INTRODUCTION
Eyal Weizman

More information about the book The Police Shooting of Mark Duggan (Cabinet Books and ICA, 2021) is available here. A roundtable organized for the book and featuring Adam Elliott-Cooper, Temi Mwale, Stafford Scott, Marcia Willis Stewart, and Weizman is here.

Mark Duggan was shot dead on 4 August 2011 in Tottenham, north London. The shooter, a firearm police officer known by the cipher V53, was part of a thirteen-person undercover unit sent to arrest him. The national police watchdog, the Independent Police Complaints Commission (IPCC), misinformed the press about the incident, saying that Duggan was killed with a gun in his hand after he shot and injured an officer. Two days later, on 6 August, when the police had still not contacted Duggan’s next of kin, Simone Wilson—the mother of three of his children—and her supporters in the local community marched on Tottenham police station, demanding answers and protesting ongoing police brutality and discrimination against Black Londoners. When the family’s calls for accountability over the killing of Duggan went ignored, the protest grew into a confrontation with the police that challenged this status quo. From Tottenham, it spread quickly throughout England. The scale and force of the police response to this countrywide uprising against police violence were overwhelming; over the next four days, more than three thousand people were arrested. In what seems in retrospect like a path toward promotion, Boris Johnson—then mayor of London and now prime minister—presided over the police repressions, while the director of public prosecutions at that time, current Labour leader Keir Starmer, ran court sessions twenty-four hours a day to administer “conveyor-belt justice” to those who had been arrested, ultimately jailing fourteen hundred people.¹ Speaking in 2014, Johnson went further still, saying that the police should have “come down much harder” and gotten “medieval immediately on those people.”²

• • •

On 4 August, officers from London’s Metropolitan Police received information that Duggan was about to pick up a gun from an address in east London. Driving four unmarked cars, the team began following the minicab in which he was traveling shortly after he had collected the gun, forcing the vehicle to stop as it entered the neighborhood of Tottenham.³ Though the officers tried to get out of their vehicles before Duggan could leave the minicab, he managed to make it onto the pavement. The details of what took place from this point on—precisely in what circumstances Duggan was shot—are contested. Only about a second-and-a-half elapsed between Duggan’s exit from the minicab and the shot that killed him, and yet this interval is one of the most controversial and consequential gaps in the recent history of London.

The shooting was not recorded on camera. The Metropolitan Police would require that body cameras be used during armed operations only following the conclusion, and as a direct result, of the inquest into Duggan’s death.⁴ The earliest videos—shot from a nearby flat and an overhead police helicopter—do not begin to capture the scene until after the incident. The police did have video cameras with them but only started recording once they began to administer first aid to Duggan, who at that point may no longer have been alive.

In 2018, Forensic Architecture was asked by the legal team representing the Duggan family to investigate the case in the context of the family’s civil suit against the Metropolitan Police. When examining a scene of police killing, Forensic Architecture typically generates a three-dimensional model in which we locate the cones of vision from available videos—recorded by police-worn cameras or security cameras, for example, as well as by passersby—each of which depicts the incident from a different angle. Such models help us study what each video shows, what remains hidden between the various footages, and what the relations are between events captured by the different cameras.

The Duggan case posed a new challenge for us insofar as the evidence overwhelmingly consisted of words: verbal descriptions provided by the thirteen police officers, each positioned at a different location across the scene, and by the minicab driver. These are to be found among thousands of pages of documents, which include statements and reports provided to the IPCC and transcripts of evidence given at the coroner’s inquest.

Our investigation focused primarily on examining the truthfulness of these testimonies. This required comparing the police officers’ accounts with our models, and other material evidence, to show what they could and could not have seen at any given moment. Importantly, it also involved textual analysis, which entailed searching in the transcripts not only for divergences, differences, and contradictions between different accounts, but also for places where separate narratives converge so closely as to suggest the coordination of versions.

Though there were no videos of the shooting, each officer recounted the incident as if they were narrating videos that had been recorded from the position of their eyes onto the hard drives of their memory. Different officers described their gazes panning, zooming, and focusing; most importantly, they spoke of their ability to pause their recollection and divide the incident into separate “freeze-frames.” This meant that our prior experience of counter-investigating police brutality using multiple simultaneous videos could finally provide an entry point for assessing the possibility or impossibility of what was being described, albeit with two caveats: committing events to human memory is not as mechanical as transcoding them onto a tape, and the testimony of a police officer delivered under the threat of prosecution might not necessarily be faithful to that memory. This is not to suggest that videos present events objectively; they too produce their own biases in the way they are framed and edited, and even in the way different lenses distort reality in different ways.

While the incident was described from multiple points of view, the testimony of V53—the officer who shot and killed Duggan—is undoubtedly the “main shot.” According to V53, he had a “lovely peripheral vision” of the scene once Duggan was on the pavement.⁵ When he supposedly first saw the gun in Duggan’s hands, his gaze, he said, shifted into “tunnel vision” and remained “focused” on the gun. To the lawyers cross-examining him at the coroner’s inquest, V53 described his recollection in these terms: “It’s like if you’ve got Sky Plus or a video recorder, it’s where you start pausing things, and in my head the world had stopped.” From one freeze-frame in particular, he claimed to have identified the gun well enough to describe it in detail: “I can see the handle of the weapon, I can make out the trigger guard, I can make out the barrel, and it’s side-on to his body and there’s a black sock covering that weapon.” In describing that “split second”—a temporal duration he would invoke more than a dozen times during cross-examination—he testified that “I had an honest held belief that he was going to shoot me.” After firing the first round at Duggan and discovering that, despite the shot, the gun was still “pointing towards my direction,” V53 fired a second time and saw Duggan falling backward.

If V53’s testimony described events as they actually happened, the gun should have been found next to where Duggan fell, mortally wounded. But it was not. It was found eight minutes later, and seven meters away, on the other side of a metal fence in an untended grassy area beside the road. Neither V53 nor any of the other officers at the scene could explain how the gun ended up in the grass. No officers said they had seen the gun being thrown, including the four positioned closest to Duggan, who asserted that they had not taken their eyes off of him at any time. This mysterious gap in the narrative during which the gun disappeared was itself presented by V53 as if he were describing a vanishing act from early cinema, or as if one frame from the series of freeze-frames that made up his testimony had gone missing: “In the course of like a split second, one second the gun is there and the next second, when I looked and reassessed, the gun is not there.”⁶

A character in a Jean-Luc Godard film famously said that “cinema is truth, twenty-four times a second.” But the smallest unit of a film is not the frame alone but also the blank interval between frames, the omission of an image for a small fraction of a second.⁷ And thus the gun that mobilized the entire events of that day in August—that initiated the original monitoring, the calling of the armed unit, the car chase, and the interception; the gun that purportedly had the eyes of the officers “glued” to it as Duggan exited the minicab, and which, according to initial reports, had shot a bullet at a police officer—vanished in the gap between two frames, only to reappear minutes later somewhere completely different.

After two years of being kept in police custody and forensic labs, the gun reappeared again, now presented as evidence to the jury at a coroner’s inquest. It was at this time that the public learned that neither Duggan’s fingerprints nor his DNA had been found on the gun. Yet after a four-month process, on 8 January 2014, the inquest’s jury concluded that Duggan’s death had been a “lawful killing.”⁸ The crucial question the judge instructed the jury to answer was: “Did V53 honestly believe or may he honestly have believed that at the time he fired the fatal shot, that he needed to use force to defend himself or another?”⁹

The legality of the use of lethal force was to be determined only in relation to whether V53 thought he perceived the gun aimed at him “at the time he fired the fatal shot.” That the judge employed the term “at the time” is no coincidence. “Split-second” decisions are evaluated in relation to “the moment” in which danger is perceived “honestly and instinctively”; context, consequences, and retrospectively obtained information are to be put aside.¹⁰ In asking the jury to render their verdict with regard to, and solely with regard to, this “split second,” the judge seems to have himself employed the filmic imaginary existing in the law, one that divides time into discrete freeze-frames. That this imaginary was already embedded in the law might well have been the impetus for V53 to deliver his testimony in such a way.

• • •

One of the best-known references for thinking about the manipulation of videographic evidence in police brutality cases is the 1991 beating of Rodney King, a Black motorist, by Los Angeles Police Department officers. When the video of the incident—shot on a Sony Handycam by George Holliday, a nearby resident—was broadcast, it ignited widespread protest. The four police officers involved were acquitted when their lawyers developed a misleading, if innovative, visual strategy. Rather than play the video continuously, they chose to show a series of freeze-frames, decontextualizing each frame and allowing the officers to point to small movements by King—such as his body twisting in pain or him trying to regain his balance—as evidence of him “resisting arrest.”¹¹ The video freeze-frames of the beating of Rodney King, like the mental freeze-frames of the killing of Duggan, cut these incidents out of a temporal continuity that includes their immediately preceding circumstances, their consequences in the future, and certainly out of the long-term history of police brutality against racialized communities.

The “split second” justification is a frequent last-ditch argument made by police forces worldwide whenever it might be demonstrated that they have used excessive force. In the US, this defense was enshrined in the 1989 Graham v. Connor Supreme Court ruling, where the court acknowledged that “officers are often forced to make split-second judgments—in circumstances that are tense, uncertain, and rapidly evolving.”¹² Government agents taking such split-second decisions are, in most states, offered “qualified” or “good faith” immunity from civil liability.¹³ In criminal trials, juries have tended to identify with the police officers and therefore accept many encounters as “tense, uncertain, and rapidly evolving,” allowing for officers’ errors and giving them the benefit of the doubt. A reliance on the “split second” defense is widespread; in addition to having encountered it in the US and the UK, we have experienced some version of it in our dealings with the legal systems of Israel, Greece, and Turkey. We have also encountered it from local politicians. Responding on LBC radio to Forensic Architecture’s request that the investigation into the killing of Mark Duggan be reopened, London mayor Sadiq Khan said: “All I can say is that it was a tragic loss of life, but I’ll tell you this, and I hope Londoners—and it’s not a comment on Mark Duggan or what happened there—will understand this, but often police officers make split-second decisions, and we saw on Friday last week officers making split-second decisions on London Bridge. And they genuinely are split second.”¹⁴

The “split second” argument is such a powerful component of defense strategies in cases of police brutality that it still needs to be invoked and even implicitly endorsed, if only to be rejected as not pertinent, when a case has nothing to do with instantaneous decision making. To take a notable recent example, Jerry Blackwell—one of the prosecutors in the trial in which Derek Chauvin was convicted of the murder of George Floyd—presented the nine minutes and twenty-nine seconds during which Chauvin slowly suffocated Floyd as precisely not a “split second” case. As he explained, “this case is not about split-second decision-making. In 9 minutes and 29 seconds, there are 479 seconds, and not a split second among them.”¹⁵

The logic of the “split second” argument is entirely preemptive. It does not refer to situations of all-out confrontation, but to those supposedly pregnant with the potential for violence to erupt. In a split second, the officer must decide which of multiple possible futures is most likely to take place. A hand reaches for a pocket: is it drawing out a gun, a license, or a phone? Will the person shoot or stop? The argument cannot rely on evidence, as the threat never materialized. Rather, it takes the void left by the absence of objective evidence as to what would have happened had the officer not reacted and fills it with arguments about the officer’s subjective sense of danger, or claims thereof. This fact makes it hard to convict officers against their word as long as they insist that they perceived a threat to their and others’ lives. Police have therefore learned to call on the “split second” argument in justifying their deployment of lethal force, and the term, together with the filmic imaginary it conjures, has been invoked thousands of times in the US.¹⁶ From the perspective of communities exposed to police brutality and murder, the “split second” is the most dangerous temporal designation: when it is invoked, one can be sure somebody has been killed or injured.

The split second thus becomes a duration in which killing might not be considered murder. The impunity and immunity it offers are reminiscent of the “space of exception” as conceived by Giorgio Agamben in relation to the concentration camp, or of the colonial frontier as described by Achille Mbembe, where the law, and in particular the prohibition on killing, is suspended.¹⁷ Instead of suspending the prohibition on the murder of a particular class of people within a circumscribed space, the legal conception of the “split second” circumscribes this exception in time.

The precise duration of the “split second” exception has, however, not been specified. Is it a half, quarter, tenth, or thousandth of a second? In fact, the split second is not so much an indication of a span of time as it is an attempt, on the part of the law, to leave elastic the threshold of perception. It is considered to be the indivisible unit of legal time or the building block of human perception, analogous to the way a freeze-frame is perceived as the elemental unit of cinematic time. Both these durations black-box, and therefore place beyond legal judgment (or film criticism), anything that happens under their temporal threshold.

Though law hasn’t done so, both science and cinema have, over the years, attempted to better understand the threshold of perception. In the nineteenth century, the duration of a “tenth of a second” was believed to be “the elemental unit of human consciousness,” describing the lag between stimulus, sensation, and response time.¹⁸ Cinema would exploit this duration in relation to “the visual threshold” needed for freeze-frames to “fuse and appear to move” continuously.¹⁹ But just like different people may possess faster or slower response times, different sensitivities of film (or digital sensors) allow for different exposure times, which in turn are components in the enabling of different frame rates: cinema traditionally works with twenty-four frames a second and security camera footage ten frames a second, while professional video cameras now record at sixty frames a second.

In another investigation into police killing, we sought to better understand the components of this otherwise legally indivisible perceptual duration. The Chicago Police Department shooting of Harith Augustus, a barber in the city’s South Shore neighborhood, was likewise defended as a “split-second” decision. Professor Tiago Branco, a neuroscientist from University College London, helped us dive into the scale of milliseconds by breaking down the killer cop’s temporality of cognition, sensory transmission, and muscle response time. It can take between a hundred to four hundred milliseconds to parse and understand sensory input as evidence of intent, he said; a further hundred milliseconds or so are needed to process the crucial decision to shoot or not.²⁰ The faster the process, the more prone to bias it is. “Instinct,” Branco pointed out, “is a short circuit to bias.”²¹ Once a decision is made, it takes approximately another fifty milliseconds for a message to travel from the brain to the muscles, and roughly fifty additional milliseconds for the finger muscles to contract and pull the trigger. The entire process of an extrajudicial execution thus unfolds in a fraction of the duration that the legal process involving capital punishment takes to arrive at and implement the death penalty. And it weighs evidence, renders judgment, and enacts it not in relation to a crime a person has committed but to one that the police officer believed at the time they might.

Responses to a sense of danger perceived, as the law puts it, “honestly and instinctively” are described as if they were primordial and unconscious, as if they were the manifestation of a natural impulse for self-preservation. But police responses are learned, acquired through teaching and training. They are also culturally and politically conditioned, and prey to racial bias and political context. The probable future from the perspective of police officers will be determined to a large extent by their sense of fear and arousal, which might themselves be conditioned by institutionally engrained racist culture and dominant “common sense.”

An exaggerated sense of fear is a constant feature in the long history of racialization. The brutal violence to which the colonized and enslaved were subjected was often argued as a matter of preemptive self-defense against the inherently violent nature of racialized people always about to erupt. To forestall their possible future violence, for which little evidence was ever needed, actual violence was applied in the present. As Marcia Willis Stewart, the Duggan family’s lawyer, says in these pages, “Usually when there’s a [police] encounter with a Black man, you hear that they’re the strongest, most violent, most aggressive.” It is in this way that within the “split second”—the molecular level of legal time—one could find the long durational history of colonialism and empire and the racist rage fueled by the sense of fear it provoked. Counter-investigating police use of lethal force, as this book aims to do, must try to look into the microscale of an incident, open the borders of its reductive temporal frames, and connect it to the world of which it is a part.

• • •

The jury in the Duggan case spent more than fifty days deliberating over the “split-second” moment and the issues surrounding it, and in the end accepted that “at the moment” V53 fired, he had an “honest belief” that Duggan held a gun. In a criminal case, consideration of self-defense does not require that such a belief be objectively reasonable, and the jury was thus never asked whether V53’s belief was realistic in their view.²² The question of how the gun got to the grass was secondary, though the jury accepted that “more likely than not, … Mark Duggan threw the firearm as soon as the minicab came to a stop.”²³ The jury’s paradoxical conclusion that V53 could be believed when he said he had seen a gun that they themselves had deemed was not there could be taken to imply that, for them, it was reasonable enough for a police officer to imagine a gun in the hand of a Black person. Their ruling assumed that a racist perception was reasonable in these circumstances, unwittingly admitting that “honesty” was in fact a confirmation of racism.

A year later, the IPCC published its final report and offered a different explanation for the paradox than had the jury in the coroner’s inquest, aligning its version of events even closer to the testimonies of V53 and the other police officers. According to the IPCC’s findings, Duggan did leave the minicab with the gun in his hand, and was aiming it at V53 when the officer fired the first shot. V53 was not mistaken in his belief. The most likely explanation for the gun’s position on the grass, they wrote, is that Duggan was in the process of throwing it when he was shot a second time. They believed V53 when he said he didn’t see the throw, despite his assertion that his eyes were “focused” on the gun. Both the inquest and the IPCC dismissed another possible explanation for the location of the gun on the grass: that police officers had taken it from the minicab and planted it there. Forensic Architecture’s investigation, as the following pages will demonstrate, arrived at the conclusion that this scenario is the most probable.

In 2018, Duggan’s family launched a civil claim against the Metropolitan Police.²⁴ Whereas in the context of criminal law the police would be able to argue self-defense if they had an honest, if mistaken, belief that their lives and those of others were in danger, in the context of civil law, it would not have been enough for their belief to be honestly held: it also had to be reasonable for them to hold such a belief. This test is sometimes referred to as “subjective objectivity.” “Subjective,” because it regards only the information available to a police officer from their perspective; “objectivity,” because it tries to assess whether claims for this perception could be supported by facts. As cases such as Duggan’s move from criminal to civil trials, the weight shifts from perceptual psychology back to the material world. It was in support of this shift that the family’s legal team approached Forensic Architecture to help spatially reconstruct the scene, otherwise obscured by a mountain of words.

Our investigation demonstrated the implausibility of both the inquest jury’s conclusion that V53 could be believed when he said he had seen a gun in Duggan’s hand and the IPCC’s determination that Duggan was holding a gun when shot. (The former examined the officer’s perception; the latter the presence of the gun). It also showed how the officers could have planted the gun in the grass, despite both inquiries discounting this possibility. The reconstruction and accompanying analysis that we carried out helped both interrogate the scene and reflect on the processes established to rule on the legality of the killing.

Aided by our report during pretrial negotiations, the family reached an out-of-court financial settlement in 2019, and Forensic Architecture subsequently worked with the activist group Tottenham Rights to present the investigation to the local community in Tottenham’s old town hall. In February 2020, we met with members of the IOPC (the Independent Office for Police Conduct, the oversight body that has in the meantime replaced the IPCC) in our office to present our methods and findings, and then sent them our report with the demand that they reopen the investigation. More than a year later, in May 2021—as this book was about to go to press—they informed us of their decision not to reinvestigate the case.²⁵

In her response to the IOPC ruling, Marcia Willis Stewart, who represented the Duggan family throughout the legal process, commented: “Unfortunately, it appears that the courage required to confront and follow up the implications of [the new] evidence remains signally lacking in the IOPC today. … Like the IPCC, the IOPC seems unable or unwilling to fulfil its responsibilities in relation to contentious deaths at the hands of the police. The consequence is … that the IOPC lack the confidence of Mark’s family and that of other families in their position.”

Despite the IOPC’s attempt to put an end to the case, however, the gun—both real and imagined, flickering between visibility and invisibility—will continue to haunt London, and especially the neighborhood of Tottenham, for years to come.

Thanks to Stafford Scott, Raju Bhatt, Daniel Machover, Marcia Willis Stewart, Elizabeth Breiner, Robert Trafford