1 Introduction

The prevalence of harmful online speech is one of the internet’s most prominent issues. One only has to sign into Twitter or Facebook, or look at the comments under videos on YouTube or posts on Reddit and one is guaranteed to quickly come across angry, hateful messages and comments against other individuals or groups. It is no surprise that citizens frequently report to come across hate speech (Siegel 2020). But not all speech is as explicit and visible as outright forms hate speech. In this paper, I focus on those online speech acts that may seem comparatively insignificant by themselves: a tasteless joke, perhaps, or a backhanded insult, in particular those directed at (members of) groups that are targeted by such speech or other prejudice more frequently. The concern with such speech acts, which, together, I will call jerkish speech, is that they may be overlooked or may escape established regulation, precisely because such speech acts are often (perceived as) relatively inconsequential. The political program of republicanism, revolving around the concept of domination, is well-equipped to deal with this risk, because domination—and the absence of it, non-domination—are explicitly dependent on hierarchies of power and subordination, both intuitively related to jerkish speech.

The question I am concerned with in this paper, then, is how one can use a republican framework to assess why a given online jerkish speech act, even if by itself of little direct consequence, is problematic, and if so, how we should respond to it based on that same republican framework. The paper aims to answer this question in two parts. In the first part, I explain how the republican notion of domination helps specify why jerkish speech is a problem: jerkish speech can be a form of micro-domination or an instantiation of systemic domination. In the second part of the paper, I explore the ways in which republicanism might respond to jerkish speech, in particular in online environments. I briefly discuss the role of and the issues surrounding speech regulation. I then suggest that Suzanne Whitten’s (2022) proposal for the cultivation of critical civility requires leveraging the central position of online platform owners to prevent it from being too demanding in online environments.

2 Online Jerkish Speech and Domination

2.1 Jerkish Speech in Online Contexts

The term ‘jerkish speech’ is inspired by Eric Schwitszgebel’s ‘A Theory of Jerks’, a blog posted originally on Aeon.Footnote 1 Schwitzgebel is concerned with what makes someone a jerk, suggesting that a large part of it has to do with disregarding the perspectives, or peerness, of others, especially those who the jerk considers to be below them in the social hierarchy. In this spirit, when defining jerkish speech, I refer to the kind of—often thoughtless—speech that signals a disregard of the status or perspectives of others, in particular those whom the speaker perceives to occupy a lower social position (on whatever basis that might be). Jerkish speech is, for example, uttered by an impatient person at those queued up before him, or by a bully in school directed at any victim, or by someone offensively snapping at their fellow passengers on a busy train—all of which remind of someone acting ‘like a jerk’. In this sense, jerkish speech is often clearly recognizable, but that is not always the case. Jerkish speech also encompasses subtle and indirect forms—think of someone making a subtle joke at the (limited) expense of someone else, or of a person disregarding the serious issues of another in order to complain about a relatively insignificant personal inconvenience.

While jerkish speech is broader than hate speech, as these examples show, jerkish speech surely encompasses much hate—in particular those forms that are diffuse, subtle, or offhanded and therefore not straightforwardly designated as hate speech.Footnote 2 Examples are a comment posted in response to a video of a woman struggling to open a bottle of wine: ‘Good luck with workplace equality!’, or what have been called ‘dog whistles’—where a given message has two possible interpretations. One of these interpretations gives little cause for concern, but the second (hidden) interpretation is aimed at particular subsets of the audience and contains a political message, often harmful (Saul 2018). Think of the phrase ‘all lives matter’, often offered in response to the ‘Black lives matter’-movement. The latter aims to draw attention to the racism faced by Black communities all over the world. The phrase ‘all lives matter’ purports to simply convey that all people are equal. Nevertheless, when offered in response to the #BLM-movement, it also questions concerns raised by members of vulnerable communities, pushing back against attempts to secure political recognition for such issues.Footnote 3

The concern with jerkish speech is that a given, singular jerkish speech act, because of its subtlety, does not need to have the immediacy and clear negative consequences that other forms of hate speech have. Jerkish speech acts do not call for violence, do not threaten and do not always clearly cause insult or other direct harms to the target. The immediate effects of jerkish speech acts by themselves are minor and perhaps even non-existent. Jerkish speech, therefore, also offers plausible deniability (Dénigot and Burnett 2020). No one was spreading hate—they were merely voicing pessimism, or personal preferences, or innocent jokes.

The fact that jerkish speech does not always cause clear and visible harms or immediate effects means that it is at risk of escaping legal and conceptual definitions of hate speech. While there is a large variety in legal systems and rationales, most legal systems that contain hate speech legislation also subscribe to an important legal principle: the ‘pressing social need’ or proportionality-principle (Brown 2015, p. 239; Webber 2021). This principle requires that any restriction of a fundamental right or freedom (in this case of speech) is justified by a sufficiently important societal interest or need, and that this restriction is proportional in relation to that need. While explicit hate speech is commonly considered to satisfy this principle to some extent, it is not self-evident that its more subtle, implicit variant is considered similarly pressing. This observation is not limited to legal discussions but extends to approaches to hate speech that are directed mostly at the harms caused by hate speech. On such approaches, whether restrictions are appropriate comes to depend on a cost–benefit analysis (McGowan 2009). The effects of jerkish speech acts are not easily expressed in costs and are therefore at risk of being overlooked.

While online jerkish speech may, at first glance, not seem fundamentally different from its offline counterpart, it is nevertheless important to assess some particular characteristics of the context and appearances of online hate speech (and by extension jerkish speech), as this will prove relevant for formulating responses to it. First, in terms of prevalence, surveys seem to show that while online hate speech is relatively rare, individuals are frequently exposed to it (Siegel 2020). This might have to do the visibility of hate speech online (Kaakinen et al. 2018) and with the fact that hate speech is more common in some online demographics. The frequent exposure that is consistently reported adds to the need to respond to problematic online speech in general. Recently, scholars working on speech detection systems have started to focus on more implicit and subtle forms of hate speech that may fall under jerkish speech (ElSherief et al. 2021). While existing methods of speech detection underperform when employed to find implicit and subtle hate speech, improvements are making it easier to detect more implicit forms (Kim et al. 2022). In alt-right communities on Reddit, 4chan and 8chan, for example, roughly half of all hate speech seems to be formulated in implicit terms rather than in explicit terms (Rieger et al. 2021), further pressing the need to engage with online manifestation of jerkish speech.

Second, hate speech—and jerkish speech—manifests itself differently in online contexts. The discussion on these differences has been focused on four aspects: the anonymity of users, the distance between speaker and listener, the ease of finding like-minded people, and the instantaneousness of speech (Brown 2018). The anonymity provided by digital media allows citizens to act and speak without fear of what others might say or respond in return, removing barriers to speech that we might experience in traditional situations. The physical distance between citizens who engage with one another online makes the effects of hate speech invisible to the speaker, inhibiting the social cues that cause empathy or politeness. The connectivity allowed for by digital media enables like-minded people to find one another and organise—including those that share a particular hatred. And, compared to traditional media, digital media allow for cheap speech that nevertheless reaches large audiences in a short time span. Whereas traditional, written works, such as an opinion piece submitted to a newspaper, require some effort (and leave time for reflection), digital media enable citizens to spread messages with little thought given. While these findings are done in the context of hate speech, they imply that similar dynamics are at play in the case of jerkish speech.

In sum, online jerkish speech is a kind of cheap speech, shared in online environments, which signals a disregard for the status of others, and includes online hate(-like) speech that is subtle, implicit and relatively inconsequential. Because of that, jerkish speech is at risk of being overlooked or of falling outside of cost–benefit approaches to speech regulation. Meanwhile, though it is challenging to detect the prevalence of jerkish speech, there are signs jerkish speech could make up a significant part of problematic speech in online contexts. Finally, the dynamics of hate speech—and by extension jerkish speech—are different in online environments, which may prove relevant for combating it. But why is jerkish speech a problem at all? A tasteless but subtle joke may be offensive, but is that all there is to it? Do we need government interference, and if so, how can we justify that? In the next section, I suggest that republicanism offers a promising normative framework with which we can approach online jerkish speech.

2.2 Republicanism, Non-domination and Harmful Speech in General

Republicans are concerned with the degree of domination that citizens are exposed to. In the paradigm case of agent-relative domination, an agent (the dominator) has an uncontrolled or arbitrary power of interference over another (the dominatee) (Pettit, 1997, 2013; Laborde, 2010). Such power is controlled when it is bound by terms set by those who are subjected to it. Power also need not be exercised for domination to occur, merely having it constitutes domination. It is only uncontrolled power that can constitute domination, however. One cannot be dominated when power is not arbitrary or uncontrolled. This allows republicans to explain why, for example, poorly-protected workers or women in sexist marriages are dominated, even if they are left to fend for themselves—their every choice is made by the leave of their dominator. Moreover, it helps republicans critically assess institutions like employment, slavery, and marriage for the roles and powers they attribute to some and take from others. According to republicans, it is up to the state to minimise domination in society by increasing robust control and by reducing arbitrary power. This is a careful balancing act, as—although democratic government power does not count as arbitrary—a more powerful and active government does increase the risk of itself becoming less responsive to control.

It is not my aim to argue that a republican perspective is the only viable one with which we might approach jerkish speech, and some of te arguments contained in the paper could also work for other perspectives. There are, however, reasons for considering republicanism to be well-equipped to deal with online jerkish speech, in particular compared with the non-interference paradigm that seems widespread. First, whereas a liberal conception of freedom as non-interference is usually quite narrow in its focus on individuals and their choices or harms that befall them, the notion of non-domination is tied explicitly to broader questions of power and hierarchy, which seem prima facie relevant for social standing and jerkish speech. In fact, ‘radical’ or ‘critical’ republicans have developed accounts of domination that are sensitive to social arrangements and status, amounting to a ‘social turn’ in republicanism (Muldoon 2022, p. 48). These accounts offer promising conceptual ways of interpreting the issue of jerkish speech in online contexts, which align with our intuitions on social freedom and its connection to social status—in particular compared to the narrow non-interference approach that emphasises the importance of freedom of speech (see also Whitten 2022).

Second, republicanism has been developed into a comprehensive political theory over the last few decades (see, for example, Honohan 2002; Laborde 2008; Lovett 2010; O’Shea 2020; Pettit 1997, 2013; Skinner 2001). This allows us not just to assess how jerkish speech must be placed in a broader social and political landscape, but also how it is related to social freedom. In turn, this helps formulating and weighing which institutional responses to jerkish speech are appropriate and consistent. Republicans, for example, are not principally opposed to extensive government involvement in matters between citizens, provided such involvement does not generate more domination than a more restrained approach would. Whether that is the case depends mostly on the degree to which the state’s involvement is democratically controlled, but it means that, in principle, republicanism allows for and even justifies more extensive government involvement than one might be able to do on the basis of non-interference. This brings nuance into the freedom of speech-debate, in which the issue of problematic speech is posited.

2.3 Social Freedom and the Secure Enjoyment of Non-domination

There are (at least) two paths that bring jerkish speech into the normative republican framework. The first path stays close to the agent-relative conception of domination and perceives jerkish speech acts as potentially compromising the secure enjoyment of non-domination—in particular when considered in aggregate. The second path latches onto a version of republicanism that is concerned with systemic rather than agent-relative domination. I will discuss both.

For the first account, it is useful to consider first how ‘proper’ hate speech could interact with social freedom. On an account proposed by Matteo Bonotti (2017), hate speech undermines the robust enjoyment of non-domination.Footnote 4 While hate speech might not interfere directly with the formal rights that all citizens share, it serves as a subtle threat that these protections are not self-evidently secure. Even if one never experiences any actual interference, the threats contained in hate speech prevent the victim from enjoying non-domination robustly. These threats might signal, for example, a lack of commitment of (part of) society to recognise a formally enshrined status, or even the intention to use political means to revoke them. This is why Philip Pettit, central to the republican revival, insists that it ‘must be a matter of shared awareness in the society that [people are guarded regardless of the will of others]’ (Pettit 2013, p. 83). Hate speech undermines that principle.

There is at least one argument, also recognised by Bonotti, that can be raised against this view: it is the state—not other citizens—that must secure non-domination through formal laws. If the state is able to ensure that legal protections against arbitrary interference are maintained even in the face of hate speech, then the threat to the secure enjoyment of such rights would fall flat. It similarly makes little sense to suggest that citizens are dominated because they are aware of the fact that there are some burglars who are not committed to upholding their property rights, while property rights are actually quite effectively protected. Hate speech that constitutes direct coercion—such as threats or harassment might do—should be regulated as a form of arbitrary power to interfere, but other forms of hate speech are not directly related to domination if they cannot violate formal or legal protections. The republican account might seem close to the liberal version in this sense: there is little to worry about if it’s just words flying around.

Bonotti suggests that this criticism misrecognises the role that social norms play in an ideal republican society. Non-domination is not merely upheld by formal or legal protections but depends on ‘widespread civility’ (Pettit 1997, p. 245). For Pettit, this is largely of instrumental importance: civil citizens are disposed to obey the laws that govern the republic, to voice their interests as they change or develop, and to be vigilant against domination by the state or by fellow citizens, serving as an extension of the mechanisms that ensure non-domination. In the absence of civil social norms, citizens are not assured of their fellow citizen’s compliance with their formal protection, nor of the effective enforcement of such protections when not under the watchful eye of authorities. As such, hate speech constitutes a problem for non-domination—even when it leaves intact formal protection—insofar it undermines these civil social norms. In other words: republicans are concerned not just with formal, legal protection, but also with the way in which these protections are supported by social means. The secure enjoyment of non-domination requires both: formal protections against arbitrary interference, and social arrangements that support these protections.

2.4 Subtle Hate Speech: Jerkish Speech as Micro-Domination

This brings us to a challenge against the idea that jerkish speech can be an issue at all: it is not evident that jerkish speech similarly undermines either formal protection or its informal support. If it cannot, then jerkish speech might be too subtle to cause domination. Outright hate speech directly compromises non-domination (e.g. threats, harassment), and we can also grant that it can undermine robust enjoyment of non-domination by violating the civil social norms that are needed to support non-domination. But jerkish speech can be quite subtle, hidden, and indirect, and usually lacks the direct effects ascribed to hate speech.

Republicans can respond that jerkish speech might contribute to domination as a form of micro-domination, a concept coined to capture instances of domination that are not particularly serious by themselves, but which, in aggregate, might still amount to significant domination. Tom O’Shea (2018) suggests that this is the lived experience of many people with disabilities, who are more or less dependent on family and caretakers in their daily lives. Many of these daily choices are minor or inconsequential—when and what to eat or whether one can smoke a cigarette, for example. Quite often they are not significant enough to be ‘contested in a court or tribunal’ (p. 136). Yet taken together, these many small choices have people with disabilities subjected to significant arbitrary power over the course of their daily lives.

Micro-domination, more generally, refers to domination that ranges over a number of choices which by themselves do not cross a certain (objective or subjective) threshold, but which are nevertheless significant in aggregate (Lazar 2021). The threshold serves as a way of prioritising particularly serious forms of domination or as a way of excluding ‘trivial’ ones. Two issues that arise a result of micro-domination not crossing that threshold are, first, that micro-domination often escapes our notice—we tend to be more concerned with extremer forms of domination—and second, that it undermines the republican reliance on contestation. Because of its comparative insignificance, micro-domination may fail to provoke a response by the victim or by bystanders at any singular point. An employer, for example, may use some uncontrolled power to influence fairly minor decisions on the workplace: he may signal that he disapproves of the way an employee dresses, or that an employee should do this or that to get into a prestigious project. Choices like this, by themselves, may not lead to action by the employee. Mechanisms of control that rely on contestation by the victim (filing complaints, lawsuits, etc.) might be perceived as being unwarranted and unnecessarily escalating.

A similar thing applies to jerkish speech, in particular in online contexts. Any singular remark might not significantly compromise the target’s non-dominated status. Consider again the example ‘good luck with gaining equal rights in the workplace’. By itself, such a comment is not likely to undermine the formal rights of non-domination of the target or of women in professional environments, nor the background social norms that would be needed for a continuous guarantee of formal protections. Similarly, in terms of contestation, such a comment might raise eyebrows and perhaps cause some pushback (perhaps it’ll be flagged as inappropriate), but it is unlikely to trigger hate speech laws or to provoke the target to put in the effort to file a complaint with relevant authorities.

Comments like these, however, are often part of a wider phenomenon, and it is in aggregate that the effects of jerkish speech must be evaluated. Jerkish speech, then, can lead to significant domination of a target if it is frequent and pervasive enough to compromise the secure enjoyment of non-domination. This might be the case if it can be shown that jerkish speech—in aggregate—undermines widespread support of the civil norms that support protections against domination.

2.5 Social Freedom: Hate Speech and Systemic Domination

A second way in which republicans can get around the challenge that jerkish speech is simply too subtle to result in domination, is by embracing critical republicanism. Recently, some authors have developed versions of republicanism that are concerned with domination as a structural phenomenon (see for example Gädeke 2020a, b; Hasan 2021; Laborde 2010; Sandven 2020). These accounts insist that domination is generally not (strictly) agent-relative but embedded first and foremost in the formal and informal social institutions of a society. On this approach, a slave is not dominated as a result of a given master having arbitrary powers of interference, but because of the underlying institution of slavery which attributes both their relative positions. In fact, it may not matter too much who the master is at any point—or whether there is one at all—as the unfreedom is already constituted by the very system of norms and practices that subordinates some and empowers others.

Based on this, Dorothea Gädeke (2020a, b) proposes to distinguish systemic domination from interpersonal domination, both of which are structurally constituted, and both of which work together. Interpersonal domination is what we know as the classic, agent-relative form of domination, where one agent wields arbitrary power over another. Systemic domination refers to how some citizens face disempowering norms and practices, which are reproduced by peripheral agents (Wartenberg 1990) or regulators (Vrousalis 2020). These disempowering norms compromise the normative authority of the target. Normative authority requires being taken seriously as someone who, apart from making statements about facts and beliefs, also makes claims about ‘normative rightness and thus to the decidedly intersubjective world of justification’ (Gädeke 2020b, p. 26). It is through systemic domination that interpersonal domination occurs, and it is in the absence of interpersonal domination that systemic domination persists. It is not necessary to have a master at any point, for as long as disempowering and subordinating norms and practices are reproduced, those targeted remain systemically dominated. Similarly, the reproduction of such norms and practices sometimes takes the form of interpersonal domination—they are two different sides of the same coin.

In the fairly recent book A Republican Theory of Free Speech: Critical Civility, Suzanne Whitten (2022) brings this critical republican approach to bear on free speech. Whitten criticises Pettit for overlooking how (external and internalised) social or civil norms themselves often contain unjust status hierarchies. The notions of assurance and that of robust enjoyment of non-domination ‘do not offer a sufficient response to those norms of racism and sexism that contribute to an unjust status hierarchy’ (p. 104). Instead, Whitten suggests, republicans must equip a different ‘normative yardstick’ to measure social liberty and equality:

The Critical-Republican-Social-Egalitarian (CRSE) account of social justice holds that, in order for individuals to enjoy an undominated, free and equal status, they must be securely and explicitly recognized as a function of standing in a relation of equality with their fellow citizens. This recognition of standing requires that individuals are recognized as discursive agents who hold normative authority as equal participants in the space of moral reasons (p. 119).

For Whitten, then, social freedom is largely a matter of a struggle for recognition and respect. Based on Darwall (1997), she distinguishes between recognition respect on the one hand, and appraisal respect on the other. The first refers to a kind of respect that ought to be extended to all persons: a recognition of them being worthy of appropriate consideration when deliberating, as persons. According to Whitten, it also requires us to take seriously another’s normative claims on social norms, as such claims are attempts to exercise normative authority. A lack of recognition respect compromises another’s normative authority, resulting in systemic domination. Appraisal respect is about evaluation of another’s character, accomplishments or acts, and at first glance seems more about esteem rather than about recognition. Appraisal also depends, however, on the degree to which we comply with social norms. As such, a lack of appraisal can turn out to be a sign of the existence of unjust social norms.

On a critical republican account the issue with jerkish speech is not merely that it might erode robust formal protections, but rather that jerkish speech is, or contains, the reproduction of social norms that deny the normative authority of its targets. Jerkish speech is a subtle signal, intentional or unintentional, that not every citizen is recognised as a full normative agent, and a signal that they are excluded from social norm-shaping processes. Moreover, jerkish speech may be a signal of disesteem (or lack of appraisal respect) for not complying with (unjust) social norms. One can think here of the category of speech that circumscribes women who are in professional leadership positions as not being ‘woman-like’. While not necessarily amounting to hate speech, such speech clearly signals disesteem for not complying with pre-existing sexist social structures.

In sum, republicans have at least two options for accounting for jerkish speech in online environments. The first is that such speech, in aggregate, risks compromising the secure enjoyment of non-domination as a formal status, or risks undermining the informal civil norms that are needed to support formal protections. The second reason flows from a critical republican perspective and is particularly relevant for jerkish speech: hate speech and jerkish speech can play a pivotal role in the reproduction of unjust social hierarchies. It might deny the normative authority of members of subordinated groups and can signal a lack of appraisal respect in response to attempts to change or reject unjust social norms that govern the social sphere.

3 Republican Responses to Jerkish Speech in Online Contexts

If we accept that republicans can account for the issue of jerkish speech in online environments (or in any environment in that regard), the question is what kind of responses are justified. I will discuss the appropriateness of two traditional republican responses: legal intervention in the form of speech regulation and promoting civility amongst citizens. I suggest that neither sufficiently leverages the configurability of online environments. A more comprehensive and context-sensitive approach includes an account of the design of online platforms and of the role their owners.

3.1 Regulatory Bans on Jerkish Speech

For a start: it is up to the state to impose an order in which citizens enjoy reliable freedom as non-domination in the relationships between each other, and to ensure that all citizens are aware of having such a status (Pettit 2013). The first and most obvious way republicans might respond to jerkish speech in online environments is by introducing regulatory bans on jerkish speech, similar to how physical violence or theft are prohibited. This draws the issue firmly into the legal-coercive sphere, where clearly codified, formal norms are enforced by the state in order to control or limit arbitrary power of interference of citizens and organisations. Most western governments outside of the United States have implemented some form of hate speech regulation, banning threats, harassment, incitement of hatred, and group libel or insult (Rosenfeld 2003). We might suggest expanding them to cover jerkish speech as well. Insofar jerkish speech does not compromise formal protections against domination directly, prohibitions might still be justified if it turns out to undermine the robust enjoyment of these protections. There are, however, three observations that are relevant when considering the appropriateness of regulatory bans in response to jerkish speech.

The first is that it is not clear whether bans on speech can effectively alleviate the undermining effect of jerkish speech (an argument also made by Robert Mark Simpson (2013) in response to Waldron’s dignity-based account of hate speech). If citizens are compromised in their robust enjoyment of non-domination because of speech voiced by others, then surely this has more to do with the underlying attitudes rather than with speech in and by itself. Banning jerkish speech would merely prevent underlying attitudes from being voiced out in the open but does little to prevent these attitudes from existing in the first place. Members of vulnerable groups would still have to face these attitudes, and if so, it is not a matter of shared awareness that they are securely protected against non-domination, nor can they count on the civility of other citizens that is needed to support such formal protections. Even if bans on jerkish speech would reduce its aggregate effects on victims, then it does so merely by preventing them from being aware of it. One could hardly argue that their enjoyment of non-domination is secure in the sense that republicans prefer.Footnote 5

A second observation is that jerkish speech, as a form of micro-domination, tends to elude regulation. A given singular jerkish speech act fails to cross thresholds that ‘true’ hate speech does not. That makes it difficult to formulate appropriate legal boundaries for speech that should be regulated, and it is improbable that jerkish speech would garner much attention over other issues. Such boundaries would need to incorporate the aggregate effect of jerkish speech in some way, as simply banning every singular potential jerkish speech act is not proportional, increasing the risk of an overly invasive and dominating government. So-called ‘fire-alarm’ strategies (Pettit 1997, p. 250), where authorities rely on victims and bystanders to report and contest incidences of jerkish speech, are also less effective when dealing with micro-domination, as the consequences of individual jerkish speech acts might not be perceived as significant enough to warrant contestation or take any other actions (Lazar 2021). This carries over into mechanisms of reporting user comments: more subtle calls for violence, for example, are reported less than explicit ones (Wilhelm et al. 2020). While jerkish speech as a wider phenomenon is of concern, the singular speech acts might remain under regulatory radars.

A third, well-known observation is that enforcing bans on hate speech is pragmatically difficult in online contexts, in particular due to the international nature of online environments posing jurisdictional issues. National approaches to hate speech diverge much and attempts to reach over borders are prone to jurisdictional and cultural conflict (Banks 2010). Considering the diversity of cultural and social norms found on the internet, and the republican concern with jerkish speech and social norms will only exacerbate these conflicts and cast further doubt on the possibility of international regulation.

These observations might not pose a definite problem for expanding regulatory bans on hate speech. One could assume, as Pettit does, that formal regulation must at least have a positive effect on the civil norms that are needed to support non-domination—provided they are perceived as legitimate (Pettit 1997, p. 253). Efforts to lower thresholds for reporting relatively inconsequential jerkish speech could make it easier for targets and bystanders to report such speech, and intensive cooperation between different governments and private actors could help overcome the difficulties in enforcing hate speech bans in a cross-border environment. Nevertheless, these observations cast some doubt on the feasibility of speech bans.

An issue not addressed yet is that of the role of the state itself. The concern here is that the state could itself play a role in perpetuating the same mechanisms of domination that eventually result in jerkish speech, and that any intervention in free speech might just as well increase rather than decrease domination. It is a serious concern that republicans are well aware of, and it can be raised against republicanism more broadly. In response, a republican could, as Philip Pettit seems to do, argue that the goal is to maximise non-domination (see Pettit 1997, pp. 102–106). Limited state intervention in free speech, even if accompanied by a limited risk of government domination, might be required in order to reduce the existence of larger, structural dynamics of domination that jerkish speech is part of—in particular in more egregious forms. This goes hand in hand with the various ways of mitigating the risk of the state domination, mostly through institutional reform. Principles like a strict separation of government power, a system of checks and balances, and an independent judicial system are meant to reduce the risk of state domination. Where these fail, civil society and critical citizens might still contest the state playing an active or passive role in perpetuating domination.

3.2 (Critical) Civility

Instead, whereas regulatory bans on speech are situated well within the legal-coercive sphere of domination, the effects of jerkish speech on informal norms of civility are perhaps better approached as part of a ‘social sphere of domination’, a term used by Alan Coffee (2015) to distinguish between a legal and a social level of domination.Footnote 6 Unlike the legal-coercive sphere, the social level of domination is comparatively ‘amorphous, intangible and fluid’ (Coffee 2015, p. 57). This also requires a different approach to the threats to freedom found on this different level. For Pettit, this boils down to the need for widespread civility discussed above, which can be nurtured through the ‘intangible hand’ (Pettit 1997, p. 254).Footnote 7 The intangible hand refers to a mechanism of motivating citizens to behave in line with civil norms because they seek the positive esteem gained by following them. Not conforming to norms of civility results in punishment in the form of disesteem. To combat jerkish speech, then, we would need to cultivate and maintain suitable civil norms that punish these forms of speech and reward behaviour that counters jerkish speech.

As discussed above, however, Whitten points out that social norms can be (and often are) unjust as part of subordinating hierarchies. Following that, disesteem is sometimes unjustly handed out to those that claim their share of normative authority over such structures. Any republican response to hate speech, Whitten suggests, cannot depend on mere legal arrangements or an intangible hand, but demands critical civility, a value that requires that the relations between citizens are relationships of ‘mutually acknowledged equal standing’ and which can be used as a yardstick to evaluate institutions and relations between citizens (Whitten 2022, p. 188). ‘Critical’ denotes that it must be understood as a process that assesses not just the relations between citizens and institutions themselves, but also engages with the background arrangements and hierarchies in order to reject and alter those that reflect systemic domination.

In more concrete terms, critical civility first comes with implied duty of citizens to have expressive respect towards all fellow citizens they come in contact with, so that all can enjoy a free and equal status. Citizens can do so by following norms of civility, through which they recognise others as possessing normative authority. This requires effort and vigilance, so that citizens with ethical disagreements can respectfully engage with each other, without apathy or denial of normative authority. In addition, citizens ‘need to be capable of determining when they have a responsibility to contest norms, as well as how they ought to do it’ (p. 196), in order to assess which normative social structures of society fail the normative yardstick. This last point is crucial to the critical approach, as it shows that the background structure of social norms and arrangements themselves are objects of contestation. It requires that institutions and citizens are aware of their role in reproducing such norms and that they are open to changing their attitudes towards them.

With regards to speech, then, critical civility implies that citizens are able to recognise not just hate speech, but also jerkish speech that contributes to (denial) of normative authority and to a reproduction of unjust social structures. Whitten suggests that next to requiring regulation of certain forms of speech, critical civility demands that citizens and institutions themselves step in. They have a (context-sensitive) duty to challenge hate speech, jerkish speech, and unjust social norms, implying a widespread change in perspective on harmful speech in general.

3.3 Critical Civility in Online Contexts

Critical civility promises to be an important part of the republican toolbox, but it is also quite demanding in terms of the motivations, attitudes and capabilities required of citizens, even more so when we turn back to the issue of jerkish speech in online environments. Whitten is mostly concerned with critical civility in general and she does not go into too much detail with respect to what that means for online contexts in particular, although she suggests that the costs of online speech are often low, and she refers to Cass R. Sunstein’s work on polarisation and the fragmentation of online environments (Sunstein 2018). In this section and in the next, I briefly consider some further implications for critical civility in online contexts, as I suspect that the characteristics of online jerkish speech are important for the feasibility and cultivation of critical civility—or at least require a context-sensitive application of its principles.

A first important element that needs to be given consideration in particular is the way in which the design of online environments directly affects the mechanisms of civil speech and the way in which this affects how online communities are shaped and interact with each other. Both are heavily influenced by conscious and unconscious choices made by online platform owners. They shape both the designs and architectures that mediate the dynamics of (jerkish) speech. In order to advance critical civility in online contexts, we need to recognise their central role. Online platforms can be approached as crucial ‘regulators’, a term used by Nicholas Vrousalis to describe agents that contribute ‘appropriately to the creation, reproduction, or perpetuation of the constitutive power dyad’ (Vrousalis 2020, p. 6). The regulator and the dyad (consisting of dominator and dominatee) together constitute a structural relation of domination: it is the regulator that creates the structural conditions for domination. More concretely, online platforms can either promote or inhibit the spread of jerkish speech and uncivil norms, and—as part of critical civility—have a duty to work towards the latter rather than the first.

Regarding how the design of online environments directly affects dynamics of speech, as discussed in the section on jerkish speech, we must take into account the impact of anonymity, distance between speaker and listener, and the instantaneousness of speech on the way individuals engage with one another online. Anonymity of users is frequently linked with es uncivil behaviour (Santana 2014),Footnote 8 the lack of direct social cues associated with online speech has negative effect on the behaviour of users (Citron 2016), and the instantaneousness of online speech removes barriers for citizens to reflect and seek actual engagement with others (Brown 2018). These effects, and perhaps other parameters of online speech, could significantly undermine efforts to cultivate critical civility and exacerbate the issue of its demandingness.

Levering the role of online platforms, on the other hand, might alleviate the issue. Various examples come to mind here, such as YouTube removing the dislike button in order to ‘create an inclusive and respectful environment’Footnote 9 and Instagram nudging users to rethink whether they actually want to post a harmful message, showing that platforms have come to realise that they play a key part mitigating problematic forms of speech. Currently, much of whether platforms decide to introduce such measures depends on their own strategies and policies, however, and some might decide not to respond to jerkish speech or not to promote critical civility—or indeed, roll back such efforts, as evidenced by the spike in hate speech after recent Elon Musk having acquired Twitter (Benton et al. 2022). Formally recognising the duty of online platforms to ensure that the designs of their platforms favour critically civil engagement rather than hateful and jerkish engagement could support widespread critical civility and prevent domination as a result of widespread jerkish speech by forcing such platforms to comply.

3.4 Critical Civility and a Multitude of Publics

The second element is perhaps more fundamental to jerkish speech: widespread critical civility could be undermined by the fragmentation and polarisation of the digital sphere. As mentioned, Whitten follows Sunstein’s concern with fragmentation and polarisation as threats to deliberation (Sunstein 2018). A key part of this concern is the ‘filter bubble’ effect (Pariser 2011), or the use of algorithms by online platforms to automatically filter which data reaches the user, in real time. As such, algorithms are often highly personalised, and the result is that individuals may come to live in their own ‘bubble’, where the only information that reaches them as listeners is aligned with pre-existing beliefs and interests.

The idea of filter bubbles relates to the notion of echo-chambers, a term often used in a more critical manner. Where a filter bubble describes a mechanism of information polarisation, echo chambers denote a certain ideological commitment that is used to weigh the trustworthiness of information. This is frequently considered as epistemically problematic and resistance to it requires individuals taking responsibility for "not allow[ing] epistemic silencing and ensur[ing] a friendly epistemic environment for everyone” (Karimov et al. 2022, p. 697). These civic duties intersect with the kind of critical civility that citizens must possess and display against jerkish speech.

From the perspective of the user as a speaker, Riemer and Peter (2021) point out that social media platforms interfere with speech through algorithmic audiencing. Not only do they determine what can be said or what reaches the user from her own perspective, but also ‘what can be heard’ and ‘by whom’ (p. 409). Social media platforms have the means to give priority to some messages over others, and their interest in user engagement frequently leads to the promotion of viral content over civil content. Moreover, rather than distributing along existing networks, audiences are made and tailored according to the algorithm. The implication is that this causes fragmentation and polarisation on a larger societal scale.

Again, online platforms, as builders of the architectures and algorithms that shape these dynamics, play a key part. In this case, however, the solution is less straightforward. In response to Sunstein’s work, Whitten suggests that critical civility requires ‘greater transparency of news sites, the protection of deliberative platforms for citizen exchange, and for the maintenance of a robust and fair common ground media that provides all citizens with a central set of resources with which to deliberate the issues most important to their society’ (Whitten 2022, p. 223). But it is not clear that this would effectively spread critical civility to the extent that individual users can confidently be free from jerkish speech every day. Even if it is possible to cultivate critical civility in the mainstream digital sphere in light of the issues raised above, disrespecting minorities might still very well find their way into the wider digital sphere. In addition, the fragmentation of online communities and the connectiveness of digital media make it more difficult to spread civil norms throughout different communities, while also increasing the likelihood of citizens coming across members of communities that are not committed to the same norms.

In fact, we have reason to think that the fragmenting and polarising effects of online platforms are even more complex. While one strand of argument and empirical research seems to support the polarisation-effect, there is also reason to believe that social media increases the exposure to viewpoints of others, increases support for and reduces prejudice against other groups, and reduces political polarisation (Barberá, 2020). In addition, there are significant differences in how individuals respond to (political) information, leading to asymmetric polarisation: ‘there are thus good reasons to expect that social media interactions may lead to polarisation among the minority of partisan individuals who are most active in discussions about politics on social media’ rather than the general public (p. 48).

Perhaps a more constructive way to approach the dynamics of fragmentation and its effect on jerkish speech would be to revisit the idea of a digital sphere as being a singular and rather unified whole. Engaging with the Australian Twittersphere, Bruns and Highfield (2015) propose to move beyond Habermas’ traditional notion of a public sphere towards countless different but co-existing and even overlapping smaller ‘publics’. The digital sphere, on their account, is not marked by either a unified whole, nor by highly fragmented communities, but rather by vertically and horizontally overlapping smaller publics based on interest or issues. These publics can exist for a long time but may also rise and fall quickly in response to specific triggers or issues. In any case, users are usually not limited to one ‘public’ but engage within many in their online activities. And, importantly, these publics ‘may follow their own logics and norms, making use of affordances of social media platforms for their own purposes, which may differ from established practices’ (p. 125).

Apart from nuancing the argument of polarisation, this could shift the perspective somewhat on what critical civility requires in online environments. On the one hand it suggests—unsurprisingly—that online platforms should employ filters and audience algorithms to promote publics where critical civility is the norm, rather than promote jerkish content that happens to sell. On the other hand, it draws attention to the possibility of placing boundaries around those publics that are uncivil (or publics where the norms of civility are matter of critical contestation). While this would not count as promoting critical civility as such—and runs somewhat counter to Whitten’s concern for fragmentation and polarisation—it could reduce some of the reach of jerkish speech and jerkish publics. It might even be the only option in cases where it is not possible or desirable to ban jerkish speech or where it is unlikely that critical civility becomes the norm any time soon, due to the issues discussed earlier. In other words, in publics where critical civility fails, targets of jerkish speech need to be protected by publics where they can count on support. A concrete example of how this could be achieved is Reddit offering users the ability to mute complete parts of its website, allowing them to effectively withdraw from publics that they find offensive or hurtful.Footnote 10

In sum, one of the most important tools within the republican toolbox against jerkish speech is the cultivation of (critical) civility. The characteristics of online environments, however, might add to the demandingness of critical civility. In online contexts, this issue can be alleviated by successfully incorporating the role of digital platform owners. Their position as regulators or mediators of the dynamics of speech allows them to (1) design platforms in ways that nudge or stimulate users to civil behaviour, and to (2) leverage the variety of different publics to promote those where critical civility is widespread and to isolate those where it is not, in order to protect potential targets from facing jerkish speech.Footnote 11

4 Conclusion

Jerkish speech, a subtle, indirect, and even inconsequential cousin of hate speech, can amount to domination in two ways: it can be a form of micro-domination or an instantiation of systemic domination. The first refers to domination that is not significant enough to be cause of concern by itself, but which is nevertheless an issue on a larger scale. Jerkish speech can also be a form of systemic domination. In this case, jerkish speech is a reflection and a reproduction of structurally unequal social hierarchies that cast the targets of speech in a subordinated position. This prevents them from participating in norm-shaping and denies their normative authority, constituting systemic domination.

Republicans are therefore committed to reducing the prevalence of online jerkish speech or, if that fails, protecting its targets. While speech regulation is an important part of a response to hate speech, it may not always be effective or appropriate in the case of jerkish speech. Promoting critical civility, as proposed by Suzanne Whitten (2022), is a promising second weapon for combating jerkish speech. In online environments, however, it is particularly demanding. Citizens need to be disposed to engage with others in deliberation without denying normative authority. Moreover, they need to critically evaluate structures of norms for traces of systemic domination.

The demandingness may be alleviated by leveraging the position of important regulators: online platforms. They should consider the effects of their designs on the dynamics of speech, and they need to promote civil content over jerkish content while making use of the diversity of online publics. Those publics where critical civility is the norm could be promoted, while those where jerkish speech is prevalent could be supressed or hidden, both in order to respectively support and protect potential targets of frequent jerkish speech. The promotion of online critical civility, in response to domination as a result of jerkish speech, depends heavily on the part played by online platforms, and warrants formal, legal recognition of this role and of the corresponding duties of online platforms.