The British Academy Brian Barry Prize Essay Why Is There No Just Riot Theory?

Abstract Given the prevalence of riots throughout human history, the lack of normative theorizing about them compared to other forms of political violence is striking. The author hypothesizes that this is due to riots' extra-institutionality. Riots are extra-public because they involve the participation of crowds, rather than institutionalized groups such as parties or social movements. They are extra-state because they violate the state's monopoly on violence. Riots are extra-legal because they constitute a form of unlawful assembly. They are also extra-parliamentary because they operate outside the normal legislative process. This article considers justifiable reasons to resist each of these foundational institutions, and proposes provisional criteria for a justifiable riot. The author concludes by urging political theorists to further examine the normative dimension of riots.

and co-creator of HBO's The Wire, David Simon issued a statement as the riots were ongoing that read in part 'If you can't seek redress and demand reform without a brick in your hand, you risk losing this moment for all of us in Baltimore. Turn Around. Go home. Please' (Taintor 2015). President Obama expressed similar sentiments when he said, 'One burning building will be looped on television over and over and over again, and thousands of demonstrators who did it the right way, I think will be lost in the discussion' (Davis and Apuzzo 2015). Both of these statements share a common concern that riots run the risk of undermining political progress that is more effectively achieved through non-violent means.
More radical thinkers have been quick to push back against condemning riots as ineffective. For instance, George Ciccariello-Maher refuted the claim that rioting only encouraged a political backlash (2015). Ta-Nehisi Coates argued that calls for the Baltimore rioters to protest peacefully in the face of violent repression by the Baltimore Police Department were hypocritical (2015). Hip hop artist Killer Mike defended the anti-police brutality rioters by arguing that riots were an effective tool for bringing about political changea view that was echoed in subsequent political analysis (Friedersdorf 2015a;Friedersdorf 2015b;Kreps 2015;Lopez 2006). These commentators argue that while riots may not be 'wise' or 'correct' (Coates 2015), they were a last resort after community members had tried and failed to gain redress for their grievances using nonviolent means. In every instance of recent anti-police rioting in the United States, the political authorities responded to some of the rioters' key demands. 1 Yet these defenders focus on riots' effectiveness rather than its legitimacy. They do not defend rioting on normative grounds.
The lack of normative theorizing in academic scholarship about riots is even more striking. There is an extensive literature in history, sociology, and empirical political science about riots and rioting. Historians have a long tradition of interpreting the changing meaning and significance of riots over time and across cultures (Rudé 2005;Thompson 1971). Political scientists and sociologists have also explored the causes of riots as well as policies to minimize and prevent them (Graham and Gurr 1979;Tilly 1976;Tilly 1983;Wilkinson 2009). Yet there is very little normative scholarship about riots. 2 There are, however, significant bodies of literature about other forms of political violence and militant resistance. For example, a long tradition in political theory, stretching back to the early modern era, considers the justifications for revolutions and resisting rulers (Finlay 2015;for example, Locke 1988;Marx and Engels 2015). Political theorists have also written normative defenses of terrorism (for example , Held 1991;Nielsen 1981). The political obligation literature spells out the conditions under which citizens may resist state authority (Delmas 2014;Klosko 2005;for example, Pitkin 1965). The just war tradition explores both when a war is justified and what is appropriate conduct in a war (for example , Coady 2008;Walzer 2006). 3 Scholars of civil disobedience have 1 This includes a US Justice Department investigation of systematic police abuse by the Ferguson Police Department, the indictment of the six police officers involved in the death of Freddie Gray and a Justice Department report on the Baltimore Police Department. 2 A notable exception is D'Arcy (2014), who I discuss below. Celikates (2016, 44) acknowledges in a footnote that riots might be politically legitimate. Shelby notes in passing that riots might 'be necessary for the ghetto poor to maintain their self-respect ' (2007, 156). Hooker argues that riots are 'an inadequate but necessary form of democratic repair ' (2016, 449). Yet none of these theorists develops a systematic theory of a just riot. After this article was accepted for publication, I was alerted to Pasternak's (2018) excellent article that develops a normative defense of rioting using criteria of permissible harm. Social theorists such as Clover (2016), Badiou (2012) and Balibar (2007) focus on the meaning and significance of recent riots rather than normative defenses of them. 3 Consider, for example, two leading undergraduate textbooks on political theory, both of which were published by the same press. One focuses primarily on domestic political theory; it has an image of a protestor confronting police on the cover, but has no chapters discussing legitimate uses of violence in domestic politics. It does, however, have a chapter on political obligation (with a brief mention of anarchy as a political doctrine) and a chapter on just war theory. See McKinnon (2014). Conversely, the textbook on normative international politics has two chapters on the ethics of war, one on the ethics of humanitarian intervention and one on the ethics of terrorism. See Bell (2010). Neither textbook covers the ethics of political protest. elaborated on the conditions under which citizens may intentionally and publicly break the law (for example , Celikates 2016;Markovits 2005;Morreall 1976;Scheuerman 2015;Zinn 2014).
While riots are similar to these other political phenomena, they are distinct in some crucial ways. First, riots are worthy of attention in their own right. Secondly, riots are not the same as revolutions: revolutions seek to replace the entire system of government with a new system, whereas riots are very localized protests of specific grievances. Thirdly, riots are not the same as civil disobedience or conscientious objection; in these cases, the law breaker intends to be convicted as part of the protest, whereas rioters often attempt to avoid arrest.
Only radical and Marxist scholars take riots seriously (Clover 2016;Fanon 1963;Lenin 1975;Sorel 2002). Yet even they do not develop a normative defense of rioting. They usually characterize normative questions as bourgeois moralism, and instead focus on riots' ability to bring about revolutionary change. The lack of normative theorizing about riots is doubly surprising given the analytical focus on riots in more empirically oriented fields, and the amount of attention devoted to other forms of political protest, violence and resistance in political theory.
This essay explores why there is no just riot theory tradition in Western political theory. I argue it is because riots are extra-institutional in four ways (1) they are extra-public because rioting crowds self-organize (they are not formally institutionalized groups such as parties or social movements); (2) they are extra-state because rioting disrupts the state's monopoly on violence; (3) they are extra-legal because they involve breaking laws concerning public assembly; and (4) they are extra-parliamentary because rioters express their grievances outside of normal political processes. For instance, Britain's Public Order Act defines a riot as follows: Where 12 or more persons who are present together use or threaten unlawful violence for a common purpose and the conduct of them (taken together) is such as would cause a person of reasonable firmness present at the scene to fear for his personal safety, each of the persons using unlawful violence for the common purpose is guilty of riot (1986 emphasis added).
Two key parts of this definitionthe presence of twelve or more persons and the use or threat of unlawful violencehighlight two of the four essential features of a riot outlined above: riots are crowd actions and involve violence. The first two sections will consider these two forms of extrainstitutionality in turn: they explore what could justify political action outside of normal democratic procedures and what could justify the use of violence. Yet even if both of these aspects of riots' extra-instituionality are justified, there are still crucial questions about their extra-legality and tendency to operate outside of normal parliamentary procedures (the third and fourth characteristics described above). These questions will be addressed in the third and fourth sections. In the conclusion I argue that because each of the four elements of a riot that is normally deemed illegitimate has well-theorized instances in which legitimate exceptions are made, there is no reason in principle that a theory of a just riot is not possible. The four types of extra-institutionality thus suggest possible criteria that could justify a riot.
The Crowd as an Extra-Public Actor The crowd has an ambiguous place in the history of political theory. McLelland observes that '[i]t could almost be said that political theorizing was invented to show that democracy, the rule of men by themselves, necessarily turns into mob rule ' (1989, 1). Political theory's bias against the crowd revolves around two main arguments. First, the elitism of most canonical political thinkers made them suspicious of crowds for lacking the expertise and leadership for effective political action. Secondly, more populist theorists suspected crowds of being too disorganized to be capable of sustained political action.
An early example of the first tendency is Plato's critique of democracy in The Republic. He worried that the people were not temperamentally suited to govern, and that demagogues would manipulate the masses into establishing a tyranny (1992, 565 d-e). Hobbes expressed a different concern. He drew a sharp distinction between the people and the crowd. 'A people is a single entity with a single will; you can attribute an act to it. None of this can be said of a crowd ' (1998, 137). According to Hobbes, the crowd only exists in a state of nature. Only a crowd can carry out a rebellion. When a rebellion takes place, the individuals participating cease to be people, and turn into a crowd; the crowd returns the people to a state of nature (1998,76). Whether the crowd leads to tyranny (as with Plato) or to anarchy (as Hobbes maintains), both lines of thinking agree that the crowd is unruly and prone to facilitating dangerous political outcomes. Even radical pluralist thinkers such as Arendt endorse this interpretation of crowds. In her analysis of the Dreyfus Affair, she distinguishes between the moba proto-fascist, extra-institutional mass movement prone to anti-Semitism and violenceand the peoplea deliberative public, capable of resisting mob rule through political action (Arendt 1973, 106-120). This historical distinction between demos, multitude, crowd and mob, on the one hand, and 'the people' on the other hand shares the belief that an institution must mediate between the masses and the government. Without a mediating institution, prior political theory maintains that the masses will either prop up a tyrant or promote anarchy.
More radical and populist theorists worry that the crowd is too disorganized to be capable of effective political action. For example, Lenin argued that riots constituted the proletariat's spontaneous resistance to capitalism (1975,36). He chastised socialists who celebrated spontaneous acts by the working class because 'the spontaneous development of the working-class movement leads to its becoming subordinated to the bourgeois ideology ' (1975, 49). According to Lenin, the crowd was incapable of accomplishing anything without the institution of the party.
In reflecting on her experiences in the Occupy Wall Street movement, Dean makes a similar argument. Left-wing social movements since the 1960s, such as Occupy, reveal a split between 'mob or people'. Dean argues that '[t]he individualism of [Occupy's] democratic, anarchist, and horizontalist ideological currents undermined the collective power the movement was building ' (2016, 4). She asserts that the collective will of the crowd can only be transformed into political power through a political party (2016, 28). These thinkers argue that institutions such as the party turn the disorganized crowd into a people, and transform the crowd's grievances into clearly articulated demands that can lead to political change.
While many theorists characterize the tumult generated by crowds as creating division in the polity, a counter-tradition celebrates disorder as a means of preserving freedom. In The Discourses, Machiavelli observes 'those who damn the tumults between the nobles and the plebs blame those things that were the first cause of keeping Rome free, and that they consider the noises and the cries that would arise in tumults more than the good effects they engendered' (1996, II.4.2 (p. 16)). His point is that tumultsmass popular disturbances including riots 4serve as an extra-institutional check on elites' power and ambition that is needed to preserve the freedom of the republic (1996, I. 4 p. 17).
According to Canetti, the crowd creates perfect equality through processes of de-individuation. Many classic crowd theorists describe the crowd as dangerous because it unleashes 'animalistic' and 'primitive' behaviors that are normally repressed by society (Freud 1793;LeBon 1896;McClelland 1989, 248-249). In Crowds and Power, Canetti upends this traditional understanding of the crowd in several ways. While not denying its potential for violence, or its vulnerability to manipulation by a leader, he develops a theory of the crowd as fundamentally egalitarian and as the first source of justice. Canetti argues that '[a]ll demands for justice and all theories of equality 4 The Mansfield and Tarcov edition of The Discourses cites Livy, Book II, 23-24 and 27-33 as the passages Machiavelli is commenting on. In II. 23-24 Livy recounts a riot started by former Roman soldiers who had been enslaved because of their debts. The result of the riot was that the Senate was pressured to convene and consider the crowd's grievance. Subsequently the consul issued an edict making enslaving a Roman citizen illegal, and prohibited the seizure of a soldier's land while he was on military service. Rioting led to the development of basic protections for Roman citizens. ultimately derive their energy from the actual experience of equality familiar to anyone who has been a part of the crowd' (1984,29). Crowds, in Canetti's theory, are ontologically distinct from individuals. When an individual joins a crowd, they lose themself and experience a sense of de-individuation. This de-individuation creates a radical equality and erases the numerous hierarchies through which societies maintain order (1984,18). 5 Canetti develops an elaborate typology of crowds to push back against the tradition in Western political thought that views the crowd as always prone to demagoguery and violence. His typology generates 280 distinct types of crowds (1984, 29-63;McClelland 1989, 302). While Canetti does not directly discuss rioting, his theory identifies three virtuous traits of the crowd: its state of pure equality, its ability to issue demands outside of formal institutions and its temporary undermining of social hierarchies. 6 His nuanced account of the crowd recognizes both its potential for destruction and its capacity for justice.
Finally, in contrast to many who argue that there can be no moral order in crowds (Arendt 1973, 106;LeBon 1896), several defenders of crowds have noted the orderliness of riots (Hobsbawm 2017, 6;Marx 1970, 27-28;Rudé 2005, 49-51;Thompson 1971, 77-79;Tilly 1983;Wilkinson 2009, 331-336). Empirical studies of riots note the tendency of crowds to selfpolice. Rioting crowds tend to act in a manner that is commensurate with the grievance that triggered the event. The crowd's activity is neither 'capricious nor random' (Gilje 1999, 7). 7 In eighteenth century England, crowds frequently protested against price gouging by 'raising a mob' to visit the local farms and estates of the wealthy and demand that grain be sold to the poor at a reasonable price. Thompson notes how orderly such a crowd was in both the articulation of its grievance and its willingness to pay for the grain, so long as it was at a price the poor could afford (1971,(107)(108)(109)(110)(111)(112)(113)(114)(115). Workers, prior to the recognition of collective bargaining rights, would use mass protest to increase wages and protest high prices. Crowds of 'Luddites' destroyed machinery to protect their jobs (Hobsbawm 2017(Hobsbawm , 6, 1952. Historians of crowd behavior have long demonstrated that crowds have their own 'moral economy' (Thompson 1971) whereby mass protest is motivated by a norm violation by the authorities. Crowds tend to behave well in protests, limiting their actions to the target of their grievances, and are generally proportional in their responses to those grievances.
The conventional reason for rejecting the crowd as a legitimate political actor is that it is irrational and does not operate through rational and deliberative institutions. Throughout the history of political theory (from Plato to Arendt) and across the ideological spectrum (from Hobbes to Lenin) there has been a strong fear of the dangers of mob rule. The crowd either promotes anarchy or tyranny, or it is too disorganized to rule effectively. When we evaluate the countertraditions that defend crowd action, however, we find four criteria by which we can assess whether a rioting crowd is behaving justly or unjustly. First, following Machiavelli, one key question to ask is Are the crowd's actions freedom preserving? Secondly, following Canetti's defense of the crowd as a field of equality, a key question to ask is Do the crowd's actions promote equality? Thirdly, in the tradition of historians and sociologists of crowd behavior such as Hobsbawm, Rudé, Tilly and Gary Marx, a key question to ask is Do the crowd's actions give voice to the grievances of marginalized groups? Fourthly, following Thompson, a key question to ask is Are the crowd's actions orderly and self-policing? If an assessment of a crowd returns an answer of 'yes' to all four questions, it is not promoting tyranny (a key concern of elitist crowd critics), and it is certainly acting in a concerted way to bring about a political end (a key concern of populist crowd critics). While these considerations alone are not sufficient to justify a riot, they are necessary criteria to justify a rioting crowd's actions. 5 He also notes that sports crowds and religious gatherings, in which a crowd is enclosed within a church or sports arena, are examples of domesticating the crowd. See Canetti (1984, 25). Contra most political theorists who see such institutionalization as a positive means of controlling the crowd, Canetti laments it as a loss of the crowd's freedom, equality and power. 6 He does discuss how crowds are prone to war and violence. See Canetti (1984, 19). But he also sees crowds as essential to social change through mass protest (which he classifies as a reversal crowd) See Canetti (1984, 58). 7 For example, attacks during the 1992 riots in Los Angeles 'had an eerie resonance reflecting the image of white police beating of Rodney King'. See (Gilje 1999, 7).

How Does a Riot Violate the State's Monopoly on Violence?
In practice, the violence 8 in riots tends to take three forms: (1) physical attacks on other people, (2) vandalism of property (both public and private) and (3) looting, which generally takes the form of rioters stealing goods from stores. This violence immediately raises two questions. First, why (and when) is violence bad? Secondly, what could possibly justify these acts of violence by rioters?
Because riots are violent political acts, they have an uneasy relationship with Western political theory tradition's normal understanding of violence. Frazer and Hutchings (2007;2009;2011a;2011b) observe that the dominant thinkers in the Western tradition perceive violence as related to politics instrumentallyeither as a means to achieve political ends (Clausewitz 1968;Machiavelli 1988;Weber 2004) or as antithetical to politics (Arendt 1970;Rawls 1999). They also identify a counter-tradition that treats violence as creative and expressive. These theorists view violence's role in politics as either expressive or constitutive of character and community (Benjamin 1978;Fanon 1963;Sorel 2002). 9 If we consider both the dominant and counternarrative positions on the relationship of violence to politics, the idea of a just riot does not fit well into any of these categories. This is most obvious in Arendt's position, which maintains that the riot, because it is violent, is difficult to justify and impossible to legitimate (1970,52). 10 In the more conventional reading of violence as a means to achieve specific ends, a crucial issue is who may legitimately wield the instruments of violence? Normally the state reserves that privilege for its agents, due to the fear that if just anyone is allowed to use violence to achieve their political ends, then the state's monopoly on violence breaks down. From this perspective, the state must keep the riot beyond the pale in order to maintain order. Conversely, the counternarrative tradition of violence and politics eschews questions of justification altogether. In this reading, the riot might be expressing the grievances of the rioter, or perhaps even constitute a new political entity (Canetti 1984) or be useful for fashioning a new sense of self (Fanon 1963). However, these political theorists do not consider the riot through the framework of justification.
Critics of riots point to acts of violence committed by rioters as justification for condemning the riot. Could anything redeem this violence? The social contract tradition suggests that two criteriawhich I call the grievance criteriajustify using violent means to resist or overthrow the government. Since rioting is less of a threat to the state's authority than an armed insurrection or revolution, anything that would justify those more significant transgressions of state authority would also justify rioting. In order for the violence of the riot to be justified, the rioters must be motivated by a significant enough grievance that it justifies their use of violent protest.
The first grievance criterion follows from the Lockean tradition of revolution. The revolution and resistance tradition recognizes that if a government violates the rights or welfare of its citizens, then the people have a right to resist and replace it with a new one (Locke 1988, 225). 11 8 I use the term violence in the narrow sense of direct violence, which is often defined as 'the use of physical force to inflict injury or to cause harm to a person or property'. See Thomas (2011Thomas ( , 1828. 9 By Frazer and Hutchings' own admission, this schema is very much comprised of a set of ideal types meant to draw out the different ways in which political thinkers in the Western tradition relate violence and politics. One of the main claims in their project is that many of the thinkers associated with the instrumental reading of violencesuch as Machiavelli, Weber, Clausewitzalso have significant features of the creative/excessive narrative of violence, but that this reading has been suppressed by later interpreters who tend to be bound by the ends-means way of theorizing how violence and politics relate (DPIR Westminster University 2014). 10 Arendt observed that 'Violence can be justifiable, but it will never be legitimate. Its justification loses in plausibility the farther its intended end recedes into the future' (1970,52). She believed self-defense was justifiable because the end was immediate, yet she doubted that the inner city rioting in the 1960s United States was close enough to the political ends of abolishing racial inequality to be justifiable. 11 My focus is here is primarily on the Lockean-liberal revolutionary tradition in which the justification for revolution focuses on a government losing its legitimacy to govern if it systematically violates its citizens' rights to life, liberty or property (Locke 1988, 225;Tully 1993, 41-43). Other revolutionary traditions justify their revolutions on different grounds. The More contemporary theorists also recognize a similar right. Rawls identifies a right of militant resistance when conditions under the basic structure are unjust (1999, 323), but does not fully state what would constitute such an injustice. 12 While it is difficult to spell out how severe a violation of the basic structure justifies resistance, one criteria for justifying a riot on these grounds is whether or not the polity systematically violated one of the constitutional triad of democracy, human rights or the rule of law. As these constitute the basic structure of a just liberal society, a polity that protects these basic structures with respect to its citizens creates a coercive relationship between citizens and the state, negating the legitimacy of the state's monopoly on violence.
Not all injustices involve violations of fundamental civil and political rights. Shelby draws upon Rawls' standard of 'intolerable injustice' to defend disobedience to the state. He notes, however, that Rawls never specifies what that limit is (2007,145). Shelby suggests that the standard should be a duty of self-respect, which is fulfilled by affirming one's equal moral worth as a person. When a society systematically violates the self-respect of a portion of its population, such as poor urban blacks in the United States, then there is a legitimate reason to protest and resist injustice. Shelby defines deviance as 'sharply divergent from widely accepted norms ' (2007, 128) and lists crime, refusing to work in legitimate jobs and having contempt for authority as examples of deviance. Gary T. Marx draws a similar distinction between two kinds of deviance nonconforming and aberrant. Nonconforming deviance is 'a thrust towards a new morality' (1970,24) and violates existing norms with an aim 'to replace them with new norms ' (1970, 24). Conversely, aberrant behaviour 'deviates out of expediency and for the momentary gratification of personal ends without seeking social change' (Marx 1970, 24). 13 These non-civil and political justifications for resistance point to a second potential criterion for justifying a riot. Rioting is certainly a deviant behavior, but if the riot involves nonconforming deviance (as, for instance, with the Stonewall riots contesting the anti-LGBT policies of the New York police), or if it expresses contempt for the authority of a social and economic system that does not afford the basic minimum required for self-respect (as in the 1960s inner city riots in the United States), then these types of deviance point towards legitimate grievances. The second grievance criterion that justifies a riot is thus: Does the polity systematically fail to guarantee conditions of reciprocity to its most disadvantaged members?
What kind of violence within a riot could be legitimate? Even a just grievance does not permit just any type of violence in a riot. A riot could be just if it is in response to political authorities' systematic violation of the rights of the rioters, but the rioters' actions would be unjust if the violence was not proportional to the oppression the rioters were confronting, or if it did not target those responsible for the grievance. The three types of violence in a riotviolence against persons, vandalism and lootingall have acceptable and unacceptable targets. If rioters use violence in self-defense, either against police dispersing a just protest or against the targets of their grievances, then this form of violence may be justifiable (cf. Brennan 2016). Conversely, if the rioters target innocent bystanders or use the riot to target individuals who did not act violently against them, then the use of violence cannot be justified. Similar criteria would be applied to vandalism. Attacking police equipment or property in a violent street clash might constitute a legitimate communist tradition justifies revolution as a necessary means of overthrowing the exploitative relationship of capitalism (Lenin 1975). The de-colonial tradition sees revolution as a necessary step in ending imperialism and colonization (Fanon 1963). The anarchist tradition argues that the state itself is an unjust form of coercion, and that revolution is necessary to end its domination over people (Bakunin 1990;Proudhon 2007;Sorel 2002). Since these radical traditions are more open to the legitimacy of rioting as a political tactic, I focus on the liberal tradition in this section as it offers the most difficult case for justifying riots. 12 Rawls (1999, 6) uses the term basic structure of society to mean 'the way in which the major social institutions distribute fundamental rights and duties and determine the division of advantages from social cooperation'. 13 Marx introduces these two kinds of deviance to distinguish between riots that have a clear grievance (such as food riots, industrial riots, prison riots and anti-police violence riots) and those that have no clear grievance, such as riots in victory (sports teams winning championships, for example) and riots during police strikes. target, but targeting bystanders' personal property can never be justified. Finally, looting could be justified if the rioters use it to redress an economic injustice, as in eighteenth-and nineteenthcentury food riots or the survival looting undertaken by the residents of New Orleans after Hurricane Katrina.
The first two forms of extra-institutionality that we have surveyed are essential to all riots. In order for a gathering to be considered a riot, it must be a mass public demonstration and it must be violenteither against property or against people. Even if a riot has a crowd that is promoting freedom and equality, enabling marginalized groups to articulate their demands, and the crowd's behavior is self-policing, and even if the violence is proportional and aimed at legitimate targets, it does not follow that the reason for the riot is justified, or that rioting in this instance is a reasonable tactic.
Not all (or even most) riots are driven by grievances. People riot to celebrate sports victories, or to persecute marginalized groups. People riot when the state's police power is withdrawn (such as during a police strike or breakdown in civil order). In none of these cases is a grievance present. Yet even if there is a legitimate grievance, one must first determine if othernon-violent and normally legitimatemeans of redressing it are possible. The need for a legitimate grievance relates to the riot's extra-legality. The requirement to exhaust legitimate means of redress points to the riot's extra-parliamentarianism. I next consider each of these in turn.

How is a Riot Extra-Legal?
A riot is extra-legal because it breaks the law against rioting. Both the riot and unlawful assembly are forms of public assembly that disturb the peace through noise and violence. Statutes that prohibit disturbing the peace criminalize public assembly. State authorities have long used such statutes to target political protestors across the political spectrum, as they are written and enforced in a way that gives police forces great discretionary power to interpret ongoing and potentially future acts as violent and threatening the public (Inazu 2017, 5-6).
There are three important aspects of the connection between rioting and unlawful assembly. First, a riot is a form of public assembly that turns violent. Secondly, state authorities can use riot laws to restrict legitimate protest. 14 Thirdly, the use of unlawful assembly and riot laws to arrest and charge protestors has a chilling effect on risk-averse citizens. For these reasons, we should not automatically assume that if protestors are charged with rioting or unlawful assembly that their actions are illegal. Local authorities' power to treat a protest as a riot means that state officials may delegitimize many legitimate protests as part of their policing and efforts to control dissent.
Political theorists generally recognize that citizens are obliged to obey the law. Four general traditions uphold this obligation: the voluntarist position, the utilitarian position, the fairness position and the morality of law position (Green 2012;Hyams 2012, 11;Smith 1972, 953). All four traditions acknowledge exceptions to this obligation. Rawls, for instance, argues that 'once society is interpreted as a scheme of cooperation among equals, those injured by serious injustice need not submit' [to the law] (1999,336). Locke maintains that if the government violates the social contract by threatening its citizens' rights to life, liberty or property, then the citizens may legitimately overthrow the government (1988,(412)(413). Bentham argues that when the law no longer maximizes the utility of a country's citizens, then citizens can disobey the law (1948,55). 15 And Raz argues that the government and the law 'is legitimate to varying degrees regarding different people ' (1986, 104) and that 'disobedience to law is sometimes justified' 14 Inazu (2017, 4-6) provides numerous examples of police forces in the United States using unlawful assembly orders to shut down protests on issues ranging from anti-police violence, to anti-war protests, to organized labour protests even in instances of non-violence. 15 Bentham (1948, 55, emphasis added) argues: 'so long as the probable mischiefs of obedience are less than the probable mischiefs of resistance: why, in a word, taking the whole body together, it is their duty to obey, just so long as it is their interest, and no longer'. (1986,101). Every tradition of political obligation recognizes that individuals may disobey unjust laws, and may engage in acts of resistance against unjust regimes. Even Hobbes, the thinker most famously associated with the absolute and unlimited authority of the sovereign, recognizes 'the Liberty to disobey' the sovereign when he or she threatens the individual (1996,151). 16 Although theorists recognize some forms of justifiable law breaking, how the law is broken matters just as much as the fact that it can be broken. Political theorists recognize at least six different types of principled law breaking as having justifiable exceptions to obedience to the law: (1) testing the law, 17 (2) civil disobedience, 18 (3) democratic disobedience, 19 (4) disruptive disobedience, 20 (5) whistleblowing 21 and (6) deviance. 22 There are crucial differences between rioting and most of the above forms of justified law breaking. First, the first four cases involve an individual consciously breaking a specific law because it is unjust: the law itself is immoral. Deviance contests the unjustness of the entire system by breaking laws that are not necessarily unjust in and of themselves. Conversely, the rioters are contesting the legal limits of public protests. They do not generally riot against laws on rioting. When they riot on grievance grounds, they are contesting some other perceived injustice. Rioting may be conscientiously motivated if it is a grievance riot. The first four forms of resistance listed above must be conscientiously motivated for the lawbreaking act to be legitimate. Many forms of rioting are not, but that would be a basis for testing a riot's legitimacy. Deviance is the sole exception to this conscientiousness test, as Shelby tends to frame justified deviance as necessitated by conditions of intolerable injusticethat is, because the system itself is manifestly unjust, those who are disadvantaged by it no longer have an obligation to abide by its laws (2007,155).
Scholars generally recognize justified law breaking if the illegal action is contesting a greater injustice (Celikates 2016, 43;Edyvane andKulenović 2017, 1361;Markovits 2005Markovits , 1898Rawls 1999, 319;Scheuerman 2015;Shelby 2007, 127). Edyvane and Kulenović argue that law breaking is 'justified when it functions to disrupt exclusionary practices that contribute to the incapacitation of citizenship ' (2017, 1360). Shelby uses the criterion of 'intolerable injustice' as a justification for law breaking and defines it as the 'constitutional essentials' in Rawls' basic framework. Shelby argues that in cases where the basic structure is inegalitarian, and the prevalence of an ideology in society is manifestly unjust, then those who are put in a position of intolerable injustice (in Shelby's case, America's 'ghetto poor') have no obligation to abide by the law (2007, 145). The criterion that determines if breaking anti-riot law is justified is whether or not the riot is protesting a fundamentally unjust state action or law. Just as other forms of justified law breaking recognize a right to resist unjust laws, a riot would be justified if it is protesting an unjust law. 16 For an argument that reconciles these two apparently contradictory positions in Hobbes' thought, see Steinberger (2002). 17 A person tests a law by challenging its validity in court through judicial review. Yet unlike other forms of disobedience, in this instance the citizen agrees to abide by the court's decision (Edyvane andKulenović 2017, 1360). 18 Rawls (1999, 320) defines civil disobedience as 'a public, nonviolent, conscientious yet political act contrary to the law usually done with the aim of bringing about a change in the law or policies of the government'. 19 In democratic disobedience, citizens break the law to address deficits in the democratic decision-making process. Whereas liberal accounts of civil disobedience focus on cases that protect rights, Markovits (2005) developed the term democratic disobedience to cover cases such as the Vietnam War protests, anti-nuclear protests, anti-austerity protests and movements like the Weather Underground, which do not centre on protections of liberal rights. 20 Disruptive disobedience is 'covert and nameless civic resistance' that disrupts institutions and practices (Edyvane andKulenović 2017, 1363). Edyvane and Kulenović developed the concept to account for forms of hacktivism, such as the Aaron Swartz leak of 4.8 million research articles from JSTOR. 21 A whistleblower publicly discloses information or actions that are illegal or unethical, often at great personal risk to themselves, their family or their career. For an examination of Edward Snowden's whistleblowing activities as a case of civil disobedience, see Scheuerman (2014). 22 Tommie Shelby defines deviance as 'sharply divergent from widely accepted norms ' (2007, 128). He considers activities by the ghetto poor such as crime, refusing to work in legitimate jobs and having contempt for authority as legitimate forms of deviance in the face of deeply unjust social arrangements (Shelby 2007, 127-128).
The easiest way to determine whether the law in question is unjust is to assess whether it contradicts a more fundamental principle of constitutional law, or the basic law of a society.
Riots can also be made illegal via a public declaration or judgment of the police. While crowds may gather to protest a grievance, it is only when the authorities declare the protest a riot that the crime of rioting occurs. Legally, this post facto constitution through labelling is most explicit in the idea of 'reading the riot act'. This expression is based on Britain's Riot Act of 1714, which empowered local officials to read out a proclamation that ordered any group of twelve or more individuals who were publicly assembled and behaving riotously to disperse within an hour. If any participants failed to disperse, they were guilty of a riot. The British Crown thus invented the concept of rioting as a crime in order to set limits on protest and dissent. The state, through its officers, decides what protests count as a riot; in making this decision, a protest turns into a riot.
Authorities can use both unlawful assembly and riot dispersal orders to control and stop the expression of dissent. Sometimes authorities issue dispersal orders knowing that they will not hold up in court simply as a means of preventing or ending a protest (Inazu 2017, 34). Anti-rioting laws are partially about controlling the basic freedoms of public assembly and speech. Sometimes these rules are either overapplied or applied in bad faith. This points to a second criterion of justified law breaking with respect to riotsare the authorities using riot law to disperse a lawful and peaceful assembly? A very high threshold is needed for authorities to invoke disturbing the peace laws to end a protest. Public safety should only be invoked to disperse a protest when the crowd directly threatens the safety of public bystanders who are not participating in the protest.

How is a Riot Extra-Parliamentary?
Democratic theory assumes that there are procedures that individuals can use to shape laws and public policies, as well as express their dissent. The normal mechanisms of democracy are voting, petitioning one's representatives in government, the free expression of ideas through mass media outlets, public demonstration and protest. When a riot expresses a grievance, it operates outside these normal parliamentary processes in two ways. First, it abandons the normal means of petitioning the government. These mechanisms are premised on the idea that a portion of the citizenry can use persuasion to bring about policy change. Persuasion consists of spoken and written words used to change a person's beliefs through a combination of reasons, rhetoric and emotion. Conversely, a riot is expressive and a form of resisting the government's authority through uncivil disobedience. Expression is concerned solely with making one's own thoughts public; it does not involve using reason to change the beliefs of others. Riots express noncompliant rage. If this rage is justified, then the riot may prompt the authorities and the public to confront the underlying injustice that the rioters are contesting. Fear, rather than reason, is a crucial component of riots. 23 Rioters use fear to intimidate bystanders, targets and the authorities in the hopes that it will compel policy change. 24 The riot is also a mode of resistance rather than a form of persuasion. It entails acting in such a way so as 'not to be governed like that' (Foucault 2007, 44). Whereas parliamentary practice assumes that groups will try to use the force of the better argument to petition a group to change its policy, riots seek to provoke change through expressive rage and militant disobedience.
Secondly, a riot is extra-parliamentary in the sense that the rioting crowd is not an organized group that fits within the normal political process. While a group may call a demonstration that breaks out into a riot (such as the Poll Tax Riot in London in March 1990), or in some instances 23 Many anti-riot statutes note that for a protest to be a riot, bystanders must be terrorized (Inazu 2017, 13, 16;Kelsey 1906, 468). 24 Not all (or even most) riots are motivated by a specific grievance. Gary Marx identifies 'issueless riots'i.e., riots where a generalized belief is absent among the rioters and the riot is not instrumental to solving the rioters' problemsas an understudied form of rioting. Examples include riots after festivals and celebrations, riots after victories by sports teams and riots when the police go on strike (Marx 1970).
(such as the Black Bloc 25 protests at the Hamburg G20 riots in July 2017) a group might actively instigate a riot, the rioting crowd does not operate in a way that is normally recognized by parliamentary processes. This is because it cannot take the form of an institution or practice that a legislature and its officials can formally recognize. It is not a political party or social movement or special interest group; it is a spontaneous organizationspontaneous not in the sense of being unplanned, but in the sense of being sui generis and temporary. 26 A group might instigate or organize a riot, or even use a riot as a political tactic, but the riot is a distinct and discrete event. Riots, like a temper tantrum or thunderstorm, are intense but fleeting. 27 They are not institutions; nor are they institutionalizable in the sense of 'an arrangement for maintaining order, resolving disputes, selecting authoritative leaders, and thus promoting community between two or more social forces' (Huntington 1968, 9). A riot does none of these things, and is not a stable or recurring pattern of social behaviour. Each riot is a unique occurrence.
Riots that enable those within the community whose grievances are either not voiced within normal parliamentary procedures or are systematically ignored through normal political mechanisms are permissible. D'Arcy argues that militant political action, such as rioting, is appropriate when it conforms to what he calls the democratic standard. The democratic standard rests on two principles. First, democracy is 'the self-governance of the people through inclusive, reason guided public discussion ' (2014, 4). Secondly, there are circumstances in which 'it is consistent with the democratic ideal to set aside discussion and apply forceful pressure through adversarial, confrontational protest ' (2014, 5). D'Arcy argues that if a community lacks an effective mechanism for expressing its grievances, then a riot is 'a kind of exit: a temporary withdrawal from attributing authority to the legal order ' (2014, 154). 28 In most instances of grievance rioting, the participants lack the ability to exercise their voice; institutions often silence these communities through their structures. We can think here of Ferguson, Missouri and the massive under-representation of African Americans in both the Ferguson Police Department and the City Council as an example of a polity failing to provide voice in a systematic way for an extended period of time. 29 Contra the claim made by many liberal critics of riots, grievance riots do work: they provoke a response from the authorities. In high-profile and large-scale riots, the public authorities normally respond by appointing a commission of inquiry that investigates the causes of the riot and proposes recommendations (D'Arcy 2014, 155). While rioting to demand an official inquiry to create policy change is not the most desirable (or even efficient) means of giving voice to the voiceless, when a political order systematically blocks all other means of airing and receiving redress for a grievance, then a riot is justified. The criterion by which one may justify a riot's extraparliamentarianism is whether or not the parliament has systematically ignored or blocked a group from receiving redress for their grievance through normal parliamentary procedures. Riots 25 The Black Bloc, contrary to many media reports, is not so much an organization as a protest tactic or formation. Black Blocs form as fringe groups within protests. Their members dress all in black and attempt to conceal their identities to avoid identification by the police. Only once on site do the members of the Black Bloc decide what actions they will engage in that day (D'Arcy 2014, 120-123). 26 I thank Douglas Dow for suggesting this distinction when commenting on an earlier draft of this article. 27 As Canetti observed, 'For just as suddenly as it originates, the crowd disintegrates. In its spontaneous form it is a dangerous thing ' (1984, 16). 28 D'Arcy develops the idea of riots as giving voice to the voiceless through an innovative reading of Albert O. Hirschman's Exit, Voice, and Loyalty. In Hirschman's classic analysis of dissatisfaction and power, participants in any organization have three possible responses if they are unhappy with the behaviour of the organization: to leave (exit), to communicate their grievances and propose possible changes (voice), or to remain silent and comply (loyalty) (Hirschman 1990). Contra D'Arcy I would argue that riots are actually a form of voice, as Hirschman observes that one feature of the state is that short of migration, one cannot exit it. 29 At the time of the Michael Brown shooting, the population of Ferguson, Missouri was 65 per cent Black. The Ferguson Police Department had fifty-three commissioned offices; three were Black, and there were two other minorities (Raab 2014). Before the shooting, the Ferguson City Council had five White members and one Black member. In the first election after the shooting, two new Black councillors were elected (Pearce 2015).
usually target local problems and specific grievances. 30 Politically motivated riots express a distinct grievance that is normally dealt with through parliamentary procedures. But if the parliamentary system either ignores or blocks the grievance, the riot is an extra-parliamentary act of last resort.

Conclusion
Political theorists do not consider riots as a legitimate form of political resistance because they operate outside four of the main institutions that most Western political theorists defend.
Considering these four forms of extra-institutionalism togetherthe fact that riots are extrapublic, extra-state, extra-legal and extra-parliamentarygenerates two conclusions. First, because there are widely accepted justified exceptions to these institutions in other areas of political theory, these reasons can also apply to riots. Secondly, because a riot is extra-institutional in four ways it needs to meet the criteria of a justified exception in each of these four institutions in order to be justifiable. This makes the threshold for justifying a riot potentially higher than other forms of political resistance, but this higher threshold does not mean that no riots are justifiable. I have identified eleven criteria that can be used to assess the legitimacy of a riot (Table 1).
These criteria are not intended to provide a simple check box exercise for assessing a riot. They should instead be used to reflect on the features of individual riots, on a case-by-case basis. Political theorists can use the criteria to determine whether an individual riot is justified or unjustified. At a minimum, in order for a riot to be justifiable it must offer legitimate reasons for disobeying each of the four institutions that are on par with the recognized forms of legitimate extra-institutionality in other practices of resistance. The more criteria a riot satisfies, the more confident we can be that it was justified.
Why would we need to make such a set of judgments? In the case of just war theory (at least in the ideal case), the argument about a war's justifiability should take place prior to its outset. Such a deliberation is unlikely in the case of riots. A just riot theory would instead provide a means of Unjust dispersal order Are the authorities using riot law to disperse a lawful and peaceful assembly? Parliament Inability to seek redress through parliamentary procedures Has the parliament has systematically ignored or blocked a group from receiving redress for their grievance through existing parliamentary procedures?

30
Compare with Tilly (1983), who observes that riots and other pre-Industrial Revolution forms of social protest such as 'rough music', seizures of grain, turnouts and field invasions were all local in nature and targeted patronized power holders. Conversely, post-Industrial Revolution protests tend to be national in scope and target autonomous power holders. Clover (2016) draws on Tilly's distinction to develop a more general theory of the decline of riots during capitalism, and their return as 'riot prime' in contemporary politics. assessing the validity of a particular riot after the fact, which would allow us to determine the appropriate response to it. At a minimum, we need such a theory to distinguish between riots after a sports team wins and those protesting police murders of unarmed citizens. Treating all riots a priori as illegitimate unfairly dismisses the grievances of the unheard, and potentially denies some of the most marginalized members of society the ability to voice their concerns. Conversely, developing a vocabulary and theory of just riots would allow us to make judgments about whether particular riots were justified, how the authorities should respond to the rioters' grievances, and how individual rioters should be punished (or excused) for their actions. It is time for political theorists to stop ignoring the fact that riots constitute a form of politics and develop a just riot theory.