Richard Spencer, a prominent representative of the ‘Alt-Right’, was punched in the face while giving an interview on the day of Donald Trump’s inauguration as president. The assailant was affiliated with ‘Antifa’, an activist network committed to combatting the rise of far-right movements and ideology (Lennard 2019). Antifa adopt a range of confrontational, coercive and sometimes violent tactics to block the dissemination of far-right discourse in the public sphere (Bray 2017). They are thus emblematic of a global wave of activists adopting a tactical repertoire that often cannot be subsumed under the category of nonviolent civil disobedience. This includes, in addition to Antifa, the Hong Kong anti-extradition law movement, France’s gilets jaunes protesters, and unrest triggered by police violence against persons of colour across a range of cities in the USA (Delmas 2018: 65).

There is, to be sure, nothing particularly novel about social movements that explicitly or implicitly adopt a ‘diversity of tactics’ in their struggles. Of more note is a subtle shift in normative political theory, where ongoing reflection on civil disobedience has been augmented—and in some cases challenged—by a surge of interest in ‘uncivil disobedience’ (Scheuerman 2019).1 The basic thought is that the nature and severity of injustice in liberal democratic contexts means that there is often no principled rationale for actors to limit their resistance to the constrained forms of conduct associated with conventional understandings of civil disobedience (Adams 2018; Aitchison 2018; Brennan 2019; Delmas 2018; Hooker 2016; Lai 2019; Moraro 2018; Pasternak 2018; Smith 2018).2 There are sometimes pragmatic or strategic reasons to accept such constraints, but these are seen as defeasible and certainly not decisive in all contexts (Meckfessel 2016). The upshot is that a range of uncivil actions—including constrained forms of violence—remain on the table.

Deliberative democrats have long since dispelled the notion that their theories of democracy leave no scope for disruptive activism and oppositional protest (Estlund 2001; Fung 2005; Mansbridge et al. 2012; Mendonça and Ercan 2015; Smith 2013). This, as I document elsewhere, has been part of a broader rethinking of the ethics of deliberative citizenship, such that it is increasingly seen as permissible and perhaps even desirable to engage in forms of action that depart from norms of respectful and inclusive dialogue (Smith 2019). This includes disruptive forms of protest as a response to inequalities, exclusions and other problems that blight democratic politics. The rise of philosophical interest in uncivil disobedience, alongside its ongoing proliferation across activist campaigns, can nonetheless be taken as a catalyst for deliberative democrats to reconsider their approaches to protest. Deliberative theorists in fact seldom attend to the diverse tactical repertoire of social movements, instead tending to situate their actions within the catch-all category of ‘non-deliberative’ behaviour (Felicetti 2018). The contrast between civil and uncivil disobedience shows how such a broad category elides important distinctions, treating as analogous forms of conduct that are quite different in their potential consequences and ethical complexion. This is arguably reflected in the failure of most deliberative accounts to address protest and activism that incorporates violence, an important thematic in many discussions of uncivil resistance. The following brief comments draw on that emerging literature to explore the issue of violence, challenging deliberative democrats to consider how they should respond to citizens and movements that utilise such tactics.

Candice Delmas argues that ‘acts of principled disobedience that are covert, evasive, anonymous, violent, or deliberately offensive are generally (though not necessarily) uncivil’ (Delmas 2018: 17). This definition illustrates how uncivil disobedience is typically defined in opposition to the more familiar category of civil disobedience, serving as a descriptor for modes of protest or activism that implicitly or explicitly reject the behavioural constraints that are constitutive of the latter (Brownlee 2012).3 It is, to be sure, a rather broad category, ranging from symbolic acts of guerrilla performance art (Delmas 2018: 43) to defensive use of lethal force against police brutality (Brennan 2019: 54). It is also contested, in the sense that the distinction between civility and incivility is the subject of intense disagreement amongst theorists (Scheuerman 2019).4 The distinction can also be highly controversial in the public sphere, as describing a campaign or movement as ‘uncivil’ rather than ‘civil’ can often be an attempt to delegitimise or discredit it.

The contrast between civil and uncivil disobedience nonetheless represents an attempt by political theorists to take seriously the diverse forms of disruptive protest employed in social movement campaigns. Deliberative democrats, by contrast, tend to be rather less sensitive to this diversity, preferring to lump all forms of protest under the nebulous heading of non-deliberative behaviour. This is a far more capacious category than uncivil disobedience, which potentially could include voting, bargaining, advertising, spectating, organising, striking, picketing, boycotting, protesting, rioting, fighting and almost anything else one might think of that does not involve participating in a mini-public or an unusually reflective legislative debate.

The lack of nuance in the category of non-deliberative behaviour can, on the one hand, result in discussions that overstate the distance between certain protests and the deliberative ideal. Deliberative theorists, for instance, have discussed disruptive actions carried out by Greenpeace (Mansbridge et al. 2012: 19) and Harvard students campaigning for a living wage (Fung 2005: 409–411). These protests might depart from strict norms of deliberative interaction, but the challenge they pose to deliberative democracy is nonetheless rather modest. These are paradigmatic cases of civil disobedience, involving citizens that abide by norms of publicity and nonviolence. Of particular importance is that they are carried out by media-savvy activists as a means of communicating oppositional arguments to multiple audiences. The protesters facilitate deliberation both by triggering reflection on the part of publics and leaving open the possibility for dialogue with their opponents (Sabl 2001). It is thus unsurprising that such ‘non-deliberative’ acts are ultimately assimilated quite easily within deliberative democracy, in keeping with Habermas’s earlier defence of civil disobedience as part of the anarchic ‘communicative processes of the public sphere’ (Habermas 1996: 383).

The capacious category of non-deliberative behaviour can, on the other hand, fail to capture difficulties posed by other kinds of protest. Deliberative systems theorists, for instance, appear to evaluate protest through a broadly consequentialist frame, including ‘coercive’ action that ‘explicitly or implicitly threatens sanctions and imposes costs’ (Mansbridge et al. 2012: 18). The resulting ‘trade-offs’ in the system might nonetheless be worthwhile if such actions ‘force the inclusion of marginalized voices or force new reasons, facts, and information into public conversation’ (Mansbridge et al. 2012: 19).

This is a plausible suggestion, but it is striking that systems theorists offer little analysis of the various forms that coercive action can take (Moraro 2013). The costs imposed on an organisation by a nonviolent sit-down protest, for instance, are quite different in kind to those imposed on its employees by targeted campaigns of harassment and intimidation (Humphrey and Stears 2006). The framework nonetheless appears to appraise such actions solely in terms of their contribution to systemic functionality, with no apparent acknowledgement that some kinds of cost levying are more problematic than others (Owen and Smith 2015: 223). The application of categories that are more discriminating than non-deliberative behaviour would, at the very least, better equip deliberative democrats to engage these concerns, cultivating an awareness of the diverse ways in which offensive, confrontational or disruptive actions impact upon the agency and integrity of actors as well as the broader systemic environment. Though it is far from perfect, the emerging philosophical literature on uncivil disobedience might prompt deliberative democrats to address this challenge.

A particularly provocative theme in that literature is the use of violence, including (a) vigorous physical attacks against persons, (b) caustic, overpowering, vigorous psychological attacks against persons, or (c) vigorous, malicious or incendiary destruction or damage of property (Audi 2009: 143). Violence is often seen as a paradigmatic case of uncivil disobedience, with nonviolence presented as a ‘sine qua non’ for civil disobedience ‘especially in the context of more-or-less functioning democracies’ (Scheuerman 2018: 137).

The use of violence has not been explicitly addressed in the major statements about protest offered by deliberative theorists, including Archon Fung’s insightful and influential analysis of ‘deliberation before the revolution’. Fung develops an ethic of ‘deliberative activism’ according to which non-deliberative tactics can be adopted in line with levels of background inequality and the extent of an opponent’s unwillingness to deliberate (Fung 2005: 405). Fung discusses certain forms of non-dialogic interaction and coercive protest, but he does not address what role—if any—violence might play in these actions. The closest he comes to addressing this issue is to say that, when confronted with ‘incorrigible hostility’ on the part of adversaries, deliberative activists are ‘at liberty to engage with them using the full array of bargaining, negotiating, purchasing, protesting, and more militant confrontational tactics that are regulated by the ordinary nondeliberative political ethics of pluralist regimes’ (Fung 2005: 412). This either permits or prohibits violent resistance depending on how we understand ‘the ordinary nondeliberative political ethics of pluralist regimes’. As critics point out, it is somewhat surprising that this sophisticated analysis ends up ‘leaving activists almost without guidance’ in circumstances that are hardly likely to be ‘an unfortunate rarity’ in existing democracies (Medearis 2015: 42).

Deliberative democracy may not require a comprehensive theory of violent protest, but it would be helpful to say a little more about its standing in activist campaigns. The first reason is that the failure to address violence diminishes deliberative democracy as a critical framework, as it limits its normative appraisal of social movement tactics. The second reason is that general theoretical discussion of uncivil resistance is diminished if deliberative democracy—one of the major paradigms in contemporary democratic theory—remains silent on this issue.5 A range of moral and strategic considerations are relevant to appraising the case for violence, but deliberative democracy should be particularly well-suited to appraising its communicative dimensions. Civil disobedience, as noted above, is often seen as a means of communicating oppositional arguments, with the constraints on its conduct defended on the grounds that it facilitates rather than frustrates transmission across the public sphere (Smith 2013). The claim that uncivil forms of disobedience—particularly those that involve violence—have similar communicative dynamics is more controversial but not uncommon.

Nathan Adams discusses a number of cases where property destruction is a crucial constitutive dimension of the message conveyed through protest, including flag burning, attacking police cars or destroying symbols of vast corporate and private wealth (Adams 2018). Ten-Herng Lai argues in a similar vein that uncivil disobedience involving modest forms of violence such as defacing or damaging statues and monuments can sometimes be communicative (Lai 2019: 103). Avia Pasternak goes so far as to offer a conditional defence of political riots carried out by socially disadvantaged groups in contexts of pervasive and systemic injustices. These acts, she suggests, should be treated as ‘communicative episodes’, during which participants convey a range of messages to police, government, and democratic publics. This includes ‘anger and condemnation of the injustice, a demand for a change of public policy, and a message of defiance of the legal order’ (Pasternak 2018: 391). The force and urgency of the message is, again, intimately related to its articulation through violent and destructive means, and would be severely diluted if participants relied instead upon conventional avenues or civil tactics. These options are, in any case, typically not available to victims of the systemic injustices that trigger political riots, who often lack the resources necessary to mobilise through organisations or make their voices heard (Pasternak 2018: 401–402).

The case for violence on these accounts does not turn solely on its instrumental role in triggering public debate or initiating political reforms. Violent acts are instead evaluated across multiple dimensions, including their impact on public debate and decisions but also, for example, their contributions to enhancing self-respect or their impacts on third parties. It is almost always the case that theorists posit limits or constraints on the use of violence in protest or resistance. Delmas, for instance, insists that all forms of uncivil disobedience, including those that involve violence, should be carried out ‘with respect for other people’s interests, including, but not limited to, their basic interest in life and bodily integrity…and their interest in protection by a stable, secure system of rights’ (Delmas 2018: 29). Adams distinguishes between ‘other-directed constitutive property violence’ and ‘other-directed constitutive personal violence’, suggesting that only the former is compatible with a commitment to treating persons as co-members of a shared political project (Adams 2018: 489). Even Pasternak’s defence of political rioting incorporates a range of moral constraints on permissible conduct by rioters. Of particular interest to deliberative theorists is that Pasternak elaborates these constraints, at least in part, through considering their relation to the communicative aspirations of riots. She thus suggests that ‘in the familiar political circumstances in which rioters typically operate, lethal harm and large scale serious bodily harm is likely to damage the riots’ success prospect, and that justified rioters should therefore moderate their clashes with the police accordingly and avoid such harms’ (Pasternak 2018: 408).

This case for restraint rests on the likelihood that certain forms of conduct will trigger fierce condemnation, which may in turn drown out the political grievances that inspire the riot. It is, of course, true that riots of any kind normally receive censure, but Pasternak is surely correct to suggest that particularly divisive acts—such as killing police officers or trashing schools and hospitals—would magnify such criticism and make it much harder for sympathetic parties to frame the grievances of rioters in a positive light. She goes on to note that social scientific studies of riots reveal that participants often do avoid such acts, using force in a manner that is discriminating and selective (Pasternak 2018: 414–415).

I have set out elsewhere a similar argument in relation to direct action, suggesting that activist tactics should be constrained in light of an ethic of responsibility toward the broader community of opinion that activists claim to represent (Smith 2018). An implication of this is that activists should moderate their use of coercive and violent tactics, while taking steps to ensure that liable or culpable agents bear the costs associated with them as far as possible.6 The failure to discriminate in this way, I suggest, may generate resentment among affected publics and create further opportunities for opponents to discredit the activists’ cause (Smith 2018: 18–19).

There remain, to be sure, important differences in the communicative dynamics that are involved in paradigmatic cases of civil and uncivil disobedience. The former, to a much greater extent than the latter, embody a commitment to treating opponents as potential discursive interlocutors. The positions staked out by the likes of Pasternak and I nonetheless show how activists can utilise violence in a way that is sensitive to its reception in the public sphere, tailoring their action to ensure that its communicative intent is more likely to be recognised.

The claim that violence has a place in the tactical repertoire of activists is one that might prompt considerable unease among deliberative democrats. This is a reasonable reaction, which reflects both general anxieties about violence and scepticism about its capacity to enhance the epistemic or inclusive quality of public debate. It is, of course, also the case that many instances of uncivil resistance are carried out by actors with little or no interest in promoting deliberation. The punching of Richard Spencer, for instance, was criticised by many because it appeared to be an ethically indefensible and politically self-defeating stunt designed to intimidate and silence an opponent (Etzioni 2017). Deliberative democrats might have considerable sympathy with this criticism; after all, as John Parkinson notes, ‘at the end of the day the deliberative movement, if we can call it that, is in large part about replacing power plays and political tantrums with “the mild voice of reason”’ (Parkinson 2012: 158).

There are, I suggest, two avenues for further research and reflection in this area. First, there is scope for empirical investigation into violence as a potential catalyst for deliberation at macro or micro levels. It should, for instance, be possible to identify and analyse case studies involving the use of violence by protesters or activists, exploring whether and to what extent such actions contribute to or detract from the promotion of values such as inclusion, epistemic reflection, empowerment, self-respect and even mutual respect. This research could also explore the extent to which the impact of violence on deliberation is dependent upon the way in which other actors in the public sphere or deliberative system respond to these acts. A core motivation behind Pasternak’s defence of political rioting is to contest the way in which politicians and media outlets frame these events as criminal outbreaks warranting blanket condemnation. This condemnation is problematic because it—perhaps intentionally—functions to divert attention from the desperate political appeal that is often implicit in such events (Pasternak 2018: 385). This point can be generalised to cover any number of cases where violence or incivility is prematurely taken as a basis for refusing to recognise actors as potential discursive interlocutors. Deliberative democrats could make a valuable contribution by setting out a principled and pragmatic case for systemic responses to violent acts that have the effect of amplifying rather than muffling their communicative content.

Second, there is an acute need for further normative reflection on the complex ethical issues that are implicated by uncivil resistance. It was suggested above that non-deliberative acts should not be evaluated solely in terms of their impact on deliberation. A coercive or violent campaign might, for instance, warrant reasonable criticism even if it realises the instrumental value of promoting deliberation across a system. The flip side is that a campaign might turn out to be defensible even if it has no impact on promoting deliberation or perhaps even reduces prospects for deliberation in the short or long term. Antifa, as noted above, is committed to preventing the normalisation of far-right discourse through a range of nonviolent and violent means. In deliberative terms, they reject any ‘meta-consensus’ that includes white supremacist ideologies as legitimate participants in the ‘contestation of discourses’ (Dryzek 2010: 112–113).7 Antifa and their supporters do not want to promote deliberation on Alt-Right claims about race and gender, contending instead that there is a moral duty not to deliberate with persons who advance these ideas (Yaure 2018: 14–15). They also argue that media-savvy Alt-Right spokespeople like Spencer are far more comfortable responding to critical questions from liberal journalists than clenched fists from Antifa activists (Lennard 2019: 13). Their general claim is that a multifaceted defence of democracy against fascist and authoritarian threats is necessary, which might include tactics that are incompatible with liberal, deliberative and perhaps even democratic values on some interpretations (Bray 2017: 148–159).8

This claim may turn out to be unsustainable as a defence of physical assault. Delmas, for instance, argues that it is often better for Antifa to pursue their aims through nonviolent, or less violent, confrontational tactics. As she puts it, ‘throwing flour and eggs—or, as PETA activists do tofu cream pie—is preferable to throwing punches, and it can work as well to impose social risks on neo-Nazis and dissuade them from publicly defending their views’ (Delmas 2018: 245). The disavowal of physical assault might also make sense on strategic grounds, as Alt-Right figures often aim to provoke violence as a means of casting themselves as victims and their opponents as aggressors.9 The literature on uncivil resistance, discussed throughout this paper, thus proposes an open-minded but critical orientation toward the use of force by activists. This is broadly the orientation that should inform study of uncivil resistance by deliberative democrats. It is not necessary to set aside well-founded concerns about violence, but it would strengthen deliberative theory if its intuitions on this subject were subject to greater empirical and normative analysis. This analysis must, at a minimum, be underpinned by a more nuanced understanding of activist tactics and protest repertoires, which are not adequately captured through the capacious and nebulous category of non-deliberative action.