Auditor Design and Accountability in Encounters between Citizens and the Police

Chapter to appear in: Security, Ethnography, Discourse: Transdisciplinary Approaches to Everyday Unease, edited by Constadina Charalambous and Emma McCluskey (Routledge) Auditor Design and Accountability in Encounters between Citizens and the Police Rodney H. Jones University of Reading Abstract The increased use of cell phone cameras by citizens to record their interactions with law enforcement officers has altered practices of policing and of talking with police. It is widely assumed that the presence of cell phone cameras can have the effect of making police officers more accountable for their actions. The degree to which this is true, however, depends crucially on how citizens and police make use of the camera as an interactional resource in the moment by moment negotiation of rights and responsibilities that unfolds over the course of the encounter. This paper provides close examination of two different encounters between citizen and police officers in the US, one involving a Caucasian motorist, and the other an African American. It focuses on how citizens used their cell phone cameras to construct different kinds of ‘auditors’ of the interaction and how the officers responded in various ways to these strategies. The analysis shows that increasing the accountability of police officers requires more than just technological means (cameras) and the ‘legal right’ to use them, but rather depends on a range of complex discursive strategies both drivers and officers engage in at the intersection of the immediate interaction order and larger systems of power and inequality. Keywords: accountability, audience design, auditor effect, communication accommodation theory, police, sousveillance Introduction On the 25th of May, 2020, Minneapolis police officer Derek Chauvin killed George Floyd, a 46 year-old African American man, by kneeling with his knee on the man’s neck for eight minutes and forty-six seconds while multiple witnesses, including some who were videoing the incident with their cell phones, begged the officer to stop. The videos of the incident that circulated on social media sparked widespread protests and renewed calls for racial justice, not just in the US but worldwide. The first video of Floyd’s death to go public was shot by a high school student named Darnella Frazier. It shows Chauvin looking down at Floyd for most of the incident as the black man pleads for his life, saying ‘I can’t breathe’ and ‘Don’t kill me’ until he becomes unconscious and the pleading stops. There are a few moments in the video, however, where Chauvin looks up and gazes directly at Fraizer’s camera with an almost nonchalant expression on his face. ‘You're enjoying it,’ one witness can be heard saying. ‘Look at you. Your body language explains it.’ There were other videos taken on that day from other angles by other witnesses, by the surveillance camera mounted outside a nearby Chinese restaurant, and by officer’s body cams, but Fraizer’s video remained the most widely circulated public document of the killing, and the moment in the video that became the still photograph that appeared in newspapers around the world was of Chauvin gazing up at his audience while his victim died. The proliferation of video cameras in the hands of private citizens has altered the relationship between police officers and the public, making incidents of police abuse like the one described above more visible and easier to document (Sandhu & Haggerty, 2017). Along with this ‘new visibility’ (Goldsmith, 2010; Thompson, 2005) has come the hope that the presence of cameras would make police more accountable for their actions after years of allegations of police misconduct being dismissed for lack of corroborating evidence (Meyer, 2005), and that this increased accountability would result in a reorganisation of the power dynamics between police and the community. Much of discussion around this new regime of visibility has focused on the power of the technology itself — the portable video camera embedded into nearly every cell phone sold today— to ‘reverse the gaze’ (Sandhu & Haggerty, 2017:80), creating what Manuel Castells (2007:238) refers to as ‘counter-power’ (see also Monahan, 2006). Engineer and inventor Steve Mann (Mann, Nolan & Wellman, 2002; Mann & Ferenbok, 2013) has famously used the term sousveillance to describe the reversal of the gaze made possible by portable technologies such as cell phones and wearable cameras. Surveillance, or ‘veillance’ from above, Mann argues, depends for its power upon being unobserved; sousveillance counters this power by making surveillance and those who practice it visible, and, crucially, by making them aware that they are visible. For Mann, this is a key difference between surveillance and sousveillance-- while those involved in surveillance often seek to remain unobtrusive, those engaged in sousveillance derive much of their power from making their presence known with the aim not just of documenting abuses by authority figures, but of preventing them, based on the notion that the awareness of being visible inevitably alters people’s behaviour (Brighenti, 2007; Foucault 1975; Thompson, 2005). The degree to which making police officers aware of their visibility actually makes them more accountable, however, is far from clear, as is dramatically illustrated by Officer Derek Chauvin’s apparent impunity before the witnesses gathered to watch his murder of George Floyd, including the electronic witness of Darnella Frazier’s cell phone camera. One reason for this uncertainty is that technologies of visibility are, as surveillance scholar David Lyon’s (2003) puts it, ‘Janus faced’, able to be used for a range of different purposes with a range of different effects. How these technologies are used depends on the interactional contexts into which they are deployed and the ways their deployment alters those interactional contexts. In other words, the camera itself does not confer power on the witness using it; rather it constitutes just one of a range of different tools with which people can negotiate power in interactions with police. As Bock (2016:15) argues, ‘most scholarship [on citizens’ use of cameras], so far, has focused on the technology and texts rather than grounded practices,’ resulting in an ‘impoverished’ analysis (see also Allan, 2013). In contrast, Bock urges us to focus on these technologically mediated interactions between citizens and police as embodied practices which unfold over time, an approach in line with Brighenti’s (2007:323) call for increased attention to ‘the relational, strategic and processual aspects of visibility.’ From this perspective, technologies are seen not as causing accountability, but as part and parcel of the ‘material orderings in which the production and consumption of accounts is instantiated’ (Neyland & Coopmand, 2014: 4), and accountability itself is seen not an inevitable by-product of visibility, but as a dynamic and contingent interactional accomplishment, something that needs to be continually negotiated over the course of citizens’ encounters with police. It might be better to think of these negotiations in terms of what Ericson (1995) calls ‘account-ability’: the capacity of different parties to create credible accounts of their activities (and elicit accounts from others) using whatever means are available to them (including what Garfinkle [1967:1] calls ‘situated practices of looking and telling’). Account-ability, says Ericson, always involves strategies of disclosure and concealment, selection and distortion, a point Goffman (1959) makes more generally about public performances of the self. More importantly, account-ability is always a matter of negotiating and organising the roles and responsibilities associated with looking and telling afforded to the different parties in the interaction, including interlocutors, involved third parties, uninvolved witnesses, and other present and non-present audiences, roles and responsibilities which Neyland & Coopmans (2014:2) refer to as ‘accountability relations.’ Finally, such negotiations rest on a foundation of normativity: the relationship between visibility and accountability is always, as Goldsmith (2010) points out, predicated on participants’ ability to maintain ‘normal appearances’ (Goffman, 1971) and ‘proper performances’ (Goffman,1959). This, of course, brings to the fore questions about what counts as ‘normal’ and ‘proper’ for particular kinds of people in particular situations, and how people act on their assumptions about normativity. To some degree, the very act of citizens videoing police officers already constitutes a disruption of normative relations of power and accountability. At the same time, it is clear that different police officers in different circumstances manage such disruptions differently, sometimes even exploiting them for their own advantages, and different citizens in different circumstances are viewed as (potentially) more disruptive than others. In this more dynamic interactional approach to the role of video cameras in encounters between citizens and police, the important question, then, is not whether or not cameras make police more accountable, but how police and citizens use cameras (along with other mediational means such as speech, gesture, and aspects of the built environment) to identify ‘relevant audiences for the discharge of accountability’ (Neyland & Coopmans, 2014: 4; see also Neyland & Woolgar, 2002) and to manage visibility and its impact on the distribution of power and the possibility for action (Fyfe & Law, 1988). Empirical studies of the ways police respond to citizens videoing them reveal a range of different strategies for managing visibility. Sometimes officers attempt to get people to stop recording them, often appealing to some nebulous law or regulation that prohibits it (see Example 1 below), and even going so far as forcing them to delete footage or confiscating their cell phones (Potere, 2012; Wilkerson, Mathew & Glazer, 2016). When police do allow themselves to be recorded, many report that they alter their behaviour based on how they think they might appear to those viewing the video footage later on, and they sometimes express frustration with the degree to which the presence of cameras changes the interpersonal dynamics between them and citizens and more generally ‘alter[s] the interactional context of policing’ (Sandhu & Haggerty, 2017:83; see also Brown, 2016; Kapok, 2014). Sometimes, however, officers see the presence of cameras as potentially beneficial and engage in strategies such as pointing cameras out to interlocutors as a way to try to control their behaviour (Goold, 2003; Sandhu & Haggerty, 2017) or producing utterances designed to shape how audiences might interpret the footage. Haggerty and Sandhu (2014:12), for example, describe how police officers are trained to repeatedly shout ‘stop resisting arrest’ when subduing suspects with batons in order to frame the use of force as justified. An important point made by Sandhu and Haggerty (2017), based on their field observations and interviews with officers, is that none of these strategies are used consistently by particular officers; rather, officers tend to choose strategies that they think match particular situations, particular kinds of interlocutors, and particular contexts of policing. Studies focusing on citizens have found similar attention to the strategic, negotiated dimensions of videoing police officers. Rather than seeing their cameras as objective witnesses to events, citizens use them both as ways to actively make visible particular aspects of what is occurring and as tools with which to negotiate interactional power with police. Bock (2016) argues, based on his fieldwork with activists involved in monitoring the police, that ‘cop watching’ is a form of ‘purposeful witnessing’ (p. 14) and ‘embodied watching’ (p. 26) in which the way participants manage their physical presence on the scene is just as important as the videos they later share over social media. Part of this involves using strategies that produce ‘hailing effects’ (Neyland & Coopmans, 2014; see also Munro 2004), gestures that function to call people to account and constrain the ways they are able to control situations. Given the suspicion with which police often view people who are recording them, videoing the police always, to some degree, involves citizens negotiating their ‘right to look’, which is, to some extent, a matter of asserting their ‘legal right’ to monitor the police and being prepared to marshal evidence to defend that right (which is why cop watching organizations often arm their members with detailed information about the laws on videoing police in particular jurisdictions). But the ‘right to look’ is not just a legal right -- it is, more importantly, an interactional right that needs to be negotiated moment by moment in encounters with authority figures, regardless of what the law says. On one hand, negotiating the ‘right to look’ depends upon who is doing the looking, who is being looked at, and how the roles of spectator and spectacle are distributed throughout the interaction, that is, what Neyland & Coopmans (2002:4) refer to as identifying and managing ‘relevant audiences for the discharge of accountability.’ On the other hand, it also depends on the ‘situated practices of looking and telling’ (Garfinkle, 1967:1) by which people produce accounts of their ongoing behaviour and make others accountable for theirs. But the ‘right to look’ is also a matter of epistemic positioning -- the right to assert that ‘what you see’ is ‘real’. Mirzoeff (2011: 1) claims that the ‘right to look’ is not just about claiming a position of objectivity, but rather of asserting ones right to a subjectivity that ‘has the autonomy to arrange the relations of the visible and the sayable’ (emphasis mine). In this paper I will explore how the ‘right to look’ is interactionally negotiated in sousveillance situations involving citizens and police using tools from interactional sociolinguistics and ethnomethodology. The data for this analysis is two videos uploaded onto YouTube by citizens involved in encounters with the police. As I will show in my analysis, the ‘right to look’ does not derive solely from the technological means to record the police nor the ‘legal right’ to deploy these means, but rather is the result of a complex range of discursive strategies that citizens and officers engage in at the intersection of the micro-history of a particular interaction and societal macro-histories of power, identity and race. Audiences and Accounts Discussion of how people manage audiences in interactions from a sociolinguistic perspective usually begin with Bell’s (1984) important insight that people design their utterances based on the kinds of people they think are listening, and simultaneously use their utterances to ‘design’ audiences as particular kinds of listeners and define situations as particular kinds of situations. The ways people alter their utterances for particular audiences, according to Bell, applies to all levels of language, including phonology, lexis, and grammar, and others have extended this to include broader discursive features of communication such as politeness strategies (Brown & Levinson, 1987), pragmatic particles (Holmes, 1995), gesture (Galati & Brennan, 2014), and other forms of non-verbal communication such as dress and hairstyle (Giles & Wadleigh, 1999). Bell’s model of audience design builds upon the theory of communication accommodation advanced by Giles (Giles & Smith, 1979; Giles, Taylor & Bourhis, 1972), which emphasizes that language is not just a matter of exchanging information but also of negotiating group membership: in social interaction people adjust how they speak, converging to the style of those they want to identify with and diverging from those they wish to distance themselves from. Apropos to the current discussion, one of the main contexts in which Giles tested and refined this theory was in interactions between citizens and police in the form of traffic stops. In such encounters, Giles (2009) observes, both officers and citizens tend to manage the interaction by accommodating to each other's communicative styles, the driver, being less powerful, usually being under more pressure to converge than the officer. The challenges for police offices and citizens in such cases, he says, is to balance being non-accommodative, which might encourage aggression, with being too accommodative, which might seem insincere or encourage complacency. But such negotiations do not occur in a vacuum; in other studies Giles and is colleagues found that the degree to which drivers accommodate to officers is related to how much they trust them, and trust in law enforcement, not surprisingly, was found to be related to drivers’ ethnicity and past experiences with police (Dixon et al., 2008; Giles et al. 2006, Hajek et al., 2006). In particular, they found that the race of individuals involved in traffic stops seemed to have an effect on the ways the parties designed their utterances and managed mutual accommodation, with people using language and non-verbal behaviour to position others in particular recipient roles in relation to specific reference groups based on personal or societal biases. Giles and his colleagues (Cherikoff Anderson et al. 2002: 21) point to social identity theory (Tajfel & Turner, 2986) to explain these effects, noting that when people deal with each other as members of social groups rather than as individuals, it is more likely for them to differentiate from each other rather than accommodate, especially ‘when the intergroup identities are embedded in their own longstanding “cultures” … be they gang, ethnic group, police community, etc.’ (p. 21). While such differentiations are rational tactics for enhancing positive identity, the creation of communicative distance can also lead to ‘misattribution, miscommunication, or even worse’ (p. 25). These negotiations are even more complex when more than just two parties are involved, i.e. multiple officers or multiple citizens. Giles (2009), for instance, discusses how the presence of third parties (for example, friends of the driver sitting in the car) can complicate the situation, putting drivers in a kind of stylistic bind, having to perform both for the officer and for their friends, who in some cases might think less of them for being too accommodative to authority– ‘while the officer clearly has the most power and the driver is under more pressure to converge than the officer,’ writes Giles (2009: 281), ‘if the driver does converge, members of her group who overhear might perceive her as a social traitor.’ Bell (1984) calls this phenomenon the ‘auditor effect’. It arises from the fact that audiences are often multiple: apart from the person one is addressing, there might also be third party hearers with different roles in the interaction, different rights when’ it comes to listening, and different effects on how people talk. Bell, drawing Goffman’s concept of ‘participation frameworks’ (1981), conceives of these other audience members positioned in concentric circles radiating out from the speaker, going from addressee, to auditor, to overhearer, and finally, eavesdropper, each having a different degree of ‘legitimacy’ in terms of their participation in the primary interaction. In speaking of the ‘auditor effect’, the kind of hearers that Bell was most interested in were ‘auditors’, people who are not being addressed but are legitimately listening. Citing a range of previous studies (e.g. Douglas-Cowie, 1978; Thelander, 1982), he concludes that auditors can have a profound effect on the way people ‘style’ their utterances, an assertion that has been confirmed in subsequent studies (e.g.Rickford & McNair-Knox, 1994). People more removed from the interaction, overhearers with less legitimate roles in the interaction, Bell posits, have less of an effect, and eavesdroppers have none at all, since speakers are not aware of their presence. Others have noted that the auditor effect is not just a matter of style, but also a matter of pragmatics. Hearers positioned in different participant roles can interpret the illocutionary force of utterances differently, and speakers can use this to their advantage. Clark and Schaefer (1992:281) illustrate this with an example invented by Searle (1969) in which a wife at a party says to the host in the presence of her husband ‘It’s really quite late’. This single utterances might be heard as an apology by the host and as a request (to be taken home) by the husband, in which case, the husband is positioned as auditor of the apology and the host is positioned as auditor of the request, with each of them likely aware of how the other is hearing the utterance. Increasingly, however, as in the case of citizens using cell phones to record their encounters with police, the actual auditors of interactions are not physically present, and who they are and what they might think can only be a matter of conjecture. The immediate auditors are the cell phones themselves. There have also been studies, of course, showing how speakers alter their speech when they know they are being recorded. Gordon (2013), for instance, has written about how participants in linguistics research sometimes orient to tape recorders as overhearers, and Haworth (2013) has studied how police officers in interrogations design their utterances both for their immediate interlocutors and for future audiences in courtrooms that might be listening to tape recordings of them. Where the phenomenon of citizens wielding cell phone cameras in their interactions with police differs both from situations considered by Bell (where auditors are conceived of as secondary ‘presences’ fortuitously introduced into encounters), and from situations where tape recorders or video cameras are more or less static electronic ‘presences’ (which audit interactions in the ‘background’), is that citizens in police encounters actively deploy their devices as interactional resources, using them to manipulate the degree of awareness officers have of being audited and the role these electronic auditors have in the interaction. Because of the control citizens have over where to point their cameras and what to do and say when they are pointing them, cell phones (and the imagined audiences behind them) can be positioned in various participant roles, sometimes as addressees, sometimes as auditors, sometimes as overhearers, and sometimes as eavesdroppers, and the different participant roles cameras inhabit can affect the ways police officers (and citizens) can be called upon to account for their actions. The presence of auditors in interactions alter not just the way people design their utterances, but, more generally, how they are made accountable for what they do and say, and how they manage this accountability. Being accountable, however, does not depend on the presence of an auditor. As Garfinkel (1967) has pointed out, people’s ability to account for their words and actions and communicate how they are making sense of others’ words and actions is the basis for all social interaction. Accountability, from this perspective, is not just a matter of coming to a mutual understanding of what is ‘acceptable behaviour’ in certain kinds of situations, but develops dynamically within and as part of social situations. It is also not just a matter of language, but includes all of the physical ways people make aspects of the situation observable and reportable (Garfinkle, 1967). The presence of cameras alters what I have called (Jones, 2020:90) the ‘infrastructures of accountability’ in encounters between citizens and police in at least three ways. First, as I said above, it provides opportunities for those wielding cameras to change the participation framework of the interaction, visibly introducing into the situation additional participants to whom officers and citizens are responsible for producing accounts. Second, it alters the organization of perception in the situation, giving those wielding cameras and those whose actions are being recorded new opportunities for making certain aspects of what’s going on visible, and thus, meaningful. Finally, and perhaps most importantly, it provides ways for both parties to communicate to each other what aspects of the situation are subject to being accounted for— that is, what aspects of the situation ‘count’. While these infrastructures of accountability are constructed moment by moment by participants as they negotiate their rights and responsibilities around looking and telling at different points in the interaction, they are built upon a foundation of social relationships that inevitably involve historically sedimented forms of inequality and sets of expectations about who has the right to look and who has the right to be seen in different kinds of circumstances, and it is often these regimes of power and visibility that are responsible for how particular interactions came to occur in the first place (e.g. why a particular driver got pulled over), and what the consequences of the interaction will be after it is finished (e.g. whether or not the citizen — or the officer— will face criminal charges). Example 1: ‘It’s my right’ The first example1 involves an attorney -- moonlighting as a Uber driver -- who was pulled over by officers from the Wilmington, North Carolina Police Department and the New Hanover County Sheriff’s Office on the afternoon of February 26, 2017. The driver, Jesse Bright, was transporting a passenger who had visited an address in a supposedly ‘high crime’ neighbourhood and was returning home. Upon being stopped, Bright was informed that his passenger was being arrested, at which point he began videoing the proceedings with his cell phone camera. This prompted one of the officers, Sergeant Kenneth Becker, to ask him to stop videoing, resulting in 1 https://youtu.be/-UQKkYWDUQ4 (still images used with permission of video owner) an extended argument between Bright and the officers, who insisted that videoing them was ‘against the law’ (Schachtman, 2017). As I mentioned in the previous section, cell phone cameras are very different from fixed recording devices (e.g. CCTV cameras) in that they allow users to actively and flexibly deploy them as interactional resources, changing what is being monitored by changing the direction in which they are pointed. In other words, the camera can be made visible as auditing the officer, the citizen, or other aspects of the scene, thereby altering moment by moment the ways citizens and officers are called upon to orient to the camera as an auditor. Although we usually think of the practice of ‘filming the cops’ as a matter of making visible what police officers are doing, for most of this video the driver trains the camera not on the police officers, but on himself as he narrates what is happening. The video begins, for example, with the driver explaining to the camera that he as been stopped and his passenger is being arrested, briefly turning the camera outward to capture the scene, as if to provide evidence for his statement, before returning it to his own face (Excerpt 1). The purpose of the driver directly narrating what is occurring rather than simply pointing the camera at the police and providing the non-present audience with a more ‘objective’ version of events, of course, is that it allows the driver to control what Mirzoeff (2011:1) refers to as the relationship between the ‘seeable and the sayable’, presenting a verbal account of the situation and choosing what aspects of the situation to make visible in order to confirm that account. In this way, the driver is able to provide both an account and an evaluation of the situation, which is sometimes expressed in words, but more often is communicated through facial expressions and gestures (such as headshaking and eye-rolling) as in excerpt 2. Excerpt 1: Camera as addressee 1 Excerpt 2: Camera as addressee 2 Another important affordance for the diver in directly addressing the camera is that it creates intimacy with the non-present audience, who is invited to experience the incident from the driver’s point of view—not just ‘as it is happening’, but ‘as it is happening to the driver’. In this regard, the camera is positioned not just as an addressee, but as a confidant and ally. Finally, by carrying on a conversation with the non-present audience, the driver positions the officers as overhearers of a conversation which is, to a large extent, being performed for their benefit. The message such performances potentially send is twofold: first of all, they highlight for officers the presence of the electronic auditor, and second they make clear to them that they are being excluded from the production of an account of the incident. Of course, officers also have means of positioning drivers as overhearers by performing conversations with other officers, or with non-present participants through electronic devices of their own. In excerpt 3, for instance, in response to the driver refusing to allow officers to search his car, the officer conspicuously calls for a K-9 unit2 to be deployed to the scene. Excerpt 3: Driver as overhearer 2 Dogs trained to sniff out drugs and other contraband. Another important way the driver uses his camera is as a pragmatic means, a tool for producing a range of different kinds of ‘speech acts’ such as warnings and assertions. In this video this tactic is particularly evident at the point in the interaction when the officer asks the driver not to record him (Excerpt 4), at which point the driver momentarily keeps the camera focused on himself, turning his head as if addressing an interruption from a bystander, and then turns his camera to focus on the officer, tracking his movements from the passenger side of the vehicle to the driver side, narrating his own actions as he does so, saying: ‘I can record you…and if you come to this side of the vehicle I can keep recording.’ Excerpt 4: ‘I’ll keep recording…it’s my right’ In this exchange, both the driver and the officer produce justifications for their actions, the driver asserting his ‘right’ to record a ‘police officer on duty’, and the officer claiming (falsely) that it’s against the law. What is more important than these conflicting justifications is the lack of accommodation between the two parties, with the officer issuing bald, on record directives using a casual form of address (‘hey bud, turn that off’, ‘step outta the car’), as well as a threat (‘turn it off or I’m gonna take you to jail’), and the driver conspicuously refusing to comply and calling on the officer to account for his actions (‘what is the law?’, ‘what are you arresting me for?’) Excerpt 5: Non-accommodation It is important to remember that such interactions, at least from the point of view of the driver, who is controlling the camera, are also designed as performances for the non-present audience. And so the accounts that the driver gives (e.g. ‘I’m surrounded by five police officers…I’m scared right now’), and that the officer gives (e.g. ‘you’re being a jerk’) are intended both as responses to their interlocuter’s challenges and as framing devices designed to shape the way the non-present auditor might see the situation. The driver’s ongoing consciousness of the auditor and ongoing positioning of the auditor as an ally in his dispute with authority figures is evident in his subsequent staging of performances designed to provoke the officers, to call them to account for both present actions and past statements, and to position the himself as the ‘reasonable’ party in the dispute. In excerpt 6, for example, the driver makes a point of asking the officer for permission to take a drink -- despite the fact that there have been no demands for him restrict his movements inside the car (as there are in the second example below)—and then turns conspicuously to the camera with a slight smirk. Excerpt 6: Powerade This performative dimension of this encounter is particularly evident in excerpt 7, where the driver calls upon another officer to account for the pervious officer’s insistence that recoding the police is against the law (‘can you tell me what this new law is…’). By doing this, the driver puts the second officer in a rhetorical bind, making him choose between contradicting his colleague or supporting him with what he probably knows is a lie. At first, he attempts to deflect the responsibility for accounting for the law to the driver, saying, ‘you said you are an attorney, is that correct?’ But after the driver continues to press for details about the ‘new’ law, the officer chooses to lie: ‘[they’ve] recently just passed it sir.’ Throughout this encounter the driver performs his scepticism through both is tone of voice and his choice of words, and he ends the conversation by turning back to face the camera and saying ‘seems like a strange law,’ as if offering to the non-present audience an evaluation of the veracity of the officer’s utterance. Excerpt 7: ‘Seems like a strange law’ In his encounters with these first two officers, the driver effectively uses his camera as an interactional resource to assert his legal rights and to position them as behaving unreasonably. Later in the video, however, an apparently more senior officer, uses a different strategy with the driver, one that is not just more accommodative, but also more sensitive to the presence of the electronic auditor (Excerpt 8). Excerpt 8: ‘I’ve never said that to you’ This excerpt begins with the officer approaching the driver and addressing him as ‘Mr. Bright’, a strategy that both communicates respect and calls the driver to account. Interestingly, the driver immediately answers by addressing the officer by his name – ‘Mr. St. Pierre’—which he apparently has read off of the officer’s badge, a move which both accommodates to the officer’s new, more respectful tone and reminds the officer that he too is being held to account for his behaviour. In the conversation that follows the officer attempts to account for the driver having been stopped with a series of questions (‘how long have you been an Uber driver?’, ‘have you been stopped a lot?’) which create the implication that the stop was justified. He then explains to him why the K-9 unit has been called and the procedures that will be followed before the driver can be ‘on his way.’ This explicit account from the officer for what is being done is delivered with a degree of professionalism not evident in the earlier encounters and results in the driver reciprocating by offering his own account of his behaviour, explaining his fear that he might be blamed for something his passenger left in the car and asking the officer, ‘can you understand my frustration?’ What is striking about this officers responses during this exchange is the way he uses metalanguage to create a clear consensus with the driver about exactly what he has said and what he has not said (‘I never said that, did I?’, ‘I asked you a simple question’, ‘Negative, sir, I’ve never said that to you’), a consensus which the officer, of course, knows will be recorded by the driver’s cell phone. Another important difference between this exchange and the earlier exchanges with the other officers is that the driver does not turn to look at the camera once, instead orienting his gaze towards the officer the entire time and even appearing to be holding the phone at a lower, less obtrusive angle. In other words, rather than being used as a tool to challenge or exclude the officer, the camera is positioned more as an objective witness, one to whom the officer has an equal opportunity to appeal with his own account of the situation. Example 2: ‘Look at this’ The second example3 involves two African American brothers, Benjamin and Ryan Brown, who, were pulled over by officers of the Colorado Springs Police Force on March 25, 2015 while driving through a predominantly white neighbourhood, ostensibly for driving with a cracked windshield. The driver, Benjamin, was immediately asked to exit the vehicle and was placed in handcuffs, prompting his brother, Ryan, to begin videoing the encounter from the passenger seat. In the moments that followed, Ryan repeatedly asked to officers to explain why they were ‘being arrested’ in response to officers’ requests that he produce his ID and exit the vehicle. The officers finally opened the passenger door, pulled him out of the car and wrestled him to the ground, all of which was captured on the video. Later Benjamin Brown was charged with driving with an obstructed view and Ryan with obstructing and resisting police officers (Roberts, 2015). While in the video analysed above, the driver did not at first explicitly alert the officers to the fact that he was recording or assert his right to do so until an officer actually told him to stop, in this example the driver calls attention to the fact that he is recording immediately and repeatedly (‘I’m recording this’, ‘I’ve got this on camera’, ‘I have my camera.. and I’m recording this, just to let you know’), using his camera as a means to deliver a warning to the officers that they are being audited and thus, presumably, will be held accountable for their actions (Excerpt 9). 3 https://www.youtube.com/watch?v=pb4DH4P-7yI (Still images used with permission of video owner) Excerpt 9: ‘I’ve got this on camera’ During this exchange, the driver alternates between addressing the officers in ways that demand that they account for their actions (‘what is the reason why you pulled us over officer’), and addressing the camera with his own interpretation of the situation (‘police officers pulled us over for no reason’). As in the first example, the utterances directed at the officers are designed also to be overheard by the camera, and the words directed towards the camera are designed to be overheard by the officers. While the camera is, for the passenger, a ‘weapon’ of sorts, pointed at police in order to constrain their behaviour, police are able to deploy their own (more literal) weapons, in this case, a taser which one officer points in the direction of the passenger window (excerpt 10). Rather than responding directly to this threat, however, the passenger continues to gaze in the direction of the camera and produces his own account of the officer’s actions (‘I’m…bein perceived as a threat’), while at the same time displaying for the camera his stationary, non-threatening body. Excerpt 10: ‘I’m being perceived as a threat’ The last excerpt highlights the fact that often in encounters with police it is as much in the citizen’s interest to document their own behaviour as it is to document the behaviour of the authorities in order to refute any possible later claims that they were acting improperly. This can also be seen in excerpt 11, where, in response to the officer’s directive ‘keep your hands were I can see them’ the passenger both produces a verbal confirmation (‘my hands are visible’) and turns his camera to document the visibility of his hands. Excerpt 11: ‘My hands are visible’ At the same time, the passenger also uses this opportunity to provide for the implied audience a broader perspective on the situation, directing the camera through the open driver side door where his brother can be seen being handcuffed (excerpt 11), all the while narrating the situation: (‘My brother is being put in handcuffs’, ‘We’re being pulled over for no reason’, ‘He still had not identified why he pulled us over’). Excerpt 11: ‘My brother is being put in handcuffs’ Slightly over a minute after the passenger starts recording the interaction, one of the officers opens the passenger door, unbuckles Brown’s shoulder strap, and begins to pull him out of the vehicle (excerpt 12). As this is happening, the passenger directs the camera at the officer who is pulling his arms and continually asks the officer to explain why he and his brother have been pulled over, why he is being arrested, and why he is being pulled out of his car. He manages to continue filming as two officers wrestle him to the ground and press his face into the pavement, and, right before one of the officers reaches over to take the phone out of his hand, he looks into the camera and says, ‘You see this? Excessive force.’ As in the other excerpts from this video, the passenger alternates between positioning the camera as an addressee and positioning it as a witness to what is being done to him. But even when he is not providing commentary directly to the camera, he designs his utterances to the police officers in a way that both provides an ongoing narration of what is happening and calls upon officers to account for their actions (‘Am I being placed under arrest?’, ‘Why are you pulling me out of my car?’, ‘I have no weapons, you have no reason to pull me out of the car’). What is perhaps most striking about this excerpt is the counter narrative that the police articulate as this is going on, saying ‘you’re not under arrest’ and ‘I’m not pulling you,’ even as the officer visibly pulls on the passenger’s arm. This verbal ‘counter-account’ of events, which contradicts both the passenger’s narrative and what appears to be actually happening, is reminiscent of Haggerty and Sandhu’s (2014:12) example of police officers shouting ‘stop resisting arrest’ while beating citizens with their batons. It highlights the fact that the ‘right to look’ and to call someone to account depends not just on what the camera sees, but also on who can claim the right to interpret what is happening, or as Mirzoeff (2011:1) puts it, who can claim ‘the right to arrange the relationship between the seeable and the sayable.’ Discussion and Conclusion Taken together, these two examples highlight a number of common strategies citizens use when deploying video cameras in their encounters with police officers. Perhaps one of the most interesting things to note is how, in both of the videos, the citizens train their cameras on themselves much more than on the police. At first this might seem counter-intuitive. In fact, most cop watching organizations advise third parties who are videoing police encounters to avoid filming the person being stopped and focus exclusively on the officers. A perspective which takes into account the role of audience/auditor design and the importance of producing and eliciting credible accounts, however, helps to explain citizens’ penchant for filming themselves. Of course, the most obvious advantage for citizens to film themselves is to create a document of their own behaviour to counter future possible accusations of aggression or illegality, a strategy that is particularly clear in the second example. Equally important, though, is the fact that, by positioning the camera as an addressee and producing for it a verbal account of the goings on, citizens are able to take control of how what is happening might later be interpreted. It also gives them the opportunity to strategically position the non-present audience either as allies (as in the first example) or as witnesses (as in the second example), and to position themselves as victims by depicting not just what is being done by the police, but what is being done to them. Finally, interacting with the camera as an addressee also gives citizens the opportunity to produce conversations designed to be overheard by officers, potentially heightening their sense of being both scrutinized and excluded. These observations point more broadly to the fact that one of the main strategic functions of the camera is, as Neyland & Coopmans (2014: 4) put it, to identify ‘relevant audiences for the discharge of accountability,’ and to position those audiences in participant roles that help to support particular versions of events. At the same time, of course, the camera is also used to document the actions of the police, and this is done by selectively pointing the camera at officers or panning to provide a broader perspective of the scene. The main advantage here is that the citizen’s control of the camera gives them control over what aspects of the situation are observable and reportable (Garfinkle, 1967). Pointing the camera outward, however, is also a communicative gesture through which citizens can direct ‘speech acts’ such as ‘warnings’ towards officers which are designed to constrain the range of action officers can take and signal to them that they are accountable for those actions. The most important moments in these encounters, however, seem to be those when both citizens and officers are subjects of the camera’s gaze and performing their interaction for the auditor, each designing their utterances in order to elicit certain kinds of verbal accounts from the other party, and each designing their accounts to be both heard (by their interlocuter) and overheard by the non-present auditor. It is in these memoments when citizens and officers directly negotiate the relationship between ‘the seeable and the sayable’, that is, between what is made observable in the situation and whose account of it ends up counting. Despite these similarities, there are also obvious differences between these two videos, not least of which is the fact that in one the citizen ends up going on his way after a bit of inconvenience while in the other the citizen ends up with his face planted in the pavement. The important thing to note, in this regard, is that there is nothing in the ways these two citizens used the various strategies discussed above that can explain this disparity. Both asserted their right to video the encounter and wielded their cameras in similar ways designed to signal to the officers the presence of an ‘auditor’. And both explicitly refused to comply with requests from the officers and called upon the officers to account for the legitimacy of these requests. The result of this was that one of the citizens (the African American one) ended up being perceived as a ‘threat’, whereas the other was simply perceived as a ‘jerk’. The wildly different outcomes can only be explained by the different assumptions about accountability that participants must have brought to these interactions. Both citizens began with the same assumption: that police officers are accountable for informing citizens why they have been stopped and for abiding by certain constitutional restrictions regarding their behaviour towards citizens (restrictions, for example, on search and seizure and on detaining citizens without charging them). In other words, the officers bear the burden of making their intentions visible. The officers in these two cases, however, appear to have approached these two different citizens with very different assumption about their accountability. At no point, for example, is the Caucasian driver ordered to keeps his hands visible, whereas this is one of the first things the African American passenger is asked to do. It is clear, as well, that the passenger himself is wellaware of the assumptions being made about him, to the point that he feels the need to use us own camera to document the visibility of his own hands. These different assumption about accountability and visibility play out particularly dramatically in the ways officers and citizens negotiate the relationship between the ‘seeable and the sayable’. In the first example, the third officer and the driver are able to negotiate a consensus account of the situation, with the driver accepting the officer’s justifications for deploying the K-9 unit and the officer acknowledging the driver’s ‘frustration’. In the second example, however, the passenger and the officers produce contradictory accounts, with the officers insisting that they are not doing things that they are clearly doing. This last point highlights one of the main reasons why technologies like cell phone cameras are not enough to make police accountable for their actions. The ability to make something visible is not enough. One must also have the ability to speak about it, to produce an account which others will regard as legitimate, the power to manage the relationship between ‘the seeable and the sayable’. And, as many African Americans who have filmed encounters with police, some of them much more deadly than the one discussed here, know, this power is almost always afforded to the powerful – the police, their attorneys, the press and politicians – rather than to the victims or witnesses of police violence. As a result, video evidence of police violence against African Americans, going back to the Rodney King case in 1992, almost never result in serious consequences for the officers . As Beutin (2017:17) puts it: As long as visual evidence of police brutality is interpreted through racialized ways of seeing, the practices of counter-surveillance and the discourse of filming the cops remain circumscribed within a larger cycle of repression that continues to reconfigure itself in ways that uphold the legitimacy of the police, and by extension, the racist state. Citizen’s encounters with police officers are moment by moment negotiations in which sometimes the presence of a cell phone camera can alter the infrastructures of accountability in the situation, changing the power dynamics between participants and constraining the range of actions available to them. At the same time, these encounters take place in the context of larger systems of inequality and injustice which inevitably place practical and discursive limitations on how certain kinds of citizens can use cameras to make police accountable. This is perhaps what is behind the unrepentant gaze that Derek Chauvin directed towards Darnella Frazier’s camera as he took the life of George Floyd, the comfortable assumption that no matter how ‘seeable’ his crime was, he and the apparatus of police power behind him would retain control over how it would be seen. References Allan, S. (2013). Citizen witnessing. Cambridge, England: Polity Press. Bock, M. A. (2016). Film the police! Cop-watching and its embodied narratives. Journal of Communication, 66(1), 13–34. Bell, A. (1984). Language style as audience design. Language in Society, 13, 145–204. Beutin, L. P. (2017). Racialization as a way of seeing: The limits of counter-surveillance and police reform. Surveillance & Society, 15(1), 5-20. Brighenti, A. (2007). Visibility: A category for the social sciences. Current Sociology, 55(3), 323–342. Brown, G. R. (2016). The blue line on thin ice: Police use of force modifications in the era of cameraphones and YouTube. British Journal of Criminology, 56(2), 293–312. Brown, P., & Levinson, S. C. (1987). Politeness: Some universals in language usage. Cambridge: Cambridge University Press. Castells M (2007) Communication, power and counter-power in the network society. International Journal of Communications 1, 238–266. Chernikoff Anderson, M., Knutson, T. J., Giles, H., & Arroyo, M. (2002). Revoking our right to remain silent: Law enforcement communication in the 21st century. In H. Giles (ed.), Law enforcement, communication, and community (pp. 1–32). Amsterdam: John Benjamins. Clark, H & Schaefer, E. (1992) Dealing with overhearers. In H. Clark (ed.), Arenas of language use (pp. 248-273), Chicago: University of Chicago Press. Dixon, T. L., Schell, T. L., Giles, H., & Drogos, K. L. (2008). The influence of race in police– civilian interactions: A content analysis of videotaped interactions taken during Cincinnati Police traffic stops. Journal of Communication, 58(3), 530–549. Douglas-Cowie, E. (1978). Linguistic code-switching in a Northern Irish village: Social interaction and social ambition. In P. Trudgill (ed.), Sociolinguistic patterns in British English (pp. 37–51). London: Edward Arnold. Ericson, R. (1995). The news media and account ability In P. Stenning (ed.) Accountability for criminal justice: Selected essays (pp. 135–62). Toronto: University of Toronto Press. Foucault, M. (1975). Discipline and punish: The birth of the prison. New York: Random House. Fyfe, G. and Law, J. (eds) (1988). Picturing power: Visual depiction and social relations. London: Routledge. Galati, A., & Brennan, S. E. (2014). Speakers adapt gestures to addressees’ knowledge: Implications for models of co-speech gesture. Language, Cognition and Neuroscience, 29(4), 435–451. Garfinkel, H. (1967). Studies in ethnomethodology. Englewood Cliffs, NJ: Prentice Hall. Giles, H. (2009). The process of communication accommodation. In N. Copeland & A. Jaworski (eds.) The new sociolinguistics reader (pp.276-286). London: Blackwell. Giles, H., Fortman, J., Dailey, R., Barker, V., Hajek, C., Anderson, M. C., & Rule, N. (2006). Communication accommodation: Law enforcement and the public. In R. M. Dailey & B. A. Le Poire (eds.) Interpersonal communication matters: Family, health, and community relations (pp. 241-269). NewYork: Peter Lang. Giles, H., & Smith, P.M. (1979). Accommodation theory: Optimal levels of convergence. In H. Giles & R. St. Clair (eds.) Language and social psychology (pp. 5-65). Oxford: Blackwell. Giles, H., Taylor, D. M., & Bourhis, R. Y. (1972). Toward a theory of interpersonal accommodation through language: Some Canadian data. Language in Society, 2, 177-192. Giles, H., & Wadleigh, P. M. (1999). Accommodating nonverbally. In L. K. Guerrero, J. A. DeVito, & M. L. Hecht (eds.) The nonverbal communication reader: Classic and contemporary readings, 2nd ed. (pp. 425-436). Prospect Heights, IL: Waveland. Goffman, E. (1959). The presentation of self in everyday life. New York: Doubleday. Goffman , E. (1971). Relations in public: Microstudies of the public order. New York: Basic Books. Goffman, E. (1981). Forms of talk. Oxford: Blackwell. Goldsmith A (2010) Policing’s new visibility. British Journal of Criminology 50, 914–934. Goold, B. (2003) Public area surveillance and police work: The impact of CCTV on police behaviour and autonomy. Surveillance & Society, 1(2): 191–203. Gordon, C. (2013). Beyond the observer’s paradox: The audio-recorder as a resource for the display of identity. Qualitative Research, 13(3), 299–317. Gregory, S. (2010). Cameras everywhere: Ubiquitous video documentation of human rights, New forms of video advocacy, and considerations of safety, security, dignity and consent. Journal of Human Rights Practice, 2(2), 191–207. doi.org/10.1093/jhuman/huq002 Hajek, C., Barker, V., Giles, H., Louw, J., Pecchioni, L., Makoni, S., & Myers, P. (2006). Communicative dynamics of police-civilian encounters: African and American interethnic data. Journal of Intercultural Communication Research, 35, 161-182. Haworth, K. (2013). Audience design in the police interview: The interactional and judicial consequences of aud Society, 42(1): 45–69. Holmes, J. (1995). Women, men and politeness. Longman, London. Mann, S., & Ferenbok, J. (2013). New media and the power politics of sousveillance in a surveillance-dominated world. Surveillance & Society, 11(1/2), 18–34. Mann, S., Nolan, J., & Wellman, B. (2002). Sousveillance: Inventing and using wearable computing devices for data collection in surveillance environments. Surveillance & Society, 1(3), 331–355. Retrieved from http://library.queensu.ca/ojs/index.php/surveillance-andsociety/article/view/3344 Meyer, R. (2015, April 29). Thank God for cell phone video cameras’. The Atlantic. Retrieved from https://www.theatlantic.com/technology/archive/2015/04/thank-god-for-cellphone-videocameras/391688/ Mirzoeff, N. (2011). The right to look: A counterhistory of visuality. Durham, NC: Duke University Press Books. Munro, R. (2004). Punctualizing Identity: Time and the Demanding Relation. Sociology, 38(2), 293–311. Neyland, D. and Coopmans, C. (2014). Visual accountability, The Sociological Review, 62: 1– 23. Neyland, D. and Woolgar, S. (2002). Accountability in action: the case of a database purchasing decision, British Journal of Sociology, 53 (2), 259–274. Potere, M. (2012). Who will watch the watchmen?: Citizens recording police conduct. Northwestern University Law Review, 106 (1), 273-316. Haggerty, K. D., & Sandhu, A. (2014). The police crisis of visibility [Commentary]. IEEE Technology and Society Magazine, 33(2), 9–12. Jones, R. H. (2020). Accounting for surveillance. Journal of Sociolinguistics, 24(1), 89–95. https://doi.org/10.1111/josl.12405 Kopak, A. (2014). Lights, cameras, action: A mixed methods analysis of police perceptions of citizens who video record officers in the line of duty in the United States. International Journal of Criminal Justice Sciences, 9 (2), 225–240. Rickford, J. R. & McNair-Knox, F. (1994). Addressee- and Topic- influenced style shift: A quantitative sociolinguistic study. In D. Biber and E. Finegan (eds.), Sociolinguistic perspectives on register (pp. 235–76). New York/Oxford: Oxford University Press. Roberts, M. (2015, May 6). Ryan Brown video: Racial profiling over cracked windshield, ACLU says. Westword. Retrieved 22 October 2020 from https://www.westword.com/news/ryan-brownvideo-racial-profiling-over-cracked-windshield-aclu-says-6688715 Sandhu, A., & Haggerty, K. D. (2017). Policing on camera. Theoretical Criminology, 21(1), 78– 95. Schachtman, B. (2017, March 8). Wilmington police officer orders Uber driver to stop filming traffic stop, prompts internal investigation. Port City Daily. Retrieved 22 October 2020 from https://portcitydaily.com/wilmington/2017/03/08/can-you-film-on-duty-police-officers-a-recentincident-may-set-the-record-straight/ Searle, J. R. (1969). Speech acts. Cambridge: Cambridge University Press. Thelander, M. (1982). A qualitative approach to the quantitative data of speech variation. In S. Romaine (ed.), Sociolinguistic variation in speech communities (pp. 65–83). London: Edward Arnold.. Thompson, J. B. (2005). The new visibility. Theory, Culture & Society, 22(6), 31–51. Wilkinson, T., & Glazer, M. A. (2016). First amendment under arrest: Photographing police in public places at issue on multiple fronts. Villanova Law Review Online: Tolle Lege, 61, 55-70. Transcription Conventions (.) (0.2) word ^ ↓↑ ↑↓ short pause (less than 0.2 seconds) measured pause stress/emphasis rising intonation falling-rising intonation rising-falling intonation