Uncivil Speech in the Social Media: Democracy, Political Liberalism, and the Virtue of Public Reason

Abstract

Initial hopes of the democratizing potential of the internet are increasingly replaced by fear that a fragmented and unedited public sphere unleashes the destructive forces of populism, hate speech, and demagoguery (Tucker et al. 2017). In social media, these forces are amplified by algorithms used to feed designed to create “audience engagement.” Thus, toxic communication in social media is no accident as postings that provoke and elicit emotional responses are consistent with the business model (Saurwein and Spencer-Smith 2021; Maréchal 2021). At the same time, operators of social media supervise and regulate postings on their platforms. Content violating “community standards” risk being demoted, flagged, or deleted, while users that do not comply with these standards are potentially excluded or “deplatformed” in the new vocabulary.

As emphasized by Gillespie (2018), content moderation is an unavoidable feature of social media. Standards for content moderation frequently target hate speech, disinformation, and violent threats. Few deny that measures taken to reduce such content are or can be justified. Less obvious is if platforms should regulate and moderate uncivil speech; words or utterances that communicate rudeness, impoliteness, insult, or hostility.1 Uncivil speech is here defined as denigrating but not hateful speech: uncivil speech and hate speech are thus mutually exclusive categories (Waldron 2013). Hence, the question is whether social media providers should delete, demote, or flag uncivil postings and comments. Are they justified in ultimately excluding from their platforms users that engage in what Meta (2017) calls “language that seems designed to provoke strong feelings”?

The value of freedom of speech is a reason to worry about regulating uncivil speech in social media. Just as there is reason to object to governments that suppress free speech in the public sphere, we should arguably oppose acts of censorship by social media platforms in the “digital public sphere” (Schäfer et al. 2015). Accordingly, critics argue that the regulatory activities of social media platforms are contrary to basic democratic values and principles. The suppression of speech in social media is a species of “cancel culture” that stifles freedom of expression and undermines the value of democratic participation (e.g., McGarvey 2022).

Against these critics, it is worth keeping in mind that content moderation has for a long time been institutionalized in traditional media (Ward 2014). Publishing and broadcasting is unthinkable without editors enforcing professional standards, codes of responsible journalism and self-regulatory nongovernmental institutions. Hence, the normative issues at stake in the digital public sphere are not altogether new. What is new is who the relevant players are. Standards of content moderation in social media are not defined by members of the journalistic profession but by corporations that apply them against ordinary citizens creating online content.

The cumulative effects on the public sphere are also different. The sense is that social media are responsible for an unprecedented “crisis of civility, a veritable war of words that distorts our public discourse” that ultimately threatens to undermine our democracy (Bejan 2017, 1). As the public sphere grows more uncivil, its democratic value quickly recedes. In a democratic society, the public sphere should be an arena for free opinion formation and “mass deliberation” among citizens. Ideally, the public sphere should contribute to keeping the democratic citizen informed, encourage critical perspectives that hold officials into account, and constitute a public arena that includes and represents the diversity of opinions and perspectives among citizens (Chambers 2012). But uncivil language is not informative, not instrumental to the accountability of public officials, and it does not represent the opinions and perspectives of citizens. To the contrary, the more nasty and “toxic” the public debate is, the less the public is likely to listen and participate in public discussions at all.2 According to Stephen Macedo (2022), we should welcome private initiatives to supervise and regulate uncivil speech that correct the dysfunctionalities of the digital public sphere.

There is consequently disagreement on how democracies should respond to uncivil speech in social media. Some argue that freedom of speech is at peril if social media platforms regulate uncivil speech. Others take the opposite view and contend that democracy is at risk if social media do not regulate uncivil speech on their platforms.

This paper asks what the democratic state should do; should it protect freedom of speech against the regulatory activities of social media corporations? Or, should the democratic state embrace and even encourage attempts by social media corporations to civilize speech on their platforms? The consequent framing of the normative problem is taking for granted the predominant framework of economic liberalism in which social media corporations operate. Whereas others argue for more radical solutions, this framing of the problem allows us to examine the viability of political liberalism for urgent normative issues “here and now.”

The conundrum is answered from the standpoint of political liberalism, which offers an account of the “moral basis for a democratic society” (Cohen 2003, 86). Political liberalism is well equipped to address the clash between conflicting institutions about the public sphere since it embraces both a “highly protective doctrine of political speech” (Nussbaum 2011) and a demanding account of public reason—the reasons appropriate for citizens in the public (Freeman 2004). Political liberalism is strongly committed both to freedom of speech and reasoned public discourse and should therefore be able to guide our judgments on the democratic dilemmas provoked by uncivil discourse in the digital public sphere. This article seeks to work out the normative implications of public reason for social media—an endeavor not yet undertaken.3

Note, however, that the normative commitments of political liberalism are scarcely unique. Freedom of speech and public reason are equally foundational to the discourse ethics of Habermas, although their “devices of representation” are distinct. While lack of space does not permit further elaborating of this point, the arguments and conclusions developed in this article would scarcely be different if they had been derived from discourse ethics.4

The argument advanced here is that the moral vicissitudes of uncivil speech scarcely justify the state intervening either to prevent citizens from engaging in uncivil speech or to prevent social media platforms from regulating uncivil speech. In the public sphere, civility is either a moral duty or a virtue of the democratic citizen. The citizenry has a moral duty, albeit not one that is legally enforceable, to avoid uncivil speech in the context of the political justification of the democratic constitution. In other public settings, civility is a moral virtue, implying that it is not morally required and certainly not legally required.

The basis for this standpoint is the distinction between the right to freedom of speech against the state and the right of associations to regulate their internal life. Freedom of speech, including the right to uncivil speech, holds against the state but not against the free associations of the public sphere. As social media platforms belong to the public sphere, they are at liberty to regulate the communicative activities of their members or users.

The conclusion is not that uncivil speech in social media is of no concern to the democratic state. While the citizen is legally free to engage in uncivil speech on social media, it is better if she does not. At stake in social media are the virtues of the citizens that undergird the social infrastructure of democracy. These virtues importantly include the idea of public reason, as manifested in reasonable political discourse. Hence, we should welcome the regulatory initiatives of social media platforms to the extent that they encourage the virtues of citizenship. In a democratic society, it is not for the state to foster the virtues of reasonableness and civility in the public sphere.

The principles of a democratic society, as defined by political liberalism, grant “priority” to the basic liberties of the individual citizen.5 Foremost among the basic liberties are freedom of speech, freedom of association, liberty of conscience and thought, and the right to vote and to hold public office.6 These political liberties have priority in the sense that they should not be sacrificed or limited for the sake of other valuable ends. Thus, “priority” does not merely convey the moral importance of these liberties; the claim is that they should “never” be infringed however important the other ends at stake may be (Cass 2021, 137). Since freedom of speech is among the political liberties with priority before other social ends, the state is scarcely ever justified in restricting speech by means of the law. Accordingly, political liberalism is typically understood as a doctrine that advocates for “unconstrained freedom of speech” (Bonotti 2015).7

The rationale for the priority of political liberties is that they represent the “social conditions” necessary for the development of the basic moral powers of the citizen, in particular, her capacity to form a conception of the good life (Rawls 1993, 293). Each citizen has a “higher order interest” in the development of her moral powers. In light of this interest, it is unreasonable to expect anyone to forgo the social conditions for the pursuit of her conception of the good even if it would confer substantial improvements to her in other respects.8

The claim that political liberties are necessary social conditions for the good life does not entail that they are sufficient conditions, for sure. However, it indicates that the denial of political liberties undermines the opportunity for citizens to identify and revise their conceptions of the good life. Without a protected right to speak, listen, and associate with others, “people will find it very difficult or impossible to develop and exercise their moral capacities” (Nickel 1994, 780). In a society where speech is suppressed by government, there is no free interplay of ideas. This reduces each citizen's understanding of what makes life good by preventing her from engaging in interpersonal comparisons of distinctive visions and practices of the good life (Melenovsky 2018). In other words, political liberties and freedom of speech have priority because they are essential to anyone's pursuit of the good life and not merely because they are essential to specifically liberal—let alone libertarian—understandings of the good life (cf. Pevnick 2015).

Now, the claim that freedom of speech is among the political liberties that have priority does not necessarily imply that uncivil speech should be protected. Is the protection of uncivil speech among the social conditions for citizens to develop their moral powers? Indeed, there is a positive case to believe that it is. Freedom of speech is justified by the higher-order interests of the citizen, and these interests extend to uncivil speech if such speech is or can be an element of a person's conception of the good life. How that might be is perhaps not obvious. How could rude, hostile, impolite, or inflammatory language be a way of life? The point is that political liberalism does not provide any criterion by which to judge the content of anyone's conception of the good life. The democratic state should consider the good life an “unknown” (Rawls 1993, 311). Political liberalism does “not presuppose the validity of any conception of the good or any set of conceptions of the good” (Barry 1994, 326, emphasis added).9

The point is not just that a democratic state does not subscribe to any conception of the good life. The further point is that a democratic state does not recognize any standard for the determination of what is—and what is not—a conception of the good life: there is no evaluative standpoint from which the democratic state can assertively deny that a citizen that engages in uncivil speech is not in the process of developing her ideal of the good life. Hence, if freedom of speech is justified by higher-order interests in pursuing the good of life, and if uncivil speech might be an element of the good life, we must conclude that uncivil speech is covered by freedom of speech. There is consequently a strong presumption against regulating uncivil speech in a democratic society.

This defense of the right to uncivil speech is controversial, for sure, and should be measured against two familiar objections. The first is that the state should be free to restrict speech if necessary to secure respect for other citizens. The second is that the state is permitted to limit speech in order to secure equal democratic participation.

Unreasonable speech is not an exercise of freedom of speech, according to Jonathan Quong (2004b). As is true for all rights and liberties, the right to freedom of speech is only designed to protect “reasonable alternatives,” which are the set of choices underwritten by respect for the freedom and equality of others (Quong 2004b, 332). Choices that impinge on the legitimate interests of others are not reasonable and therefore not “covered” by freedom of speech (Yong 2011).

The fact that unreasonable speech is not covered by freedom of speech does not entail, as Quong is quick to point out, that the state can legally restrict unreasonable speech as it sees fit. Legal bans on speech are justified only if supported by sufficiently strong reasons. As understood by Quong, the state is morally justified to limit unreasonable speech only if it poses a direct threat to the “essentials of a liberal democratic regime” (Quong 2004b, 324).

However, the relevant question for us is whether uncivil speech constitutes “unreasonable speech” as defined by Quong. Speech is unreasonable only if it articulates “unreasonable objectives” such that it denies others’ “basic freedom and equality” (Quong 2004b, 333). For verbal statements to express unreasonable objectives, they must either explicitly or implicitly be claims about the rights and freedoms of others.

Now, even if claims that implicitly or explicitly reject the rights and freedoms of others may be expressed in conjunction with uncivil language it does not follow that uncivil language is itself unreasonable. Rude, offensive, or mean language are not statements about the rights and freedoms of others, though it may certainly be used in conjunction with such statements. But when uncivil language figures in the context of speech that attacks the rights and freedoms of others, the unreasonable part of that speech is the claims made about their rights and freedoms. Uncivil language is not unreasonable per se since it does not include claims about the rights and freedoms of others.10 The verdict is consequently that uncivil speech does not include “unreasonable objectives.” Quong's insistence that unreasonable speech is not protected by freedom of speech is for this reason irrelevant to uncivil speech.

The second objection against legal rights to uncivil speech is that such speech undermines the democratic rights of other citizens. Following Andrew Reid, hateful speech has “second-order silencing effects” that deprive others of “de facto capacity to participate in political life” (Reid 2020, 188). These harms generate pro tanto reasons in favor of a ban on hate speech. These reasons are not necessarily conclusive, of course. There are pro tanto reasons also against the state imposing a ban on hate speech. One of them is the danger of the state “overreaching” and, perhaps unintentionally, suppressing the legitimate interest of the citizen.11 Yet, in balancing the reasons for and against there is likely to be some point where the reasons in favor of a ban prevail. A democratic society should consequently accept that legal limitations on “some hate speech” can be legitimate (Reid 2020, 187).

Again, uncivil speech is not necessarily hate speech. Yet, in Reid's analysis, the nature of speech is immaterial. The only thing that matters is whether a particular piece of speech is silencing others such that they are unable to participate in democratic institutions on equal terms. Conceivably, speech does not have to be either unreasonable or expressed in propositional form to produce these effects. Hence, if it is true that uncivil speech contributes to the silencing of others, there is a pro tanto reason for a legal ban on uncivil speech.

It is far from clear, however, that political liberalism allows the state to limit speech due to silencing effects. Recall that political liberties should be granted priority before other socially valuable ends. If freedom of speech is among the political liberties that should be granted priority, the democratic state is not entitled to restrict freedom of speech even if there are reasons in favor of so doing. The fact that others are “silenced” by uncivil or hateful speech is certainly one consideration in favor of limiting the right to engage in uncivil or hateful speech. But if freedom of speech enjoys “lexical” priority before other ends, silencing effects are never sufficient to justify legal limits on speech.

The priority of political liberties does not imply that they are absolutes, of course. According to political liberalism, the basic rights of the citizen are permissibly subject to legal limitations if necessary to protect the “system of rights.” This is the case only in the event of a “constitutional crisis” where the coercive powers of the state can justifiably be used to suppress speech if “necessary to prevent a greater and more significant loss to these liberties” (Rawls 1993, 356; Nussbaum 2011). Although legal limitations on uncivil speech are theoretically consistent with this condition, the conclusion is that the legal protection of uncivil speech should normally be guaranteed.

The notion that a democratic society should protect rights to uncivil speech does not imply that citizens are entitled to uncivil speech everywhere. The political liberties apply only to the “basic structure” of society, primarily constituted by the state and its legal institutions. Importantly, the basic structure does not include the associational life of the public sphere.12 The public sphere, or the “background culture,” includes a variety of associations: churches, universities, the media, political parties, unions, and other gatherings that are created to promote the interests of their members.13

The limited scope of political liberties means that, for example, the right to vote and the right to freedom of speech have no application in the internal life of associations. There is no constitutionally protected right to vote in churches, universities, the media, political parties, or trade unions. Similarly, these associations are free to regulate speech in spaces subject to their control as they see fit. The associations of the public sphere can legitimately impose sanctions on members who express opinions contrary to the values of the association.

The claim that associations can permissibly regulate their own internal life is a necessary precondition for the existence of communities of “belief.” In a democratic society, the associations of the public sphere are not “guided by any one central idea or principle, whether political or religious” (Rawls 1997, 768). They should consequently be free to regulate the doctrines of belief upon which they are founded. Unless they are, citizens would lack the capacity to unite and form associations based on their shared comprehensive doctrines of the good life (Alexander 2008). From the point of view of political liberalism, associational freedom is a necessary precondition for the effective enjoyment of freedom of conscience (Fischer 1997, 32).

Political parties, churches, and so on, are thus at liberty to exclude members on the basis of ideological, religious, and other “comprehensive” doctrines. The right to exclude members on these grounds effectively entails the right to regulate speech and to sanction (e.g., exclude) members who engage in speech contrary to the doctrines espoused by the association. The right of associations to regulate speech is illustrated by the fact that standards of acceptable speech are omnipresent in universities, schools, legal professions, voluntary associations, and corporations (Cortina et al. 2019).14

Social media platforms are scarcely communities of belief, of course. They are, first of all, private corporations. Following Gillespie (2018, 18), they are “complex institutions” run by commercial interests, though not exclusively so. This observation does not change the fact that corporations are not part of the basic structure of society, however. Just like private associations and the media, corporations belong to the background culture of society. From the vantage point of political liberalism, corporations have more in common with voluntary associations than with the state (Phillips and Margolis 1999). Just like associations, corporations are at liberty to define “final ends and aims” on the basis of the comprehensive doctrines of their owners or members (Singer 2015).15 In contrast to the state, associations are free to define their own purposes.

The claim that the rights of corporations are derivative of interests in freedom of association can be taken to extremes, however. Following Messina (2023), corporations with expressive interests that dislike the views or beliefs of an employee should be able to fire (“disassociate from”) her.16 Just like the Catholic church should be able to exclude a priest who is no longer a Catholic, a corporation should be able to sack an employee with views disliked by the owner.

That is not the position defended here, however. The argument in this section is not that corporations should be granted the same rights as voluntary associations. After all, corporations are legal entities introduced by the state to serve the public good. The state should then be able to regulate the activities of corporations to ensure that they align with the public good as defined by the democratic process (Rönnegard and Smith 2024; Ciepley 2013).

The point is rather that the right to uncivil speech—a right enjoyed against the democratic state—does not hold against corporations because, like associations, they are free to define their own purposes. This is important as it implies that freedom of speech does not provide a justification for the state using the law to prevent corporations from regulating uncivil speech among employees or members.

A final objection is that the analogy to associations is misleading because social media platforms are not mere corporations that are providing services. In effect, they are suppliers of communicative infrastructures (Plantin et al. 2018). Corporations responsible for infrastructure in telecommunications or electricity clearly should not be free to exclude users just because they engage in uncivil speech. Similarly, social media platforms should not have the right to exclude citizens from the communication infrastructure.

It is doubtful, however, that social media platforms are providers of infrastructure. In contrast to users of infrastructure, the users of social media platforms agree to “terms of service” that indicate an active choice to join a community. In addition, the value of infrastructure is not necessarily compromised by public ownership and regulation, whereas the value of social media platforms would be. From the point of view of political liberalism, there is a value to the pluralism of content, moderation strategies, and technologies that is unlikely to materialize if social media were run by the state.

The claim that the democratic state should not abrogate uncivil speech does not entail that the citizen is morally free to engage in uncivil speech. The legal right to speak is consistent with the moral duty not to. In fact, political liberalism stipulates moral limits on speech as defined by the idea of public reason.

Public reason refers to the standards of practical reasoning that apply to the citizen when she deliberates together with others on the constitutional essentials of a democratic society. In these contexts, the citizen has a “duty of civility” to speak in conformity with the “political virtues of reasonableness” (Rawls 1993, 224). The duty of civility is a moral, nonlegally enforceable duty that applies to the members of the high courts, the parliament, and the higher echelons of the public administration. It also applies to the citizen when she votes in general elections. Voters are to think of themselves “as if they were legislators”; they are to vote on the basis of judgments about the “most reasonable” laws and policies (Rawls 1997, 769).

The idea of public reason includes standards of reasoning as well as standards for factual judgment. To speak in accordance with the cannons of public reason is to make arguments that are consistent, valid, and truthful. Speech that complies with public reason is reasonable as it can be understood and, therefore, accepted by others. Since the state is legitimate only if the constitutional essentials are grounded in reasons that are reasonable, and since public reason defines the standards for reasonableness, the citizen has a moral “duty of civility” to comply with the standards of public reason (Rawls 1993, 253). Public reason is a standard for reasonably acceptable political argument.

It is not clear if the duty of civility excludes uncivil speech. The very expression “duty of civility” suggests that it does. Yet, the criteria that determine the content of public reason may not be applicable to words and utterances that are considered uncivil. Consider, for example, name-calling, which is reportedly the most common form of uncivil speech (Coe et al. 2014). Is name-calling prohibited by standards for reasonable argument as specified by public reason? Note that to call someone by a foul name is not to introduce a reason for or against a proposition. Name-calling is exclamation, neither reasoning nor asserting facts: the statement “You scumbag!” has no truth value. Hence, if public reason is a set of standards for reasoning and factual judgment, and if name-calling is neither, public reason does not apply to name-calling.

Nonetheless, it appears odd to conclude that public reason is consistent with the use of foul words, name-calling, and inflammatory rhetoric. After all, public reason should be manifested in the discourses of high-court judges, government officials, and elected representatives, and we do not expect any of them to engage in uncivil speech. This conclusion is corroborated by Rawls’ “test” of when a statement is consistent with public reason. To test a statement, we should ask ourselves how it would “strike us as presented in the form of a Supreme Court opinion? Reasonable? Outrageous?” (Rawls 1993, 254). The answer is, plausibly, that name-calling and other forms of uncivil speech are unreasonable as we would find it outrageous among the judges of the Supreme Court. But on what grounds?

If the claim that name-calling and other forms of uncivil speech are contrary to public reason cannot be verified by the standards of public reason, the basis for this claim must be sought elsewhere. The justification is, I propose, that uncivil speech is contrary to the political values that justify public reason.

Reciprocity is the preeminent political value of a democratic society and is obtained when the basic structure of society is justified on terms that are reasonable for all citizens. Reciprocity is the “intrinsic normative and moral ideal” of liberal political legitimacy (Rawls 1996, xliv). Accordingly, the political value of reciprocity is achieved only if the coercive powers of the state are justified by reasons that “others can reasonably accept” (Rawls 1997, 771). Now, since public reason defines the conditions for reasonable justification, public reason is a necessary feature of a society that realizes the value of reciprocity: the “role of the criterion of reciprocity” is “expressed” in public reason (Rawls 1997, 771; Freeman 2004, 2040).

The upshot is that public reason is not merely a standard for reasonable political argument. Public reason is also a condition for the realization of the moral ideal of reciprocity among the members of society. Reciprocity requires that the members can distinguish between the moral requirements that apply to them in the “capacity of citizen” and the rights and duties that apply to them as private persons (Freeman 2004, 40; O'Neill 1997, 418). To act as a citizen is to adopt the standpoint from which public reason is acknowledged as the currency of reasonable justification.

Accordingly, public reason is not a mere constraint on the deliberations of the citizen as it defines the standpoint of the reasonable citizen who is engaged in political justification. The point is that the reasonable citizen is fully committed to providing reasons that others can reasonably accept. There is no place for name-calling or other forms of uncivil language among reasonable citizens, not because uncivil speech fails to pass the test of public reason, but because the sole aim of the reasonable citizen is to offer reasons that others can accept. Public reason is the mark of the reasonable citizen who provides a political justification of the basic structure of society in order to secure the political values of reciprocity. Consequently, reasonable citizens avoid uncivil speech because it violates the political value of reciprocity.

The claim that uncivil speech is contrary to public reason is not sufficient to conclude that citizens should not engage in uncivil speech on social media. The reason is that the scope of public reason is limited: The duty of civility applies only to the “public political forum,” defined as the context where decisions are made on the constitutional essentials of the state. Since constitutional decisions are made only by elected representatives and judicial institutions, and indirectly also by voters, the domain to which the duty of civility applies is to be understood narrowly. On this narrow reading, the moral duties of public reason do not extend to the public sphere: “the idea of public reason does not apply to the background culture with its many forms of nonpublic reason nor to media of any kind” (Rawls 1997, 768). Provided that the background culture includes social media, it follows that the duty of civility does not extend to discourses in social media—the citizen is under no duty to avoid uncivil speech in social media.

It may be objected that public reason should apply “broadly,” not just to speech in the public political forum (Pariente 2020, 108). The premise for this argument is that the liberal principle of legitimacy applies to the coercive powers of the state in general. The standards of public reason that define the content of reasonable political justifications should be adhered to whenever the state enacts “binding, coercive law” (Quong 2004a, 237f.). Not all forms of political speech are effective in deciding coercive law, of course. But if political arguments are arguments about the coercive powers of the state, and if public reason applies to all justifications of the coercive powers of the state, it follows that all political arguments should align with the standards of public reason. The claim defended is that public reason should guide debates on all “problems of everyday democracy” (Quong 2004a; Pariente 2020).

The broad interpretation of public reason can be challenged upfront by denying that the liberal principle of legitimacy applies to coercive decisions of the state in general (Hodgson 2012). However, even if the broad reading is accepted, it is unclear that public reason extends to political discourse in social media. Though discourse in social media is frequently political, not every argument on the coercive powers of the state is a justification for actual exercises of coercive power. Political arguments can be part of political justifications, of course. That might be true, for example, of the arguments made in parliaments. It nevertheless remains true that arguments about the coercive powers of the state are often distinct from the justifications provided for the state's exercise of coercive power.

The point is that even on an expanded reading of the liberal principle of legitimacy, the requirement of reasonableness and public reason applies only to political debates that serve as justifications for the coercive powers of the state. Since debates in social media are not justifications of actual exercises of coercive power, there is no basis for the claim that public reason applies to political arguments in social media. In sum, there is no duty of civility in social media.

Yet, there may be other duties that do apply to discourses in the public sphere. This possibility has gained increasing attention in times when democratic institutions are challenged by populist and extremist movements. The democratic citizen arguably has a duty to engage in “counter-speech” when extremists use social media to circulate lies and conspiracy theories that undermine trust in democratic institutions (Cepollaro et al. 2023).

Following Badano and Nuti (2018), political liberalismestablishes a moral duty to engage with unreasonable doctrines in the public sphere. The survival of democracy ultimately depends on the willingness of the general population to be reasonable such that an overlapping consensus can be created among people with conflicting moral and political convictions. Therefore, the stability of the democratic order is jeopardized by populist movements that circulate a variety of unreasonable doctrines.

The moral duty defended by Badano and Nuti is to “press the unreasonable” and to make unreasonable persons “change their mind.” The democratic citizen has a moral duty to confront advocates of unreasonable doctrines and to eventually induce them to be more reasonable (Badano and Nuti 2018, 157). The duty applies if unreasonable doctrines “risks leading society towards a real threat to the stability of liberal institutions” (163).

The relevant point is that the duty to press the unreasonable is not limited to the public political forum as it also extends to “informal venues” that bear no direct connection to the decisions about the coercive powers of the state (Badano and Nuti 2018, 158). The conclusion is that political liberalism offers a rationale for a moral duty to engage with unreasonable persons in the public sphere. While social media is not mentioned, there is no reason to doubt that digital encounters fall within the purview of the “informal venues” to which this duty applies.

Does the duty to press the unreasonable entail the duty to press others not to engage in uncivil speech? The answer is perhaps obvious given that reasonableness is defined as the predisposition to offer reasons that others can reasonably accept. Since there is no place for uncivil language in a reasonable justification, as argued in the previous section, it follows that a person “pressed” into being reasonable should also avoid uncivil speech.

Recall, however, that a reasonable citizen should avoid uncivil speech only when she engages in political justification. According to political liberalism, public reason is the voice of the reasonable citizen in the public political forum. Thus, even if there is a moral duty to “press the unreasonable” in the public sphere, it does not follow that public reason applies to the public sphere. Discharging the duty to press the unreasonable is to make an effort to convince other citizens of the value of reciprocity and public reason in political justification. But in discharging that duty, we are not acting on the duty of civility as we are not in the context of providing a political justification. Moreover, the unreasonable person that is pressed into being reasonable is realizing the value of public reason such that she will abide by public reason when offering political justifications in the public political forum. That, again, is distinct from being reasonable in the public sphere, let alone in social media.

Political liberalism is committed to freedom of speech, including legal rights to uncivil speech. Yet, social media platforms are part of the background culture of society and not subject to the principles of political justice. Platforms are thus free to regulate speech as they see fit. Moreover, although the citizen has a moral duty not to use uncivil language in the public political forum, no such duty applies in the public sphere. There is no general democratic duty to avoid rudeness, name-calling, and offensive language.

However, these claims about the duties and rights of the citizen are consistent with the claim that uncivil speech in social media and other domains of the public sphere is bad. The basis for this conclusion is that a democratic society affirms an ideal of citizen virtue and that it is a mark of a virtuous citizen to resist the temptation to engage in uncivil speech. An account of virtue does not specify norms that should guide either institutions or individual behavior. An account of virtuous citizenship grounds judgments about the valuable state of affairs, not judgments about the norms of action that apply to either individuals or institutions (Jubb 2016).17

The thesis is that political liberalism embraces an account of the virtues of citizenship according to which civil speech in the public sphere is of great political value—albeit not a duty. To defend this view, we need to show that it is explicitly endorsed by political liberalism, that it is consistent with the other commitments of this doctrine, and that uncivil speech is in fact contrary to the virtues of the good citizen.

The virtues of the democratic citizen are explicitly affirmed in political liberalism under the “cooperative virtues of political life.” The idea is that a well-ordered democratic society “encourages” the “virtues of reasonableness and a sense of fairness, and a spirit of compromise and a readiness to meet others halfway” (Rawls 2001, 116; Rawls 1993, 163; Gursozlu 2014; Dagger 2013).18 Reasonableness is a “highly desirable” feature of the citizen; it represents an ideal of the democratic citizen (Rawls 1993, VI §2; Pariente 2020, 110).

The virtue of reasonableness is distinct from the duty of reasonableness. As argued earlier, the citizen is under a moral duty to offer a reasonable justification in the political discourses that are directly pertinent to justifications of the constitutional essentials. The virtue of reasonableness, on the other hand, is a desirable feature of the citizen; it is realized not by the citizen discharging her duty of civility but by the citizen being virtuous in everyday settings.

Hence, the virtue of reasonableness is not limited to the context of constitutional essentials. The claim is that the cooperative virtues “constitute a great public good” when “widespread in society” (Rawls 2001, 195; Rawls 1997, 769). In other words, political liberalism offers an evaluative theory of the good citizen that is distinct from the theory of the duties and rights of the citizen. While the duties of public reason are limited to the public political forum, the virtues of the citizen know no such limits. The cooperative virtues of the citizen apply in all domains and extend to the public sphere and social media.

However, it is not immediately clear why citizens who endorse cooperative virtues would not engage in uncivil speech. It is conceivable that the citizen recognizes the value of being reasonable and to “meet others halfway” in social encounters generally, while she does not recognize any reason to withhold rude or hostile language. To substantiate this potential objection, consider that the virtues of the citizen are specified as “cooperative,” exemplified by the traits such as “fidelity” and “trust” (Rawls 1971, 355), indicating that they apply to situations where people are acting for mutual benefit. Citizens cooperate on reasonable terms when they temper their self-interests and when they take into due consideration the interests of others. The point is that cooperative virtues are constraints on behavior rather than on speech. It seems perfectly consistent to be a cooperative citizen and yet indulge in uncivil language.

The phrase “cooperative virtues” is misleading, however. The term “cooperative” alludes to the idea of society as a venture among self-interested individuals. However, as demonstrated by Barry (1995), the basis for a democratic society is not that of mutual advantage. A democracy is, instead, based on the moral claim that respect for others necessitates a political justification that others can reasonably accept. It is “cooperative” in the sense that all members are respected as free and equal. The cooperative virtues are consequently not merely behavioral predispositions that constrain self-interests; they also comprise the attitudes mandated by the principles of respect for others as moral persons. Ultimately, the cooperative virtues are manifested in the willingness to propose terms for cooperation that others can reasonably accept. This idea is mirrored in the claim that “impartiality” is among the cooperative virtues (Rawls 1971, 472). The same point is conveyed by the claim that the “political virtues” include “civility and tolerance” (Rawls 1993, 194). Public reason should, in other words, be regarded as a virtue of the democratic citizen.

Again, the virtue of public reason is distinct from the moral duty of civility. The moral duty to adhere to public reason does not apply to the public sphere as it would undermine the freedom of the individual to pursue the good life, as defined by our personal comprehensive religious, ideological, or philosophical doctrines. In a democratic society, the public sphere is not subject to either legally or morally defined constraints on speech; “Nonpublic reason is the reason appropriate to individuals and associations in society” (Rawls 2001, 92).

Yet, the claim that the citizen is morally and legally free to act on nonpublic reasons in the public sphere is fully consistent with the claim that public reason is a virtue of the citizen. There is consequently reason to reject the notion that public reason finds no application in the public sphere. Though it is commonly asserted that adherence to norms of public reason is mandatory only in debates that pertain to the regulation of the basic structure (Peter 2007), this view ignores the possibility of nonmandatory norms of public reason. If public reason is a virtue of the citizen in the public sphere, there is no moral requirement that she adhere to it, although it is desirable that she does. The virtuous citizen accepts the norms of public reason in all encounters in the public sphere because she believes in reciprocity, which is a basic political value of a democratic society. When “firm and widespread,” the virtue of reasonableness “is one of the political and social roots of democracy” (Rawls 1997, 768).19 Reasonable discourse in social media and the public sphere is bound to lend “strength and vigour” to democratic institutions and to the political liberties of the citizen.

Should a democratic society praise or condemn efforts by private corporations to “civilize” the interactions between citizens in social media? The issue is confounded by the fact that the democratic principles at stake pull in opposite directions. As citizens of a democratic society, we have interests both in freedom of speech and in a well-functioning public sphere such that it keeps citizens informed, holds public power into account, and reflects the interests and opinions of the population.

Political liberalism is able to guide our judgments by separating the distinct normative issues embedded in the consequent “paradox of the public sphere.” First, political liberalism grants priority to the political liberties of the citizen and recognizes uncivil speech as protected speech. Yet, basic political liberties should be protected against the state: associations are free to regulate their internal operations for the same reason as citizens are free to regulate themselves. Second, political liberalism affirms the idea of public reason and the moral duty not to engage in uncivil speech in the context of political justification. Yet, the moral duties of the citizen apply exclusively to the public political forum and do not extend to speech in the public sphere generally.

Third, political liberalism affirms the value of public reason in the public sphere and social media. A feature of the democratic citizen is that she accepts the virtue of civility according to which it is desirable to avoid speech that is contrary to reasonable terms of social cooperation. Since civility is a virtue of the democratic citizen, incivility in social media and other domains of the public sphere is cause for concern. It does not follow, however, that the state should enforce the virtues of public reason. Respecting the rights of the citizen in a democratic society, the state can only hope that the citizen and her associations—including corporations and social media platforms—come to acknowledge the virtue of public reason in public discourse. The hallmark of a just and free society is that the state recognizes and has faith in the responsibilities of its members and associations. A democratic public sphere should ideally be a place where the “intangible hand flourishes” (Pettit 1999, 257). Thus, democrats should salute the individual citizen who shows restraint in the public sphere and who freely accepts reasonableness as a standard for public reasoning; democrats should for the same reason welcome and encourage social media platforms that constrain uncivil speech and that incentivize reasonable dialogue in the social media.