August 2, 2021
From Nonsite

Rachel Watson responds to Kinstler’s article here.

In February 2020, an exhibition titled Forensic Architecture: True to Scale opened at the Museum of Art and Design at Miami Dade College. It was the first major U.S. survey of the work of Forensic Architecture, a research agency based out of Goldsmiths, University of London, which specializes in mobilizing architectural and forensic techniques to conduct global human rights investigations. The Miami exhibition featured thirteen investigations conducted by the agency since its founding, in 2010, by the Israeli-British architect Eyal Weizman. Curator Sophie Landres described the exhibition as a pedagogical opportunity, stating that the work of Forensic Architecture (FA) “teaches us what tools are available to verify truth claims amidst our currently murky and disorienting media environment.” Those tools included augmented reality, 3D modeling, machine learning, and audio analysis, all of which were used in the creation of the investigations on display. The language of the exhibition press release suggests that these tools make for unimpeachable witnesses: “Amidst what Founding Director Weizman refers to as the ‘dark epistemology’ of a post-truth media environment, [FA’s] meticulously researched, carefully analyzed, and conscientiously presented evidence demonstrates that it is both possible and imperative to verify what is really happening in the world around us.”1 If the contemporary media environment is characterized by a “dark epistemology” that peddles in lies and untruths, FA’s methodology is invoked as a means of securing verification and certainty.

Weizman was prevented from attending the exhibition opening, on February 19, after a U.S. embassy official informed him that a national security algorithm had identified him as a security threat.2 The exhibition, which had also been called a “retrospective,”3 was soon shuttered due to the Covid-19 pandemic. The thirteen featured exhibits included the agency’s investigations into a drone strike in North Waziristan, chemical weapon attacks in Syria, and a police killing in Chicago, among others. They would remain accessible to the public in the form of videos on the FA website, but viewers would not be able to encounter them as artistic installations. The difference between the two viewing experiences is significant: On the website, a video detailing the North Waziristan drone strike shows how FA used smuggled video footage of the strike’s aftermath to reconstruct the path of the missile. At the exhibit, visitors might have approached a scale model of the room where the missile was deployed, positioning themselves at the scene of the crime.

The Miami retrospective marked ten years since FA was founded, the end of a decade in which the agency made its name as both a phenomenon in the art world and a celebrated yet controversial actor in media, law, and human rights. Over the past ten years, FA has collaborated and been commissioned by NGOs, private families, attorneys, newspapers, and international organizations like the European Center for Constitutional and Human Rights (ECCHR). It has submitted evidence to the United Nations, to the Inter-American Court of Human Rights, the Israeli High Court, to German regional courts, and other legal forums, with varying success. All of its work aims to mobilize “new ways of using forensic sciences” to identify how violations of human and environmental rights leave identifiable marks upon the built environment.4 Yet the forensic imaginary that it has popularized remains dramatically undertheorized. This essay aims to articulate the terms, accounts, and descriptions of Forensic Architecture preliminary to a theory of the forensic imaginary, as well as the group and its methods.

A decade ago, Weizman began publicly advancing his understanding of the theory that undergirds his agency’s discipline and practice. “[I]n the field of international law, but also in general political culture, we might have entered a stage when we have become more attuned to the communicative capacity of things, of things speaking, if you like, between themselves and to us,” he said in a 2010 interview.5 By “things speaking,” Weizman was referring to the idea that evidence can seem to speak for itself. A transformation was underway: the creation of international tribunals, coupled with the advance of forensic technologies had “accelerated a shift … from testimony to evidence, from speech to medical data, and sometimes from the account of living people to the testimony of forensic anthropologists on behalf of bones and dead bodies.”6 Courts had begun deprivileging witness testimony in favor of more empirical forms of evidence, such as cell phone data, DNA samples, and satellite footage. Weizman calls this transformation the “forensic shift,” a development he sees as characterized, in an idiosyncratic way, by its own ethics, aesthetics, and rhetoric, and accompanied by a rising cultural interest in the figure of the “forensic-detective.”7 If the twentieth century had brought the “era of the witness,” then, he prophesied, “perhaps the beginning of the twenty-first century will come to be known, above all, as ‘the era of forensics.’”8

The Miami exhibition created an occasion to reappraise these claims, to consider whether a “forensic shift”—or, a deepening of public faith and commitment to the notion of object testimony— has in fact occurred, and if so to what end. Forensic Architecture’s ascent coincided with the increasing prevalence of digital forensic evidence, a development that prompted legal actors to re-evaluate their approach to collecting, preserving, and analyzing evidence. Some legal scholars advocate that digital evidence should be held to equivalent standards of admissibility as peer-reviewed DNA evidence; to that end, practitioners have been trying to develop formalized standards for the admission of digital evidence.9 From 2013 to 2017, the U.S. National Commission on Forensic Science worked to “promote scientific validity, reduce fragmentation, and improve federal coordination of forensic science.”10 The commission began evaluating digital forensics only in 2015, and when it was disbanded two years later one of its primary recommendations was that further research be undertaken on the subject.11 In December 2020, the release of the Berkeley Protocol on Digital Open Source Investigations, published jointly by U.C. Berkeley’s Human Rights Center and the United Nations, attempted to remedy this problem by proposing the first formalized procedures for the use of open-source evidence in international legal proceedings. Launched as part of an event commemorating the 75th anniversary of the Nuremberg Trial, the Protocol “describes the professional standards that should be applied in the identification, collection, preservation, analysis and presentation of digital open source information and its use in international criminal and human rights investigations.”12

Forensic Architecture approaches the processing and presentation of information in a strikingly opposing manner: its investigative methodologies reshape and transform familiar forms of evidence, like spoken testimony and crime scene imagery, such that they cease to be isolated pieces of evidence and become “evidence assemblages,”13 a network of relations and recollections packaged as a coherent whole. This process, as Weizman articulates it, is rhetorical first and aesthetic second. The word “forensic,” Weizman writes in his most recent book, Forensic Architecture (2017), originates from the Latin forum, a “chaotic and multidimensional domain” where justice, politics, and trade were practiced all at once.14 He seems to conceive of forensics in an Aristotelian sense, as a branch of persuasive speech “concerned with making a judgment,”15 where persuasiveness rests in part upon the character of the speaker. In the same 2010 interview, Weizman suggests that forensic architects work primarily as “translators” who help inanimate objects articulate legal claims:

Forensics was part of rhetoric. Rhetoric, of course, is about speech, but forensics does not refer to the speech of humans but to that of objects or things. In forensic rhetoric, objects address the forum. Things need, however, a “translator” to interpret and mediate their speech. Because the thing speaks through, or is “ventriloquized” by, its translator, the object and its translator make a necessary and interdependent duo. To refute a legal/rhetorical statement, it is enough to refute one of the two: to either show that the object is inauthentic or that its interpreter is biased.16

There are various points raised here that one might object to. One might challenge the idea that objects “speak” like human beings, insofar as their “speech” can be “ventriloquized” or “translated.” Or, one might ask, if the “object” and its human “translator” are an “interdependent duo,” then what happens to a person’s speech or meaning when the object disappears? Can they no longer mean what they thought they meant? Or, what if other people also feel the need to “interpret” the object—is their connection any less necessary? One might also query what it means to “refute” an object’s “speech” if the speech was not intended by the object. In what sense could such a refutation make sense? Finally, one might point out that the two potential options—inauthenticity or interpretive bias—are irrelevant if objects cannot be said to have intentional speech.

Be that as it may, Weizman’s approach aims to ensure that, because evidentiary assemblages are composed of artifacts from many independent sources—for example, cell phone videos from a crowd of onlookers—one inauthentic artifact does not impugn the whole collection. It also, Weizman claims, creates a particular kind of social formation: “With each new investigation, a new community of praxis is woven from the meshing of its divergent viewpoints,” he writes. “In socializing the production and dissemination of evidence, it ultimately establishes an unlikely but fundamental commons in which the production facts constitute the foundation of an expanded epistemic community of practice built around a shared perception and understanding of the world.”17 Forensic Architecture claims to mobilize its own ethics, aesthetics, and rhetoric in the pursuit of forensic knowledge. Weizman and his collaborators have suggested that their methodologies operate outside of, or adjacent to, traditional legal procedures, and lend “new material and aesthetic sensibilities” to legal conceptions of proof.18 The result is that Forensic Architecture’s evidentiary assemblages have an uncertain, ambivalent relation to spaces of legal deliberation even while they act upon them. This essay will take a brief look at how this ambivalent relation structures two Forensic Architecture investigations.

1. Killing of Harith Augustus

Still from Forensic Architecture’s “The Killing of Harith Augustus-Milliseconds.”

The Miami retrospective included Forensic Architecture’s investigation into the 2018 police killing of the 37-year-old Harith Augustus in Chicago, produced in partnership with the Invisible Institute for the Chicago Architecture Biennial. Augustus was shot five times during what police called an “investigatory stop,” according to the investigation homepage.19 Forensic Architecture used footage of the killing released as a result of a Freedom of Information Act suit to analyze the shooting at different timescales: milliseconds, seconds, minutes, hours, days, years. Each scale revealed a new dimension of the crime, indicating how the actions of a single officer escalated the confrontation. The initial footage is impartial: Forensic Architecture’s virtual recreation fills in the remaining scenes, restaging the confrontation using digitized, faceless figures. It re-enacts the killing, rewinds it, and re-enacts it once more.

The exhibition debuted at the January 2019 biennial while the Chicago Civilian Office of Police Accountability’s (COPA) investigation into the killing was ongoing. The reconstructive videos had been produced for public consumption; Forensic Architecture’s efforts to reanimate the inanimate, to restage the conditions and context of police killings, might be seen as an attempt to counteract law’s tendency to flatten, minimize, and contain the constellation of causes and effects that constitute a crime. But at the biennial, FA and the Invisible Institute decided not to include videos analyzing the shooting in the public installation. “Over the course of the project, we were confronted by an apparent contradiction between the necessity of looking and the difficulty of showing,” the agency said in a statement. “We became increasingly concerned about presenting graphic scenes of police violence against a black man in the context of an exhibition, about the danger of foreclosing other ways of engaging with the life of Harith Augustus by repeatedly showing his last moments, and about inflicting difficult images on visitors who had not consented to view them.” The videos—or rather, the evidence assemblages—would be posted online instead. “The fact that the visual investigation is not presented in this gallery is an attempt to open a space for critically important discussions about police violence and the politics of representation,” the statement explained.20

It was a striking decision on the agency’s part, an acknowledgement of the limitations of its approach to staging and circulating its investigations, several of which analyze police killings. The choice to post the videos online, where they can be infinitely viewed and infinitely shared, suggests that a paradoxical logic is at work. If keeping the footage out of the exhibition hall preserves space for careful discussion of questions of racism and representation, then the implicit conclusion is that these discussions cannot or will not take place online, where most if not all of the agency’s investigations are posted even while they circulate in art exhibitions and legal fora. The agency’s decision to remove the footage from the Biennial exhibit reflected their concern about “the difficulty of showing,” and suggested that there must be further critical inquiry into the affective afterlives of forensic assemblages, the uncertain relation between video and viewer, and the perhaps limited utility of the forensic gaze.

In September 2019, nine months after the Biennial closed, COPA released additional footage of the killing. In March 2021, the office concluded its investigation and sent recommendations for further action to the Chicago Police Department, which is conducting a separate inquiry into the killing.21 COPA’s recommendations will become public once the police investigation has concluded.

2. Grenfell Tower

<img data-attachment-id="14883" data-permalink="" data-orig-file="" data-orig-size="936,648" data-comments-opened="1" data-image-meta="{"aperture":"0","credit":"","camera":"","caption":"","created_timestamp":"0","copyright":"","focal_length":"0","iso":"0","shutter_speed":"0","title":"","orientation":"0"}" data-image-title="Image 2" data-image-description="

Still from Forensic Architecture’s “Grenfell Media Archive: Call for footage.”

” data-medium-file=”×208.png” data-large-file=”×415.png” loading=”lazy” class=”wp-image-14883 size-large” src=”×415.png” alt width=”600″ height=”415″ srcset=”×415.png 600w,×208.png 300w,×532.png 768w,×256.png 370w,×325.png 470w, 936w” sizes=”(max-width: 600px) 100vw, 600px”>

Still from Forensic Architecture’s “Grenfell Media Archive: Call for footage.”

In June 2017, Grenfell Tower, a high-rise council-owned apartment building occupied largely by low-income, immigrant families, erupted in flames. Sparked by an electrical malfunction in a “large fridge-freezer” on a lower-level flat, the fire became deadly when it spread to the building’s exterior cladding, which contained chemicals that acted as fuel and caused the flames to spread “rapidly up, down and around the building.”22 The building’s architecture—one narrow stairway and a single set of elevators, surrounded by combustible plastic panels—accelerated the disaster. Seventy-two people died as a result of the fire.23

In March 2018, Forensic Architecture announced that it was launching a new project: the agency would reconstruct what happened at Grenfell Tower, and solicited cell phone images and videos of the conflagration from the public. Grenfell was precisely the kind of crime scene that FA had been created to investigate. “Our aim is to create a powerful and freely-available resource for members of the public to explore and better understand the events of the night of the fire,” the agency explained in a news release. Additional funding for the project had been provided by The Victoria & Albert Museum. “These pieces of footage will become a continuous ‘3D video’ of the fire, mapped onto our architectural model of Grenfell Tower. The model will allow the user to investigate the fire, and will sit within a web platform which will ultimately act as a freely available public resource.”24 The resulting “media archive,” which is still to be released, would provide a comprehensive public record of the event.25

By the time FA put out its call for footage, a public investigation of the event already existed in the form of the Grenfell Tower Inquiry, a government panel commissioned by the prime minister to investigate the causes of the disaster and identify responsible parties. The Inquiry would post documents, witness testimony, photographs, and footage to its website; it was a formal legal proceeding, with a dedicated team of counsels and solicitors. Under the “Evidence” tab on the Inquiry’s homepage, one can scroll through cell phone footage, depositions, and photographs submitted to the court, organized according to “witness category,” date, and type. “Witness Category” contains a field of acronyms: BSR (Bereaved, Survivors, and Residents), LAS (Life Administration Services), MET (Metropolitan Police), KCTMO (Kensington & Chelsea Tenant Management Organisation). Here, too, witnesses become “exhibits,” their testimonies and cell phone videos arranged into a uniform field of squares.26 What Sandra Ristovska calls the “thinness” of visual evidence—its flatness, incompleteness—is palpable as one clicks between the isolated video fragments.27 Several witnesses understandably began to point their cell phone cameras at the flames only once their physical safety had been secured. They capture smoke pouring into the hallways, burning debris falling to the ground, rescue crews approaching the windows with their ladders. The Inquiry’s unaltered presentation of these evidentiary fragments at once flattens and enlivens, preserving the “thinness” that acts as an index of the person behind the camera, the human behind the screen.

The FA media archive of the Grenfell disaster was formed outside of, or perhaps adjacent to, this public legal forum. It is a counter-archive, a product of what Weizman calls “counter-forensics,” a forensic practice that aims to operate as a form of political resistance, to “reverse the forensic gaze and to investigate the same state agencies—such as the police or the military—that usually monopolize it.”28 Counter-forensics, akin to the filmic practice of “shooting back,”29 challenges law enforcement authorities’ monopoly over the “facts” of a crime. The Grenfell Media Archive is accompanied by a disclaimer that hints at this uneasy relation, stipulating that the agency would “support the BSR (Bereaved, Survivors, and Relatives) and their representatives in a number of ways in their ongoing pursuit of accountability and transparency in relation to the Grenfell Tower fire, in the context of the Grenfell Tower Inquiry and related legal avenues.”30 The agency thus began creating its alternative repository of evidence; out of the many fragments of testimonial media objects it received, it would create a single, seamless whole, another evidentiary assemblage. These assemblages fill narrative holes in their source material, uncover missing plot points, and “establish relations, between, say, digital photographs, material ruins, remains of ammunition, and human testimony,” as Weizman puts it.31 This approach is framed as a technique that liberates human and nonhuman witnesses, a way of liberating both human and nonhuman witnesses from the straitjacket of legal procedure. In the still-to-be-released final product, the crowdsourced footage of the Grenfell Tower fire will be patched together and projected back upon a virtual rendering of the burning building. The real Grenfell Tower burned for twenty-four hours; its digital twin will burn online continuously, indefinitely.


4.&  Eyal Weizman, The Least of All Possible Evils: A Short History of Humanitarian Violence (London: Verso, 2017), 114.
5.&  Eyal Weizman and Tina Di Carlo, “Dying to Speak: Forensic Spatiality,” Log, no. 20 (Fall 2010): 125.
7.&  Weizman, “Dying to Speak,” 126.
8.&  Weizman, “Only the Criminal Can Solve the Crime,” 13, 15.
9.&  Jill Slay et al., “Towards a Formalization of Digital Forensics,” in Advances in Digital Forensics V, ed. Gilbert Peterson and Sujeet Shenoi (International Federation for Information Processing, 2009), 37–47. See also Rodney McKemmish, “When is Digital Evidence Forensically Sound?,” in Advances in Digital Forensics IV, ed. Indrajit Ray and Sujeet Shenoi (International Federation for Information Processing, 2008), 3–15.
13.&  Eyal Weizman, Forensic Architecture: Violence at the Threshold of Detectability (New York: Zone Books, 2017), 58.
14.&  Weizman, Forensic Architecture, 65.
15.&  Aristotle, On Rhetoric: A Theory of Civic Discourse, trans. George A. Kennedy (New York: Oxford University Press, 1991), ch. 1, book 1.
16.&  Weizman, “Dying to Speak,” , 126.
18.&  Eyal Weizman, “Introduction: Forensis,” in Forensis: The Architecture of Public Truth, (London: Sternberg Press, 2014), 9.
27.&  Sandra Ristovska, “The Rise of Eyewitness Video and its Implications for Human Rights: Conceptual and Methodological Approaches,” Journal of Human Rights, 15, no. 3 (2016): 356.
28.&  Weizman, Forensic Architecture, 9.
29.&  Faye Ginsburg, “Shooting Back: From Ethnographic Film to Indigenous Production/Ethnography of Media,” in A Companion to Film Theory, ed. Toby Miller and Robert Stam (Oxford: Blackwell, 1999), 295–322.
31.&  Weizman, Forensic Architecture, 58.