Numbers [are…] essential and insufficient, dehumanizing and reparative, necessary and complicated.
—Diane Nelson, Who Counts? Footnote 1
On December 18, 2021, the New York Times released more than 1,300 declassified civilian harm assessments from Combined Joint Task Force—Operation Inherent Resolve (CJTF-OIR), which were obtained through several Freedom of Information Act requests and subsequent lawsuits.Footnote 2 The files run to more than 5,400 pages and contain various CJTF-OIR memoranda instructing officials to evaluate and enumerate coalition-caused civilian casualties.Footnote 3 The documents also contain over 1,300 classified assessments that CJTF-OIR officials conducted into coalition airstrikes in Syria and Iraq between 2014 and 2018, including 216 civilian harm claims that coalition officials deemed credible and more than 1,100 cases that the coalition dismissed as non-credible. These confidential assessments served as the basis for the numerical information that the CJTF-OIR released in its Monthly Civilian Casualty Reports, which claim that at least 1,395 civilians were killed and more than 353 civilians were injured during this period.Footnote 4
Casualty figures are essential to the ethical and legal debates surrounding armed conflict.Footnote 5 These counts provide the “raw material” necessary to evaluate specific strikes and determine whether civilian impact remains within acceptable legal bounds. However, counting casualties is rarely straightforward. Azmat Khan—who led the New York Times investigation—noted that the confidential assessments revealed a startling disconnect between coalition promises and coalition practices, with flawed intelligence and failed targeting causing considerable harm to Syrian and Iraqi civilians.Footnote 6 At the same time, she also argued that the confidential assessments raise concerns about how CJTF-OIR officials count civilian casualties and whether this information is being used to hold the CJTF-OIR accountable when civilians are harmed. She discovered cases that were dismissed because CJTF-OIR officials had misread coordinates, mistranslated locations, or missed important evidence.Footnote 7 Moreover, she noted that the confidential assessments seemed not to hold the coalition accountable: no wrongdoing is identified, no disciplinary action is recommended, and fewer than 12 percent of assessments recommend further investigation.Footnote 8
In our contribution to this special section, we will examine coalition efforts to enumerate civilian harms, focusing specific attention on the rationale behind the count, the policies that outline why and how the count should be conducted, and how the count was actually conducted. Numerical information is often seen as essential in ethical debates because it supplies the raw material that is required to ascertain what is happening and whether it is acceptable.Footnote 9 In armed conflicts, accurate casualty figures are considered a critical tool when it comes to accounting for the harms inflicted upon civilians, so that militaries can be held accountable if these harms violate established norms and laws.Footnote 10 While we agree that these numbers are indispensable, we will also contend that these numbers are not sufficient when it comes to ethical considerations because this numerical information tells us little about how civilians experience these harms. Moreover, we will argue that the categories used to enumerate civilian casualties can work to conceal the indirect, cumulative, and reverberating harms that civilians experience—such as long-term injuries, exposure to infectious diseases, and damage to the built environment—that can be much harder to measure.
The first section will outline the rationale behind the CJTF-OIR counting, drawing on the declassified documents to show that coalition officials were more concerned with military success than civilian protection. The second section will outline how the CJTF-OIR counted civilian casualties, focusing specific attention on the categories used to organize this information. In addition to undercounting civilian casualties, we will argue that the coalition’s counting methods imposed a narrow understanding of what counts as civilian harm. Our final section will turn to a different count, assessed by the pre-eminent London-based not-for-profit organization Airwars and also conducted during this conflict. In contrast to the coalition, which prioritized martial considerations above ethical concerns, Airwars orients tracking and enumerating civilian casualties in order to hold militaries more accountable. Furthermore, we will outline how Airwars foregrounds the narratives behind these numbers to outline how civilians experience various wartime harms. Our contention that numbers are necessary but insufficient shifts the focus to victim-centered narratives, which can be used to contest a dehumanizing logic that reduces civilians to mere statistics, enabling a more relational response to the harms inflicted on the civilian population.
Ethics and Enumeration
Various moral norms and legal codes regulate the violence inflicted in armed conflict, including specific rules that are supposed to protect the civilian population.Footnote 11 Article 51 of the Additional Protocols to the Geneva Conventions, for example, stipulates that “the civilian population as such, as well as individual civilians, shall not be the object of attacks.”Footnote 12 Article 51 also prohibits indiscriminate attacks that are not directed at a specific military objective, as well as attacks that are “expected to cause incidental loss of civilian life… which would be excessive in relation to the concrete and direct military advantage anticipated.”Footnote 13 In addition, Article 57 stipulates that belligerents should take constant care during planning and executing an attack to spare the civilian population, including doing everything feasible to verify that their targets are legitimate and “tak[ing] all feasible precautions in the choice of means and methods of attack with a view to avoiding, and in any event to minimizing, incidental loss of civilian life, injury to civilians and damage to civilian objects.”Footnote 14 There is also a broad consensus within the literature on civilian harm that counting casualties is essential when it comes to observing possible violations of international humanitarian law and ensuring that militaries can be held accountable when civilians are harmed. After all, we cannot be certain that belligerents are discriminating between combatants and noncombatants when directing their attacks and ensuring that the incidental harm to civilians is not disproportionate to the anticipated gains if it is unclear whether civilians were harmed in an attack, and how many civilians were harmed.Footnote 15
While international humanitarian law does not require belligerents to count civilian casualties, several theorists have raised concerns when belligerents refuse to do so, warning that belligerents might be seeking to obfuscate the harm inflicted on civilians in order to camouflage immoral or unlawful actions, including possible war crimes.Footnote 16 Other theorists have suggested that this reluctance further indicates that belligerents do not value the lives that have been extinguished.Footnote 17 When belligerents have refused to count civilian casualties, external organizations have often intervened to document and disseminate this information, with several organizations collaborating to create standards for recording civilian casualties and to harmonize data collection across different organizations and different conflicts.Footnote 18
This section will argue that militaries have tended to prioritize martial considerations above moral considerations when counting civilian casualties, often using these figures for consequence management rather than holding themselves accountable.Footnote 19 In 2008, for example, the International Security Assistance Force (ISAF) in Afghanistan established a Civilian Casualty Tracking Cell (CCTC) to collate information from the various regional commands in a centralized spreadsheet.Footnote 20 Yet there was a martial imperative underpinning this endeavor: maximizing the effectiveness of the attacks. ISAF officials were not concerned with civilian harm per se but with the risk that civilian harm could become an impediment to mission success.Footnote 21 Anxious that civilian casualties were alienating the local population, ISAF officials imposed several restrictions on combat operations to minimize civilian casualties alongside several post-incident consequence management techniques, which included ex-gratia payments to Afghan civilians (or their surviving family members) and official apologies.Footnote 22 CCTC data was essential to these consequence management techniques because it enabled ISAF officials to respond to allegations it considered to be credible and to refute allegations it considered to be noncredible.Footnote 23 In their directives, ISAF officials were clear that the purpose of counting civilian casualties was not focused on making the coalition more accountable.Footnote 24 As General David Petraeus explained, “Every Afghan civilian death diminishes our cause. If we use excessive force or operate contrary to our counterinsurgency principles, tactical victories may prove to be strategic setbacks.”Footnote 25
CJTF-OIR officials were also keen to count civilian casualties (“CIVCAS”) during Operation Inherent Resolve and ascribed particular importance to consequence management in their policies on reporting and responding to this harm. On September 27, 2017, CJTF-OIR officials circulated a memorandum warning that “the incidence of CIVCAS can… temper or completely blunt the benefit of counterterrorism efforts [because…] both credible and not credible CIVCAS allegations can harm the relationship between coalition forces, the civilian population, and host-nation governments, and can negatively impact national and international opinion.”Footnote 26 The memorandum argues that CJTF-OIR officials require timely and accurate information on civilian casualties so that the coalition can begin a “consequence management [process…] to mitigate the impact of CIVCAS on the civilian population, the international community, and the mission.”Footnote 27 The memorandum outlines the reporting procedures that coalition units are expected to follow to ensure that CJTF-OIR officials receive the information that is needed. It stipulates that coalition units have twenty-four hours to submit a First Impressions Report (FIR) when civilian casualties are suspected to ensure that CJTF-OIR officials have basic information about potential incidents. After receiving the initial allegation, the CIVCAS cell—similar to the CCTC established in Afghanistan—must complete an Initial Assessment (IA) to determine whether the claim is credible.Footnote 28 The memorandum instructs CIVCAS cell officials to consult coalition strike logs when conducting their IA to confirm whether CJTF-OIR actions were conducted at this time and at those coordinates. CIVCAS cell officials can recommend closure in their IA if there are no corroborating actions in the strike logs or if “the allegation provides insufficient information as to time, location and detail.”Footnote 29
When CIVCAS cell officials can locate a corroborating strike in the coalition logs, the command responsible for conducting the strike is required to complete a more detailed CIVCAS Credibility Assessment Report (CCAR), which should state whether the claim is considered credible and outline how this determination has been reached.Footnote 30 The CJTF-OIR memorandum outlines the criteria that should be used to assess claims, stating that claims should be listed as credible “if the [Action Officer] determines, based on all reasonably available information, that it is more likely than not that a civilian was killed or injured as a result of coalition action.”Footnote 31 It states that all individuals in the strike area should be listed as civilians “unless positively identified otherwise”—but the memorandum does not outline the criteria that should be used to make a positive identification, or even define who qualifies as a civilian.Footnote 32 The standing definition that coalition troops are supposed to use when determining status is a “person who is not a member of the military and who is not taking a direct part in hostilities against a party to the conflict in an armed conflict.”Footnote 33 Yet several commentators have warned that the negative definition means that other demographic markers—including race, gender, and age—are often invoked in practice to differentiate between combatants and noncombatants (see below).Footnote 34 Once the CCAR is complete, consequence management can begin: the CJTF-OIR Public Affairs Office may choose to issue a statement confirming or disputing the allegation and, if necessary, ex-gratia payments can be disbursed to the victims.Footnote 35
While there are provisions in the memorandum that allow a commander to initiate an internal investigation, as noted above, only a minimal number—fewer than 12 percent—resulted in disciplinary action.Footnote 36 This further highlights the tendency to use investigations in ways that seek to bolster the legitimacy of coalition violence, even justifying civilian casualties that might have resulted. Shaped by operational concerns and avoiding the strategic set back of “losing hearts and minds,” the effort to assess casualties was more directly related to ensuring that from a military perspective, civilian casualties were minimized to ensure support by the civilian population for coalition activities. In other words, the emphasis in coalition assessments was to ensure that violence was used in ways that mounting civilian casualties would not derail the mission.Footnote 37 As Thomas Gregory has argued elsewhere, the impetus was less a humanitarian goal than a martial one to determine how force could be optimized—minimizing civilian casualties in circumstances where they could derail international and local support for the mission.Footnote 38
This emphasis on martial—rather than moral—considerations is also apparent in the CJTF-OIR Monthly Civilian Casualty Reports. Unlike classified CCARs, these monthly reports are issued as press releases and intended for public consumption. Each report follows the same basic format, listing the various allegations that coalition officials have assessed to be credible or noncredible alongside other numerical information, including the square kilometers and approximate population “liberated… from Daesh oppression.”Footnote 39 Coalition officials also used these monthly reports to reiterate their commitment to minimizing civilian casualties.Footnote 40 Although the CCARs submitted to the CIVCAS cell contain detailed information about what happened during these attacks, the monthly statements released to the public resorted to boilerplate language with sparse details about individual cases. For example, the classified CCAR on a strike in Mosul—which was conducted on January 13, 2016—describes how four adults and four children were observed moving in the rubble after the “first munition detonated on the target building [but…] following the detonation of the second munition, only three persons can be observed moving in the vicinity of the target and it is likely that although they survived [these persons…] were injured as a consequence of the strike.”Footnote 41 The CCAR also contains nine redacted images from the strike, a link to a shared drive with the full-motion video, and extracts from the chat logs where individuals involved in the attack discuss what they think they are seeing on their screens.Footnote 42 None of these details are evident in the CJTF-OIR “Monthly Civilian Casualty Report,” which simply states that “during a strike on ISIS fighters in a house it was assessed that eight civilians were unintentionally killed [and that…] during post-strike video analysis civilians were identified near the house who were not evident prior to the strike.”Footnote 43
Likewise, another confidential CCAR describes an airstrike on a suspected ISIS machine gun position located in a residential building, which was conducted on May 15, 2017.Footnote 44 The CCAR notes that local residents contacted the Joint Coalition Coordination Center in Erbil to report that five civilians were inside an adjacent building when the target was struck and that three perished in the attack. The CCAR describes—in quite graphic detail—how falling rubble trapped the victims inside their basement and how survivors used mobile phones to request assistance. Iraqi authorities were unable to provide assistance as ISIS militants were still in the area but “some neighbors… heard their cries and on or about 18 May were able to rescue two of the civilians.”Footnote 45 The CCAR notes that three civilians died in the attack, stating that “dehydration was likely the principal cause of death given that they were trapped in hot conditions for over 72 hours.”Footnote 46 These more distressing details were omitted from the CJTF-OIR “Monthly Civilian Casualty Report,” which covers the entire incident in a single sentence: “during a strike on an ISIS fighting position, it was assessed that three civilians in a building adjacent to the target building were unintentionally killed.”Footnote 47 The confidential CCARs tend to contain a much richer narrative about each incident while the Monthly Civilian Casualty Reports, which are intended for public consumption, are largely concerned with reporting the statistics, stripping individual accounts down to the bare minimum—often a few vague, non-descriptive, and sanitized sentences—in order to extract the basic numerical information.
These figures allow CJTF-OIR officials to collate the information into several broad categories: the number of reports that (i) have been received, (ii) have been assessed, (iii) the coalition considers to be credible, (iv) the coalition considers to be non-credible, and (v) still need to be assessed.Footnote 48 One problem with these short, itemized summaries is that the more distressing details tend to disappear from view because there is no discussion about how the victims experienced the harms inflicted upon them. Another problem is that CJTF-OIR officials were able to use these reports to reaffirm their commitment to civilian protection without reflection on their conduct in individual cases, including cases where mistaken judgements directly led to the killing of civilians. While it is reasonable to expect monthly reports to present more of an overview than an in-depth analysis of specific cases, our argument is that these reports fail to communicate the moral gravity and gruesome reality of killing civilians. Taken as a whole, the monthly reports thus also undermine public faith in the process of counting and reporting civilian casualties, especially as they systematically omit questions of accountability and disciplinary action. For example, the section on credible incidents in the “Monthly Civilian Casualty Report” often began with a generic statement insisting that coalition strikes were lawful, with the CJTF-OIR taking “all feasible precautions… to reduce the risk of harm to civilians.”Footnote 49
The itemized summaries then follow the same basic technical format, listing the date, the location, and the target before stating that the incidental harm to civilians was regrettable but unintentional.Footnote 50 Again, no mention is made of specific accountability mechanisms or instances where disciplinary action has been taken. This raises some important questions about the relationship between ethics and enumeration because the existing literature assumes that counting civilian casualties can help to ensure that militaries are held more accountable for their actions, but CJTF-OIR counts seemed to be more concerned with legitimizing coalition violence than interrogating possible problems.Footnote 51 When considering efforts to enumerate civilian casualties, it is important that we remain attentive to the rationale behind these counts because we might discover, as this section has shown, that certain counts prioritize martial considerations above moral considerations.
Enumeration and Erasure
Whereas the previous section focused more on the rationale behind these counts, this section will detail the mundane processes and practices that CJTF-OIR officials use to enumerate the harm inflicted upon civilians to trace how these bureaucratic procedures circumscribe what casualties are counted and how these casualties are counted.Footnote 52 We focus specifically on the categories that CJTF-OIR officials have used to count civilian casualties to show how these categories work to render certain harms more visible while erasing other civilian harms from view. One way to illustrate this issue is to draw attention to discrepancies between different counts. According to CJTF-OIR estimates, coalition forces have killed 1,454 civilians and wounded 402 civilians across 360 separate incidents since the conflict commenced.Footnote 53 Airwars numbers are markedly different from those of CJTF-OIR. It estimates that coalition forces have killed more than 8,222 civilians and wounded at least 5,906 civilians across 2,985 separate incidents, which suggests that civilian deaths are five times higher than coalition estimates and civilian injuries are nearly fifteen times higher.Footnote 54 The stark difference between the Airwars estimate and the coalition count owes to significant methodological and interpretational differences.
Airwars monitors open-source information, including social media and local news sources, to identify civilian harm claims, which it then cross-references with other sources in order to provide a provisional assessment. Allegations are graded as “confirmed” when a specific belligerent has accepted responsibility for harm to noncombatants in a specific incident. Allegations are graded as “fair” when there is a reasonable level of public reporting on an alleged incident from two or more credible sources. All sources agree on what happened and this is often coupled with detailed biographical, photographic, or video evidence. Crucially, this grading also “includes likely or confirmed actions by a belligerent in the near vicinity for the date in question.”Footnote 55 Allegations are graded as “weak” when the allegation is based on a single—albeit reputable—source (with biographical details and visual evidence) and it has been confirmed that international strikes were taking place in that area and on the date in question. Airwars grades allegations as “contested” when there is a reasonable level of public reporting but there are disagreements about who is responsible for the attack or how the harm to civilians occurred. Airwars “discounts” those claims when its researchers can demonstrate that the victims were combatants rather than noncombatants, no other parties were responsible for inflicting this harm, or that no civilians were harmed in this incident.Footnote 56 Its published figures—including those listed above—are limited to incidents that have been confirmed or assessed as fair. It is also important to note that referrals from Airwars triggered approximately 70 percent of the coalition assessments conducted during Operation Inherent Resolve, including 40 percent that the coalition listed as credible.Footnote 57
Both counts raise important epistemic questions about how civilian harm is enumerated—how it is counted, how it is categorized, and how it is communicated—and how the different methodologies that have been developed to count civilian casualties delimit which civilian casualties are counted.Footnote 58 Yet there are also significant differences in motivation. Coalition officials were counting civilian casualties to assist with consequence management, using this information in an effort to prevent civilian casualties becoming an impediment to mission success. Airwars counts civilian casualties because it wants “to reliably and independently document the human cost of war in order to promote a more peaceful world where human lives are acknowledged and taken into account.”Footnote 59 Airwars is a member of the Casualty Recorders Network, which was established to develop and disseminate clear standards for recording casualties to ensure that all conflict-related deaths are recognized.Footnote 60 Airwars uses its civilian harm data to ensure belligerents acknowledge the full human toll their actions take on the civilian population and have measures in place to minimize human suffering while providing victims with tools to hold belligerents accountable for their actions.Footnote 61 We will return to the divergent motivations that underpin these counts in a moment, but for the time being we want to focus on how methodological differences determine whether specific casualties will be counted and how these casualties are categorized, and how these methodological differences can result in significant discrepancies between the different counts.
In some cases, these discrepancies result from simple errors. The New York Times reviewed the 1,300 confidential assessments that it obtained from the Pentagon and discovered that basic errors were concerningly common in the coalition count. There were cases where officials muddled coordinates, misspelt names, and ignored conflicting evidence.Footnote 62 For example, one allegation was dismissed as non-credible on the grounds that:
The geographic locations of “Siha” and “Bawabat Al Sham” do not correlate with known districts of West Mosul or the surrounding area. In the absence of further particularizing information or evidence, this allegation should be treated as not having the potential to be credible.Footnote 63
However, reporters from the New York Times were able to locate the west Mosul neighborhood using a tool as simple as Google Maps, with a slight spelling variation (“Sihah” rather than “Siha”).Footnote 64 Although CJTF-OIR officials have not commented on specific errors that have been identified in these reports, they have acknowledged that there might be some mistakes and inconsistences in their database and have committed to working alongside external organizations—like Airwars—to correct these mistakes.Footnote 65
It would, however, be a mistake to think that the sizeable discrepancies between the different datasets can fully be attributed to specific mistakes in individual assessments. As noted above regarding Airwars, discrepancies are often the result of significant methodological differences, which has a significant bearing on how these counts are conducted.Footnote 66 Ryan Goodman, for example, argues that the “implausibly low number of civilians that the Pentagon admits to killing in Iraq and Syria [reveals…] a fundamental flaw in the way the United States military screens claims of civilian casualties.”Footnote 67 Goodman argues that there are several factors that might explain these discrepancies, including the fact that CJTF-OIR officials lack on-the-ground access to the blast site. He also claims that the evidential threshold required to initiate a CCAR is too high, warning that the coalition is “obviously going to miss a multitude of cases in which civilians were killed.” As an initial screening device, he concludes, this threshold has become a “stunning, systemic flaw.”Footnote 68 Likewise, when Airwars compared these classified assessments against its own internal archive, it discovered several instances when CJTF-OIR officials had requested information from Airwars to assist with their investigations but then proceeded to ignore this information, even when it came with precise coordinates, satellite images, and witness testimonies.Footnote 69
In her book The Seductions of Quantification: Measuring Human Rights, Gender Violence, and Sex Trafficking, Sally Engle Merry suggests that numerical “indicators are appealing because they claim to stand above politics, offering rational, technical knowledge that is… objective, scientific, and transparent.”Footnote 70 Although this statistical information might sound attractive, she warns that the “process of translating the buzzing confusion of social life into neat categories that can be tabulated risks distorting the complexity of social phenomenon, [stripping them…] of their context, history, and meaning.”Footnote 71 Despite promising objective statistical information, she argues that quantification fails in at least two specific ways. On the one hand, it camouflages the political considerations that shape what information is collected and how this information is presented, including the social processes involved in constructing the categories used to organize this information.Footnote 72 As she explains, “Rather than objective representations of the world, such quantifications are social constructs formed through protracted social processes of consensus building and contestation.”Footnote 73 On the other hand, she argues that quantification conceals the interpretative work that goes into deciding whether a specific incident should be included in the count and how this item should be categorized. She warns that “the ambiguity of the categories, errors in counting, missing data, and lack of commensurability disappear in the final presentation.”Footnote 74 Her work is useful when it comes to thinking about the CIVCAS cell data collected during Operation Inherent Resolve because it enables us to draw attention to how the categories CJTF-OIR created to organize the information flowing into their spreadsheets delimit what civilian harms are legible as civilian harms—what can be counted and what will remain uncountable.Footnote 75
In order to understand how the coalition enumerated civilian harm, we examined both the CJTF-OIR policies that governed how these counts ought to be conducted alongside the CCARs that show how these policies were enacted. We focused specific attention on the categories that the CJTF-OIR constructed to collate the information that was flowing into CIVCAS cell spreadsheets and the interpretative work that went into categorizing specific incidents. In the previous section, we introduced the CJTF-OIR memorandum outlining the rationale behind the coalition count, but the memorandum also outlines the definitions that CJTF-OIR officials should use when assessing claims. The CJTF-OIR memorandum outlines what constitutes a CIVCAS, defining it as “the death or injury of a civilian, who has not lost [their] protected status.”Footnote 76 The memorandum also defines what can cause a CIVCAS, limiting it to “direct military action, to include Close Air Support (CAS), Close Combat Aviation (CCA), Direct Fire (DF), Indirect Fire (IDF), and other actions directly linked to security operations.”Footnote 77 In doing so, the memorandum limits civilian harm to isolated moments when specific civilians were killed or wounded as a direct consequence of identifiable coalition attacks, which obscures the indirect, reverberating, and cumulative harms that are connected to the conflict but cannot be directly linked to specific attacks.Footnote 78 As we show below, these definitions are not neutral, but have a significant bearing on which harms can be enumerated and how. As Derek Gregory puts it, “Spaces of constructed visibility are also always spaces of constructed invisibility.”Footnote 79
The memorandum also seeks to redraw the distinction between combatants and noncombatants, reminding us that these categories are not fixed and stable entities with clear and consistent definitions.Footnote 80 The CJTF-OIR memorandum defines the civilian in negative terms, as “persons who are not combatants [or taking…] direct part in hostilities.”Footnote 81 Yet, the memorandum then supplements this basic definition in the template for CIVCAS cell officials to use when reviewing specific incidents. The template suggests that CIVCAS cell officials compare the “slant” identified in the pre-strike collateral damage estimate with the “slant” that is visible in the post-strike video footage. It is written as three numbers separated with slashes (hence, the term “slant”)—the first indicating adult males, the second indicating adult females, and the third indicating children (for example: “3/1/2”).Footnote 82 As Helen M. Kinsella explains, the distinction between combatants and noncombatants is often secured with reference to gendered and racialized assumptions about who is innocent and who is involved, with adult males categorized as combatants and adult females categorized as noncombatants—irrespective of their actual status.Footnote 83 One particular problem with this simplistic “slant” approach is that all adult males are at serious risk of being considered combatants merely because of their perceived age and gender—and if they are considered combatants they would never be counted as civilians even when their status is merely inferred rather than concretely known. In this respect, the slant does much more than describe who is a combatant and who is a noncombatant: it produces certain populations as combatants and noncombatants, which has a significant bearing on how individuals are categorized in the coalition counts.Footnote 84 Discussions in the CCARs also reveal the ambiguities that haunt these categories, with officials unclear whether adolescents should be categorized as adults or children.Footnote 85 None of this is visible in the Monthly Civilian Casualty Reports because all these doubts and uncertainties—which amount to potentially thousands of civilian lives extinguished—tend to disappear in the final presentation.Footnote 86
Critics have suggested that such methodological flaws and definitional deficiencies result in the CIVCAS cell undercounting civilian casualties, leading to sizable discrepancies between coalition and non-coalition figures.Footnote 87 In addition to the above, Oona Hathaway and Azmat Khan have identified problems with CJTF-OIR reporting policies and practices, including an over-reliance on low-quality video footage, access to incomplete or inaccurate strike logs, insufficient linguistic and cultural expertise on the target area, and a reluctance to reassess cases when new evidence emerges.Footnote 88 In her reporting for the New York Times, Khan discovered cases that were dismissed as non-credible because the casualties were not visible in the overhead reconnaissance footage because the victims were buried in the rubble or concealed under the smoke, both of which are common following aerial strikes.Footnote 89 She also identified cases that were dismissed as noncredible during the initial assessment because individual strikes had not been recorded in coalition logs or because the incorrect coordinates had been listed. In at least one case, officials dismissed a claim as noncredible because coalition records showed multiple strikes at the specified location within the specified timeframe, but the claim did not pinpoint the exact strike that inflicted the harm.Footnote 90 CIVCAS cell officials also often lacked the cultural expertise to translate information submitted in the local vernacular—such as geographical landmarks—into information that could be used to search coalition records. In one case, local witnesses reported that two civilians were killed and three civilians were wounded in an airstrike in Al-Bab Al-Gharbi, which is a historical landmark in Hit, Iraq. The CIVCAS cell dismissed the claim as noncredible because officials were “unsure where Al-Bab and Al-Gharbi are located”; they had assumed that the term referred to two separate neighborhoods rather than a historical landmark.Footnote 91
However, even when individual incident reports were marked as credible, CIVCAS cell officials were often undercounting the civilians harmed in these attacks because the victims were not visible on their screens, or their bodies were miscategorized as combatants. Khan argues that confirmation bias—such as the presumption that all adult males might be insurgents—led to the coalition targeting innocent civilians and then listing the victims as combatants.Footnote 92 In one incident, the coalition struck a suspected ISIS explosives factory outside Raqqa after operators spotted white bags—thought to be ammonium nitrate—inside a walled compound. When locals reported that nine civilians were killed in the strike, the coalition reviewed the strike and determined that the building was in fact a cotton mill, the secondary explosions witnessed at the scene were reflections from other buildings, and two civilians were killed. However, the coalition maintained that the target was legitimate based on a news report that ISIS controlled cotton production in Syria.Footnote 93 Khan points to other incidents where civilian casualties were not recorded because officials could not discern them on the screen, including one case where a man was observed dragging an “unknown heavy object” from a building that was later reassessed to be “a person of smaller stature” (likely a child).Footnote 94 We also identified several cases where CIVCAS cell officials were unable to determine an exact number and so listed the casualties as a range, but then listed only the minimum number in their monthly report. For example, in one CCAR, CIVCAS cell officials concluded that “it is impossible to determine the exact number [but…] it is reasonable to assess that between 5 and 12 civilian casualties occurred.”Footnote 95 Yet, when the case was listed in the CJTF-OIR “Monthly Civilian Casualty Report,” it was reported that “five civilians were unintentionally killed”—possibly seven victims disappeared from the count.Footnote 96
It is important that we can see the discrepancies between different counts and understand how these discrepancies might occur, but different counting methods can also erase certain civilian experiences from view. As noted above, coalition policies and procedures are concerned with civilian deaths and civilian injuries that can be attributed to specific coalition actions, but this definition obscures “the cumulative harm” that arises during a conflict, what Fionnuala Ní Aoláin defines as the composite, aggregate, collective, and layered harms that civilians might experience during a given conflict.Footnote 97 Counting dead and wounded bodies is one metric that can be used to measure the death and destruction inflicted during a conflict. Although casualty counts can detail the physical harms that are prevalent in any given operation or conflict, they say little about how families and communities experience these losses—simply listing wounds and causes of death provides little insight into how civilian victims must live with the long-term effects (including the chronic pain, the mental anguish, and economic deprivation). Nor does it provide information about the burden that the death or disability of a loved one places on those around them.Footnote 98 Additionally, while the CIVCAS cell counts dead bodies, it does not count damaged buildings, so the coalition database cannot tell us how coalition strikes have rendered homes uninhabitable, water supplies undrinkable, medical supplies unusable, and the coalition cannot enumerate the “aggregate impacts of protracted armed conflict on civilian populations, including individual and societal trauma, socio-economic degradation, severe and prolonged infrastructure decline, and the accumulation of civilian death across repeated attacks over time.”Footnote 99
In short, though seemingly mundane, bureaucratic procedures account for the vast discrepancies between coalition data and external counts. More than a neutral set of technical parameters, however, practices can enforce a restrictive definition of harm—one that recognizes only discrete incidents of death and injury. By tethering “valid” harm to specific airstrikes, these policies systematically ignore the indirect, reverberating, and cumulative devastation that defines the lived reality of conflict.
Numbers and Narratives
We have advanced two distinct but interrelated arguments about coalition counts. On the one hand, we have suggested that these counts were conducted in accordance with a martial logic that was more concerned with consequence management than holding the coalition more accountable for the harm caused to civilians. On the other hand, we have suggested that the policies and processes that govern these counts have worked to render certain harms more visible in debates about the conflict while obscuring those harms that do not fit within prescribed coalition categories. In this final section, we turn our attention more explicitly to the Airwars count mentioned above. As we have noted, Airwars prioritizes recognizing coalition victims rather than rationalizing coalition violence. Airwars tracks civilian harms across several conflicts in order to hold militaries accountable, ensure militaries center civilian protection measures in mission planning, and, crucially, provide civilians with the evidence needed to pursue justice for the harms that have been inflicted upon them.Footnote 100 It is important to contrast the CJTF-OIR count with the Airwars count for three reasons. First, we will outline the distinct imperatives behind these counts to draw attention to the politics involved in counting civilian casualties. Second, we will contest coalition claims that the CIVCAS cell can provide clear, objective, and impartial numerical information about the conflict, showing instead how its policies and practices can circumscribe what casualties are counted and how these casualties are categorized. Finally, we will show how organizations like Airwars seek to move beyond numerical indicators to foreground the narratives behind their numbers, showing us that numbers are necessary but insufficient.Footnote 101
As noted above, there are sizeable discrepancies between CJTF-OIR data and Airwars figures. Airwars has tracked between 19,226 and 29,834 alleged deaths from 3,011 coalition operations in Syria and Iraq. Airwars estimates that between 8,256 and 13,352 civilian deaths can be categorized as confirmed or fair and between 2,397 and 2,821 civilian deaths can be categorized as weak. Airwars has discounted 694 to 1,077 civilian deaths and listed the remaining 7,879 to 12,584 as contested.Footnote 102 Our intention here is not to adjudicate whose data are more accurate, but to reflect on how numerical indicators circumscribe what harms are seen and how these harms are seen. As noted earlier, numerical indicators—when examined in isolation—tell us little about the civilians who were impacted in these attacks or how these civilians experienced the harms inflicted upon them. One thing that differentiates the Airwars count from the CJTF-OIR count is its commitment—whenever possible—to name the victims, share their images, and describe their lives with the specific intention of humanizing the dead.Footnote 103 Airwars has also collaborated with other organizations to visualize the harms inflicted during coalition attacks and to draw out some of the narratives behind their numbers.Footnote 104 One example is the interactive website that Amnesty International created using Airwars data, which visualizes the death and destruction inflicted during a four-month assault on Raqqa in 2017, focusing on the stories of individuals and families rather than abstracted numbers.Footnote 105 During this assault, Lieutenant General Stephen Townsend—who commanded CJTF-OIR at the time—insisted that the coalition applied “rigorous standards” when selecting targets and made “extraordinary efforts” to protect the civilian population, challenging humanitarian actors to “find a more precise air campaign in the history of warfare.”Footnote 106
Airwars estimates that at least 1,600 civilians were killed in coalition airstrikes during this period.Footnote 107 Using this data, Amnesty International was able to visit more than 200 strike locations, interview more than 400 witnesses, and name more than 1,000 victims.Footnote 108 Unlike the CJTF-OIR Monthly Civilian Casualty Reports, which catalogs the credible cases in a few short, formulaic, and sanitized sentences, Amnesty International’s interactive website stiches together audio, images, and video to create a more immersive experience. For instance, Amnesty International uses a 360-degree panoramic image to document an attack on a five-story residential building that resulted in at least thirty-two civilian deaths, including twenty children.Footnote 109 The website enables visitors to stand amid the rubble, looking up at the exposed framework, broken concrete, and twisted steel. A video shows an interview with a man named Maher Hammada, who describes losing his wife and their two children—alongside another four relatives—in the attack.Footnote 110 There is another 360-degree panoramic image to illustrate a coalition strike on a building in Harat al-Badu, which killed Nahla Hammud and her three children.Footnote 111 A video shows a survivor walking around the rubble, describing how victims were blasted from the collapsing building. We see another survivor holding images of the deceased, who had to be buried in the garden because the fighting prevented bereaved families from transporting them to local cemeteries.Footnote 112
These visual narratives are significant because they draw attention to the cumulative harm that eludes the categories on which the coalition relied when seeking to enumerate civilian harm. This goes beyond simply undercounting the dead and injured in an immediate sense. In his work on “slow violence,” for example, Rob Nixon argues that we are accustomed to thinking about civilian harm in visceral terms: the mangled bodies and mutilated corpses that remain when munitions explode. While such visually arresting scenes capture the most immediate violence that is inflicted during these attacks, it obscures the violence that does not have such an immediate effect, such as the toxic remnants these munitions leave behind, which continue to deplete both people and place long after the initial detonation, and even after the conflict subsides.Footnote 113 Alongside work on the slow violence inflicted during these conflicts—and its deferred effect on the civilian population—is a growing interest in the harm that accumulates from multiple or prolonged attacks on the same location, which cannot be attributed to isolated attacks or individual events. Repeat attacks on the built environment can leave homes uninhabitable and critical infrastructure—power, sanitation, communication, healthcare, and so forth—unusable, exacerbating existing issues and causing excess deaths among civilians.Footnote 114 Yet, these cumulative and collective harms are often uncounted because these harms are not direct and cannot be ascribed to a specific incident.
The interactive visualizations that are used on the Amnesty International website draw attention to the slow and cumulative violence that is often obscured within the numerative frame. Alongside the panoramic photographs depicting the violence that coalition airstrikes have inflicted upon the urban environment are individual narratives describing how civilians experience these diverse harms. One account describes how a woman named Rasha Badran was forced to flee her home after three artillery shells slammed into the building where she lived.Footnote 115 Rasha had returned home during the conflict to care for her wounded relatives in need of medical treatment but was forced to shelter with her neighbors as her home was obliterated in the attack. After one month of the assault on Raqqa, the hospital was no longer able to treat its patients, and resource shortages meant that residents had to leave the area.Footnote 116 Once again, Rasha and her family were forced to flee, but the building where they sought shelter was also bombed, killing everyone except Rasha, her husband, and his cousin—who were both wounded in the attack. She recalls that:
My husband was the most seriously injured—he had a head wound and blood was pouring from his ears. It was dark and we could not see anything. We called out but nobody else answered; nobody moved. It was completely silent except for the planes circling above.Footnote 117
Rasha managed to escape, but she lost thirty-nine relatives during this period, including her one-year-old daughter, Tulip, who they buried under a tree.Footnote 118
Another video on the website features a man named Mohammed Aswad, who describes sheltering inside a basement during a coalition airstrike on his neighborhood, which reduced his building to rubble.Footnote 119 Mohammed could hear his nephew Mahmoud screaming outside the basement door, but he was unable to help him because his nephew was trapped under a concrete pillar. Once the bombing subsided, Mohammed was able to moisten Mahmoud’s lips with water but was unable to remove the concrete that was pinning him down, so he went searching for other relatives. Mohammed discovered several dead bodies—including his brother and another nephew—but the others were not visible under the rubble.Footnote 120 These accounts are difficult to read and distressing to watch but these narratives provide an important supplement to an enumerative framework that reduces human beings to mere statistics, which conflate these singular experiences into standardized categories that can be counted, categorized, and compared. These narratives enable us to understand how civilians experience these harms: the desperation that Rasha must have experienced moving between shelters as bombs rained down from above and the anguish she must have experienced while burying her one-year-old child; the terror that Mohammed must have experienced knowing that his nephew was trapped outside during an airstrike, and the horror he must have experienced uncovering dead relatives buried amidst the rubble. These narratives are also more concerned with who these individuals are (singular beings, with unique and deeply traumatic stories to tell) rather than reducing them to what these individuals are (dead/wounded, combatant/non-combatant).Footnote 121 These narratives also enable victims of war to tell their own stories, to assert their human rights, and to insist on dignity and freedom in the face of their denial. “We are not numbers,” asserts a group of youth writers from Gaza who refuse war’s reductionistic frames of quantification.Footnote 122
Conclusion
This article has examined CJTF-OIR efforts to enumerate civilian harm during Operation Inherent Resolve in order to draw attention to some underexamined martial considerations about counting civilian casualties and their ethical implications. The literature on civilian protection tends to presume that counting civilian casualties can enhance ethical deliberations because it can provide concrete, impartial, and unambiguous information about the harms inflicted on civilians, which can be used to assess whether militaries acted in accordance with the established moral norms and legal codes.Footnote 123 Our position is that counting civilian casualties is necessary but insufficient. Accurate numerical information is needed to capture the harms inflicted in a conflict, to ensure that victim experiences receive some recognition, and to hold militaries accountable when their actions violate the rules. Yet, this article has drawn attention to other ethico-political considerations that are often overlooked in these conversations, including the rationale behind certain counts, the interpretative work needed to conduct these counts, and how a reliance on numerical indicators can conceal certain harms. We argued, for example, that CJTF-OIR policies on recording civilian casualties were imbued with martial considerations about effective consequence management techniques, with accurate information needed so that coalition officials could contest those claims it considered to be non-credible. At the same time, we argued that the categories that the CJTF-OIR constructed in its policies and procedures—and the interpretative work that went into counting, categorizing, and collating these casualties—worked to obscure certain civilian harms. We argued that these concerns were evident in both the discrepancies between the coalition count and the Airwars count and the fact that certain harms seemed to elude enumeration, including the indirect, reverberating, and cumulative harms to civilians.
Our arguments about enumeration should not be seen as an argument against enumeration, not least because some militaries have stopped enumerating these harms. On December 21, 2023, the Pentagon issued new instructions on civilian harm mitigation and response, which seemed to formalize various measures trialed in Operation Inherent Resolve, including tracking civilian harm.Footnote 124 The instruction stated that a Civilian Harm Assessment Cell (CHAC) should be established whenever a conflict commences to collate information on civilian casualties in order to assist with consequence management, and even investigations into possible war crimes.Footnote 125 The instruction also outlined provisions for a Civilian Protection Center of Excellence inside the Pentagon, which would analyze the data flowing into CHAC spreadsheets to locate possible patterns or spot possible trends.Footnote 126 Though the instruction continued to prioritize martial considerations above humanitarian concerns—insisting that “the protection of civilians and civilian objects is fundamentally consistent with the effective, efficient and decisive use of force”—they at least bring civilian harms into the consideration of war operations.Footnote 127 In March 2025, however, the Pentagon announced that the Civilian Protection Center of Excellence would be terminated as the Trump administration set about removing what it viewed as administrative impediments to achieving mission success.Footnote 128 The Pentagon has not confirmed whether the entire instruction on civilian harm mitigation and response will be abandoned, but it has removed the individuals who were charged with collecting this information and, at the time of writing, it has stopped releasing data on civilian casualties.Footnote 129
This refusal to count is concerning because it suggests that the Pentagon has little interest in civilian protection. Indeed, the current U.S. Secretary of Defense, Pete Hegseth, has expressed public distain for rules of engagement and international law, which are the cornerstones of civilian protection. He has also suggested that what he calls the “warrior ethos” has been unnecessarily constrained by military lawyers who are charged with giving legal advice on military targeting operations and crucial considerations of military necessity and proportionality.Footnote 130 In the last decade or so, the minimization of civilian casualties has become an ever more important component of war fighting, but from Israel’s war in Gaza to the American and Israeli assault on Iran, it seems we may be witnessing a weakening of the norms surrounding civilian protection. Now Hegseth speaks not of minimizing civilian casualties but rather: “Maximum lethality, not tepid legality.”Footnote 131 Counting civilian casualties might be insufficient, as we have argued, but it is a vitally important necessary first step, and the dismantling of the infrastructure designed to protect against and count civilian casualties is an extremely concerning development.Footnote 132 Those concerned with reducing civilian harm are now even more reliant on the data that is produced by organizations such as Airwars. These NGOs are invaluable as we seek to engage more meaningfully with the civilian populations caught up in these conflicts and to understand their embodied experiences of both the direct harm that is done to them and their communities as well as the indirect, cumulative, and reverberating harms to their lived environments.
Open access
