‘They attacked the city’: Security intelligence,the sociology of protest policing and the anarchist threat at the 2010 Toronto G20 summit

Introduction

International summits have become routine spaces for protests, where a convergence of social movements rally against corporate globalization (Seoane and Taddei, 2002). Protest policing at global summits has also become common – the G8/G20 summits in Huntsville and Toronto (in the province of Ontario, Canada) during June 2010 were no exception. The meetings drew sizeable crowds of demonstrators and a massive police presence, becoming the largest policing operation in Canadian history (with extensive military involvement). Protest events spanned the course of a week, culminating in large marches scheduled for Saturday, 26 June 2010. Despite the massive police presence, a Black Bloc contingent broke from the march and snaked through Toronto’s business district for almost two hours while engaging in property destruction that targeted the windows of corporate retail establishments. Four police cars were set ablaze. Yet, the Black Bloc were not confronted by police (eventually dispersing of their own accord) and only hours later did police begin widespread arrests in numerous protest zones. In what became the largest mass arrests in Canadian history, over 1100 people were swept from the streets in chaotic scenes, then held captive in a makeshift detention centre. After the summit, most criminal charges stemming from these arrests were withdrawn or stayed without explanation.

Many of the circumstances pertaining to the policing of the Toronto G20 remain unknown. Media reports first focused on the four burnt out police cars and property crimes, and later detailed the police crackdown that followed the Black Bloc actions. Media outlets and legal and civil rights watchdogs publicized cases of police aggression and widespread rights violations (CCLA, 2010; Marin, 2010) but investigations have been stymied by what critics have called a ‘blue wall of silence’ (Yang, 2011a). ¹ Many individuals have taken to civil remedies, including a 600-plus member class action suit filed by those detained during the summit. In the face of public criticism, police agencies have rationalized their behaviour by dwelling on the anarchist threat. Toronto Police Chief Bill Blair has been most forceful in his lampooning of the anarchists, claiming that ‘when they couldn’t attack the summit they attacked the city’ (Robson, 2010). Because windows were smashed and police cars were burned, Chief Blair reasoned the litany of arbitrary searches, arrests and detentions documented by civil liberties groups would not be subject to second-guessing or external investigations. Blair insisted police measures were taken to ‘keep the criminals out’ of the summit. Yet these remarks belie a conflation of anarchism and criminality as well as a sophisticated effort combining local policing agencies and national-level security intelligence agents that targeted and criminalized persons considered to be insufficiently institutionalized in their protest actions.

Our purpose is to investigate the dynamics of contention and control related to policing at international summits. Scholars have argued that the sociology of protest policing must comprehend the undercurrents of unease and confrontation among divergent social actors as well as the patterns of protest and police responses leading up to such events (Cunningham, 2003; Gorringe and Rosie, 2008). Sociologists have commented on the transition from protest policing based on escalated force models to negotiated management to hybrid forms of ‘intelligent control’ and strategic incapacitation (de Lint and Hall, 2009; della Porta and Reiter, 1998; Gorringe and Rosie, 2008; King and Waddington, 2005; Noakes and Gillham, 2007). Recent protest policing literature has also begun to assess the role of intelligence in policing practices and training, an insight that we extend in this article by examining police preparations for the G8/G20. We examine how policing strategies enacted during the Toronto G20 summit were rooted in intelligence programmes developed years prior and, having been generated over long-term periods, had a significant influence on police training.

This article contributes to sociological analyses of protest policing by illustrating how security agencies focused on one particular threat category: the anarchists. We detail how, in advance of the summit, anarchists emerged as the predominant threat category in police training and operations. Drawing from the results of Access to Information Act (ATIA) requests, we detail the covert intelligence programme targeting suspected anarchist groups and provide an analysis of the frontline officer training programme. Waddington and King (2007) note that sociologists of protest must begin to study the surveillance practices that go on before protests and how this monitoring shapes protest policing. We show how intelligence gathering shaped the police actions that occurred at the Toronto G20 summit. We outline the material consequences that arise from surveillance practices related to protest policing, including the making up of group identities (e.g. the anarchists) deemed to be threatening, and how these designations are used by police. Demonstrating how intelligence is mixed with police training and operations in the form of what we call a threat amplification spiral, we discuss how intelligence practices that consistently emphasized the ‘anarchist threat’ resulted in sweeping police repression and willingness to arrest ‘anyone wearing black’ (Kraus, 2010; McNeilly, 2012: 90; Porter, 2010) at the Toronto G20.

This article has three parts. First, we use documents obtained through access to information requests to detail the intelligence campaign targeting anarchist groups. Second, we show how intelligence informed G20 frontline officer training and created conditions for threat amplification. We detail how this training differentiated between ‘anarchists vs protestors’ and produced a caricature of anarchism. Finally, we relate these materials from the G20 to protest policing literature. We discuss the concepts of ‘intelligent control’ (de Lint and Hall, 2009) and strategic incapacitation (Gillham, 2011) and argue that these provide viable sociological frameworks for understanding contemporary protest policing at international summits and the surveillance that informs these projects.

Method of enquiry

Producing data on policing practices and surveillance has become increasingly complicated post- 9/11 (Earl, 2009). Intelligence and security agencies such as the Royal Canadian Mounted Police (RCMP) and Canadian Security Intelligence Service (CSIS) are not forthcoming with documentation on past protocol or future plans for protest policing. The same can be said of security intelligence and policing agencies in the USA, UK and elsewhere. Other methodological strategies are required to gain access to such research sites. In Canada, requests under the Access to Information Act (ATIA) are one way to learn about government agencies and their practices. While access requests are subject to procedures that prevent, delay and redact disclosure, the ATIA can provide timely materials that speak to processes that would be otherwise hidden from public scrutiny. Although these materials are sometimes fragmentary, tactical and sustained use of the ATIA can produce valid data about typically closed-door government agencies (see Larsen and Walby, 2012; Walby and Monaghan, 2011a, 2011b). The materials addressed below demonstrate both the fragmentary nature and contemporary relevance of using the ATIA as a means of data production in the social sciences.

Materials reviewed stem from a collaborative research project that followed the G20 meetings in Toronto. As a result of the unprecedented police actions and arrests, a group of media members, researchers and activists collaborated on a project to coordinate ATIA requests, as well as provincial and municipal Freedom of Information requests. This group, the G20 Research Group, has provided these materials in an online format to the public and legal representatives. ² Requests focused on the federal RCMP, the provincial Ontario Provincial Police, as well as municipal police forces that were merged into the networked command structure for the G8/G20. Documents included financial records, operational plans, emails and memos and briefing documents. We focus on operational plans, threat assessments and training materials. ³ We supplement our textual analysis by drawing from exploratory interviews with people arrested or detained during the Toronto G20.

Security intelligence, protest policing and G8/G20 summits

Police training and the anarchist threat

As conservative counter-terrorism scholars (Borum and Tilby, 2005) note, it is common for police in countries that host G8 and G20 meetings to have an operational definition of ‘anarchists’ for protest policing efforts. A key indicator in this definition is propensity towards violence, ostensibly signalled by black garb and other stereotypical indicators. This attempt to operationalize ‘anarchists’ was evident in the police preparations for the Toronto G20 (see also McNeilly, 2012: 262).

Materials discussed below originate from an officer training course designed for the G20 summit. The ‘G20 Face to Face – Front Line Officer Training’ was a three-hour session for ‘all officers assigned to front line duties at the G20’. The training was ‘designed as a reinforcement of the cognitive skill set required of all officers who are being deployed to the G20 Summit’. As a preparatory programme for frontline officers, it offered an outline of their mandate, legal powers, and provided a backgrounder on upcoming events. The session also provided tactical training, where ‘officers from all participating Police agencies will examine: Crowd Management protocols, CBRN [chemical, biological, radiological and nuclear incidents] response, Operational Considerations and practical Gas mask application in concert with defensive and frontline tactics’. The delivery methods included lectures, memos, scenarios and practical exercises. This material on anarchist indicators sheds light on how police begin to think about dealing with perceived threats at protests.

At the beginning of the training session, participants were asked to pose questions to themselves regarding their duties and objectives in preparation for facing a hostile public during the protests. For the exercise, participants were also given talking points in order to provide responses to these questions:

Police participants were cautioned to understand ‘everyone has the right to protest [but] be mindful of agitators in the crowd’. To assist with the identification of so-called agitators, an outline was provided on how to recognize anarchists. As part of a slide entitled ‘Protestors vs. Anarchists’, the presentation distinguished between lawful protestors and anarchist criminals. It read: ‘Protestors are there to be heard and to express themselves. Anarchists may have similar goals (or not) but choose distinctly different methods.’ It stated bluntly that ‘an anarchist is someone who believes that governments should be abolished as unnecessary and are willing to use a lawless action to attain this goal’. De Lint and Hall (2009: 254) use the idea of ‘coding for militancy’ to refer to such translation of intelligence into police protocol.

This presentation outlined characteristics that frontline officers should rely upon to spot anarchists in crowds. It defines the ‘black bloc as a temporary collection of Anarchists that represent a contingent in a protest march’. Black Bloc members are described as having an ‘average age [of] mid teens to early twenties but they have the occasional 30–60 year old anarchist/Blk Blc’, adding: ‘Some, not all, Anarchists will dress in all black clothing or carry them in a bag to be used later on in the protest. It acts to identify/visibility/solidarity/mutual aid. Safety equipment is often worn to protect from lawful removal by police’ (emphasis in original). Finally, the presentation concludes: ‘To help with articulation if there are to be protests please go home or at your work station and google Black Bloc, look at their behaviours, clothing, and listen to their reasons.’

This police training translated the intelligence material from JIG-PIIT into operational protocol. Officers are cautioned to observe grounds for reasonable suspicion during protests. They are told to look for ‘individuals carrying “water bottles” that do not appear to be water inside’, ‘abnormally heavy knapsacks’, or smell for gasoline. The presentation then provides possible relevant Criminal Code sections to detain suspects exhibiting these characteristics. The presentation also warns officers that ‘Anarchists will attempt to bait you into a verbal and perhaps physical altercation. They are hoping for a media sound bite or an escalation of activities or Violence.’ Finally, officers are warned of the context in which the anarchist Black Bloc will act:

Expects lots of noise. Whistles. Yelling. Air horns, fireworks, LRAD (Long Range Acoustic Device). Noise is useful in keeping enthusiasm and drive in tiring situations. The listening of music can create sudden drives of energy and joy within the bloc where none exists, and can send contradicting message of power to the enemy in battle. Black Bloc bagpipe players could bring immense joy to many, as does radical cheerleading to others.

All these indicators are taken as proof of the intelligence forecasts, that these protestors are not ‘sufficiently institutionalized’ (de Lint and Hall, 2009: 47). As recent ethnographic works have demonstrated, anarchism has become an increasingly broad and widespread influence within the Global Justice Movement (Graeber, 2010; Juris, 2008). With its emphasis on collective decision-making and participatory democracy, anarchism has been adopted by a wide range of social movements, demonstrated most recently with the Occupy Wall Street movement (Berrett, 2011). Yet the training programmes developed for frontline officers produce stereotypical caricatures of violent anarchists and the Black Bloc.

The Black Bloc itself is best described as a tactic rather than a movement, a point acknowledged after the protests by Toronto Police Chief Bill Blair (Byrne, 2010). Although the Black Bloc tactic has become associated with property destruction, damage to property is not its main purpose. Having emerged from the 1980s squatters’ movement in Germany, the Black Bloc is primarily a defensive strategy emphasizing mutual protection. Unlike strategies of non-violent civil disobedience that involve individuals sacrificing their bodies in order to be passively arrested, Black Blocs convene ‘a mass of equals, each of whom will risk arrest only to prevent their comrades from being arrested, or to rescue them’ (Graeber, 2010: 418). Of course, the Black Bloc cannot be separated from acts of property destruction. These militant actions are what Jeffrey Juris calls ‘performative violence’. Juris (2005: 416) details how activists stage ‘spectacular violent performances, partly, to gain access to the mass media’. These symbolic confrontations communicate an opposition to capitalism, the state and corporate globalization. Yet, Juris (2005: 416) warns that police and government officials manipulate images of performative violence, ‘decontextualizing and reinserting them within narratives that frame protestors as dangerous criminals or terrorists’. Ethnographic accounts from the Global Justice Movement reveal the deeply divided opinions that participants share about what constitutes violence (Graeber, 2010; Juris, 2008). While many of the debates about ‘violence’ involve important distinctions, enactments of performative violence can reaffirm the antagonisms between demonstrators and police, serving to intensify use of physical violence by police.

While debates regarding notions of direct action, property destruction and state violence illustrate diversity within the Global Justice Movement, we highlight the processes of categorization and the resulting threat amplification cycles. Literature on communications and social movements has demonstrated how media framing of protests creates folk devils out of misrepresenting anarchist politics (Rosie and Gorringe, 2009). It is not only media representations of performative violence that are used to create a caricature of anarchism. The caricature is produced in a similar way by intelligence agencies that create profiles of categorical suspicion, which later inform protest policing and training. Here we emphasize the process of threat amplification where intelligence produces categories by simplifying issues and adding stereotypes (e.g. ‘violent’) to the categorical identities. Subsequently, these categories are translated for the purpose of police training and applied during police operations in ways that feed back into intelligence and reinforce the original assessments and stereotypes.

Unlike the direct and premeditated effects of covert intelligence that targeted a small group of suspected ringleaders detailed in the earlier section, the impacts of G20 police training programmes had wide-ranging repercussions. Over 1100 individuals were arrested. While 330 charges arose from these arrests, only 66 cases have proceeded with all the others being stayed, dismissed, or withdrawn (O’Toole, 2011). Moreover, only 32 cases have resulted in guilty convictions. As a result of the widespread arrests, lack of civil protections and poor detention conditions, victims of police violence initiated a 600-member class action lawsuit. Many individuals have given accounts of how police justified their arrests because they were wearing hoodies, or backpacks. One such account is the arrest of Michael Puddy. Puddy was arrested during a police kettling operation late on Saturday, 26 June, while he was heading to a punk rock concert downtown (R v. Puddy: para 20). At trial, four police witnesses testified they knew of no accounts of Puddy behaving violently. The arresting officer, Sgt Ash Awad, testified he did not witness the defendant’s original detention, nor did he hear of any reasons for Puddy’s apprehension. In light of the criminal charges that Puddy ‘obstructed police’, Awad said he did not see the defendant obstruct anyone. Explaining the arrest, Awad said that Puddy was dressed in black clothes and sported a ‘Mohawk’-style haircut (R v. Puddy: para 9). Awad testified that he had been briefed on the ‘Black Bloc’, and Puddy matched the characteristics of this group due to his black clothing. Another officer, Detective Philip Chung, described Puddy as ‘dressed like’ a member of the Black Bloc (R v. Puddy: para 17). On these grounds, Chung recommended that the Crown seek a detention order. However, in court, Chung conceded that he had no information specific to the defendant’s conduct when he drafted the synopsis and that he ‘generalized’ from reports for other arrestees when he described the circumstances of the defendant’s arrest (R v. Puddy: para 17). This is the culmination of what we call threat amplification, when categorical suspicion is produced by vague characteristics of supposed criminality presented in intelligence reports and officer training, which are then represented as evidence and grounds for detention. Puddy was arrested, strip-searched and held for two days. He was released on a $25,000 recognizance with conditions that required him to live with his girlfriend’s father as his surety, under ‘house arrest’ among other stringent terms. The charges were only defeated on 11 August 2011 – more than one year after the G20.

While only one account among thousands, Puddy demonstrates how threat amplification translates intelligence on the ground. When threat assessments are coded into categorical identities, they become operationalized through circuits of training and further intelligence work. While passing through the circuits, categories of threat become increasingly blurred and more like stereotypes. On the ground, threat amplification often results in coercive and reactionary policing practices. When police were processing arrests, we see that the characteristics of threat become ‘generalized’ to justify a host of charges that are divorced from any particular actions: simply fitting the stereotype sets criminal proceedings in motion. Puddy also underlines the significant human costs of policing through categorical suspicion. The key role of intelligence in this process requires that sociologists of protest policing account for how threat categories are produced, applied and amplified.

Discussion: Enfolding security intelligence into protest policing

Since the ‘Battle of Seattle’, studies of protest policing strategies have noted a movement away from the ‘negotiated management’ models of the mid-1990s. While contemporary protest policing continues to incorporate elements of accommodation, scholars have noted the limits of the negotiated management framework for explaining international summit policing (de Lint and Hall, 2009; Noakes and Gillham, 2007; Noakes et al., 2005). Starr and Fernandez (2009) also claim that, despite dramatic changes post-Seattle, academic literature has failed to keep up with the complex dimensions of repression and creative resistance.

Transformations to policing strategies are made especially evident by the prominence of the ‘Miami Model’ that has proliferated in the United States (Starr and Fernandez, 2009). Emerging from FTAA meetings held in 2003, the Miami Model involves actions that deliberately attempt to stop gatherings through excessive force, and to deter and punish those who remain committed. It involves the implementation of new (often temporary) legal measures to criminalize activists or limit movement, and indiscriminate arrests (often including clearly marked safety personnel such as medics, legal observers and media). Militarization of protest policing post-Seattle involves the use of the iron fist without the velvet glove, and research is only beginning to examine the profound reorientation away from negotiation and accommodation (Starr and Fernandez, 2009). This is particularly accurate when it comes to certain protest groups who refuse to negotiate with police on the grounds that police use these interactions for intelligence gathering. Moreover, since radical social movements (Fitzgerald and Rodgers, 2000) – such as ‘new new social movements’ (Graeber, 2010) – are not based on hierarchical organizing and have no institutionalized leadership, police often refuse to recognize them as legitimate. Negotiated management strategies have thus become secondary to coercive options. However, our point is that an exclusive focus on force does not account for the analytical and empirical importance of surveillance that informs protest policing, or the construction of threat categories in intelligence.

De Lint and Hall (2009) have offered a corrective to the literature on protest policing and crowd management with their idea of ‘intelligent control’. Demonstrating how hybrid policing models have emerged since the Seattle demonstrations, de Lint and Hall (2009) argue that some aspects of protest policing have become more routinized (e.g. the labour movement) while policing of other protest groups has become militarized. The reason why protest policing has moved towards militarization is because police services ‘see strikes as highly predictable events’ (2009: 215) whereas protests organized by new social movements groups ‘are seen as less predictable’ (2009: 216). The conjoining of conciliatory techniques with crowd management is what de Lint and Hall refer to as ‘intelligent control’. Intelligent control requires: (1) the appearance of a liaison function to negotiate with and accommodate perceived leaders; (2) the use of surveillance to produce actionable intelligence; and (3) militarization. While de Lint and Hall include repression of protests in their framework, the idea of ‘intelligent control’ stresses the pre-emptive aspects of policing that develop actionable intelligence ahead of protests.

The production of actionable intelligence is facilitated through intensified information gathering, and more information sharing between agencies. De Lint and Hall (2009) argue that municipal police in Canada are coming to rely more on intelligence produced by organizations at the national scale of policing and security (e.g. RCMP, CSIS), which our case study affirms. Intelligence is thus crucial to police in their efforts to disaggregate elements of the movement and disrupt activities that are perceived to be outside their control.

Gillham (2011) has pushed insights into protest policing further with the idea of strategic incapacitation (see also Noakes and Gillham, 2007). He argues that strategic incapacitation is based on: (1) the use of surveillance and information sharing as a way of assessing threats and managing risks before protests; (2) the use of pre-emptive arrests before protests and weapons such as tasers during protests to disrupt certain groups of people who are deemed risky; and (3) the cordoning off of city spaces to isolate certain groups and contain protestors who are deemed to be ‘non-institutionalized’ by police. As with the idea of intelligent control, the notion of strategic incapacitation elevates the importance of intelligence as a way of monitoring and pre-empting certain mobilized groups. However, the idea of strategic incapacitation denotes the sense of profiling involved in our case and protest policing at international summits in the early 21st century.

Not only do processes of ‘intelligent control’ and strategic incapacitation mix repression and accommodation based on pre-emptively constructed categories, but intelligence practices make up the identities of those being monitored. Finding an ‘anarchist’ agitator in the crowd becomes a self-fulfilling prophecy, as the indicators in the police training are shaped by intelligence classifications. The case of Puddy demonstrates this quite clearly. The process of categorizing some protestors as anarchist starts with the assembly of indicators collected from intelligence practices. These indicators are then incorporated into training instructions that policing agents use at protests to inform their ways of controlling people and space. Anarchists are treated as threats prior to and during protests, but the very idea of threat – and who is designated a threat – is put together through intelligence. There is a feedback effect that we refer to as threat amplification connecting representations of anarchism, the categories of intelligence agencies and the training of frontline officers. As the emphasis on anarchists accelerates through the loops of intelligence networks, police on the ground fixate on traces of the anarchist threat. The result is that wearing black clothing, or flying a black flag, or merely being on a bus travelling to a protest, becomes embedded with categorical suspicion.

Earl and Soule (2006) argue that police are afraid of losing control and legitimacy as a result of poor planning or the actions of well-organized protestors. The escalation of intelligence in protest policing is part of an effort to keep control and surveil organizers long before and after demonstrations. Local agencies responsible for protest policing at international events increasingly incorporate security intelligence from other scales of security for more targeted monitoring and interventions. While this might not be the case when policing small protests (Rafail, 2010), international events place extensive pressures on local police (Ericson and Doyle, 1999) requiring multidimensional projects that involve negotiation, force, but also surveillance and intelligence campaigns.

Conclusion

The 2010 G20 in Toronto affirms that changes in protest policing noted in other countries have taken hold in Canada, with practices of ‘intelligent control’ and strategic incapacitation visibly entrenched in police activities at the G20. Policing agencies combined long-term intelligence projects with overt repressive tactics during the event. In planning for the G8 and G20, policing and intelligence agencies were aware of the potential for mass mobilization, the effectiveness of which was not lost on police who planned their projects with this struggle for consensus in mind:

2010 has the potential to become a year which will be used by Left Wing extremists to widely disseminate their ideological message. It is important for law enforcement to be vigilant of potential criminal activity that could result in a convergence between ideologies and individual personalities. It appears at this stage that this convergence is emerging as an extremist crux for the year 2010. (File 3551: 191 emphasis added)

While events on Toronto’s streets seemed spontaneous, intelligence agencies had identified and surveilled targeted groups that presented ostensive threats to public order far in advance. The intelligence campaign fixated on and produced an identity of the criminal anarchist at the centre of the ‘extremist crux’ that justified the multi-agency surveillance project and also informed the training that G20 frontline officers received. The synergy between intelligence and policing leads to a process that we have referred to as threat amplification, which creates threat categories that become operational for police during demonstrations.

Policing for the Toronto G20 was ‘intelligent control’ insofar as it was intelligence-led policing before protests and a combination of coercion and accommodation during the demonstrations. It was also strategic incapacitation insofar as the intelligence effects obstructed activists before the protests and facilitated snatch squad actions. We have gone further to argue that this enfolding of security intelligence into local policing has intensified the practice of making up threat categories. Sociological literature on protest policing should account for how intelligence and threat amplification have become central components of hybrid strategies that are employed in protest policing operations at international summits.