- Top of page
- Indicators: Measuring violence against women
- Discussion and future directions
- References cited
Science and technology studies can help to unveil invisible modes of power that are embedded in the ways conflicts are known, debated, and resolved. Legal forms of adjudication, reporting systems used by international commissions, and data gathering on the part of governmental and nongovernmental agencies shape how conflicts are fought out on the ground and in policy arenas. Assumptions about evidence, categorization, adjudication, and measurement privilege certain forms of suffering over others, even as they omit phenomena that defy categorization. Using two examples—a global survey of violence against women and a U.S. government initiative to defer the deportation of certain undocumented immigrants—we bring insights from science and technology studies to bear on sociolegal phenomena. In so doing, we highlight tensions between measurement and invention, visibility and invisibility, and objectivity and discretion that are intrinsic to new forms of governance. We thus examine what measurement initiates and precludes, the reactive and proactive nature of technologies, and how new practices reproduce established techniques.
In its most insidious form, power is embedded in social processes that render its workings invisible and thus preclude resistance or challenge. Social problems such as gender violence, entrenched poverty, and mass displacements of persons demand responses from both policy makers and the public at large. Responding, however, requires making problems visible, defining their scope, identifying aspects that can be remedied, and implementing and evaluating solutions. Technologies of knowledge, such as the audit (Power 1999; Shore and Wright 2000; Strathern 2000), are therefore used both to assess and to shape social realities. Though they appear neutral or apolitical, these technologies produce and reinforce hierarchies between what is “knowable” and what is not. As a result, knowledge systems are part of conflicts rather than extrinsic to them.
In this article, we draw on science and technology studies approaches to reveal these invisible modes of power. Intersecting with governmentality studies (Foucault 1991; O'Malley 1999; Rose 1996, 1999; Rose et al. 2006), anthropological and related work in the science and technology tradition (Cole 2002; Mertz 2011; Riles 2001, 2006) has identified techniques embedded in the ways conflicts become known, debated, and resolved. We argue that fully understanding the political dimensions of social conflict requires attending not only to the substantive issues involved but also to these knowledge systems and their assumptions. Paradoxically, the act of measurement can produce both the supposedly preexisting phenomena being measured and the world in which these phenomena come into being (Callon et al. 2002; Espeland and Sauder 2007; Espeland and Stevens 1998, 2008). A national census, for example, constructs the very categories by which the population is known. And, as Andrea S. Ballestero (2012:241) shows, systems of audit, designed to manage and supervise, can be appropriated as modes of self-definition that create, rather than merely document, facts about selves. Likewise, measurement technologies can produce certainties out of ambiguous and contested situations. Measuring requires fitting people, objects, and events into categories in ways that obscure ambiguity (e.g., as happens in the use of gender as a binary category), thus constructing a façade of certainty over blurred social boundaries. Yet measurement systems rarely report to their publics where categories come from or what they leave out. Measurement systems make some things visible and others invisible. For instance, if work is defined by receiving wages, women's unpaid labor in domestic settings is inevitably uncounted and therefore ignored (Waring 1999).
Thus, these seemingly apolitical techniques can transform conflicts of interest into questions of knowledge and existence. Political struggles may become issues of evidence and data in ways that privilege numbers, hard evidence, or efficiency over substantive justice—and, for these reasons, enumeration and documentation is sometimes resisted (Muehlmann 2012). For example, evaluating undocumented immigrants' legalization cases focuses attention on the kinds of evidence necessary for individuals to qualify for legal status rather than on the conditions that lead them to immigrate without authorization. Likewise, carrying out a survey of the frequency of rape by interviewing people in a relatively public setting that discourages them from revealing their victimization will undercount the problem and minimize it as a political issue. In both cases, a technology for producing “truth” shapes the way problems are understood and managed.
We argue that, to understand the political dynamics of conflict, attending to the features of these knowledge systems is as important as examining the violence, injustice, and merit that such systems assess. Although the pain that they cause defies narration (Scarry 1985), violence and injustice become socially known through naming, counting, and adjudicating. Without in any way downplaying the violence and injustice that these systems are designed to combat, we seek to foreground what anthropological analyses usually background, namely, technologies of measurement and assessment that are increasingly a component of administrative “problem solving” (for an exception, see Muehlmann 2012).
The forms of measurement, quantification, and evaluation that we take up in this article are part of a broader move toward what has variously been called “new governance,” “experimentalism,” and “results-based management”: essentially, a shift from a command-and-control strategy of governance to collaborative, consensus-building discussions focused on problem solving and improvement (Simon 2004:11–28). This model encourages learning and innovation, constant collaborative revision, and the participation of multiple stakeholders (De Burca 2010; Rose 1991; Simon 2004). The use of evidence organized by guidelines, standards, metrics, and performance evaluations is essential to decision making. In contrast to earlier systems, which relied on rules and punishments, new governance relies extensively on “soft law” in that it shapes behavior by establishing standards; requiring individuals, groups, corporations, and even nations to report on how they have met them; and comparing results (De Burca 2010; Trubek and Trubek 2005). For example, countries that fail to meet targets or that are ranked below others on key indicators are to be “shamed” into improving their records (see also Maurer 2005; Trubek and Trubek 2005).
The quantification involved in such calculations takes multiple forms: counting instances of particular types of phenomena within populations, reconciling different “measures” in the creation of an index (e.g., a “Corruption Perceptions Index”; see Transparency International 2012), producing a ranking (which is ordinal and relational),1 and determining ranges and intervals (such as the interval of time that someone has lived in the United States). Conflicts are thus embedded in the production and assessment of particular sorts of records.
The techniques of knowledge production associated with new governance are fundamental to bureaucratic forms of power. While common under neoliberalism, use of metrics and audits is also found under Chinese state socialism (Kipnis 2008). Some countries have even developed their own metrics, such as Vanuatu's Alternative Indicators of Well-Being, that evaluate development through customary practices rather than economic markers (Republic of Vanuatu 2010). Though seemingly mundane, metrics perpetrate a form of symbolic violence by rendering disparate phenomena commensurate, a process that requires reconciling or omitting elements of experience that do not fit categories through which data are analyzed (Espeland and Stevens 1998). Grouping disparate things and people according to common traits inevitably ignores some features while privileging others, thus cutting off more “messy” and therefore incommensurable aspects of persons, objects, and events. U.S. census makers' struggles to generate meaningful categories for designating race and ethnicity vividly illustrate the difficulty of adequately measuring complex identities (Lee and Bean 2010). And, of course, nonstate entities sometimes deploy numbers in inventive and even oppositional ways, as Occupy Wall Street did with its famous 99% (Verran 2010). Likewise, by claiming to be the “67%” who will not qualify for legalization under proposed DREAM Act legislation, some undocumented students have challenged dichotomous contrasts between “deserving” high-achieving students and more “undeserving” youth who have been convicted of crimes (see pablop 2011).
Though they are not the same as physical injury or personal displacement, measurement practices create knowledge that may affect these forms of violence. On the one hand, data gathering contributes to transparency and thus makes it possible for advocates to deploy data in ways that can advance social justice in an effort to prevent violence. Efforts to understand the “truth” of social problems and political conflicts may reduce violence, as when a truth commission report makes state violence visible and strengthens societal opposition to repression (Wilson 2001). On the other hand, data gathering is connected to surveillance and social control. Measurement systems may exacerbate structural violence, as when the denial of legal status to undocumented populations forces long-term separation of families or leads people to immigrate without authorization, along dangerous routes where many die (Menjívar and Abrego 2012).
Measurement systems require data, yet phenomena such as corruption or the rule of law are difficult or expensive to measure directly. Consequently, measurement systems often rely on proxies, or indirect measures. Documents, survey results, expert opinions, and administrative records approximate and stand in for social realities. For example, while a policy maker may wish to know the extent of human rights consciousness in a country, direct measures of such a nebulous phenomenon are typically absent. So, instead, he or she may count the number of people who receive training in human rights consciousness, a kind of administrative data that is far more readily available. Similarly, it is difficult to determine how many people experience interpersonal violence but far easier to count how many contact the police, even though such data have long been recognized as seriously undercounting violence. Likewise, to assess whether an immigrant has been continuously present in the United States, an adjudicator may rely on proxies of presence, such as receipts or school records, in the absence of direct knowledge of a person's history. Proxies inevitably shape the way measurements construct knowledge.
Whether direct or indirect, the development of metrics makes exercises of discretion appear definitive and unassailable, thus enabling a decision or calculation (which could have come out otherwise) to take on characteristics of an object (that has force and can circulate further). In their analysis of the historical emergence of the concept of “objectivity” in mid-19th century science, Lorraine Daston and Peter Gallison (2007:35) show that achieving an “objective view” required effacing oneself and leaving as little space as possible for discretion. The considerable interpretive work entailed in measurement can thus be hidden under the claim to objectivity.
The temporal processes of both knowledge production and law can be imagined by the metaphor of archaeology. We use this metaphor to refer to the way that data, documents, statutes, court cases, notices, and records are excavated by practitioners from layered and overlapping sites and used to make decisions at particular historical moments (see also Merry 2004).2 Both law and systems of measurement are artifacts that are constructed at least partially out of preexisting material, and, in this sense, they participate in the citational practices that are a characteristic of language itself (Butler 1993:220).3 Paralleling analysis in science and technology studies of the practice of politics, we argue that law, similarly, “comes into existence through a large number of material arrangements and technical objects” (Asdal et al. 2007:44). Crafting a statute, writing an opinion, creating a file, or issuing a document entails entextualization, that is, excerpting elements of other texts, documents, or records and redeploying them in a new case or context (Bauman and Briggs 1990; Coutin 2011; Richland 2008; Urban 1996). These redeployments invoke texts that have already been deemed authoritative, make use of agreed-on language, ensure that a new policy applies to a previously delineated population, and occur as part of corrective law-making cycles (Halliday and Carruthers 2007; Riles 1998). Each instantiation of law therefore builds on prior instantiations, even as each survey or analysis of data as well as its presentation rely on past templates, surveys, and data-analysis strategies. Both law and measurement systems are, in a sense, the residue of prior negotiations, a residue that leads forward as well as into the past. In both cases, there can be considerable inertia as procedures and categories become naturalized. As science and technology studies scholarship indicates, over time, such systems divide up the world in ways that come to seem natural and real, for example, specifying the point at which a fetus becomes human (Jasanoff 2011) or the sign that marks a disease (Mol 1999).
Attention to such layering is critical because the documents and texts of which both law and measurement are composed do not remain confined to a single historical stratum. Rather, past models and templates are brought forward as techniques of resolving new problems, while legal artifacts produced in one forum can reshape the claims being staked in another (see also Mather and Yngvesson 1980–81). Such temporal incorporation may reproduce particular inequalities but also has the potential to give rise to reinterpretation and change (Coutin 2011). Legal and technical knowledge systems thus make particular versions of social reality visible, even as they obscure others, rendering them unknowable. We therefore ask: What is set in motion and what is precluded through the act of measurement? How are technologies reactive and proactive? And do how techniques, once established, continue to influence forms of knowledge through which governance occurs?
To address these questions, we consider two examples from our own research: first, efforts to assess the prevalence of gender violence around the world, and, second, the Deferred Action for Childhood Arrivals (DACA) request process that the Obama administration recently implemented as a means of enabling undocumented students to remain temporarily in the United States. These two examples differ in key respects. For example, the production of gender violence indicators requires aggregating individual-level data, whereas the DACA request process studies requestors' histories and thus entails individuation. At the same time, by bringing these practices together here, we seek to illustrate common forms that emerge in administrative solutions to social problems and conflicts. As our case studies show, the work entailed in creating and assessing records achieves a temporary solution—the development of gender violence indicators and the awarding of DACA status to undocumented students—even as it also defers (explicitly in the name Deferred Action for Childhood Arrivals) the actual prevention of gender violence, the removal of undocumented immigrants, or the award of permanent legal status to another moment. After presenting these two examples, we conclude by outlining how the approach taken here might inform future work in the anthropology of conflict.
Indicators: Measuring violence against women
- Top of page
- Indicators: Measuring violence against women
- Discussion and future directions
- References cited
Since the 1990s, there has been an explosion in the use of quantitative measurements known as “indicators” that claim to describe the world. Indicators produce general knowledge for various publics and, at the same time, directly or indirectly enter into political decision making. Some determine which countries receive aid from the United States or which ones are sanctioned for failing to act aggressively against trafficking. Large-scale global databases not only provide useful information about global patterns but also are used by policy makers for governance. Broad standards are converted into measurement systems such as grades or human rights compliance, and responsibility for conformity with the standards lies with those being governed (O'Malley 1999; Rose 1989). This use of statistical information is fundamental to new forms of governance that base political decisions on evidence. As such knowledge acquires political salience, it becomes important to understand how it is produced and what it means. A genealogical approach that traces how indicators are put together, how their categories are formed, what their underlying theoretical framework is, and what forms of advocacy shape the selection of measures and proxy data for those measures is the best way to understand how indicators produce knowledge. Unpacking the process is important because this knowledge often serves as a means of governance and an adjunct to law.
One example of the development of indicators is the effort, in response to a global movement in place since the 1980s, to measure violence against women. Although there are multiple initiatives, here we discuss one being promoted by the United Nations, through the UN Statistical Commission (UNSC), made up of government representatives. In response to a major report from the UN secretary-general in 2006 that called for greater global action against violence against women, including collecting better data on the extent of the problem, the UN General Assembly (2006) asked the UNSC to develop indicators to measure such violence. In 2007, an expert group met to develop possible indicators. It included academics with substantial research experience on violence against women, representatives of UN regional bodies, NGO representatives, representatives of the UN Statistical Division (the secretariat of the UNSC), and representatives of countries that had some survey experience (UN Division for the Advancement of Women 2007). The invited paper, written by a sociologist from the United Kingdom with considerable experience doing surveys in that country, proposed a set of twelve or so indicators (Walby 2007). Even though, internationally, violence against women is defined broadly to include many forms, such as domestic violence, sex trafficking, and honor killing, this group proposed a more limited definition focusing on physical violence, sexual violence, intimate partner violence, and certain harmful practices (female genital mutilation and early marriage) (UN Commission on the Status of Women 2008:7; UN Division for the Advancement of Women 2007:25–27).
These expert group recommendations were taken to the UNSC, which appointed a committee, the Friends of the Chair (FoC), made up of country representatives, to propose a set of indicators for it to review. In 2009, the FoC presented its first draft of recommended indicators, lifted directly from the 2007 expert group meeting. At the UNSC meeting to consider this draft, which Merry attended, about forty-five minutes was spent in open discussion of what should or should not be included on the list of indicators. Focused explicitly on how violence against women was to be defined and measured, the discussion implicitly was about which local practices should be included. Representatives from member states expressed their views about the proposed list of indicators. For example, some wanted to drop female genital cutting because it was too sensitive an issue, while others wanted to delete child marriage because their country's legal age of marriage was under 18. China suggested including emotional violence. Thus, after some public debate and contestation, the UNSC charged the FoC with finalizing the list of indicators.
The subsequent meeting of the FoC committee took place in Mexico in late 2009. The committee consisted of about fifteen government members, mostly statisticians. They retained the basic set of proposed indicators but, in response to the UNSC discussion, dropped child marriage. They reinstated female genital cutting and psychological and emotional violence. The 2009 FoC report, with its list of indicators, was on the agenda of the annual meeting of the UNSC in early 2010. It was accepted without further discussion and was not discussed at the commission's 2011 and 2012 meetings, even though it was listed on the agenda each time. A guidebook for carrying out surveys on the basis of the indicators, intended for use by national statistical offices, was released in 2013 (UN Statistics Division 2013).
The list of core indicators approved by the UNSC defines violence against women much more narrowly than is accepted within the global women's movement.4 It measures the incidence of physical and sexual violence in intimate and nonintimate relationships plus female genital cutting and emotional and economic violence. It leaves out state violence by police and the military, harassment and threats, stalking, female foeticide, isolation, intimidation, and many other forms of violence, such as violence against men. It categorizes violence against women by type of relationship, type of violence, severity of violence, and frequency. The complexity of women's relationships, their histories, the interplay of love and fear, and the role of kinship and residence patterns disappear. Instead, a series of relatively discrete and apparently objective categories are used to measure this phenomenon across class, national, and cultural divides.
One story of domestic violence illustrates the gap between this quantitative accounting and the experience of violence and dramatizes the way numerical representations can flatten violence in comparison to narratives. It comes from a document written by a woman named Dora (a pseudonym) who had experienced violence and was seeking redress from the court. Dora was living in a small town in Hawai‘i during the early 1990s, at the beginning of the battered women's movement. She was in her early twenties, a mainland white woman from a middle-class family, had two years of college and an adequate family income, and was married to Sam, a Samoan man. Like many other battered women, Dora turned to the courts only after enduring years of violence by her husband. She wrote the following account in 1992 as a request for compensation as a crime victim. Merry also talked to her about her experiences.
Sam and I have been together for almost five years. There has been abuse on and off for the first few years. This past year has been the worst, it got to the point where he would beat me at least once a day and for about four weeks he beat me two or three times a day. It was so hard living with him. I have no family out here, only myself and our son. I lived in constant fear of Sam, never knowing of his coming here, afraid of what he was going to be like. Sam has threatened me with guns, spear guns, knife on one occasion. He would drag me down the hill by my hair, rip my clothes off of me, smash pans over my head. We had to replace or fix all but two doors in our house because he threw me through the other doors.
There was so much constant abuse it seemed like it would never end. Many times I thought that when I died it would be because my husband killed me. I was afraid to have him arrested because I knew he wouldn't stay in that long and I thought that he would kill me when he got out. Finally, on May 31, 1992, I couldn't deal with it. We were driving home from Hilo, my husband was sitting in the back of our truck. I was driving because Sam was too drunk. We were driving down the road and he reached through the back window and grabbed my face, scratching my face, then he tried to choke me and I felt that if he got open the door he would kill me. I looked over at my son in his car seat. He was frightened, screaming, crying and I knew I couldn't put up with this terror any more. I managed to drive away when he got out of the back to open my door. I just wanted the hell that my life had become to end.
As this story indicates, violence against women should be viewed in the context of a range of relationships, events over time, and meanings of masculinity and femininity, intertwined with ideas of race, class, region, and so on. Dora attributed Sam's violence to his Samoan ethnicity, for example, and felt especially vulnerable because she was far away from her family. Such systems of kinship and meaning are excluded when domestic violence is counted, yet this simplification is essential to the production of comparative knowledge.
At the same time that the UNSC was developing its measures, the Economic Commission for Europe and the Economic Commission for Latin America and the Caribbean together developed a survey module to measure violence against women globally. They adopted a broader vision of violence that incorporates social isolation and emotional abuse as well as acts, their frequency, and the injuries they produce. The model for this survey was one developed by the WHO in the early 2000s and published in 2005. The WHO survey adopted a feminist framework that emphasized women's safety during the interview, provided support and privacy for those who talked about their violence, and maximized disclosure of violence by posing questions in a range of ways. In contrast, the UNSC initiative was not inspired by a concern about the issue of violence against women but by its overall mandate of enhancing the statistical capacity of countries, particularly those with limited resources for data collection and analysis. Its goal was to produce a manual that national statistical offices could use to carry out surveys of violence against women. Issues of safety and disclosure were not major concerns. The European and Latin American–Caribbean initiative relied on social science scholars with extensive knowledge of gender-based violence and survey techniques developed for measuring it, whereas the UNSC, and its secretariat, the UN Statistical Division, relied on its experience designing health, housing, and victimization surveys and developing guidelines to assist national statistical offices in carrying out surveys.
These groups took quite different approaches to the problems of commensuration and data collection. The European and Latin American–Caribbean initiative sought to define gender violence broadly and maximize disclosure of violence while the UNSC–UN Statistical Division project was intended to develop national statistical capacity and create a mechanism that governments would use. The first had a more feminist political stance concerned with maximizing disclosure, and the second had a more statistical orientation, endeavoring to develop clear categories and measuring only what was countable and what governments were willing to count. Clearly, the process of creating such surveys has political dimensions. The knowledge they produce will inevitably differ.
One of the indicators in the UNSC plan was the severity of violence. This meant that it was necessary to develop a system for classifying violence as moderate or severe. Distinguishing severe and moderate forms of violence in a way that can cross boundaries of nation, religion, ethnicity, and class is extremely difficult. Yet global surveys require categories that can travel across such cultural borders while remaining commensurable. This situation creates a paradox: The survey categories need to be translated into local terms to measure local ideas and behavior accurately but need to retain their universal meanings to make comparisons possible across these borders. The categories must refer to the same thing wherever they are used even though the underlying phenomenon being measured manifests itself differently in different places.
Richard Rottenburg (2009) has explored this tension in his analysis of processes of translation and code switching in the field of development planning. Technologies of representation and inscription produce the facts of this domain (Rottenburg 2012:3; see also Rottenburg 2009:177–200). In development work, a generic language, or “metacode,” capable of crossing cultural boundaries is juxtaposed to a “cultural code,” a language imagined as the province of particular sociocultural groups. The technical language of the metacode assumes that reality can be represented without distortion and that it can exist alongside distinct cultural codes. The metacode is, of course, also culturally situated but is pragmatically treated as universal to make the development project possible (Rottenburg 2012:4). Constant switching between the codes occurs in development planning as the various parties negotiate using the metacode while attributing differences to the cultural code. Merry (2006) found a similar pattern in UN discussions about gender, in which participants assumed a universal understanding of gender identities while relegating differences to the domain of culture.
Rottenburg (2009) argues that the “technical game,” that part of development planning that relies on the universal representations of the metacode, is thought to be outside social and cultural frames of reference. Recourse to the technical game is unavoidable in epistemologically heterogeneous zones such as development projects. In a similar way, global systems of measurement assert metacodes that claim universality even though they are products of the distinct national histories of their creators. Measurement metacodes circulate globally as technically neutral but universal technologies without reference to their culturally specific origins. Yet, at the moment of measurement, those carrying out data collection translate them into cultural codes.
By examining the efforts of the FoC committee to define “severe physical violence” in contrast to “moderate physical violence,” it is possible to trace the construction of a universal set of categories, a metacode, out of a culturally varied set of practices. In tackling the problem of distinguishing severe and moderate violence, the FoC committee considered three alternatives. One is to measure acts. A widely used survey instrument provides a respondent with a list of actions to tick off, from pushing to threatening with a weapon. Using this list, already ranked by severity, it is possible to separate acts that are severe and those that are moderate. This approach defines severity in terms of the act itself, such that, for example, threatening with a gun is more serious than a slap. It ignores fear, injury, and the experience of the victim. It seems objective since it requires little interpretation by the victim. It is, however, a culturally embedded list. For example, at the FoC meeting, Bangladesh suggested the following additions to the list: “Burn/Acid throwing, Dropping from higher place, Smash the finger or hand, Needle the finger, Kick in the abdomen, Hit in the teeth, Murder (Femicide).” Its proposed list of psychological violence criteria similarly suggests a locally distinctive set of categories: “Non-response to queries, Pressure for dowry, Threat of separation, To marry other women in addition to existing wife, Do not pay attention to the children, Expel from the house, Disregard the opinion of the females in household decision making, To compel to do hard work during pregnancy, To rebuke for giving birth to female child” (Ali Mollah 2009:16–18). Several of these acts of gender-based violence are particular to Bangladesh or South Asia generally.
In fact, the apparently universal list is based on research in advanced industrial societies and the categories that make sense in that cultural context. Use of such a list is the prevailing approach in previous research, most of which has been developed in these countries. The UN Statistical Division official overseeing the process of constructing the survey instrument did a background study of all existing surveys on violence against women. He located 64 surveys, mostly done in Europe, North America, Australia, and a few countries in Latin America. The use of a list of acts is commonplace in these studies, with about 90 percent of the surveys in Europe collecting data on the following forms of physical violence: pushed/grabbed/shoved, kicked/bit/hit, hit with something, choked, used or threatened with a gun or knife (UN Division for the Advancement of Women 2007:12). The measurement of acts is based on modifications of the widely used Conflict Tactics Scale, developed in the United States, which lists acts in ascending severity and coerciveness (Straus 1979:78). This approach is the same as that adopted by the large number of national demographic and health surveys carried out by the U.S. company Macro International. It was used by the local women's center Merry studied in Hawai‘i in the early 1990s. Thus, the list has become a metacode despite its origin in a particular cultural context. Over time, it has become an “immutable mobile” with claims to universal applicability. While there is no clear consensus about how to group the entries on the list into moderate and severe forms of violence, they are ranked in terms of severity so that drawing a line in the middle is a common approach.
A second approach to measuring severity considers injury. This, again, poses problems of condensing variety into a single category. For example, even the apparently simple measure of seeking medical attention varies greatly by rural–urban and class factors, such as the availability of hospitals, transportation systems, costs, public medical care, and so on. At the FoC meeting, Bangladesh suggested another approach to universalizing this category, the criterion of bleeding: “Severity in case of physical violence may be considered for those incidents, which cause injury to the victim, bleeding from any organ of the body may also be considered severe type of physical violence. The other incidents may be termed as ‘moderate'” (Ali Mollah 2009). Developing a metacode for injury that acknowledges this diversity is difficult.
A third approach to defining severity is to ask the woman about her level of fear. Australia, for example, did a survey that included questions about whether a woman changed her daily pattern of activities because of fear. Anxiety or fear was measured by its effects on a series of daily activities such as work, social or leisure activities, and child care and on whether the woman felt the need to install a home security system. Yet the final report from the 2009 FoC meeting expressed concern that fear is “subjective” and therefore hard to measure with certainty. Many at the meeting preferred more objective-seeming data, such as the nature of the act and the severity of the injury. However, relying on measures of acts or injuries rather than a person's fear displaces the decision about whether the violence is severe from the person who experiences it to someone who does not.
At the conclusion if its meeting, the FoC ultimately went with the established list of acts modified by whether an injury occurred. The conclusions and recommendations of the 2009 meeting used the scale of acts plus or minus injury to distinguish moderate and severe violence:
- Moderate physical violence consists of the following categories that did not result in bruises, cuts, broken bones or need for medical treatment or hospitalization:
- Hit with something
- Kicked, bit or hit
- Pushed, grabbed, shoved
- Threatened to hit
Severe physical violence consists of the first four categories listed above that resulted in bruises, cuts, broken bones or need for medical treatment or hospitalization, plus several other categories:
- Hit with something with aggravated consequences
- Kicked, bit or hit with aggravated consequences
- Slapped with aggravated consequences
- Pushed, grabbed, shoved with aggravated consequences
- Threatened with knife, gun, other weapon
- Assaulted with knife, gun, other weapon
- Other [UNSC 2009:9–10; see also UNSC 2010]
This template ignores diversity to create commensurability and neglects fear and experience in favor of the apparent objectivity of acts and injuries. The report concludes that fear is too variable to systematically measure, depending as it does on national cultural circumstances, and therefore too subjective to use in the international setting. It attempts to bring the emotional consequences of violence into the analysis but does not specify how they are to be measured.
The conclusions and findings of this meeting were to be folded into a set of guidelines for use by national statistical offices around the world. A consultant with experience in international violence against women surveys drafted a report that was then circulated to the members of the FoC. After suggestions were received, it was revised and recirculated. The final draft of the guidelines is not yet published, but the UN Statistical Division statistician overseeing the project told Merry that, despite the logic of the system of categorizing moderate and severe forms of violence, consultations led to the conclusion that the model would not work across countries. Instead of requiring this model, the UN now includes it as one option among several that governments might select, but the guidelines allow countries to decide for themselves. Merry asked how it would be possible to draw comparisons across countries without a shared system of categories, and the statistician said, “We will do our best.” Commensurability gave way to sovereignty.
In contrast to these systems of classification, the ethnographic observations of domestic violence support groups that Merry (1995a, 1995b, 2009) carried out in a small town in Hawai‘i during the 1990s suggested that women talk about violence in terms of the history of the relationship and a long series of events, both violent and nonviolent, that humiliate, frighten, and injure them in a wide variety of ways. The act of violence was less important than the violation of sense of self, repeated insults and humiliations, threats to children and pets, and excessive demands for money that these women experienced. It was these events that concerned the women who spoke and were the focus of their talk, rather than the specific acts of violence they endured, as Dora's story showed as well. Such nuanced and contextualized understandings of violation cannot be revealed by a list of acts, and they are not amenable to study through a cross-national survey. Even though measures of feeling and perception provide a better account of how victims experience violence, they appear less objective and amenable to comparison than a list of acts. Focusing on a specific and ranked set of acts makes measurement easier and comparison more possible. It usefully reveals the extent of violent acts. However, by relying on the list of acts, the “objective” comparison loses critical information on the experience and the perspective of the person herself. It is a proxy for the experience of violence. It effaces the very phenomenon these measures seek to expose: the extent to which women's everyday lives and experiences are affected by the possibility and practice of violence. This is the paradox of quantification.
These examples show how constructing categories for measurement is a political process that requires making choices among alternative ways of putting things together and labeling them. The process can be, but is not necessarily, subject to debate and contestation. Some categories, such as those of the Conflict Tactics Scale, become established through repeated use by experts, so that they become taken-for-granted, seemingly natural building blocks of knowledge. As they become natural, their power becomes invisible. To discern such invisible modes of power, it is critical to focus on the process of knowledge production itself, particularly commensuration (see Espeland and Stevens 1998). This is a technical way of transforming complex specificities into entities that can move across boundaries with fixity, as immutable mobiles. The technical process is itself historically and culturally produced, of course. It develops a universal set of categories and codes over time on the basis of experience, testing, revision, and retesting. Experts working on commensuration confront the pragmatic problem that things are always different but that some overarching framework is necessary to make comparisons, both for social theory and to navigate the world. Actors, networks, geopolitical concerns, debates at the United Nations, social science expertise, templates, documents, forms, past practices, previous surveys, and funding sources all contribute to the formation of these mechanisms of knowledge production over time. Measurement systems are a quiet mode of power. Like other forms of power, they have the capacity to strengthen the control of dominant groups over subordinate ones but also to expose practices of control and to make the suffering and disadvantages of vulnerable groups visible to dominant ones. Measurement is a technology with power, harnessed by those able to use it.
- Top of page
- Indicators: Measuring violence against women
- Discussion and future directions
- References cited
Like the formulation of international gender violence indicators, the U.S. Deferred Action for Childhood Arrivals, or DACA, process provides a lens through which to consider the new governance systems developed to address intractable social problems. In the United States, intense debates have arisen over whether to deport or legalize the some 12 million undocumented immigrants living here, including long-time residents who immigrated as children, have become assimilated, and know little of their countries of origin. Over the past decade, there have been, on the one hand, record-breaking marches in support of comprehensive immigration reform and, on the other, increased adoption of state and local measures designed to make the undocumented so unwelcome that they “self-deport” (Betancur and Garcia 2011; Johnson and Hing 2007; Kobach 2008; Varsanyi 2008, 2010). Students have been at the forefront of advocacy efforts, and, through marches, vigils, and hunger strikes, have been promoting passage of the federal DREAM Act, which would grant residency and, eventually, citizenship to qualifying undocumented college students (Abrego 2008; Gonzales 2008).
In June 2012, during an election campaign in which President Obama was roundly criticized for failing to deliver on his promise of immigration reform, U.S. Department of Homeland Security (DHS) Secretary Janet Napolitano created DACA by means of a memo. The memo stated that prosecutorial discretion would be issued in favor of “productive young people” (DHS 2012:2) who had entered the United States prior to the age of 16, had lived here continuously for five years, were either in school or high school graduates, had not been convicted of crimes, and were under the age of 30. Those who met these criteria would be granted temporary work authorization. The stated rationale for this exercise of discretion was “to ensure that our enforcement resources are not expended on these low priority cases but are instead appropriately focused on people who meet our enforcement priorities” (DHS 2012:1). The memo instructed the U.S. Citizenship and Immigration Services (USCIS) to “establish a clear and efficient process for exercising prosecutorial discretion, on an individual basis, by deferring action,” with the action in question being removal (DHS 2012:2). This “clear and efficient process” was to be developed and implemented within 60 days. This rushed deadline led to a flurry of form creation and procedural innovation that some immigrant rights advocates jokingly referred to as “DACA-mentation” and a “DACA-nami.”
What makes the DACA process particularly appropriate for our consideration here is that this discretionary remedy developed within an area of the law that is already relatively unfettered. As Coutin, Justin Richland, and Veronique Fortin have noted elsewhere (in press), immigration falls under the “plenary powers” doctrine, according to which the U.S. Congress and the executive branch have the authority to exercise their political will in ways that are “not subject to be controlled by the judicial department of the government” (Lone Wolf v. Hitchcock, 187 U.S. 553 ). Though noncitizen residents are subject to the Constitution, according to the notion of plenary power, the U.S. government has the discretion to decide which noncitizens to allow on its shores and which to remove. Historically, this discretion has been exercised in multiple ways, sometimes treating long-time residents as acquiring equities that demand legal recognition and other times allowing Congress and the executive branch to act according to their own pleasure, with few limits (Motomura 2006). Increasingly, these competing impulses have resulted in the creation of temporary statuses that grant individuals the right to remain in the country and to work but that do not permit them to hold the full panoply of rights afforded to legal permanent residents and citizens (e.g., the right to vote, petition for family members, and exit and reenter the country). DACA is a case in point. As the final paragraph of its authorizing document states, “This memorandum confers no substantive right, immigration status or pathway to citizenship” (DHS 2012:3). DACA is therefore intrinsically ambiguous—it allows individuals to remain in the country but does not grant immigration status.
Our account of DACA derives from fieldwork that Coutin has been conducting over the past two years at a Los Angeles–based nonprofit that provides legal services to Spanish-speaking immigrants. During summer 2012, she was present as this organization geared up to provide DACA services. She attended trainings on applicant screening and DACA request procedures, observed and assisted in the preparation of requests, attended planning meetings, and collected secondary documents, such as memos and lists of frequently asked questions posted on the USCIS website. These experiences gave her a unique opportunity to see how the DACA process was created as well as insight into the dilemmas that confronted applicants and their advocates before the process became routinized.
Just as the international agencies that developed systems to measure gender violence drew on preexisting systems, so too the creation of DACA built on preexisting legal artifacts, including Temporary Protected Status, elements of the proposed DREAM Act legislation (such as the proposed age and presence requirements), and the use of deferred action and prosecutorial discretion in other legal contexts. This scaffolding thus enfolded multiple historical struggles into DACA in ways that may not be immediately apparent. Temporary Protected Status was created via the 1990 Immigration Act after a decade of solidarity work on behalf of Central American asylum seekers fleeing war and persecution in their homelands. It was an intermediate solution that granted these war refugees the right to remain in the United States but did not grant them permanent status, the right to travel internationally, or the ability to petition for their relatives. It thus kept them in a state that has been characterized as legal “limbo” (Mountz et al. 2002) or “liminal legality” (Menjívar 2006). The Temporary Protected Status application form, I-821, was used as the basis for developing the DACA request form, which is numbered I-821D. Likewise, deferred action has also been granted to domestic violence victims who self-petition under the Violence Against Women Act. In an attempt to liberate battering victims from dependence on their batterers for the purpose of determining immigration status, wives of abusive spouses may apply for green cards. Under some circumstances, their petitions are approved, but a visa is not yet available for them or their children. In such cases, the women and their dependents are granted deferred action, that is, the ability to remain in the country until their visa becomes available. Prosecutorial discretion has also been exercised in individual cases in which, for humanitarian reasons, immigration officials opt not to remove someone who is otherwise deportable.
DACA is thus indebted to the struggles of undocumented students, asylum seekers, battering victims, and other immigrant groups. As an immigration “remedy,” though, it incorporates the tenuousness and ambiguity of the benefits extended to members of these groups. According to the USCIS website, “An individual who has received deferred action is … therefore considered by DHS to be lawfully present during the period deferred action is in effect. However, deferred action does not confer lawful status upon an individual” (USCIS 2013b).5 Though the distinction is important for legal purposes, how it is possible to be lawfully present but without lawful status is not particularly clear. This phrasing is reminiscent of the circumlocution resorted to by the U.S. Supreme Court in a 1901 insular case, in which one justice deemed Puerto Ricans to be “foreign in a domestic sense” (Burnett and Marshall 2001).
The ambiguity of the remedy that DACA provides—lawful presence without lawful status—makes requestors' decision about whether to make themselves visible by applying complex. On the one hand, securing work authorization means obtaining an identity document, specifically, a work authorization card, which, despite not officially conferring status, can, in practice, be taken as an indication of belonging. Being awarded DACA therefore has considerable potential to improve recipients' lives and to provide a measure of legal recognition. On the other hand, DACA is temporary. Those who receive deferred action are only authorized to remain in the country for two years and, as is clearly stated in the authorizing memo, are not placed on a path to a permanent status. To allay their concerns that becoming visible to the U.S. government could lead to removal, USCIS has assured requestors that information in their request forms will only be used to initiate immigration enforcement proceedings if they pose a security risk or misrepresent their eligibility. Furthermore, even though questions about parentage are routine on U.S. immigration forms, DACA does not require requestors to report their parents, who may be undocumented. Yet the request does constitute a record, and, hence, requestors need to be cognizant of the possible legal ramifications of revealing particular information, such as details about their mode of entry. For instance, individuals who falsely claim U.S. citizenship are permanently barred from legalizing, so an individual who reports having entered the country by falsely claiming U.S. citizenship could be subjected to this bar. And assurances about the uses to which information will be put can ring hollow, given that immigration law itself is subject to change. At the same time, regardless of the outcomes of individual requests, the large numbers of requestors who materialized in immigration offices at the process's outset drew attention to the plight of youth raised in the United States without legal status. The mere act of requesting deferred action therefore makes the population of undocumented students visible. In the first six months of the program's existence, 407,899 DACA requests were filed (USCIS 2013a).
Like the commensuration required in assessing gender violence, proving that DACA requestors met the process's requirements meant repurposing disparate documents in ways that rendered them commensurate. One of the biggest challenges facing advocates at DACA's outset was demonstrating that applicants had resided in the United States continuously since June 15, 2007, five years prior to the date that the DACA memo was issued. Demonstrating presence required assembling a documentary record covering this five-plus-year period. To be useful as proof of presence, documents had to be dated and show the DACA requestor's name and address. School records, bank statements, medical records, loan payments, and receipts for purchases met these criteria and became “proof of presence,” even though applicants sometimes worried that other information that these documents contained—such as negative comments teachers might have written on students' report cards—would be used against them. Advocates put these disparate documents in order by year, from 2007 or earlier to the present. This ordering was designed to demonstrate continuity, but it also made the gaps between documented dates visible.6 For example, a DACA requestor might have school records for the 2007–08 academic year but no record over the summer of 2008. Or a requestor might have graduated from high school in June 2010 and then have a gap of six months until a January 2011 visit to a medical clinic. Advocates, who had decided to prepare only requests that stood a good chance of success, debated what size gap might be considered unacceptable by immigration authorities—two months? Three months? Six? Responding to frequently asked questions on September 14, 2012, one month after the program began, USCIS set the standard as follows:
It is helpful to USCIS if you can submit evidence of your residence during at least each year of the period. USCIS will review the documentation in its totality to determine whether it is more likely than not that you were continuously residing in the United States…. Gaps in the documentation as to certain periods may raise doubts as to your continued residence if, for example, the gaps are lengthy or the record otherwise indicates that you may have been outside the United States for a period of time that was not brief, casual or innocent. [USCIS 2012]
Given the impossibility of creating a gapless documentary record, then, one with gaps that were neither too lengthy nor filled in by evidence of being outside the country stood in as an acceptable proxy. Of course, many who could not meet this standard might have been present continuously but simply lacked proof.
The legal violence (Menjívar and Abrego 2012) of the distinctions drawn by immigration categories is made evident by the experiences of individuals who are similarly situated but who do not qualify. On the third day after the DACA program began, Coutin was in the offices of a Los Angeles nonprofit that was attempting to serve the hundreds who had appeared at its doors in search of assistance. As she had attended a DACA training, staff asked her to help screen potential requestors to ensure that they met the eligibility requirements. One of the individuals she met was a young woman from Ecuador. Handing Coutin the passport on which her entry date was clearly stamped, the woman explained that she first came to the United States when she was fifteen and a half. Looking at her birth date, which was also visible on the passport, and the entry date, Coutin quickly did the math in her head and then again on paper. According to the passport record, this young woman had entered after her 16th birthday and was therefore ineligible for deferred status. When Coutin explained this, the young woman responded with disbelief. Her mother, who had accompanied her, insisted, “Surely there is some flexibility,” and the young woman asked incredulously, “You mean that if I had entered the United States three months earlier, I would qualify?” Unfortunately, that was the case. The violence of arbitrarily drawing a line that excluded some undocumented youth from even the limited remedy of DACA was apparent.
Yet power also works through the process that qualifies individuals for immigration remedies. Just as the definition of gender violence privileges certain forms of abuse and certain victims over others, so too do legal definitions make certain aspects of requestors' experiences central and others irrelevant. For example, in August 2012, Coutin assisted a high school student she refers to as “Ignacio,” who was applying for DACA with his mother's help. Ignacio's mother seemed to think that because Ignacio had entered the country on a tourist visa rather than as an undocumented immigrant, he was a strong candidate for DACA. In reviewing Ignacio's documentation packet, though, a legal worker identified a small and correctable evidentiary flaw. Ignacio needed to provide evidence that he was currently enrolled in school. The 2012–13 school year had not yet begun, so Ignacio's most recent school record was from June 2012. A legal worker suggested waiting until September 2012, when school would start, and then obtaining a letter from the school. Ignacio's mother responded by pointing out that getting this document would mean additional appointments, thus requiring more days of missed work and school. Furthermore, this document struck her as unnecessary, given that her son had been enrolled in high school at the end of the prior year and the documentation that he had obtained was consistent with school practices. Though Ignacio's mother agreed to return with the requested documentation, this incident highlighted the gap between her own understanding of merit and the legal worker's professional responsibility to document the specific facts needed to qualify for DACA.7
Although DACA was designed to benefit young people who had already been productive, rather than to shape future behavior, one of its effects has been to encourage individuals who are otherwise eligible but who dropped out of school to reenroll. DACA thus illustrates the “responsibilization” associated with both new governance practices and neoliberal policies (see also Varsanyi 2008). As Nikolas Rose, Pat O'Malley, and Mariana Valverde note, such practices entail enacting “technologies of the self” in which subjects “produce the ends of government by fulfilling themselves rather than being merely obedient” (2006:89). Thus, rather than being sanctioned for failure to attend school, undocumented migrants who are requesting DACA take it upon themselves to complete their educations. Coupled with this “responsibilization” is a displacement of the substantive judgment that an application for legal status might normally generate. Instead of determining whether requestors qualify for residency or for an interim status (such as asylee) that could lead to residency, DACA merely evaluates whether requestors are low priority for removal. Those whose requests are not granted have no right to appeal. DACA is thus purely administrative. As an exercise of prosecutorial discretion, it is officially an instance of government benevolence rather than a right that accrues to undocumented youth. For these reasons, students who have been lobbying the U.S. government for passage of the DREAM Act have welcomed DACA as an important first step but have also vowed to “continue to push for the DREAM Act and a legitimate pathway to citizenship for all immigrants” (Macedo and Cabrera 2012). By making the youth who lack status more visible and by granting them work authorization documents that confer some degree of legitimacy, DACA may further this goal while providing an interim remedy for undocumented students.
Discussion and future directions
- Top of page
- Indicators: Measuring violence against women
- Discussion and future directions
- References cited
Clearly, there are significant differences in the way forms of evidence, documentation, and survey data construct knowledge that, in turn, shapes decision making in the two arenas we have considered here. Our first case, for example, is international and our second located in the United States; one entails measuring a societal phenomenon and the other evaluates the records of individual immigrants. Yet across these differences, we find six themes that suggest similar ways that knowledge is constructed and used in these fields and that link to the ways scholars working on “new governance” discuss the role of information in governance decisions.
First, in both of these cases, the technology of standards, forms, evidence, and measurement is built up over time, with the templates of the past serving to guide developments in the future. This temporal scaffolding produces a stratigraphic relationship between early efforts and later ones, what we have referred to as “an archaeology” of law and governance. As they are brought forward in time, the same laws or texts are given new meaning. Indeed, as Theodore M. Porter (1995) points out, although categories of enumeration may be highly contingent at first, once they are in place they become extremely resilient (see also Bowker and Star 1999).
Second, in both situations, technical approaches that focus on forms, procedures, and the organization of data into categories are used to grapple with situations of conflict. Rather than satisfying the demand that violence against women be reduced or that unauthorized immigrants be allowed a path to citizenship, these technologies are interim measures that postpone substantive resolutions into the future. Immigrants are not to be deported or legalized now but might be later; gender violence will not be eliminated now, but knowledge about it may help in the future.
Third, despite their apparent objectivity, these techniques of knowledge and decision making entail interpretive work and discretion. Even though the list of criteria for measuring violence looks objective, and even though apparently subjective phenomena such as fear or stalking are not included, there is still considerable interpretive work involved in determining what kind of violence a woman experiences or whether it is moderate or severe. Likewise, the DACA requirement that an applicant document evidence of presence in the United States seems clear, yet discretion is exercised in determining the adequacy of evidence. Over time, information technologies transform fragmentary and interpretive knowledge into more fixed and unambiguous forms. In the case of measuring violence against women, this is done by the magic of numbers and, in the DACA case, by the notion of documentary evidence itself.
Fourth, both measuring violence against women and assessing an immigration claim depend on practices of commensuration: of rendering the diversity of individual circumstances and experiences similar and therefore comparable. In the case of gender violence indicators, making different things similar enough to compare them requires stripping away context and history to zero in on a relatively simple dimension of the phenomenon. In the case of DACA, geopolitical histories that have led people to immigrate to the United States without authorization are not part of the assessment; rather, individual histories are evaluated to distinguish “productive” and “deserving” young people from “undeserving” undocumented immigrants. Power is embedded within these methodologies in ways that are not always obvious (Espeland and Stevens 1998). Phenomena are not intrinsically commensurate but are made so by creating equivalences, categories, and distinctions. Violence victims are to be counted in the same way, despite their obvious differences, even as gender violence is distinguished from other violent acts. DACA request procedures evaluate requestors against a single standard even as DACA itself is premised on the essential incommensurability of the undocumented population in the United States. The law, its authorizing memo states, is not “to be blindly enforced without consideration given to the individual circumstances of each case” (DHS 2012:2).
Fifth, both systems promote and reward responsibilization, a core dimension of neoliberalism, although in different ways. For example, the DACA exercise welcomes those undocumented immigrants who have been responsible, attended school, and remained in the country. Similarly, the use of standards for behavior that is monitored through data collection processes invites those who are monitored to voluntarily conform to the standards. Human rights monitoring, in general, seeks to establish standards that countries need to follow to be understood as human rights compliant. The use of standards that contain measures allows those being monitored to decide for themselves when their performance is adequate.
Sixth, an important effect of these techniques is rendering some populations and problems visible and others invisible. Counting the frequency of violence against women makes the problem known and worth attending to; narrowing the scope of the measurement renders some forms of violence, such as harassment and humiliation, invisible. Documenting those deemed worthy of deferred action renders them visible and therefore possibly encourages their ultimate admission as legal permanent residents at the same time that it erases those who, for a variety of reasons, as simple as not being the right age, are ineligible for DACA.
Enumerating these similarities suggests that the processes that undergird comparisons—building comparability of data and measurable categories—have political and social implications. To understand measurement, it is necessary to examine these processes of interpretation and categorization. Who is doing the interpreting? What are their interests? Who can speak with authority about which categories to use and what to count? Who is even consulted about relevant categories and counts? Who sits around the planning table? How are regulations and memos that guide those attempts to meet legal standards developed? Is there feedback between the measurers and the measured, the authors of policy and the populations at issue? What enables measurement systems to be scaled up or down, from individual to population, case to aggregate, locality to nation? And what roles are played by those who use information systems, those who are targeted by and are subjects of such systems, and those who advocate for policy changes? Whose expertise and experience shapes the commensuration process? How does the influence exerted by different parties—advocates, officials, applicants, targets—compare?
We suggest that anthropological analyses of the sorts of conflict that new, evidence-based governance techniques attempt to manage require attending to the knowledge systems through which “problems” are identified and made known in the first place. What gets measured and counted depends on what can be categorized and coded. That which is too hard to categorize and classify may not be discerned at all. The experience of gender violence, for example, confounds classification because it is an interior, relatively subjective phenomenon. Since it is so hard to measure and count, it is typically gauged by something more objective, such as injury or type of physical hit or blow. Likewise, the complex social ties forged by long-time yet undocumented residents disappear in the face of arbitrary cut-offs based on age and arrival dates and cannot be substantiated in their entirety. The demands of encoding thus loop back and shape classification since the need to sort and classify determines what categories are usable and what forms of equivalence will work. In the cases we have discussed, encoding and assessment lead to an emphasis on apparently objective phenomena, such as actions or dates, rather than on apparently subjective ones, such as fear or sense of belonging.8 This means that some aspects of experience are made visible while others are not. Battered women's senses of fear, vulnerability, and humiliation are hard to encode when constructing indicators because of problems of creating equivalence across individual differences and cultural forms. The need for objective categories for comparison and for clear legal standards means that the very things surveys want to examine or that adjudicators attempt to measure—how extensively women's lives are affected by violence or which undocumented youth are productive—do not get counted. Experiences that cannot be made commensurable or that cannot be documented cannot be measured. In effect, such experiences are disappeared from the analysis. And in this disappearance, categories are naturalized such that power too is rendered invisible. By foregrounding the techniques involved in such seemingly mundane acts of calculation, anthropologies of conflict can make visible alternative accountings that demand rather than defer justice.