply impossible for the uncle to still successfully perform the illocutionary act of marginalising the White boyfriend. However, that conclusion seems hasty. For, there are other forms of authority that do not depend on licensing via bystander silence. And it is to one of these other forms that we now turn.
V Informal social authority
A third way ordinary people (i.e. people without formal authority) could successfully perform hate speech acts is by obtaining certain types of informal social authority associated with having a high status, power, or influence within a social group, where this special standing is not related to positions of formal authority (e.g. a boss, a judge). Of course, some kinds of informal authority might be gained in more prosaic ways, such as if in a conversational setting speakers simply presuppose they have authority and nobody else in the conversation objects (Langton 2018a; forthcoming). But I want to focus on a particular sort of informal authority, namely an informal social authority that is grounded in a person’s special standing in a social group.
Now it might be objected even at this early stage that informal social authority, like licensed authority, is ultimately dependent on some sort of act of bystander authorizing, suggesting that this putative third approach could be indistinguishable from an explanation which appeals to bystander silence. After all, surely it must be true that for someone to obtain informal social authority relevant others must recognise that authority. Indeed, one could justifiably assert that social authority partly consists in the recognition by others within a social group of a fellow member’s authority. Some people might think therefore that the proposed third answer collapses into the second answer.
However, informal social authority is not reducible to licensed authority. For one thing, the two forms of authority have different grounds or sources. In the case of licensed authority, bystanders furnish the speaker with the authority to perform the relevant hate speech acts by dint of their silence. For example, bystanders might lend their own prerogative to issue verdicts and make decisions about other people. These bystanders could be anyone in the sense that they need not be known to the speaker and need not belong to the same social group as the speaker. In the case of informal social authority, by contrast, what authority the speaker comes to possess is grounded in their high status, power, or influence within a social group. Who else is in the group and the relationships and power dynamics within the group matters crucially.
What is the nature of this high status, power, or influence? I do not have space here to provide a comprehensive account but can offer some general observations. One is that high status, power, or influence within a social group is a matter of things like good reputation, being favoured, being taken seriously, social capital, sway, control, and so on. Depending on the social group in question, the degree of informal social authority someone enjoys could be a function of various underlying factors common to social group relationships and power dynamics. These could include wealth, occupation, age seniority, gender hierarchy, epistemic standing, practical ability or prowess, physical strength, attractiveness, perceived trustworthiness, similarity, likeability, role or position, moral reputation, and so forth. Which of these factors is the greater determinant of people’s informal social authority is likely to depend on the nature and purpose of the relevant social group. It is a further question what sorts of things might qualify as social groups. 16 But it seems relatively straightforward to say that the family barbeque either constitutes some sort of social group or else is a venue for the coming together of a social group (the family and assorted others).
I believe this third answer to the authority problem has greater explanatory heft as applied to examples of reverse hate speech than the previous two answers (norm-governed social practices and licensed authority). Returning once again to the benchmark case of the family barbeque, it seems reasonable to imagine that the uncle does or could have the informal social authority to successfully perform the illocutionary act of marginalising the White boyfriend. Moreover, this would appear to hold true even if the uncle cannot count on the presence of a secondary rule or general norm G* that supports his enacting of a primary rule or specific norm S* that the boyfriend should not be made welcome, and even if he cannot count on the silence of others in the moments following his utterance of the phrase ‘Hey whitey, we don’t want your kind here’. In other words, his informal social authority seems capable both of explaining his successful performance of the relevant hate speech act and of explaining how that performance could survive an abscence of general norm G* and a lack of bystander silence.
The uncle’s informal social authority seems to be jurisdictional in the sense that it is an authority held over a certain audience, in a particular place and time (see also Langton 2018a, 138–41; Lepoutre 2023, 541–2). But aside from its jurisdictional quality (a meta quality), what is the actual source or grounds of this informal social authority? It might derive from various sources. First, he could be the owner of the home where the barbeque is taking place. Second, he is male and older than other family members. Third, he might have a well-paid job. Fourth, he might play an especially important role in the family, for example, as someone who looks out for elderly relatives and helps to raise children including his nieces and nephews. Fifth, he might have epistemic standing such as if it is widely known he studied for an undergraduate degree in African American studies and was formerly a member of an influential Black power movement and family members perceive him as a credible source of knowledge. Finally, he might have a good moral reputation such as if family members see him as a good person due to his work in the community. Some or all of these factors combined could help to give the uncle an elevated position in the family pecking order in general and at the family barbeque in particular (whose attendees are the relevant social group for the purposes of the analysis).
In addition to this direct explanation, I propose the third answer to the authority problem can also indirectly explain how the uncle is able to successfully perform a reverse hate speech act. Sometimes hate speech acts fail or misfire because other people block the speaker’s performance of them (Langton 2018b). However, these other people also sometimes face barriers to blocking hate speech acts. I believe the informal social authority of the speaker and the potential lesser authority of the would-be blocker can help to explain the presence of some of these barriers and the subsequent failures to block. Now I do not mean to deny that members of oppressed groups on average face greater barriers to blocking hate speech acts than members of oppressor groups (Brown 2015, 258–62; Delgado and Stefancic 1997, 120, 1996, 480–1; Embrick and Henricks 2013, 198; Lawrence 1990, 455–6; Matsuda 1989, 2357; McGowan 2009, 402–4). But it would be naive to assume that members of oppressor groups cannot face barriers of their own within certain situations of reverse hate speech, even if these barriers tend to be more situational and temporary in nature.
Consider once again the family barbeque example. In theory, the White boyfriend could try to block the uncle’s act of marginalising him by deflecting, challenging his authority, speaking from a position of security about his own ethnicity, or by presupposing that he is welcome at the barbeque, for example. (5) Bottom line is I love your niece man, so we family, I’m your White boy. (6) Listen man, nobody else here feel the way you do. (7) You don’t get to decide who’s welcome here. (8) Hey, don’t do me like that, not all us White boys are the same you know. (9) You can say that to me, but at the end of the day your niece wants me here and she don’t care that I’m white. (10) The fact is I am welcome here, I’m gonna grab myself another beer.
But in practice the boyfriend might refrain from responding in these sorts of ways for a variety of reasons that pertain to the informal social authority of the uncle and the impact this has on the boyfriend given the situation in which he finds himself. First, the boyfriend may be conscious of the fact that by convention home ownership gives people the prerogative to make decisions about what happens in their own homes, including who is or is not welcome (‘My house, my rules’). Second, the boyfriend might recognise the uncle is held in high esteem within the family whereas he is an ‘outsider’. Third, the boyfriend could be experiencing the family barbeque as a hostile or intimidating setting and may fear an escalation into a violent confrontation of one against many. Fourth, maybe the boyfriend wants to make a good impression, so would not want to risk making a scene or embarrassing his girlfriend by backtalking a senior member of the family. Fifth, perhaps the boyfriend is aware of the reputation of the uncle as a ‘good person’ and wants to impress his girlfriend by demonstrating his own ability to turn the other cheek. Finally, he might understand that the family sees the uncle as being knowledgeable on issues of race relations and Black history and that the uncle has very publicly predicted the relationship will fail, and so the boyfriend may think it pointless to debate the uncle and, moreover, that he would rather not risk inadvertently fulfilling the uncle's prophecy.
Of course, some barriers people face to blocking hate speech acts are generic and do not have anything specifically to do with background social hierarchies or power dynamics (Simpson 2013). But it deserves mention that some of the reasons I have proffered as to why the White boyfriend might be on the back foot during the family barbeque are reverse images of familiar reasons why Black people may stay silent and not counter-speak or otherwise attempt to block hate speech acts targeting them. An important difference – itself a form of racial inequality – is that Black people will regularly face barriers to blocking hate speech acts, whereas for White people these barriers are much less common and more likely to be highly situational and temporary.
VI Conclusion
I end the paper by making several general observations, replying to potential objections, and drawing out some practical implications. To begin with the general observations, the first is that informal social authority is both contingent and contextual. It is contingent in the sense that the uncle only has it because of his position within the social group. The little brother lacks it, which is another reason why when he repeats the uncle’s phrase ‘Hey whitey, we don’t want your kind here’, it seems less likely that this would constitute the successful performance of the illocutionary act of marginalising the boyfriend. Moreover, informal social authority is contextual in the sense that the uncle is unlikely to enjoy the same authority in other social contexts and with regards to other social groups. For example, if the uncle were to attend a social event at a nearby White majority golf club and utter the same phrase-type ‘Hey whitey, …’, he would almost certainly lack informal social authority and more likely than not his utterance would misfire in the sense that he would fail to perform the intended illocutionary act of marginalising another person on grounds of race. Indeed, it is likely the White people at the golf club could easily block the intended act of marginalisation (‘How dare you say that!’; ‘This guy is such a joker’; ‘We shouldn’t judge him too harshly for his inappropriate words, he’s probably suffering from Black rage’). This observation underscores the precariousness of reverse hate speech acts, how dependent they are on circumstances. But just because they are more vulnerable to failure, does not mean they do not count as hate speech acts when they do succeed. More importantly, it is still worth making the argument that, by and large, successful reverse hate speech acts are more likely to be successful in virtue of the speaker’s informal social authority than either norm-governed social practices or licensed authority (via bystander silence).
My second observation picks up on an important further dimension of (reverse) hate speech not discussed in this paper: perlocutionary effects (Brown and Sinclair 2023, 61–63, 130–65). When assessing whether the family barbeque vignette is a bona fide case of hate speech, and evaluating how serious the case is, it is surely relevant to ask whether the uncle’s use of the offending phrase and perhaps also the little brother’s repetition of it could have a range of perlocutionary effects – for example, making other members of the family see and treat the boyfriend as socially unwelcome and as a result making the boyfriend feel that he is socially unwelcome. Arguably, even the little brother’s (unauthoritative) repetition may contribute to a climate of marginalisation that not only makes the boyfriend feel unwelcome but could increase the likelihood of other negative behaviour, including the boyfriend being asked to leave the family barbeque or deciding to remove himself or remain silent in the face of the abuse. 17 A further undesirable effect of the uncle’s words could be the undue influence it has on the little brother, who is himself a vulnerable party because impressionable and not yet intellectually equipped to critically assess what his uncle is saying. 18
My third observation starts with recognition that I have only discussed three ways in which ordinary people might obtain the capacity or authority to perform hate speech acts. Other ways might be relevant in some situations. 19 For example, some speakers can successfully perform hate speech acts against targeted groups by themselves occupying positions of formal authority (institutional, political, legal) (see McGowan 2012, 128–32; Brown 2015, 85; Brown 2018, 8; Brown 2019a, 166–9; Brown and Sinclair 2020, ch. 7; Brown and Sinclair 2023, 60). Likewise, in some situations of reverse hate speech, speakers might rely on formal rather than informal authority (Brown and Sinclair 2023, 162–3). 20 Reflecting on these possibilities, I endorse a pluralist answer to the authority problem: different speakers, different messages and styles of hate speech, different kinds of wrongful or harmful hate speech acts, different sorts of targets of hate speech, different wider social contexts, and different speech situations or discursive microenvironments require different answers to the authority problem (ibid., 59–60). In particular, I believe that each of the three answers to the (or an) authority problem discussed in this article (norm-governed social practices, licensed authority, and informal social authority) could potentially be sufficient, given exactly the right set of circumstances, to explain how speakers could successfully perform wrongful or harmful hate speech acts, albeit these circumstances are less likely to occur in the case of the first two answers, but also that none of these three are necessary (given the potential applicability of other possible explanations).
My fourth observation is that whilst the successful performance of some reverse hate speech acts involves both authoritatives and the situational reversal of standard power dynamics which exist within the wider society, this does not mean that all instances of reverse hate speech share these two properties. There are reverse hate speech acts to which the authority problem does not apply simply because they are non-authoritatives, meaning speakers do not require authority to successfully perform them. Likewise, there can be reverse hate speech acts that do not necessarily involve any situational reversal of power dynamics. Consider situations in which members of oppressed groups perform illocutionary acts of prejudging, stereotyping, misrecognising, or essentialising members of oppressor groups (Brown and Sinclair 2023, 130–65). Sometimes these illocutionary acts can be expressives, assertives, or behabitives, meaning they have to do with expressing, asserting, or adopting certain beliefs, attitudes, appraisals, or behaviours towards others. Such categories of illocutionary acts are non-authoritatives since they can be performed even by speakers lacking authority. What is more, they do not necessarily depend on any reversal of standard power dynamics.
Having made these general observations, I shall now reply to three potential objections. The first objection is that, even if correct, my analysis is politically dangerous. We live in an unjust world where White men are members of dominant, powerful groups. If it turns out that concepts of hate speech, including the idea of recognisable hate speech acts such as marginalising or inciting, can be applicable even to speech targeting White men, this could unwittingly enable White men to rely on forms of whataboutism so as to deflect from their own hate speaking. Handing them a conceptual tool to paint themselves as the ‘victims’ gives yet further power to the powerful and potentially undermines decades of insights and understandings into the nature of racism and misogyny, for example.
However, my response is to point out that there is no analysis available here that is danger free. Consider the downsides of the alternative path of simply trying to ignore reverse hate speech: ‘a blanket and peremptory refusal to consider as hate speech even words that may in fact bear several semantic and pragmatic similarities to prototypical hate speech can smack of double standards’ (Brown and Sinclair 2023, 159). A peremptory refusal to accept the possibility of reverse hate speech might lend succour to (some) White men’s retaliatory rejection of the very concept of hate speech as ‘liberal tyranny’ or ‘woke fascism’ – a dialectical consequence that is unlikely to make (some) White men reflect more deeply about their own language towards oppressed groups and that is unlikely to help protect victims of traditional hate speech in the long run. If Internet companies and governmental institutions were to say there is such a thing as hate speech but also tell certain people that it is impossible for them to be victims of hate speech, then the natural response on the part of (some) of those people will be to disbelieve that anyone really is a victim of hate speech and to denounce these companies and institutions as biased and untrustworthy. At least if society acknowledges the possibility of reverse hate speech, then it opens up the possibility of a conversation involving concepts people share, even if their experiences and concerns are also different in important ways. Acknowledging that reverse hate speech can exist might also help to restore or grow public trust in companies and institutions tasked with the work of combating hate speech (ibid., 25–26). Arguably, hate speech is ultimately defeated only when White men also accept and internalise social norms and moral principles against hate speech and buy into what companies and institutions are doing with their various content policies and laws on hate speech. For those companies and institutions to flatly reject the very idea of reverse hate speech looks counter-productive. Rather than accept and internalise these norms and principles, White men would travel further down the opposite road of resisting companies and institutions that ‘police’ hate speech.
The second objection picks up on my nuanced argument that some, although by no means all, examples of reverse hate speech can involve situational reversals of standard power dynamics. The objection is that, contrary to my nuanced argument, all cases of reverse hate speech must necessarily involve reversals of standard power dynamics: if there is no reversal of standard power dynamics, then something cannot be a genuine example of reverse hate speech after all. This objection trades on Matsuda’s original stipulation that hate speech involves attacks against historically oppressed groups but redraws this insight as being not so much about historically oppressed groups as about power dynamics, including situations in which standard power dynamics have been reversed.
My response to this objection, however, is that it is quite normal for competent language users to apply the term ‘hate speech’ even to speakers who are simply expressing, asserting, or adopting certain beliefs, attitudes, appraisals, or behaviours towards other people based on the possession or perceived possession of protected characteristics, irrespective of what power dynamics might be in play and in what directions. It is conceptually baseless to assume that using the term ‘hate speech’ in this way only makes sense if there is a power dynamic in play between the speakers and the groups of people they are talking about. Reverse hate speech acts may still count as hate speech acts even when the acts being performed amount to prejudging, stereotyping, or misrecognising, and even when there is no reversal of standard power dynamics in the situation. Furthermore, this objection conflates power dynamics in general with the more specific notion of historical oppression and in a way that risks trivialising the latter. Even if obtaining informal social authority is a form of power, it is not necessarily the same thing as being an oppressor. What is more, pace Matsuda, there is no reason to assume that hate speech is by nature unidirectional: hate speech is multidirectional, and this applies to the ordinary as well as legal concept of hate speech.
A third potential objection is that the category of reverse hate speech acts is really a placeholder for the notion of righteous hate speech (Brown and Sinclair 2023, 166–194) insofar as righteous hate speech always takes the form of reverse hate speech acts and reverse hate speech acts are always righteous hate speech. Of course, Matsuda (1989, 2361–3) might argue that righteous speech simply cannot be hate speech in her sense of the term. But let us set that definition aside for the sake of the objection. The objection is that if reverse hate speech is simply a placeholder as described, then arguably the true source of the authority of speakers to perform reverse hate speech acts lies not in informal social authority but in the righteousness of what they are doing (e.g. speaking truth to power, retaliating, voicing grievances, fighting injustice).
However, I reject the main assumption underpinning this objection. Consider cases of righteous hate speech in which although the speaker sincerely believes the target of speech is up to no good or poses a threat, the speech is not performing a reverse hate speech act (precisely because the target is not an oppressor group). Examples might include the use of slurs like ‘gold digger’ and ‘ex-con’ which seem to pick out people who are members of historically oppressed and discriminated against groups. Conversely, a woman might post on social media ‘All men are scum’ and manage to successfully perform certain reverse hate speech acts against men such as stereotyping, prejudging, or inciting hatred, but in a way that lacks righteousness such as if the attacks are not reasonable, not in good faith, and not proportionate (Brown and Sinclair 2023, 174–5).
Furthermore, even if some examples do involve both reverse hate speech acts and righteous hate speech, this does not automatically mean the authority of the speaker lies in the righteousness of the attack. Imagine that the woman who posts on social media ‘All men are scum’ is a well-respected member of a group of radical feminists who met at college. Her high status among the group derives from the fact that she achieved the highest grades in her college classes on feminist theory and from writing for the college newspaper on feminist politics. She also obtains credibility from the fact that her views chime with the prejudices of her peers. Arguably, her authority to perform hate speech acts lies not in the righteousness of the acts per se but in her epistemic authority and informal social authority. In other words, the objection fails because a person’s authority is not necessarily determined by, or reducible to, her righteousness. The righteousness of someone’s communicative attack is a matter of having a just cause, being reasonable, speaking in good faith, and proportionately, which are moral as much as social facts (Brown and Sinclair 2023, 174–5).
I turn finally to the practical implications of the arguments made in this paper about reverse hate speech and the (or an) authority problem. The authority problem exists not merely in an ivory tower. Social media platforms remove vast amounts of hate speech and in order to do so must consider what does or does not count as hate speech. This involves not simply checking for types of speech but also tokens, that is, whether particular posts are instantiations of types of speech they designate as disallowed hate speech. In carrying out these checks it seems reasonable to assume that staff (and the automated detection systems they utilise) must pay heed to whether they are removing content because of what it says (locution), the acts it performs (illocution), and/or the effects it has (perlocution). Arguably, it is not merely a superficial difference that Meta has a community standard on ‘hate speech’, 21 whereas X has a policy on ‘hateful conduct’. These labels provide clues to substantive differences in the definitions they provide: Meta’s community standard underscores dehumanising speech and imagery, generalisations that state inferiority, and harmful stereotypes, whereas X’s policy highlights threats, harassment, and inciting harm. 22 The crucial point here is that if social media platforms are at least sometimes classifying and removing hate speech because of the illocutionary acts being performed, then looking into how the acts are performed, including with what authority, must inevitably be part of their practice. It is no surprise that both Meta and X’s public-facing policies make repeated mention of ‘context’.
The arguments of this paper also have implications for the legal concept of hate speech. For example, McGowan (2012, 143) presents her answer to the authority problem as not merely an abstract conceptualisation but also as something that could inform legal definitions of hate speech. She even suggests that if legislatures and other lawmaking bodies adopted legal definitions of hate speech that focused on the performance of illocutionary acts such as enacting or inciting a discriminatory policy against a racial group (e.g. ‘Your kind are not allowed in here’), then courts could look into the circumstances of both the speech situation and wider social environment to work out if the speaker possessed the requisite capacity, power, or authority to perform the relevant acts including based on norm-governed social practices of racism. In a similar vein, I believe courts could also consider other sources of illocutionary authority including informal social authority. This is not simply a matter of deciding whether the high status, power, or influence of the speaker was an aggravating factor in his or her offending behaviour (i.e. determining the seriousness of the offending behaviour) (Brown 2015; Brown 2025; Brown and Sinclair 2020). It would involve defining or redefining elements of the relevant criminal or civil offences in terms of certain illocutionary acts and then relying on assessments of the defendant’s informal social authority (and other relevant forms of authority including formal authority, for example) to determine if he or she had the authority needed to perform those acts in the first place (i.e. establishing guilt or innocence).
Now this type of speech pragmatics assessment might not be suitable for all hate speech laws, but it could make sense for those whose core elements involve illocutionary acts. For example, some civil torts contain elements that could be defined or redefined in terms of certain illocutionary acts, such as degradation or humiliation, for instance (see Brown 2018). Speech pragmatics assessments of speaker authority could also be relevant to criminal offences such as incitement to hatred, where the nub of the offence is an illocutionary act (e.g. stirring up hatred). Such assessment might also be relevant where the offence of incitement to hatred is defined or redefined in terms of other elements which are themselves also illocutionary acts. Consider a potential definition of the criminal offence of incitement to hatred that casts the offence in terms of a person using speech acts of threatening, subordinating, degrading, humiliating, defaming, or negatively stereotyping another person or group of people with intent thereby to stir up hatred against them on grounds of protected characteristics. When seeking to identify whether speakers had the requisite capacity, power, or authority to perform the relevant acts, courts could look to a range of things including informal social authority. To clarify, I am not seeking to defend such a definition of the criminal offence of incitement to hatred. There is a real risk that such a definition would be overbroad (see Brown 2025). Rather, I am merely suggesting how courts could be involved in assessing speaker authority, hypothetically speaking. At present, the England and Wales stirring up hatred offences, as set out in the Public Order Act 1986, are couched in terms of either ‘threatening, abusive or insulting words or behaviour’ (Part 3) or ‘threatening words or behaviour’ (Part 3A). But presently prosecutors do not need to demonstrate how and why the speaker had the requisite capacity, power, or authority to perform the illocutionary acts of threatening, abusing or insulting. They need only prove that a reasonable person would deem the words or behaviour to be threatening, abusive or insulting (Brown 2025). This reflects a general legal approach to interpreting statutory language whereby undefined terms are given a plain or ordinary meaning. Then again, arguably a reasonable person would only regard certain words or behaviour as threatening, abusive or insulting if the circumstances befitted that description. Perhaps the authority of the speaker, informal as well formal, either does or should factor into judgements about what a reasonable person would say about given words or behaviour being threatening, abusive or insulting in the circumstances.
