Strategic ignorance and the legitimation of remote warfare: The Hawija bombardments

How must we understand and conceptualize the rationales and repercussions of remote warfare? This article contributes to scholarship on the ontology of remote war by analysing how Dutch officials engage with responsibility for the bombardment of an Islamic State weapons factory in Hawija, Iraq in 2015 under Operation Inherent Resolve. It observes that the main feature of Dutch officials’ accounts of Hawija is their diverse claims to not knowing about civilian casualties. Official narratives shifted from denial to secrecy to strategic ignorance. Bridging work on secrecy from the field of critical security studies with work on strategic ‘unknowing’ from ignorance studies, we propose a new take on the Foucauldian notion of ‘regimes of truth’. The regimes of truth that emerge to justify shifts to remote warfare – that it is riskless, precise and caring for civilian others – rely not merely on secrecy and denial but on feigned and imposed ignorance about casualties. Whereas denial can be disproven and secrecy has an expiration date, ignorance is more elusive and open-ended and hence politically convenient in different ways. Deliberate unknowing does not just postpone investigation and accountability but fundamentally and indefinitely obstructs it and thus sustains the regimes of truth for future remote wars.

covert human costs of overt action (Cormac and Aldrich, 2018). Our argument extends what Walters (2020: 61, 60) has called the 'new secrecy research' that reflects a turn towards secrecy as a 'field of power relations that merits theoretical scrutiny'. Bringing in ignorance in this field adds a new thread to what Toom (2020) has theorized as 'ontologically dirty knots', a notion that allows us to analyse events that -like the Srebrenica genocide in Toom's analysis and the Hawija bombardment in ours -are defined by the denial, secrecy and controversy that come with disputed 'body counts'.
We present our argument as follows: The next section further introduces the Hawija case as exemplary of dominant regimes of truth about remote warfare. It concludes with the observation that denial, secrecy and ambiguity are central to these regimes of truth yet insufficiently conceptualized and operationalized in the relevant academic debates. The subsequent section develops a conceptual framework that constitutes the article's main theoretical contribution by merging relevant insights from security studies and ignorance studies that allow us to trace agency and interests in our investigation of epistemic politics. In the section that follows, this framework is applied to the political afterlife of 'Hawija' in Dutch politics. In the article's final section, we reflect on the political and intellectual potential of contesting strategically imposed ignorance in debates on remote warfare.

Debates: Remote warfare and regimes of truth
The strikes on Hawija led to a large secondary explosion. During a control flight it was clear to the pilots that a whole district had been destroyed. On 15 June 2015, the then Dutch Minister of Defence (MoD), Jeanine Hennis-Plasschaert, received a classified Civilian Casualty Credibility Assessment report from the commander of the anti-IS Coalition, the US Central Command (CENTCOM). Herein, CENTCOM acknowledged that the claims being made by open sources Reuters and the Red Cross were credible: over 70 civilians died in the incident. 2 Yet, on 28 June, Hennis informed Parliament that although she could not disclose any exact information about Dutch airstrikes due to 'national, operational and personnel security', she could relay that 'as far as known at the moment, the Netherlands had not been involved in any instances of civilian casualties caused by airstrikes in Iraq'. 3 Four and a half years later, when Dutch journalists finally uncovered that the Netherlands was responsible for the Hawija attack, the Dutch Prime Minister, Mark Rutte, not only claimed he had no recollection of being informed about the incident by Hennis, but that the number of 70 civilians casualties was just a 'rumour ' and 'irrelevant' and 'there is no way of knowing how many civilian casualties perished'. 4 The Hawija strikes were one of the 2100 missions Dutch F-16s flew across Iraq and Syria between October 2014 and June 2016 and between January and December 2018. The Netherlands continues to contribute to OIR in targeting processes and financing and training local actors, such as the Iraqi forces and Kurdish Peshmerga. The Netherlands was not alone in choosing to wage war against IS remotely. The UK, France, Belgium, Denmark and Australia also joined the US-led Coalition 'to destroy ISIS' by using 'coordinated airstrikes, training and equipping local security forces, and targeted special operations' (McInnis, 2016: 2). Since its establishment in 2014, OIR has engaged in over 34,000 strikes, firing over 100,000 bombs and missiles across Syria and Iraq. 5 The West's contribution to OIR is one of many recent examples of how remoteness in all its modalities -from distancing to outsourcing -has become a characteristic feature of warfare. From the 2011 NATO bombings in Libya to the US Africa Command's counter-terrorism training of Ugandan soldiers, remote warfare is characterized by a shift away from boots on the ground. It involves drone and air strikes, while on the ground small (private) military training teams assist local forces to fight and die on behalf of Western interests. Although these strategies each have long genealogies, they are now combined in a way that facilitates the execution of violence by Western defence ministries without the 'exposure' of their own military to opponents under the condition of mutual risk. This spatial reconfiguration of war has been conceptualized in a range of ways from 'risk-transfer war' (Shaw, 2005) to 'democratic warfare' (Sauer and Schörnig, 2012), and more recently as 'liquid warfare' (Demmers and Gould, 2018) and 'vicarious warfare' (Waldman, 2021).
What we analyse here is not the assembling and outsourcing of warfare as such. Instead, we contribute to a resurgent scholarship on the evolving ontology of war (Barkawi and Brighton, 2011;Demmers et al., 2020;Bousquet et al., 2020). This body of literature analyses how actors assemble authoritative knowledges on war. With regard to remote warfare, Demmers et al. (2020) argue that although distant, riskless and often highly secretive, remote war, like every war, does not remain insulated from the machinations of propaganda. With remote violence, war, truth and power thus remain intimately related, but in new and particular ways.
Such relations between knowledge and violence can be better understood with the help of Foucault, who, in his interview on Power and Truth, summarised his understanding of the connection between the two by offering the following definition of a 'regime of truth': Truth is a thing of this world: it is produced only by virtue of multiple forms of constraint. And it induces regular effects of power. Each society has its regime of truth, its 'general politics' of truth: that is, the types of discourse which it accepts and makes function as true; the mechanisms and instances which enable one to distinguish true and false statements, the means by which each is sanctioned; the techniques and procedures accorded value in the acquisition of truth; the status of those who are charged with saying what counts as true. (1980:133) This concept has since been enthusiastically applied across disciplines, fields and case studies (for a recent application, see Abdel-Fattah, 2019), also with regard to secrecy and regimes of (post-) truth (Krasmann, 2018). The explicit link with war has been made more recently by Demmers et al. (2020), who used the 'regime of truth' to interrogate OIR. They show that those conducting the operation contrast the brutal violence perpetrated by IS with the surgical precision with which its strongholds were targeted. Coalition commander Stephen J Townsend (2017), for instance, challenged anyone to find a more 'precise air campaign in the history of warfare . . . The Coalition's goal is always for zero human casualties'. Such statements point out how the constant application of new smart technologies and proportionality principles allow for a new form of 'perfect warfare', which saves lives of both Western military personnel and friendly civilians on the ground (Demmers et al., 2020). It presents past, present and future remote killings as 'ethical' (Schwarz, 2016), a 'moral act of care' (Chamayou, 2015: 108) and a form of 'humanitized violence' (Bonds, 2019).
In line with Brass's (1996) observation that one should not only focus on the interpretative processes before but also after violent practices, Demmers et al. (2020) observe that this 'perfect war' narrative, like any regime of truth, has been contested with counter-truths. Not satisfied with how Western militaries assess the number of civilians killed by their airstrikes, monitoring organizations, such as Airwars and Amnesty International, have developed new remote sensing techniques to count the number of civilian casualties from Western airstrikes. For OIR, claims about the 'true' number of civilian deaths after 35,000 airstrikes are astonishingly disparate. Airwars's current conservative estimate is that between 8317 and 13,190 civilians have likely been killed in Coalition actions. The Coalition itself currently 'confirms' only 1417 non-combatant deaths. 6 Individual Coalition members have an even worse track record in conceding to civilian harm, with the Netherlands for years being one of the most active yet least transparent members of the Coalition. When confronted with civilian casualty claims, Western Coalition members often deny their existence, rarely share information about strikes in the name of security, hide behind the Coalition to defend non-disclosure, and often go as far as claiming that it is altogether impossible to make any unambiguous claims about the number of civilian casualties (Airwars, 2018). For us, what is crucial is that the same distance is portrayed as surmountable when legitimizing 'precision' targeting, but insurmountable when confronted with the inconvenient counter-truth of civilian casualties in their wake. This resonates with Stampnitsky's (2020) observations that revelations and knowledge are not directly proportional to the 'amount', proximity or even the quality of evidence.
The Dutch government's sanctioning and discrediting of the counter-truths on the civilian deaths in Hawija, then, is not an exception but the rule among its Western Coalition allies. Hawija thus offers insights into the constructions and contestations of regimes of truth about remote war. More specifically, it points to the centrality of forms of epistemic politics -the political dimensions of claiming or denying knowledge (Aradau, 2014) -in these processes, which involve denial and secrecy but also appear to go beyond them, conjuring more fundamental claims of unknowability about civilian casualties to maintain an image of a perfect war. The next section offers a conceptual framework to further interrogate such epistemic politics in the context of remote warfare.

Critical security studies and the contentions of secrecy and accountability
The essence of remote warfare is that the number of military personnel needed for and involved in attacks is minimized and kept at a physical distance. This material and geographical distance generates political and epistemic distance. The intangibility of events in the 'push-button conflicts' that constitute remote warfare, Kearns notes, position 'witnesses as spatially, intellectually and morally distanced from them, delimiting the ethical import of the violence inflicted upon casualties ' (2017: 20, 14). The confusion and ambiguity that are implicated in such distancing are essential to upholding the practice of remote warfare as well as its legitimation.
Consequently, one of the central questions in the critical analysis of remote warfare is the issue of how relevant war actors handle incidents of mass civilian casualties that might undermine the dominant regime of truth that remote warfare is particularly precise and hence justified. Such questions on the epistemic and discursive dimensions of warfare and their moral implications should always be central to critical security and conflict studies, but particularly so when the legitimacy of the type of war at stake so specifically hinges on the alleged minimization of 'collateral damage' and the fiction of a 'clean' war.
The two main answers to these questions on 'inconvenient truths' presented by the field of critical security studies are the related practices of denial and secrecy. An emerging field of 'secrecy studies' has taken the realization that 'states, if engaging in practices that violate international norms, will operate with secrecy and denial as "the normal state of affairs"' as the starting point for critical conceptualization (Cohen, 2001: 249;Stampnitzky, 2020: 2; see also Cormac and Aldrich, 2018). Denial refers to refuting knowledge claims, in this case that civilian casualties exist. Despite declarations of an impending post-truth era, this is a rather crude epistemic strategy that is hard to maintain in times of increasing datafication and democratized information technology (Aldrich and Richterova, 2018). Denial mostly serves as a tactic of diversion or stalling, with the initial denial buying time to develop more sophisticated ways in which to manage problematic situations.
Such 'management' usually revolves around the notions of secrecy, as in withholding knowledge, and securitization, which construe not, or only partially, publicly disclosing available information as being in the national best interest (Carson, 2018). Yet while secrecy is a central feature of remote warfare, Kearns concludes that it has been 'insufficiently conceptualized ' (2017: 13). In what follows, we take this up and engage with the work on 'secrecy as an affective mode of communication and a technology of power' that critical security scholars offer (Walters, 2015: 288; see also Cormac and Aldrich, 2018;Horn, 2012). Because while presenting secrecy as being in the public interest might be accepted as legitimate in particular situations, it is never self-evident. Secrecy, after all, goes against both political and legal understandings of the desirability of transparency, without which accountability, and ultimately democracy, are unattainable. The tension between performative democratic obsessions with transparency (Stampnitzky, 2020: 3) and the notion that states have a unique right to secrecy makes that secrecy needs permanent legitimation. This is often produced through processes of securitization, which position secrecy as central to the 'epistemology of security policy' (Walters, 2015: 288). In response to particular crises, publicity is construed as an existential threat that merits the exceptional measure of secrecy (Walters, 2015). This produces an iterative logic: secrecy is key in the response to crises and crises are simultaneously essential for the legitimation of secrecy. Cormac and Aldrich (2018) importantly add that 'secret' operations are often 'open secrets' of significant communicative value: strategically visible while simultaneously vigorously unacknowledged (see also Carson, 2018: 10;Krasmann, 2018).
Thus, despite modern democratic imaginations to the contrary, transparency and secrecy are not binary opposites (Birchall, 2011). What matters, therefore, is understanding the ways in which remote war actors negotiate the continuum of knowledge and information production and dissemination within and beyond their military assemblage.

Ignorance studies and the functionality of feigning and imposing unknowledge
What is striking in the discourses of Dutch officials about the bombardment on Hawija is not just denial -as reflected in initial claims that there were no civilian casualties -and secrecy -as evident in statements that the existence or number of civilian deaths could not be disclosed due to security concerns. What stands out is a more existential refutation of the possibility of knowing; denying that they (can) know of casualties altogether. Although such ignorance is rooted in practices of denial and secrecy, it also goes beyond them in important ways. We therefore consider it a distinct form of epistemic politics that demands a conceptualization beyond the centrality of secrecy that is routine in security studies' approach to matters of intelligence and classification.
Critical security studies recognizes that 'both knowledge and non-knowledge are constitutive for the formulation of security policy' (Daase and Kessler, 2007: 430). The emerging field of ignorance studies offers relevant handles to take our thinking on the functionality of denying or evading or impeding knowledge further. Ignorance here refers to various forms of not-knowing ranging from doubt to silence, ambiguity and uncertainty. Ignorance studies builds on multidisciplinary work on 'structural amnesia' (Arendt, 1958) and 'states of denial' (Cohen, 2011) to establish three key principles: that ignorance is pervasive, that it is socially constructed, and that it can be advantageous (Smithson, 2008: 209).
For our analysis, we draw specifically on the strands of ignorance studies that foreground the (political) functionality of ignorance by investigating what actors do not know, claim not to know and aspire not to know (Code, 2014;Croissant, 2014;Gross and McGoey, 2015;McGoey, 2012McGoey, , 2019Smithson, 2015;Stel, 2019). This work, often inspired by feminist and postcolonial studies of discursive and institutional silencing, crucially revisits relations between power and knowledge and theorizes the strategic nature of ignorance (Sullivan and Tuana, 2007;Tuana and Sullivan, 2006;Wylie, 2008). The notion of agnotology merits particular attention here and will inform our reading of ignorance studies below (Proctor and Schiebinger, 2008). Agnotology is specifically concerned with the study of 'agnogenesis', the process of generating or maintaining ignorancetracing its agential dimensions and studying how it becomes institutionalized.
Concretely, ignorance studies, specifically agnotology, can advance our understanding of the role of not-knowing in Dutch engagement with Hawija -and more broadly in what Demmers et al. (2020: 243) have called the capacity of remote war 'to ward off political questions' -by offering insights in how and why ignorance can be strategically produced. In understanding how not-knowing can be an effective strategy to accrue or maintain power, the distinction between defensive and what we will here discuss as offensive ignorance is informative. For McGoey (2019: 3), this distinction gets at the continuum between avoiding liability for past mistakes and generating support for future political initiatives. As such it underpins our argument that strategic ignorance is essential not merely to retaining the legitimacy of previous strikes but to upholding regimes of truth that enable future remote wars.

Defensive and offensive ignorance
Invocations of denial, secrecy and ignorance do not necessarily occur in linear or separated fashion. As the following conceptualization of defensive and offensive ignorance shows, denial and secrecy can precede but are also often part of ignorance claims. Indeed, shifting between various forms of epistemic politics is itself part of the imposition of ignorance.
Defensive ignorance refers to actively avoiding knowledge of a particular issue; to deliberately maintaining ignorance through evasion. It applies to situations where people arguably could have known but chose not to, resisting rather than pursuing knowledge (McGoey, 2019: 28). This might entail burying reports, disregarding inconvenient information and shelving potentially relevant questions. Defensive ignorance also refers to denying knowledge; to feigning or claiming ignorance. Often, these forms of ignorance overlap. Authorities might know a bit about something and choose to deny this and to not want to know more (Stel, 2020). Here, one can for instance think about European authorities' positioning towards coastguards' 'pushback' practices of migrants on the Mediterranean, which are extensively documented but routinely disregarded. This also relates to the assembled nature of state systems, which entails inevitable forms of tacit ignorance where different sections or levels of an organization do not know -and often prefer not to know -what other components are doing. In these dynamics, political ignorance is facilitated by institutionalized administrative anonymity and bureaucratic discretionary power to deflect and decentralize so that connections between 'battlefield and boardroom' are legitimately untraceable (Kearns, 2017: 15).
Distinct from this is something we term 'offensive ignorance', which refers to the ways in which authorities impose ignorance on others, by obstructing their access to information or generation of knowledge. This can happen through the myriad ways we have identified above in relation to repertoires of secrecy legitimized through security logics. Examples of this are the sabotaging of freedom of information requests or the preventing of access to specific sites and spaces. By conceptualizing such practices as 'offensive ignorance', what ignorance studies adds to this existing understanding of the political utility of secrecy is the insight that the creation and reproduction of secrecy and ambiguity are often facilitated through claims of unknowing: defensive and offensive ignorance are crucially related.
As authorities feign and maintain their own ignorance, they obstruct the production of knowledge by others. In our analysis below, we are primarily interested in the instances in which offensive ignorance ('you don't get to know this') is legitimized through instances of not just denial ('you don't get to know this because it never happened') or secrecy ('you don't get to know this because that would compromise security'), but defensive ignorance ('you don't get to know this because we don't know it and we can't know it').
This interplay between defensive and offensive ignorance can be read through McGoey's (2019) idea of 'ignorance pathways' that link 'macro' and 'micro' manifestations of not-knowing and draw attention to the ways in which the individual ignorance of people in positions of authoritywhether 'real' or pretence -structures access to information and knowledge of other societal actors and the broader public. This manifests itself in, for instance, the absence of not just governmental investigations, but -through this absence and other forms of obstruction -the undermining of journalistic or academic research.
It is here that our interpretation of ignorance studies touches upon the notion of regimes of truth outlined above. Offensive ignorance is an expression of power in that it shapes the general politics of truth in a society (Foucault, 1980: 133; see also Aradau, 2017: 331). The maintained and professed forms of authorities' defensive ignorance offensively shape the components of the regimes of truth -mechanisms, sanctions and status -that structure societal knowing and not-knowing. Authorities' denial of knowledge hampers other actors' 'mechanisms' for distinguishing between true and false statements as the very existence of a knowledge base for such a distinction is disavowed or put on hold. It produces particular 'sanctioning' logics as any formulation of unambiguous claims is discredited by pointing to the (deliberately upheld and potentially indefinite) unknowability professed by relevant authorities.
Strategic ignorance forces stakeholders beyond the war assemblage into a subject position 'focused on uncontrollable ephemerality' (Kearns, 2017: 20). This in turn has profound effects on the 'status' of both authorities and those seeking to scrutinize them. The former might effectively delegitimize the latter by casting any claim to knowledge or truth as problematic in light of the not just secret but unknown nature of a particular phenomenon -in our case the occurrence of civilian deaths in the Dutch bombardment on Hawija. An agnotological perspective suggests that, as McGoey (2019: 315) puts it, 'elite power functions through strategic ignorance -through the ability to select which voices to acknowledge and which to dismiss'. In this sense, ignorance is preemptive to untruth. Where denial often takes the form of lying, and secrecy manifests itself in the silence between 'truth' and 'lies', the combined mustering of defensive and offensive ignorance averts having to lie. This leads us to ignorance studies' take on why strategic not-knowing works. Here, the short answer is that claiming ignorance -and thereby partially imposing it -helps authorities avoid liability, responsibility and ultimately accountability. Protecting their 'ignorance alibis' can help authorities maintain and strengthen their positions of power (McGoey, 2019). The analytical challenge, then, is not so much to account for why not-knowing works, but to empirically locate its strategic nature. As McGoey (2012: 559) provocatively articulates it: 'The pyrrhic challenge for scholars of ignorance is to prove the existence of something for which the very ability to evade detection is a key criterion of success'. There is no quick fix here. The key heuristic exercise proposed by agnotology is to systematically interrogate something that is presented as a passive state -being ignorant -as a potentially active move -ignoring as evading (Smithson, 2008: 210).
While such a move is likely strategic, it might be unhelpful to see it as necessarily intentional as such intentionality supposes a consciousness that might often be absent as the result of powerful processes of cognitive bias. In addition, as Rappert (2012: 43) has convincingly shown, a narrow ambition to 'unmask' strategic unknowns as the result of 'individual or institutional duplicity' risks projecting onto actors an often unwarranted 'coherence in action and a competence in ability'. Moreover, even where such purposeful duplicity is likely to play a role, it may be inherently impossible to 'prove' (Proctor and Schiebinger, 2008). As Wedeen (1999: 6) has compellingly shown in her analysis of the 'ambiguities of domination', it is ultimately 'impossible to get into policymakers' heads and come away with exact knowledge of why they do what they do'. In searching for strategic dimensions of ignorance, then, ignorance studies points us to what might be thought of as the political functionality of ignorance, where the tracing of interests -partisan or personal, political or material -directs analysis. This comes with a focus on contentions as it is in the resistance against claims of not-knowing that passive states of ignorance can be revealed as active modes of ignoring.
For our attempt to understand the role of unknowing in Dutch accounts on the Hawija bombardment and the ontologies of remote warfare that this might be indicative of, an agnotological perspective is helpful because it enables us to understand the ways in which ignorance can sustain regimes of truth in ways that denial and secrecy cannot. As we will demonstrate with reference to our case study in the next section, feigning ignorance ('we don't know') involves imposing ignorance ('you can't know') that entails obstructing investigation and accountability potentially indefinitely. This surpasses mere denial ('it's not true') or secrecy ('we can't tell you'), which respectively allow for counter-truths or eventual disclosure. Whereas denial can be disproven and secrecy has an expiration date, ignorance is both more elusive and more open-ended and hence politically convenient in different ways.

Methodological and analytical reflections
In studying not so much the implementation of remote warfare but its legitimation and the regimes of truth that ensure its continuation, we follow Kearns's (2017: 14) observation that 'covert strikes are rarely documented as they are enacted, but instead are primarily known through snippets of information, rumour and debris'. We draw on these discursive (non-)traces of strikes to understand how such regimes of truth are constructed, contested and upheld.
This 'discursive debris' (Kearns, 2017) consisted of a range of primary sources, including video footage of Dutch parliamentary debates with the Ministers of Defence and the Prime Minister, parliamentary papers, (declassified) Coalition civilian casualty reports, governmental and nongovernmental policy documents, press releases and media reports. In line with the agnotological prescription to explore claims about passive conditions of being ignorant as potentially active moves of ignoring, in our analysis of the assembling of unknowledge in Dutch politicians' engagement with Hawija, we inevitably put a premium on contestation. Without relentless contention the known unknowns we now study would have remained epistemologically nigh unapproachable unknown unknowns (Daase and Kessler, 2007; see also Stampnitzky, 2020 andWalters, 2015: 289). The challenge here is to explore not just what is missing but also what is misleading (Cormac, 2017: 170). Our ability to sift through the debris and contestation was enhanced by holding 15 round tables with legal experts, investigative journalists, political stakeholders and NGOs (Airwars, Amnesty, PAX and Open State Foundation) working on the Hawija case between 2017 and 2020. 7 Mirroring previous conceptualizations of war actors as assemblages (Demmers and Gould, 2018) and following Aradau (2017: 328), we approach our analysis of the discursive debris left by the Hawija strikes as reflections of a process of assembling knowledge and unknowledge. In this process, we leverage the conceptualization on the dialectic between defensive and offensive ignorance outlined above. Through an iterative coding process of our source material, we aim to locate, attribute and explain (i.e. establish the functionality of) specific instances of offensive ignorance -which include denial and secrecy -and are particularly interested in the cases where these are produced through or legitimized by defensive ignorance. Tracing the genealogy of claims of unknowledge and establishing the audiences at which they are primarily directed works as a form of 'encircling' (De Goede et al., 2020: 9) that allows us to render these instances visible as a form of epistemic politics: rather than primarily seeking to uncover that which is ignored, which has already largely been done by investigative journalism, we focus on the practices, narratives and rites of the process of continued ignoring.
We take to heart Rappert's (2012: 42) warning that questioning claims of ignorance almost inevitably produces new forms of ignorance. Despite attempt to minimize these, traces of teleological reasoning or ambiguous terminology may shape our analysis as well. As discussed above, however, foregoing the quest to establish intent at all costs may dampen these risks. Moreover, while 'reading back from "the facts" in order to decipher who is trying to produce ignorance has severe limits' (Rappert, 2012: 55), too much is at stake not to nevertheless try (Cormac, 2017: 170). Our objective in this academic article, therefore, is not to 'uncover a singular, hidden truth' Luscombe, 2020 cited by De Goede et al., 2020: 4). Rather, we explore the power dynamics enabled by the epistemic politics involved in strategic unknowing in contexts of remote war. Our analysis can be read as a 'thick description' of what Stampnitzky (2020) has conceptualized as a fundamentally contested process of first 'exposure', the release of information, and gradually 'revelation', the more collective realization of such exposure.

Analysis: Hawija and the making and maintaining of regimes of truth through strategic ignorance
This section applies the concepts outlined above to the process through which various (war) actors enacted, contested and upheld different modes of knowledge and non-knowledge regarding the Coalition bombardment on Hawija to demonstrate the article's core claim that ignorance is politically convenient in different ways than denial and secrecy are in upholding regimes of truth on remote warfare. We show that whereas denial and secrecy put checks and balances on hold, ignorance dissolves them.
Our analysis starts after Dutch F-16s were no longer deployed in OIR, in the midst of the Dutch government's assembling and reassembling of defensive and offensive ignorance regarding the civilian harm caused by its aerial bombardments. Our entry point into this assemblage is the work of four journalists working for the Dutch news outlets NRC and NOS. On 18 October 2019 they revealed that two Dutch F-16s were in fact responsible for one of the deadliest Coalition airstrikes, namely the 2015 attack on Hawija (Schippers and Versteegh, 2019).
Up until this point in time, the airstrikes in Hawija and its estimated civilian casualties had been public knowledge, but the identity of the belligerent remained a secret. Two days after the strikes occurred in 2015, when questioned about the Hawija bombardment and its aftermath, Lieutenant General Hesterman stated 'we haven't seen any evidence of civilian casualties so far, but we'll conscientiously look into it as we do with every allegation.' 8 During the ensuing six months, both Airwars 9 and the UN (2015) reported that an estimated 70-150 civilians were killed. While the US CENTCOM civilian casualty investigation report on Hawija remained classified, in December 2018 the then spokesman of the Coalition, Colonel Sean Ryan, did explicitly admit in an on-therecord email to the NRC reporters Boon and Versteegh (2019) that 'The strike [. . .] unfortunately killed 70 civilians'. Yet when questioned about who was responsible for the attack, the broader Coalition and the Dutch MoD kept the identity of the belligerent under wraps, the former by claiming it was up to individual countries to claim responsibility for attacks and the latter by legitimizing its secrecy in the name of maintaining 'national, personnel and operational security' (Airwars, 2018).
In response to this ambiguity, through open source intelligence and interviews with key military personnel, the Dutch reporters managed to pinpoint that it were Dutch F-16s that had bombed Hawija by ruling out the possibility of any other Coalition member's involvement (Schippers and Versteegh, 2019). Two weeks after the publication, the Dutch Minister of Defence, Ank Bijleveld, admitted in a letter to Parliament that the Netherlands was indeed responsible for the Hawija attack. In the letter she refers to the journalists' unveiling of the lines of responsibility. She argues that considering the Netherlands no longer conducts airstrikes, she could now reveal 'the exact dates, times and estimated civilian casualties' of those incidents in which 'the Netherlands themselves have concluded that civilian casualties can be surely or most probably regretted as a direct consequence of the Dutch weapon deployment' without posing a 'threat to security'. 10 We see here that faced with undisputable evidence, the Dutch MoD could no longer legitimize secrecy vis-à-vis a military deployment that had come to an end and could therefore no longer postpone disclosing responsibility for the Hawija attack.

Shifting discourses from 'who bombed?' to 'who died?'
In the same letter to Parliament, however, Bijleveld reassembles defensive and offensive ignorance about the number and nature of the 70 casualties that occurred. First, she highlights that CENTCOM is the only Coalition member that has 'the capacity, expertise and intel to investigate the credibility of allegations'. 11 Bijleveld thus defends her own ignorance about the exact number and nature of the civilian causalities by underlining that her Ministry has not developed the ability to generate knowledge about civilian casualties themselves. Instead, it depends on its US ally and the commander of OIR, CENTCOM, whom is conjured as exclusively having the necessary techniques and procedures to count the bodies of the dead.
Secondly, she alleges that while according to 'CENTCOM's cited open sources, around 70 casualties occurred [. . .], including both IS fighters and civilians', 'the ratio between IS fighters and civilians could not be established after the event'. 12 Bijleveld emphasizes that even for CENTCOM it is hard to know the true number or nature of the casualties, 'due to a lack of on-theground investigations, the fact that the dead get buried quickly in Muslim communities, incomplete population registers, and [. . .] the difficulty of distinguishing retrospectively between IS fighters and civilians'. Bijleveld thus not merely passes the buck but undermines the reliability of CENTCOM's methods (and thus anybody else's) to produce unambiguous knowledge on the quantity of the casualties. In other words: We can't know how many people died exactly. Moreover, the identity of the casualties is questioned. Here she shifts the onus of proof to the casualties themselves, suggesting the victims should have (posthumously) proven their civilian innocence. The implication here is: We can't know how many of them didn't deserve to die. She thus creates ambiguity along two axes: the quantity and quality of the harm done.
Finally, Bijleveld explains that if CENTCOM does find credible civilian casualty cases they immediately inform the Coalition members responsible for the attack: 'The country in question then decides independently what it does with that information', but the CENTCOM report itself is classified. 13 This account shows how the Coalition offensively obstructs anybody from outside the military assemblage gaining access to CENTCOM's civilian casualty reports by keeping them classified. So, while the Dutch MoD could no longer deny liability for the attack on Hawija, by strategically (re)assembling defensive and offensive ignorance between and beyond the Coalition members, acknowledgement for the severity of the civilian harm done was nevertheless circumvented.
Inadvertently, and for the interrogation of strategic ignorance crucially, by explaining how the Dutch MoD received the classified CENTCOM civilian casualties report on the Hawija attack on 15 June 2015 (13 days after the attack), Bijleveld laid bare that her predecessor Hennis had in fact misinformed Parliament when on 23 June 2015 Hennis stated that 'there are no known cases of civilian casualties caused by Dutch bombardments'. 14 Indeed, not only did Hennis at the time deny knowledge of civilian casualties, she underlined why it was uncalled for to question the perfect nature of the remote war, stating that 'the (targeting) process is so precise that we have not had any incidents so far', adding the unprompted and in hindsight painful embellishment that 'it is not as if you can destroy a whole suburb or district [. . .] because of the smart weapons I was just telling you about'. 15 At the time, Hennis's denial of civilian casualties and the concomitant reinforcement of the regime of truth on remote warfare was a form of 'active denial'. Such denial served a clear purpose: on 19 June 2015, the Dutch cabinet decided to prolong participation in OIR. Denying knowledge and thus feigning ignorance about the civilian deaths that had occurred in Hawija just weeks earlier prevented inconvenient parliamentary questions and possible public outcry. This allowed Hennis to pursue a continuation of the Dutch involvement in the Coalition.

The political functionality of 'unknowing' the number and nature of civilian deaths
Hennis's denial, however, did not suspend inconvenient questions indefinitely. The inability of her successor Bijleveld to reassemble Hennis's lie led to a major uproar within Parliament and to three parliamentary debates on the topic at the end of 2019. 16 Herein Parliament not only underlined that Hennis had violated its right to information in 2015, but also that both Bijleveld and Prime Minister Rutte must have known about Hennis's strategic denial of probable civilian casualties and thus were also responsible for misinforming Parliament.
In the three debates, Rutte and Bijleveld responded to these allegations with both offensive and defensive forms of ignorance. Rutte stated: 'The truth is that I have absolutely no memory of being informed in June 2015; I cannot guarantee that it did not happen but I do not have a memory of it.' 17 Bijleveld relayed that when she came to power in 2017 she was informed 'in a general sense about the main features of the Hawija bombardment and the likelihood of civilian casualties [. . .], but not in detail'. 18 She backed up her choice to keep herself and Parliament ignorant, by stating that even if she had known more she could not have shared this information with Parliament due to 'operational, personal and national security '. 19 This refers to the fact that in the same year she entered office, the Dutch government decided to re-enter its F-16s in the Coalition from January 2018 onwards, after having suspended their contribution in June 2016. The inconvenient truth about the civilian casualties in Hawija would not have worked to advance support for the MoD's plea to return to air war as it risked undermining the regimes of truth that such support crucially hinges on.
Finally, both Bijleveld and Rutte repeatedly claimed -contrary to the information that the two Dutch journalists received from the Coalition spokesperson -that their official contact 'high up' at CENTCOM had reconfirmed to them that the 70 civilian casualties were not part of CENTCOM's official civilian body count (at the time 1377). 20 Rutte stated that the number 70 was based on 'open sources' and therefore was a 'rumour' and 'irrelevant' and CENTCOM had told him personally that 'there is no way of knowing how many civilian causalities perished'. 21 Hereby Rutte both undermined the status of these 'open sources' -Reuters, the Red Cross, the UN and Airwars -and sanctioned any claims about the number of civilian casualties by pointing to the unknowability professed by CENTCOM who, in his view, was the only authority with the status and techniques to count the bodies of the civilian deaths and therefore also the only authority that could profess if and how the number of casualties was knowable at all. 22 This combination of offensive and defensive ignorance was enough to appease the majority of Parliament, as Bijleveld survived two and Rutte one vote of no confidence. Parliament did, however, demand that Bijleveld ask CENTCOM for 'clear and undisputed communication on paper' about whether the 70 civilian casualties were counted in their official statistics and to declassify its Hawija civilian casualty report. 23 Three months later, Bijleveld informed Parliament that the US Deputy Under Secretary of Defense had now confirmed to her that CENTCOM had undeniably been counting the 70 civilian casualties in their official body count for three years after all. 24 This de-assembly move was prompted by the fact that just a week later, the classified CENTCOM investigation into the Hawija bombardment became public knowledge through a Freedom of Information Act request submitted by the same Dutch journalists in the United States. The declassified report that the Dutch MoD received in June 2015 revealed that CENTCOM acknowledged the 70 civilian causalities as credible immediately after the attack. With this report now out in the public domain, the United States could no longer stand in as the Netherlands' ignorance alibi without undermining its own status and authority: their epistemic alliance was broken.
This led to another 160 parliamentary questions and a fourth debate. Once again, Bijleveld justified her own and Parliament's past and present ignorance about the number of civilian casualties in Hawija by stating that she herself had not read the classified CENTCOM report, that the Dutch MoD could not have shared the content of the report earlier with Parliament because it was classified, and that she regretted that she and Rutte had repeatedly been wrongly informed by their contact within CENTCOM about the official body count. 25 Finally, she emphasized that even if the 70 were part of CENTCOM's official body count, it was still not possible for anybody to know how many of those casualties were civilians and how many were IS fighters. This final claim reiterates how the unknowability of the number and nature of the casualties is used as a subsequent line of defensive ignorance. First Bijleveld and Rutte chiefly claim ignorance and unknowability about the number of civilian casualties and, once this has become untenable, they claim ignorance and unknowability about the nature of the casualties, questioning the exact numbers not so much of casualties, but of civilian casualties. This feigning of offensive and defensive ignorance allowed Bijleveld to survive a final vote of no confidence.
This account illustrates how offensive ignorance ('you don't get to know this') vis-à-vis the Dutch Parliament was justified by the Dutch government, first through instances of denial by Hennis in 2015 ('you don't get to know this because it never happened'), then through secrecy by Bijleveld in 2017 ('you don't get to know this because that would compromise security'), and finally through defensive ignorance by Bijleveld and Rutte in 2019-2020 ('you don't get to know this, because we don't know it, because it can't be known').
Reflecting on the war context in which these various forms of ignorance were upheld sheds light on their political functionality. Hennis's denial of civilian casualties in the immediate aftermath of Hawija allowed her to maintain parliamentary support to continue the war and be promoted to the position of UN representative for Iraq before inconvenient counter-truths came out. Bijleveld's secrecy allowed her to restart an air campaign in 2018. Rutte's and Bijleveld's subsequent assembled offensive ignorance about the number and nature of civilian casualties allowed them to maintain their position of power when they could no longer construct ambiguity about the Netherlands' responsibility for the attack. Finally, their ongoing ability to sanction the techniques used by CENTCOM (and any other source) to create valuable knowledge on the true identity of the casualties innately undermines the status of victims who seek justice for the attack.
Our analysis ultimately reveals how 'knowing what not to know' can serve the interests of polities beyond the specific historicity of 'this war'. The political significance of strategic ignorance is that it produces a subject position that is focused 'not on the infliction of violence but on the struggle to comprehend intangible state practices' (Kearns, 2017: 14). During the 22 hours of debate, parliamentarians indeed struggled to comprehend how Coalition members created knowledge on civilian casualties, how they shared this information within the military assemblage, and what they recollected of this. This prevented the debate moving on to pose questions about the logic of the violence itself, such as: Why are we waging a war that is so remote that we cannot distinguish between whether we have hit our enemy or a civilian? How can we then claim that these wars are precise? What kind of image does this leave of Western 'liberation'? Should we not reconsider our contribution to this and the next remote war? This marginalization of violence, by generating 'a struggle to comprehend', thus prevents the questioning of dominant regimes of truth on remote warfare and sustains its power to legitimize the next war.

Reflections: Interrogating agnogenesis
Our analysis reinforces the ontological perspective on war as the always uncertain (re)working of meaning, truth and order through violent means. What we have flagged as particularly interesting about remote war is its capacity to avert political questions on how it has transformative effects on politics and society at home and abroad. Through the imposition of strategic ignorance that is strengthened by distance, remote military violence seems to (re)produce and sustain regimes of truth that apparently defy scrutiny, debate and questioning.
But what does such evasion through ignorance mean for contention and accountability? The strategic ignorance we investigated here, after all, has as its main aim to 'manipulate domestic elites and public opinion' (Carson, 2018: 5). Security and conflict analysts have already been grappling with the question of how to 'proceed when political actors invoke national security and enact closure so as to minimize their need to furnish public rationalizations and justifications for their actions' (Walters, 2015: 288; see also Carson, 2018). But how to analytically and politically resist a war assembly's strategic ignorance? This question becomes even more pertinent since critical and constructivist work on epistemic politics has been blamed for 'doing the intellectual groundwork for the current rise of post-truth' (Aradau and Huysmans, 2019: 41).
First, contention is not impossible. As we have reflected above, it is only through the various journalistic, watchdog and political challenges by civil society actors to Dutch politics that we have been able to access our case study. The 'ambient accountability' demanded by such collectively extracted transparency should not be underestimated (Aldrich and Richterova, 2018). As Walters (2015: 289) points out, the fact that a policy domain is 'governed according to various conventions of secrecy' does 'not necessarily serve to diminish its public profile'. This means that contention in the face of denial, secrecy and unknowledge can never be a generic 'call for clarity or precision' or a 'simple increase of knowledge, data and transparency' (Aradau, 2017: 338;Daase and Kessler, 2007: 431). Rather, as we have demonstrated, it might be seen as a targeted investigation of the political functionality of ignorance claims that could expose which fundamental questions are avoided and sidelined in the sometimes ritualistic tug between respective proponents of secrecy and transparency. Second, that contention is possible does not mean it will happen. The main battlefield now is against the voluntary ignorance of political constituencies in the Global North, who are complicit in the assembled unknowing we described by their consent for forms of warfare that cater to fatigue of war as boots-on-the-ground rather than of war as such (Cohen, 2011;Cormac and Aldrich, 2018). What might be fruitfully highlighted in this struggle against tacit public ignorance is that the danger of remote warfare is not simply military and its costs not merely humanitarian. They are ultimately fundamentally political. Indeed, the most potent potential for popular contention, we believe, appears at the nexus of revealing death and destruction 'over there' and the erosion of democracy 'over here'. The 'blowback' of remote warfare is more pervasive than terrorist attacks (Waldman, 2021). Denying not just the existence of civilian casualties but the ability to know about them effectively undermines institutional checks and balances and democratic accountability -in the realm of war but potentially also beyond.
We should recognize the asymmetrical struggle for information about casualties as a mirror image of the asymmetrical nature of remote warfare itself. But if we hope to prevent ignorance about today's wars legitimizing tomorrow's, we should not be blinded by this mirror to the extent that the quest for information about violence replaces the questioning of the logic of violence.