Gendered Genocide: The Socially Destructive Process of Genocidal Rape, Killing, and Displacement in Darfur

Abstract

Accounts of mass atrocities habitually focus on one kind of violence and its archetypal victim, inviting uncritical, ungendered misconceptions: for example, rape only impacts women; genocide is only about dead, battle-aged men. We approach collective violence as multiple, intersecting forms of victimization, targeted and experienced through differential social identities, and translated throughout communities. Through mixed-method analyses of Darfuri refugees' testimonies, we show (a) gendered causes and collective effects of selective killing, sexual violence, and anti-livelihood crimes, (b) how they cause displacement, (c) that they can be genocidal and empirically distinct from nongenocidal forms, (d) how the process of genocidal social destruction can work, and (e) how it does work in Darfur. Darfuris are victimized through gender roles, yielding a gendered meaning-making process that communicates socially destructive messages through crimes that selectively target other genders. The collective result is displacement and destruction of Darfuris' ways of life: genocide.

Forgotten Victims of Genocide

For three months, the planes kept bombing us but we stayed. Then, the Army and janjaweed came. They came with vehicles and horses…and they started shooting. Everyone started running, and I left the village. I was separated from my children.

Men were targeted for killing. Women were raped. If they resisted, they were killed. The children were not killed but were abducted.…Cattle were taken. Food was taken.…[C]attle were killed by bombing.…[H]uts were burned. All furniture and belongings were stolen.…The village was completely destroyed.

They were yelling, “Kill the Nuba! Kill the Nuba [black slaves]!”

— testimony from Darfur

Darfur's atrocities have reached their second decade (Reeves 2014). Analysts confirm Government of Sudan (GoS) forces and Arab janjaweed militias have killed hundreds of thousands of Black Africans (Hagan & Rymond-Richmond 2009; Prunier 2008; U.N. Security-Council 2008). Still, the international community cannot agree it is genocide.

Gender-selective killing, sexual violence, and forced migration typify socially destructive processes of genocide (see Jones 2009; Maybury-Lewis 2002; Seifert 1994). Many argue these crimes can constitute genocide (Hagan & Kaiser 2011a; MacKinnon 1994), and international law concurs (Genocide Convention 1948; Prosecutor v. Akayesu 1998). Yet even experts who accept these arguments in theory have trouble understanding how—and therefore when—such violence is genocidal. Absence of a genocide label reduces humanitarian and political aid, changes legal requirements, hinders public and scholarly understandings, and all but invalidates survivors' experiences.

This misunderstanding stems from a dangerous tendency across disciplines to focus single-mindedly on one kind of victimization and one kind of victim. Like most conflict-related discourse (Cohen 2013; Jones 2009), accounts of Darfur's atrocities concentrate on killings—disproportionately of “battle-aged” men. Women, children, the elderly, and even surviving young men are forgotten. They are commonly treated as irrelevant to genocide—as “merely” tortured, beaten, raped, and left to die from malnourishment and disease (e.g., International Commission of Inquiry 2005; Schabas 2008). At best, they are misremembered: simplified into survival rates, so they seem “only” evidence of “ethnic cleansing” or “overzealous counterinsurgency”—though neither actually precludes genocide.

Meanwhile, advocates who are concerned with refugees and sexual victimization distance themselves from genocide conversations (e.g., Médecins Sans Frontières 2005; Copelon 1995). Not questioning the hegemonic primacy of homicide in the discourse simultaneously diminishes their arguments' power and subordinates their subjects to “real” concerns about killing. Moreover, understanding Darfur's multifaceted atrocities as genocide is impossible without the holistic approach these specialists could help provide. Legal authorities like William Schabas (2008; 2000:174) thus flatly declare it “unrealistic and perhaps absurd to believe that a group can be destroyed…by rape and similar crimes.”

Scholars likewise fail to recognize collective violence as varied, interacting experiences filtered through gender and other identities. Groundbreaking research on the causes of conflict-related rape neglects interactions with other victimization (Cohen 2013; Green 2004); excellent scholarship on displacement considers only relationships with “overall” violence (Davenport et al. 2003); and useful studies of Darfur's atrocities analyze murder and rape independently (Hagan & Palloni 2006; Hagan et al. 2009). Besides “undermin[ing] disciplinary knowledge production,” treating any subgroup as representative of all victims hinders “legal thinking…and struggles for social justice” (Cho et al. 2013:787). It inadvertently encourages artificial separation of the raped, the displaced, and the killed—ultimately impeding understandings of genocide.

Genocide is not just mass murder. Dead, battle-aged men are not genocide's only victims. Genocide by definition intends to destroy an entire social entity, and it does so through multiple forms of systematic victimization. Immediate death, or “extermination,” is thus only one aspect of genocide targeting one subset of victims. Remaining group members face the more subtle process of “elimination,” physical and social conditions designed to destroy entire communities, groups, and nations (Hagan & Kaiser 2011a, 2011b).

Others make such arguments, but it is difficult to imagine social destruction in the abstract—and how it can be intended without complete extermination. We use a concrete example to show empirically the process of genocidal social destruction. Understanding genocide requires holistically considering extermination alongside elimination as interacting mechanisms that socially destroy through group members' shared experiences. It also requires recognizing victimization as differentially targeted and experienced through gender, age, and other identities.¹

We thus join interdisciplinary theory on killing, rape, displacement, and conflicts with a gendered, social lens. First, we elaborate the social nature of genocide by building on Hagan and coauthors (2008; 2011a) and Shaw (2007), clarifying how sexual violence and displacement can be genocidal. Next, we use Seifert's (1994) and Jones's (2000) feminist approaches to hypothesize that gender-selective targeting of genocidal violence (a) relies on perceptions of women as bearers of culture and community, and men as a culture's default members and users, and (b) produces a gendered meaning-making process whereby men receive a message of elimination through sexual victimization of women while women receive that message through killings of men. Copelon's (1995) and Schabas's (2000) asocial accounts of genocide, displacement, and rape provide null hypotheses.

Then, we use mixed methods to analyze the experiences of a genocide's survivors. We qualitatively analyze testimonies alongside quantitative measures of displacement over time. We focus holistically on Darfuri men's and women's experiences of gender-selective killing, sexual violence, and displacement. Our results show that Darfur's genocide operates as multiple forms of gender-selective violence that interact through victims' collective experiences to produce social destruction.

Genocide: Intentional Social Destruction

Victims' differential experiences are fundamental to understanding genocide because, as Shaw (2007) and Card (2003) argue, “genocide” means intentional destruction of a group target.² The term originated from the Armenian and Jewish Holocausts, crimes that could not be encapsulated by “mass murder” or any existing label because they targeted the collective entity itself (Lemkin [1944] 2008). Hence, “[g]enocide is a denial of the right of existence of entire human groups, as homicide is the denial of the right to live of individual human beings” (U.N. 1947).³

Unfortunately, genocide is rarely understood as intending social destruction—because scholarship neglects how it functions and is experienced. Studies reveal that genocide results from contextual forces (Abusharaf 2010) combined with escalating ethno-sectarian tensions (Mann 2005; Straus 2006) that activate collective mobilization against a defined social group (Hagan & Rymond-Richmond 2008; Campbell 2011). Processes from genocidal intent to social death are unexplored.⁴

Just as groups are socially constructed, they must be socially destroyed. Although groups may expire peacefully or unintentionally, genocidal destruction emerges at the “nexus between [intentional] destruction of collective ways of life and institutions and bodily and other harm to individuals” (Shaw 2007:106). Individual, micro-level victimization translates to the macro level, fundamentally altering the “shared meanings” and “collective representations” defining a way of life (e.g., Eliasoph & Lichterman 2003; Wimmer 2008).

Thus, any act intended to cause radical transformation (or obliteration) of a social order has genocidal intent. “Even [extermination] can be viewed as extreme means to the primary end of social death”: a process of destroying members who maintain collective representations (Card 2003:63). Yet, no commonly recognized genocide has involved only extermination; they all incorporate multiple patterns of victimization (e.g., Maybury-Lewis 2002). We therefore anticipate a second process that translates victimization into social destruction: through intragroup communication and meaning-making, victims interpret the group's experiences as damaging to its collective representations (see Janoff-Bulman & Frieze 1983; Eliasoph & Lichterman 2003). Even killings can thus eliminate: through survivors.

Finally, since collective violence is targeted and experienced based on social identities (Jones 2009; Kruttschnitt & Macmillan 2006), we would expect differential experiences and meaning-making between subgroups. The key to understanding genocide—and elimination as genocidal—is understanding how perpetrators translate socially destructive intentions into differential targeting, and how victims collectively interpret these differential experiences as socially destructive (Figure 1).

Figure 1. Genocide as Social Destruction, Occurring through Dual, Experiential Processes of Extermination and Elimination

Multidimensional Experiences of Genocide

“The United Nations has come too late. We have already died.” (emphasis added)

Atrocities reflect their origins. Darfur's Zaghawa, Fur, Masaleit, and other Black African tribes were farming communities with stark gender roles: men as dominant decision-makers protecting women as reproductive, cultural bodies (Abusharaf 2006). Since the 1600s, the Fur Sultanate's hakura land-use system privileged certain “Black African” farmers over others, especially nomadic “Arab” herders—sparking conflict over arable land (de Waal 2007b). Neither Black African nor Arab, however, were salient identities until Sudanization policies spread and Omar al-Bashir seized the presidency in 1989 (de Waal 2005; Prunier 2008). Combined with desertification and famines, Sudanization intensified and racialized clashes over food and water. When the al-Bashir government, assisted by Libya's al-Gaddafi, armed the Arab janjaweed militias, Black Africans became targets for destruction.

The resulting genocide uses gender- and age-selective extermination alongside elimination through anti-livelihood crimes. Typically, after janjaweed and GoS forces bomb, slaughter young men, and rape women, they plunder food, poison wells, and burn whole villages—tactics that are ongoing (see Figure 2). Darfuris predictably flee. Millions hide in valleys and “bush” and are ultimately forced into concentrated camps. On the road, they face further attacks. In camps, multitudes die from dehydration, starvation, disease, exposure, and attack-related injuries, largely because Sudan obstructs humanitarian aid efforts (e.g., Reeves 2014; U.N. Security-Council 2008). The result: “the very foundation of village societies has been blasted to bits…[and] whatever notions and significances people attached to their practices are being shattered” (Abusharaf 2006:69).

Figure 2. Physical Destruction of Darfur's Villages by 2004 and by 2009

Conflicting Interpretations: An Empirical Question

Underlying arguments that Darfur's atrocities are nongenocidal are one-dimensional ideas of violence as impacting only immediate victims in a social vacuum, and of genocide as indiscriminate killing. Feminist lawyer-activist Rhonda Copelon (1995) thus claims “genocidal rape” is not an empirical phenomenon. Dismayed by all rapes, she denounces any distinctions between them:

[T]o emphasize as unparalleled the horror of genocidal rape is factually dubious and risks rendering rape invisible once again…We must examine critically the claim that rape as a tool of [genocide] is unique, worse than, or incomparable to other forms of rape in war or in peace—even while we recognize that rape coupled with genocide inflicts multiple, [cumulative] harms. (1995:199)

Any “distinctive characteristics,” she argues, “do not place genocidal rape in a class by itself” but merely reflect the additive effects of gender and ethnicity (205). Copelon focuses on criticizing legal “rankings” of rape, but does so by seeing rape and genocide as merely “double jeopardy,” not intersectionally constitutive.⁵

Similarly, Schabas (2000:200) artificially distinguishes “ethnic cleansing” from genocide by arguing it intends to “displace a population…not destroy it”—wrongly assuming one intent precludes the other. This logic recurs. Alex de Waal (2005:xix) concluded Darfur's atrocities are “genocide” in the legal sense of “systematic campaigns against ethnic groups with the intention of eliminating them,” but not “Genocide (capitalized) in this sense of the absolute extermination of a population.” A U.N. Commission on Darfur (2005:4, 100) simultaneously recognized that “sexual violence ha[s] been used by the janjaweed and Government…as a deliberate strategy” to displace, but still decided “the policy of attacking, killing and forcibly displacing members of some tribes” shows only “the intent to drive the victims from their homes, primarily for purposes of counter-insurgency warfare.”

These arguments miss multiple ways displacement, rape, anti-livelihood crimes, and selective killing can constitute genocide. Displacement can kill through exile, what the U.N. (1947) terms “slow death”:

If a state systematically denies to members of a certain group the elementary means of existence enjoyed by other sections of the population, it…condemns them to death at the end of a medium period instead of to a quick death in concentration camps. (25-26)

Displacement, therefore, can further the same genocidal process murder does: destroyed group membership. “Slow death” applies particularly in Darfur (Prunier 2008), where systematic attacks on food and water without humanitarian aid are predictably lethal. Because “slow” killing occurs in all genocides, any “discussion of genocide… should not therefore focus solely or even principally on deliberate attempts to massacre entire societies” (Maybury-Lewis 2002:45).

Sexual violence can likewise destroy group membership. In the Rwandan genocide, women were raped until death (actually making it extermination) or purposefully given AIDS to eliminate their group (Sharlach 2000). Bosnia's atrocities likewise involved rape as physical elimination: by displacement into “slow death,” by damaging Muslim women as reproductive vessels, and by purposefully giving them “Serb” babies (MacKinnon 1994). Women were sometimes kept in “rape camps” and impregnated in a society where fathers' ethnicity determines children's. Darfur's genocide mirrors these patterns: Black Africans report sexual mutilation, forced abortions, abductions into sexual slavery, and declarations that they will carry Arab babies (Hagan & Rymond-Richmond 2009).

Less recognized is elimination through social interpretations of victimization. Rape victims, for instance, “suffer grievous harm to their bodies, minds and ethnic identities” while being socioeconomically ostracized or rendered “unmarriageable,” so that “mass rape can destroy a substantial part of a group and thus constitute genocide” (Scheffer 2008; see also Human Rights Watch 2005; MacKinnon 1994). Socially destructive rape also predictably causes displacement, a key chain elaborated in the next subsection.

Still, whether such crimes can destroy group life is an empirical question hinging on how groups interpret them. We thus present the first systematic analysis to test how these crimes socially destroy. To do so, we build on Hagan and Kaiser's (2011a, 2011b) operationalization of elimination.

Measuring Social Destruction: Displacement as Both Mechanism and Effect

Complete social obliteration is arguably impossible. Every genocide leaves some damaged remnant of group life; even complete extermination would leave cultural and institutional residue with social reality for others. Consequently, we defined social destruction as obliteration or radical transformation of collective representations and institutions. Documenting successful destruction could therefore involve extensive histories, interviews, and observation to chart damage and change to the social order. Yet, genocide only requires intent to destroy, not successful planning and implementation (Schabas 2000), so even such comprehensive evidence would not measure failed genocidal attacks.

Hagan and Kaiser (2011a), however, show mass displacement can empirically measure successful and unsuccessful genocidal attacks. In addition to causing social destruction, displacement is a predictable outcome of genocidal attacks for three reasons. First, even when not fatal, forced migration may mark incomplete extermination; mass killing is a prime cause of flight (Gnyawali 2005). Accordingly, Schabas (2000) concedes survival cannot disprove genocide. Second, since victims flee diverse threats besides killing, attempted elimination through denial of food and other necessities incites flight (Davenport et al. 2003; Gnyawali 2005). Anti-livelihood crimes are especially important with Darfur's scarce resources; although killing is significant, the primary predictors of Darfuris' flight are attacks on food and water, and also arson (Hagan & Kaiser 2011a).

Most importantly, displacement can result from successful social destruction. Violence causes flight through perceived threats to personal integrity (Davenport et al. 2003). Sociologically, we should expect collective interpretation of such threats. “People do not make decisions in isolation—rather, they rely on the information available in their environment to make decisions…include[ing] input from others” (Davenport et al. 2003:43). Darfuris almost always flee in groups after attacks on communities'—not just individuals'—food and water: damage to familial and communal integrity (Hagan & Kaiser 2011a). Displacement occurs because collective ways of life are harmed.

Mass displacement thus indicates social destruction as cause or consequence (and genocide if intent is present). The converse, however, is untrue: social destruction can still occur without displacing. Other outcomes are possible. Thus, mass displacement is a conservative measurement of genocidal attacks.

Gendered Genocide: Hypotheses about Social Processes of Elimination

Hagan and Kaiser (2011a) argue socially mediated elimination is key to what makes any victimization socially destructive. They do not, however, unpack victims' experiences to investigate how genocide works: through differential targeting and collective meaning-making, the key processes in Figure 1.

In the science of crime, we must remember all crime is gendered (Britton 2011). Gender strongly predicts criminality. Victims are chosen largely through perceived vulnerability (associated with femininity) or threat (masculinity). Victims' experiences are inundated with gendered meaning-making (Heimer & Kruttschnitt 2006). Even “victimless” crimes have gender-specific effects that derive from and reproduce gender roles (Steffensmeier & Allan 1996). In short, since crime is gendered through perpetrators' characteristics, their decision-making, and its effects, every study of crime should consider gendered patterns. With men's and women's experiences collapsed, Hagan and Kaiser find some victimization—including sexual violence—nonsignificant in predicting elimination. Since rape is so prominent in Darfur's atrocities, their findings invite deeper analysis.

Theories anticipate mass rape in conflicts like Darfur's (see Ferrales & McElrath 2014). Classical approaches view rape as expressed masculine dominance in ultrapatriarchal societies, which fits Darfur's context (MacKinnon 1994). Wartime rape is also likely when violence targets minorities (Green 2004), especially when groups expect to live apart and independently afterward (Wood 2006). Even theories that rape furthers social cohesion in unstable perpetrator groups apply to the volatile combination of janjaweed and GoS (Cohen 2013). Nonetheless, these theories cannot account for the (genocidal) character of sexual violence without addressing resultant meaning-making.

Ruth Seifert (1994) presented five (untested) hypotheses about mass rape after Bosnia's genocide, and two address elimination.⁶ First, she argues, “Rapes in wartime aim at destroying the opponents' culture” (62). Despite assumptions about wartime honor, women are frequently raped. Since rape is actually costly and difficult to perpetrate, however, it is not caused simply by male urges (Cohen 2013). Instead, Seifert suggests women are targets of strategies intending to destroy groups because they are responsible for children and therefore perceived as keepers of culture. Thus, wartime rapes are “culture-destroying actions with strategic rationale” (Seifert 1996:62)—rationale that is by definition genocidal. Just as racial intentions are found in epithets Darfuris hear during attacks (Hagan & Rymond-Richmond 2009), perpetrators' speech can reveal gendered intentions to destroy Black African culture by raping women:

H ₁: Genocidal sexual violence is intended to eliminate groups by targeting women as bearers of community and culture.

Second, Seifert argues, “In belligerent disputes the abuse of women is an element of male communication,” symbolizing humiliating defeat of men in their inability to protect “their” women (1994:59). She suggests this logic was behind Serbian trucks returning visibly pregnant rape-abductees, and the blame rape victims faced during WWII. “At the heart” of the issue, Seifert argues, “is the outcome for men, not the suffering of women” (1996:59; Abusharaf 2006). Thus, unlike H ₁ about perpetrators' intentions, this hypothesis investigates rape's communication effects: meaning-making by victimized men. In the context of genocide and H ₁, we anticipate a message of elimination, that men are now unable to protect “their” culture, “their” women. Through this symbolic destruction of group life, we see a counter-intuitive, specifically gendered expectation:

H _2a: Sexual violence in genocide contributes to elimination, and therefore displacement, of men.

If Copelon and Schabas are correct, however, that genocidal rape is empirically equivalent to “normal” rape, H _2a would be false. Urban gang rape, familial rape, and other non-genocidal rapes are considered crimes against women that terrorize women (e.g., Card 1996). We certainly would not expect men's displacement in response:

H _2b: Sexual violence in genocide contributes to displacement of women.

These hypotheses capture intent and effects of sexual violence as a mechanism of elimination, but in conceptualizing gendered experiences of genocide, they are incomplete. Too many authors conflate gendered violence, sexual violence, and violence against women. For example, Copelon (1995:207) states, “Emphasis on the gender dimension of rape in war is critical to…surfacing women as full subjects of sexual violence in war” (see also Sharlach 2000). Such arguments encourage dangerous assumptions that only women experience sexual violence, that only men perpetrate it, that only women interpret it, and that women experience no other victimization (Ferrales & McElrath 2014).

This logic is the counterpart to conceptualizing genocides as centered on killing, the most hegemonically masculine form of genocidal victimization that occurs most selectively to ethno-sectarian men (see Cohen 2013). Men's victimization besides murder is often elided (Jones 2009); torture, beatings, rape, starvation, and disease are considered weaker, less honorable, and less important (Britton 2011).

We therefore question whether one gender's direct victimization accurately characterizes either that gender's experiences or patterns in that victimization type. Numerous Darfuri men suffered anal rape, but they self-report in even lower proportions than women do (Hagan et al. 2009). Similarly, women were more frequently killed than raped. There are gender-selective patterns—rape typically targets women; killing typically targets men—but, given H _2a, even rape of women can victimize women and men. Unpacking violence, then, requires examining each gender's victimization and meaning-making.

Nevertheless, since so few men directly experience sexual violence, other crimes must factor more prominently in women's elimination; Jones's (2000) work suggests killing, the genocidal crime that most selectively targets men.⁷ If sexual violence, the most women-selective one, manifests as communication to victimized men, killing likely communicates to women. Seifert's logic that women's rape signifies men's inability to protect them applies correspondingly; men's deaths signify women's lack of protection:

H _3a: Killing in genocide contributes to elimination, and therefore displacement, of surviving women.

As with H ₂, one-dimensional assumptions (that genocidal violence is indistinguishable from other violence) suggest the opposite:

H _3b: Killing in genocide contributes to displacement of surviving men.

We also consider a hypothesis corresponding to H ₁, theorizing perpetrators' intent behind men-selective killing. If women are targeted as bearers of Darfurian culture and community, men may be targeted as creators of culture and users of community, as the “generic” Black Africans (Card 2003). Many assume (battle-aged) men are targeted because they are “dangerous and threatening” (Carver 2004:285; Jones 2009). Just as assuming rapes result simply from male urges, it is simplistic to assume men are “preemptively” killed as threats; in genocide, targeted groups are already designated threatening. Gender-selective killing comes from deeper, gendered motives. Moreover, in Darfur's genocide where resistance is sporadic and futile (attackers have guns and bombs; defenders spears or tools), perpetrators are unlikely to perceive any Darfuris as actual threats. Instead, they socially construct men as dangerous based on widespread gendered assumptions. “Hegemonic masculinity” equates men with “people” and “citizens,” with their experiences considered the norm and women's the exception—inviting accounts of mass violence to “efface the male,” treating violence against men as ordinary (Jones 2000; Cohen 2013). Thus, men are socially constructed as the “default” person, the only being capable of exercising a will, possessing economic capital, providing sustenance, participating politically, or defending the group:

H ₄: Genocidal killing is intended to eliminate groups by targeting men as the default member and protector of the group.

Finally, even less-selective crimes manifest through gender. Our hypotheses thus far are consistent with patriarchal meanings one might expect anywhere in today's (ubiquitously patriarchal) world: men “possess” women and through them culture, so only men can protect those possessions. Correspondingly, we expect anti-livelihood crimes to disproportionately affect men, the users and protectors of “their” homes, provisions, and other property:

H ₅: Anti-livelihood crimes in genocide disproportionately contribute to elimination, and therefore displacement, of men.

Neither Copelon's nor other theories suggest anti-livelihood crimes would especially impact women, so we exclude this alternative. Our data also provide little information about intentions behind anti-livelihood crimes, leaving this matter to future researchers.

Darfur Testifies: The Atrocities Documentation Survey

Case-study approaches (e.g., Mann 2005) reveal macro-level context but not micro-level experiences or their relationship to communal outcomes, the crucial nexus for explaining gendered processes of social destruction. We integrate historical context with individual testimony and community-level statistics.

In summer 2004, the U.S. State Department constructed a survey to investigate the Darfurian crisis: the Atrocities Documentation Survey (ADS). The ADS was designed to empirically assess crimes by sampling 200,000 Darfuris who fled to Chad (see Figure 2, Map A). They sampled clusters proportionally by ethnicity and size of informal villages (which approximated originating settlements), with random walks of clusters and random sampling of adults within households, yielding 1,136 respondents (Howard 2006).

The ADS includes data on displacement and 34 other crimes committed by GoS and janjaweed forces. Testimonies include dated information about events experienced or witnessed from March 2003 to August 2004, including decisions to flee. Although respondents are all externally displaced, many report multiple, prior internal displacements. Several studies have found ADS respondents demographically representative of all displaced Darfuris (e.g., Hagan & Rymond-Richmond 2008; 2009).

Our analysis includes 930 respondents who fled from the 22 village clusters with 15 or more respondents and were able to recall dates of their experiences (Table 1).⁸ Reflecting disproportionate killings of younger men, the sample is about 60% women with an average age of 34 for women and 42 for men. The Fur, Zaghawa, and Masaleit are the most numerous ethnic groups. The Fur, who faced the most lethal victimization, are comparatively fewer.

Table 1. Descriptives, Darfuri Victims of Genocidal Attacks

The data provide controls for “push” and “pull” factors of displacement (Davenport et al. 2003). Darfur's hakura system grants land ownership to certain tribes, especially the Fur and Masaleit; other “Africans” (e.g., the Zaghawa) have partial rights and are therefore semi-nomadic (de Waal 2007b); and herders within tribes and entire “Arab” communities have inferior grazing rights (often sparking conflict). Ethnicity and occupation, therefore, are crucial controls to captures physical capital and lifestyle differences (Bohra-Hishra & Massey 2011). Since land-use rights are associated with patronage-based political power, they measure human capital, too. Also relevant for human capital are education level and household size (wife/child counts are status symbols). Nonfarming occupations are more complex: besides community/religious leadership (indicating status and civic commitment), they are all considered unworthy, lower-class work. Because business ownership may still involve physical capital, we measure it separately. The data include no measures of social capital, a principal migration predictor not yet comprehensively tested in forced migration (Bohra-Hishra & Massey 2011). Finally, given arguments that Darfur's atrocities are “merely” counterinsurgency, we control for nearby African rebel groups.

Together, respondents reported 4,147 dates of significant events—each including multiple crimes (Table 2). We exclude attacks targeting refugees on the road; 3,522 events precede their decisions to flee. The most frequent were bombings (41.7%), killing (33.9%), targeting of food or water sources (28.1%), and destruction of homes or entire villages (24.8%). We treat missing persons as killed, because respondents report considering them dead when contemplating flight. To control reporting bias, we disaggregate self- and other-reports of sexual violence (Hagan et al. 2009). Because event variables are binary rather than victim counts, they estimate victimization severity conservatively.

Table 2. Descriptives, Events Reported in Genocidal Attacks

Erring toward underestimation, we use only self-reports of displacement. ADS sampling interviewed one member per household (preventing intra-household dependence) and did not ask about other members' experiences, so we measure individuals' displacement decisions. Since decisions are affected by personal and others' victimization, other variables include self- and other-reports. Thus, our analyses measure the effects of community victimization on the individual decision to abandon home.

We use Cox proportional event-history analysis to measure displacement risks over time (Lin & Wei 1989). Continuous-time survival models allow likelihood measurement of daily influences of events on a hazard rate (of displacement), while accounting for temporal dependence and censoring when respondents join (are first attacked) or leave the study (are externally displaced). Additionally, Cox semi-parametric models do not impose a relationship between time and hazard rate, assuming only a proportional, constant effect for each covariate. We stratify each model to control for fixed effects of district-level variance within Darfur. Since Figure 3 (curves displaying daily displacement risks) confirms gender-consistent patterns even without proportionately constrained hazard rates, any differential results reflect gendered responses to community victimization.

Figure 3. Risk of Displacement during the Darfurian Genocide, by Gender

Additionally, we use qualitative ADS testimonies (anonymously indexed) to explain genocidal elimination more holistically. Interview teams included interpreters, trauma and refugee specialists, investigators, international prosecutors, and genocide scholars. Experts on sexual violence and linguistics trained interviewers to record exact language, create comfortable, private environments, and probe victimization in relation to settlements and attacks rather than personal experiences.

Interviews were close-ended and directed specifically toward potentially genocidal crimes, so testimonies are limited to direct answers and information interviewers considered relevant. Nonetheless, they provide considerable confirmatory and clarifying data. Because the crimes of the GoS and janjaweed are ongoing, our analysis is in the present tense.

Gendered Crime and Communication in the Genocidal Process

Genocidal Intent and Targeting

As Figure 1 shows, the process from racialized intent in Darfur (see Hagan & Rymond-Richmond 2009) to social destruction begins with perpetrators differentially targeting victims. Testimonies support H ₄ that perpetrators target men for extermination because they perceive men as the group's true members and protectors.

During attacks, the janjaweed and GoS purposefully hunt for men. Often, they “ask[] the women, ‘Where are the men?’ and beat them if they refused to answer” (516). Respondents recognize perpetrators do so because they selectively consider men threatening: “The women who were raped, said the military was asking, “Where are your men? They belong to the S.L.M.'” (697). Afterward, women are released or simply tossed aside, as one Masaleit woman (487) was after being forced to watch her husband and brother-in-law killed:

Arabs were laughing and…pushing a gun butt into my shoulders, saying, “Ha, ha, old woman,” while I was crying. I heard one wanted to kill me, but the other said, “No.…We can just tie her up. We need guns for the young men.”

Often, attackers express intent to kill “everyone.” One Zaghawa woman (372) saw her father and brother killed, witnessed her other brother's legs broken, and was stripped, beaten, and told, “We will kill all the people here and take the property.”

“All these people belong to the S.L.M. We will kill everyone here!” (692)

“We will continue killing you until you all leave this country.” (162)

Yet, perpetrators do not kill everyone. Unless they accidentally kill mostly battle-aged men, “everyone” refers to a class of person. Killing “everyone,” actually means killing “everyone who counts as anyone”: men. Men are “Nuba” or “people,” but women are “wives of Nuba” or “women”:

“You are Zaghawa women. Your men, all of them are Torraborra [trouble, rebels].” (756)

“You are mother of Torraborra. We will kill all the people who live [here]. We will not leave anyone. We will kill children, everyone.” (169)

“You are black people's wives…but now you have to bear white people's child.” (791)

Although victims and perpetrators need not share identical cultural meanings (one way genocidal intentions could be unsuccessful), historically linked groups likely have related perspectives. Victims' perspectives thus contain similar masculine-centric language, especially when asked about subgroups targeted or spared from harm. Their answers habitually refer to hegemonically masculine “targeting” and “harm.” One Masaleit woman replies,

At the Wadi, they came in cars and started shooting. They shot 10 people. They seemed dead.…They focused on the men and shot men. I think they shot the men so that they could take the women. (104)

A Fur woman (692) says, “Men were targeted. Some women were hit at random. But men were targeted and shot.”

Conversely, respondents routinely discuss women's rape when asked who is spared. The same respondent (692) reports, “Women were not shot. But five were rounded up and taken away and raped.” One Zaghawa man (630) even explains, “All people were killed. Children were slaughtered, women were put together and raped.” Like the perpetrators, Darfuri victims conceptualize men as the only true “people,” the only tribe members able to act for it. It is no wonder perpetrators perceive only men as threats requiring extermination, leading to the hypothesized gendered asymmetries:

There was four months of bombing. [On the day I left,] there were first more bombings. The huts were burning. Soldiers were targeting the men and grabbing women. There were mounted guns, and they were holding AK-47s and other guns. The janjaweed were killing people with guns and slitting throats with knives. Only men were targeted for killing. Janjaweed killed [about 55] men and [10] male children. Janjaweed raped the young girls. If the girls ran, they were shot. If they didn't run, they were caught and raped. The women were dragged outside the village to be raped. More than 50 of them. After the rapes, many women were taken to the hospital. Some died. Women who were raped told me that the soldiers said, “People from this area are against the government in Sudan. We will kill the men and rape the women and clean this area.” (698, Zaghawa woman)

ADS testimony likewise supports H ₁ that the GoS and janjaweed target women as perceived carriers and reproducers of culture. First, perpetrators indicate specifically targeting women based on their ability to carry offspring. Respondent 489 saw 32 family members and villagers killed while her sister and three others were gang-raped. Her sister told her “that Sudanese and Arab rapists said to them, while raping them, ‘We rape you to make a free baby, not a slave like you!’” Another woman (491) testifies, “The Sudanese soldiers said to some of the girls while raping them, ‘You Nuba are all Black, but we want to make Red babies.’” This pattern parallels the Bosnian atrocities. Darfur's genocide frequently incorporates rape-abductions (15% of events include both), and perpetrators express intent to destroy the Black African group by impregnating the women.

Second, when sexually brutalizing women, perpetrators express clear intent to displace Darfuris. Respondent 287, a Masaleit woman, testifies,

I was running after [a bomb killed most of my family] carrying my baby and my 3-year-old daughter. Two pickups, Toyotas, followed me with soldiers. I was taken and raped by 10 soldiers.…They threw my baby boy near a tree, my daughter was crying and trying to come to me and they kicked her away.… I was bleeding and could not walk. They [raped] me for nearly three hours. I was laying there while my village burned.…The soldiers said that President Omar al Bashir has sent them to do this, kill and rape and drive you from your land.

Another Masaleit woman (259) reports,

16 of us women were caught and raped there.…[One] was raped vaginally with her breasts slashed and a deep cut on her thigh. They shoved a stick in her vagina. [Another] was very pregnant at the time. Four Government of Sudan soldiers held her hand and feet. They took turns—vaginal rape. Shoved a stick far inside of her, until the baby was dead. They slashed her breasts. [Another], age 14, was raped by 4 persons—Government soldiers. Vaginal and anal rape. They shoved a stick in her vagina. Two girls…died after the gang rape, bleeding badly from breast cuts and their vaginas. They brought five local men they caught. Castrated them. They died from the bleeding. Five horsemen had me. Four held me down, raping me one after another. They took my clothing. Vaginal rape—oral rape.…They were laughing and shouting at all of us, “If you like this, stay in Sudan—if you don't, go to Chad.”

One woman (207) reports witnessing her 14-year-old sister raped and abducted: “We told them not to take her and they said, ‘You have two choices: one is you accept that we take her and go, and the second is that we kill her right here.’” A Zaghawa woman likewise says,

After we ran away, 3 of us girls went back to see and they caught us and took us for 7 days. They raped us each night, two men came and raped us each night. They beat us with their hands, slapped us. The three of us were in one house and they would come and take us and rape us, then put us back together. They kept asking where the men are. They said, “Why did you come back? You were supposed to go away.” After 7 days, they told us, “Go.” (163)

As we explain above, sexual violence intended to dislocate is by definition intentional social destruction. Even absent deliberate killing, ethnic impregnation, and anti-livelihood crimes that further slow death, perpetrators expressly mean continual threats of sexual violence to prevent Darfuris from maintaining life in their communities—i.e., to socially destroy.

Women report similar threats during other victimization:

The soldiers beat me with a belt and a whip, and they slashed my body. They tore off my clothing and left me.…I was beaten so badly I could not move. The attackers said, “What are you doing here? You have no land here. You must leave and go like others to Chad.” (554)

I was beaten and whipped when they were stealing my cattle. They were saying, “You black people are slaves and can't live here. You must leave this land.” (307)

I was struck five times with a cane as I ran from the village. By one Arab soldier. He grabbed me, called me “Nuba,” and…told us to leave the village and go to Chad. (378)

One man chased me on a horse and was whipping me.…He said, “Go away, you donkeys! Go away!” (207)

They said, “We want to eliminate you from this village and for you to go to Chad.” (82)

Of course, reports of “only” beatings may indicate undisclosed sexual violence; interviewers noted such suspicions (e.g., for 554 above). Yet, even if only 9% of events include sexual violence (Table 2), testimonies make clear from GoS and janjaweed threats to “leave” that they do not simply target women for enjoyment (Cohen 2013), but rather as a strategy of eliminating:

Alhayli Harin, [a] Janjaweed leader, was a local commander who went into villages and commanded locally…. I heard him say to capture beautiful girls and kill the boys, even small boys. Those women were all raped. After the rapes, the women told me that they said, they should go tell their families that they will all have to leave Darfur. The girls were all before marriage and one was about four years old. She died from the rape. (496, Masaleit man)

Collective Experiences

Contrary to assumption, no rule states genocidal violence must manifest as intended; victims can interpret it as undamaging or socially destructive through unanticipated avenues. Thus, both H ₁ and H _2a would be true even if perpetrators intend to displace women by sexually victimizing them as bearers of culture but men actually flee in response. H ₁ investigates perpetrators' acts and intentions (genocide), while H _2a concerns victims' interpretations and responses (successful social destruction).

The ADS's first substantive question asks directly about displacement: “Why did you leave your village?” Yet, respondents hear an introductory invitation to discuss violence generally. Model I of Table 3 shows displacement occurs most frequently after targeting of food and water (β=2.29, p < 0.001) and killings (β=2.01, p < 0.001). Figure 4 shows consistently high risks of these crimes on any given day, especially in mid-2003 and early 2004 (see Appendix 1). Hence, we might expect answers to this question to focus on these events.

Figure 4. Risk of Genocidal Victimization during the Darfurian Genocide

Table 3. Cox Proportional Hazard of Displacement in Darfur, Mar 2003–Aug 2004 (Stratified by District)

Exponentiated coefficients.

^a Age and “Other ethnicity” included but not printed.

^* p < 0.05, ^**p < 0.01, ^***p < 0.001.

Responses, however, reveal gendered scripts about violence. Consistent with violence being socially organized around masculinity, they often begin with killing:

Because the attackers—they killed my sons. So I took my other children and we left. (129)

I left because the Arabs attacked—they were killing people and stealing things. We heard of other villages being attacked, but this is the first time we were attacked. (156)

I left [my village] in November 2003 after bombing killed my son. We went to [another village] and stayed two months. We left [there] on the day of the attack. Military vehicles came with strong guns, killing people. We ran away. (681)

Some simply answer, “They killed us,” speaking directly of communal destruction (438).

Respondents also report generically, “The Arabs attacked” (157), or, “The attack started at 5 a.m. I awoke to the sound of guns…” (553). Respondent 354 simply says, “People came by airplane, horses, and guns. How could I stay?” Likewise, they report ubiquitous bombings—occurring months before, during, and after ground attacks, chasing refugees as they flee. A Zaghawa man (427) responds, “The main reason [I left] was the nearly daily bombings of my village,” but like most respondents, despite its “terrible psychological effect on all of us, especially the children and even the animals,” he stayed until ground troops attacked. Respondents 690 and 166 relate,

There was six months of bombing.…People began hiding in the mountains.…[Then,] the Sudanese military came and killed with machine guns. The janjaweed killed by knife.

The planes kept coming and bombing but we kept staying. But then the Army came in cars with machine guns so then we ran, leaving everything behind.

Because Darfuris endure these ceaseless bombings, they do not predict flight.

Conversely, Darfuris are less at risk for attacks on food and water, but they are the most statistically significant cause of displacement (Figure 4). The few who actually answer the question explicitly say they cause elimination. One Masaleit woman (381) reports,

Once it was dark, I decided to venture back to the village to see what had happened.…[M]y hut…was burned to the ground. I couldn't salvage anything. My four cows, three donkeys, and seven sheep were also missing. The Arabs must have stolen them.… There was nothing to stay for and so I left for Chad the same evening.

Another (337) ordered her losses this way: “They killed our men, burnt our houses, stole our livestock—so we had to run.” Still, they rarely mention anti-livelihood crimes—or sexual violence—in initial descriptions.

Many consider their reasoning self-evident. Respondent 576's interviewer noted “laughter as although this is a stupid question,” followed by, “They want to kill me!” Respondent 196 is typical; he endured four attacks before fleeing, never explaining what was so special about the last one (the first to include rape).

Rather than unimportance, respondents' reticence to discuss sexual violence likely reflects hegemonic masculinity. Ideas that rape is humiliating or the victim's fault for not safeguarding her (or his) body erase it from discourse (Britton 2011). Indeed, despite its relatively low risk throughout Figure 4, rape in Darfur is strongly associated with severe violence and explicit racial intent (Hagan & Rymond-Richmond 2009), and when the survey later brings it up, respondents describe sexual violence as critically important.

If so, rape may only be statistically nonsignificant for Hagan and Kaiser (2011a) and Model I because victims' experiences are gendered. In Models II and III, we separate by gender; Model II represents men's victimization while Model III represents women's. Model IV shows all respondents with interaction terms revealing collectively gendered experiences.

Intragroup Meaning-Making

Models II–IV support hypotheses derived from Seifert and Jones over those from Copelon and Schabas. Model II discounts H _3b: men who witness killing are not more likely to flee (p > 0.1). Actually, Model IV shows that killing decreases men's hazard of displacement by almost half (p < 0.05). Conversely, men who witness sexual violence are almost 1.8 times more at risk for flight (p < 0.001; in Model II, β=1.33, p < 0.05), supporting H _2a. Men are displaced primarily by victimization of women and communities, not themselves (Figure 5).

Figure 5. Darfuris' Gendered Risk of Displacement from Sexual Violence and Killing

Testimonies illuminate the meaning-making underlying this pattern. That certain subgroups perceive messages of elimination from particular crimes suggests they interpret only some violence as radically antagonistic to collective representations—and some as undamaging (see Eliasoph & Lichterman 2003). Put another way, the primary mechanisms communicating social destruction involve (gendered) expectations of normality.

Men repeatedly view killing as routine. One Masaleit man (379) matter-of-factly observes, “[S]ome of Dawy's soldiers came into the pastures outside the village and took 80 cows…and killed four people. All men. I think they were testing out reaction, but there was nothing that we could do to defend ourselves.” Another (584), shot twice, shrugged: “The attackers have been killing and looting since 1997.” One Masaleit man (122) who experienced eight attacks says it best: “In this village, many people died but I did not see them. I heard that a lot died, but I do not know a number. Killing is normal.”

Conflicts between Darfur's “African” farmers and “Arab” nomads are recurrent and escalated as food and water became scarcer (de Waal 2007b), making killing routine and expected. Hence, respondent 172 comments, “There have been problems with Arabs coming to out village for over four years and taking our cattle.” Since the role of warrior and defender is traditionally masculine, it makes sense that men in particular view homicidal violence as consistent with their way of life (Carver 2004).

Conversely, men only abnormally experience or even hear about rape; women only reticently discuss it. Especially in Darfur's ultrapatriarchal society, masculinities cause women to minimize rape's criminality, internalizing rather than divulging (Britton 2011). Muslim women may even be blamed for “allowing” rape (failing to protect their offspring-producing bodies or, as Seifert would say, their culture) and thereby automatically divorced and ostracized. Thus, when confronted with public spectacles of mass rape, Darfuri men respond oppositely from their routinized reactions to killing:

I saw ladies in the village (as I lay wounded) being raped right in front of everyone, even their fathers and their children. By janjaweed. They would catch them and do this terrible thing…We could do nothing. Nothing. (258)

A Zaghawa man (630) whose wife was raped in front of their sons testifies,

Two of my sons have become mentally unstable, and my wife was very traumatized. Sometimes they don't make sense when they talk….I have seen many women raped. Too many to count. I saw some of them naked. I saw soldiers take them and put the butt of their rifles in their sexual organs.

This alarm and distress increases men's chances of fleeing. Respondent 629 says, “I was not physically harmed, but I saw my female relatives stripped naked and taken away naked in cars by soldiers. …Then, I ran away because I couldn't stand to see the women in my family.” Rape tends to accompany torture, beatings, mutilations, and other severe victimization (Hagan et al. 2009), crimes men may also view as nonroutine.

Women respond to victimization correspondingly. Model III contravenes H _2b while supporting H _3a (Figure 5): women do not abandon home after sexual violence (p > 0.1)–or rape-abductions (Appendix 2)–but killing increases their flight risk by 2.5 times (p < 0.001). If anything, Model IV shows sexual violence encourages women to stay (β=0.78, p < 0.05). The most ready explanation for elimination of women through killing but not sexual violence parallels men's perceptions of normality: Darfuri women interpret rape more as a known, expected event while being unused to eyewitnessing killing.

Darfuri women never call sexual violence “normal,” but unlike killing, which is routinely accepted and honorable for men, rape is routinely unacceptable and dishonorable for women. An elderly Zaghawa woman (209) explains,

At that time, the Army was taking over the well water, so no women except the old women went…But some of the women who lived far away had to come for water, and then they would be captured and kept for two days or three days or…eight days even. After the Army went, those girls left but they did not ask for help, because they could not go to a village not their own and tell they had been raped.…After the Army left, some old women went to see what had happened…but they couldn't say anything to the girls because they knew what happened.

Women in many cultures view rape as “standard operating procedure,” an expected outcome of domineering masculinity (Britton 2011:82), explaining how Darfuri women routinize it.

They generally do not, however, face killing in the gender-segregated course of Darfurian life (de Waal 2007b). Thus, men never report shock or fear from killings, but women do:

I did not want to see the bodies—I just wanted to cross the border…there are many soldiers in [town], so I was very afraid…I cannot describe their faces; I was too afraid to look. (202)

While fleeing the attack, it felt like running on dead bodies there were so many dead and injured. (4)

Attackers were shouting.…I was in a panic and did not recognize or register what they said. (563)

People ran. They didn't take men but shot them only. They killed the men, but the women [were] always tortured and frightened. (87)

Antinovs attacked the village with 300 men on foot with green, yellow, and blue uniforms. The airplane dropped bombs, the victims of which were dismembered. I fled my village in fear. (740)

I was in the village when the bombs fell; I had just given birth to a child. I was scared. (455)

Although these answers may simply be more acceptable for women, there is reason to expect this pattern; their fear and panic are socially scripted. Research shows women internalize rather than panic after rape (Britton 2011). Yet, Darfuri women report fleeing directly because of seeing men slaughtered:

I was at home when the attack began. A man entered the house. They shot my husband even before entering, as he was carrying furniture outside. Then, they looted the house. I took my children and ran away. I panicked when my husband was shot and ran all the way until I reached the Chad border. (574, Masaleit woman)

Socially Destructive Meanings

We have shown how victims receive messages from collective violence, but not yet how they interpret content. In Darfur's genocide, each gender differentially receives the same message: elimination.

The most prominent mechanism producing such meanings corresponds to the gendered logics Seifert describes: women keep culture; men create and use it. Consider Table 3's evidence of H ₅: attacks on food and water, while significant in every model, are far stronger for men in Model II—increasing their flight risk by 4.5 times (p < 0.001)—than for women in Model III—increasing their risk only 1.6 times (p < 0.01). Model IV shows a less dramatic divergence (β_men=2.33, p < 0.01; β_women=1.69, p < 0.01). Arson targeting buildings besides homes also significantly increases men's hazard of displacement (β=1.98, p < 0.01) but not women's (p > 0.1). Figure 6 captures these effects. Unexpectedly, arson targeting homes shows no gender-selective effects (Appendix 2).

Figure 6. Darfuris' Gendered Risk of Displacement from Anti-Livelihood Crimes

Overall, Table 3 shows men are most at risk for displacement when “their” property is threatened: food, water, buildings like market stalls and granaries, or “their” women. Women are most at risk when their provision (food and water) or providers (men) are targeted. Despite the ADS's imperfect, event-focused approach to victims' experiences, many interviews confirm such meaning-making:

I went back to the village later. …I saw about 55 corpses of villagers. …Then, I discovered that all my livestock had been stolen and my two shops had been looted. Houses in the village were destroyed. There was nothing left for me.…I had already taken my family to Chad…and now I went back to join them. (384, Masaleit man)

They killed my daughter.…They beat my three sons. They took away my sons…to make them heard [sic] my cattle.…I have four wives, and only two are here with me. The other two were raped by the janjaweed.…I saw this start but then had to run and hide… .I am here and have nothing any more, and even my other wives are not here with me. (786, African man)

Darfuris emphasize being told rape victims now “belong to” the perpetrators. One man (575) testifies, “They also said that they would kill as many Masaleit as they could and the rest would never live there again. They also said, ‘We will take your women and make them ours. We will change the race.’” Another (533) reports hearing, “We will kill all the men and rape the women. We want to change the color. Every woman will deliver red. Arabs are now the husbands of these women.” Women remember, “Following the rapes, the military would say…, ‘This is a new wedding day for you’” (690).

After interpreting victimization as radically antagonistic to their ways of life, Darfuris flee: our measure of elimination. GoS and janjaweed intentions to socially destroy succeed:

My brother was killed and my house has been burned. My cousins, all of them died, so I left, alone. (78, Zaghawa woman)

Alternative Meanings and Caveats

Testimonies most frequently reveal narratives of property and provision, but other patterns exist. Darfur's Black Africans live in farming communities that focus on tribe and family. Genders contribute to household welfare differently, but each concentrates on mutual survival. Thus, Darfuris facing genocidal violence worry especially about their families.

Sometimes, men stay behind while women and children flee. One woman testifies, “Our sheikhs told us to protect our children. On Monday, we left.” A wounded man (157) says,

The Arabs attacked. They attacked 2 times… The first, they entered the village, and I saw two men killed, they shot them. Everyone left the village.…With the second attack, they killed 19 men and burned the village. At that time, there were no women in the village, they had all run away with the children.

Another (220) relates, “We ran away and took our women and children to a safe place. After one week, the men went back…to bury the bodies.” Similarly, respondent 591 says, “We were afraid for furniture, so we kept men to protect it.…After that, [women] ran from the village.” One woman (251) laments, “My husband was wounded—I had to leave him in the village.” After numerous attacks, respondent 434 reports men fled specifically to “ensure our families were safe.”

Thus, Darfuris resist social destruction through gender roles: men stay to maintain the way of life, while women ensure the children's (culture's) survival:

Woman and children run before men. We ran. I carried my children and hid outside the village.…Men inside the village were without weapons. They said that they couldn't leave, that they must die there.…The village was completely destroyed. The ground was full of bodies. (597, Zaghawa woman)

These roles indicate partially structural explanations for gender-selective victimization. By staying, men more likely perish while women survive to testify. Likewise, gendered displacement partly follows from pre-attack separation of genders: men visit market and tend crops; women shepherd and fetch water. One woman states, “The first attack, they…came into the village to look for the men. The men were all at market in another town.”

Structural practices also influence patterns during attacks. Being more mobile, men more likely experience (a) multiple attacks and (b) attacks that do not affect their community's homes and provisions. When witnessing attacks on other villages, men often report only shooting and killing. They do not inspect damage to other communities' structures and provisions, nor would such damage likely increase their displacement risk. Respondent 568 reports, “Nine months before the attack [on my village], 51 people were shot by janjaweed in the market and nine were wounded…The market is held once a week.…It is a 45-min walk from [home].”

Our findings are generalizable. Overall patterns recur in mass atrocities (Mann 2005). Seifert's (1994) hypotheses, based in Bosnia, hold in Darfur despite dissimilar contexts. Anchored in patriarchal cultural logics, our hypotheses likewise broadly apply in today's universally patriarchal world (Jones 2009; Britton 2011). Still, extent of generalizeability is an empirical question, for example, where perpetrators hold highly patriarchal norms compared to victims'.

Lastly, it is important not to essentialize. Through rigid categories, statistics can depict gender roles as distinct and inflexible, whereas gendered boundaries and behaviors are quite fluid. For instance, despite conventional wisdom, our data contain almost no examples of Muslims ostracizing raped women. Many men who witness or know about women's rapes continue to support them, remain married to them, and defy every behavior discourse expects of them. Additionally, our results could reflect (unreported) paralytic responses to victimization, a predictable powerlessness where fleeing after genocidal attacks is impossible without encouragement from others. Nevertheless, attacks are highly gender selective, creating gendered patterns that victims respond to and reify. In this sense, we truly see a gendered genocide wherein perpetrators' intentions and actions, and victims' meaning-making and responses are all produced by and reproduce gendered thought and behavior.

Conclusion: The Gendered Process of Genocide

In a gendered society, all crimes are gendered. Since all modern societies are gendered, so are all crimes. Genocide is no exception.

Feminists have long argued that criminologists study highly gendered phenomena (Kruttschnitt & Macmillan 2006; Steffensmeier & Allan 1996). It should be unsurprising that criminal behaviors and experience reflect gender roles, and that purposively analyzing gender substantially changes the picture of crime. This study reminds us that fully understanding crime—especially genocide—is impossible without a gendered, multidimensional lens.

Our results show that the most significant events predicting Black Africans' displacement during the Darfurian genocide are crimes targeting collective livelihoods and other genders. Women's flight risk is best predicted, respectively, by killing (primarily of men) and attacks targeting food and water. The crimes that most increase men's risk are attacks on food and water, sexual violence (primarily of women), and attacks on buildings besides homes. If anything, sexual violence prompts women not to flee, as does killing for men. As Seifert (1996) theorizes, men interpret sexual violence as communicating they cannot protect “their” women. Likewise, our data show women interpret killing as communicating loss of men's protection and provision. In both cases, gender-coded meanings and anti-livelihood crimes signify social destruction of Darfuris' ways of life. Successful genocide involves dual processes: immediate killings of group members—“extermination”—combined with crimes that cause slower, more subtle destruction—“elimination.”

Similarly, testimonies show perpetrators' genocidal intentions filtered through gender roles to produce gender-selective rape and killing: they target women for elimination as carriers and reproducers of Black African culture and community, and men for extermination as the group's default members and protectors. Whether the GoS and janjaweed plan particular gendered outcomes, their intent is to socially destroy the group through gender-selective violence, and they succeed in doing so.

Scholars and other experts routinely misunderstand socially destructive intentions and results by single-mindedly conflating genocide with mass murder. Schabas (2000), Copelon (1995), and others adopt such one-dimensional understandings. Genocide, however, targets a social entity—the ethno-sectarian group itself. Destroying a collective entity requires more than just killing; it means socially destroying its membership and way of life, by inciting collective meaning-making that radically transforms institutions and collective representations. Although complete extermination could arguably accomplish group death, every commonly recognized genocide in history includes widespread elimination. Yet, largely because of these misunderstandings, the world allows both to continue in Darfur.

Displacement, sexual violence, anti-livelihood crimes, and selective killing are therefore essential mechanisms of genocide. Despite Copelon's and Schabas's assertions that there is no such thing as genocidal rape, our results demonstrate otherwise. Genocidal rape is empirically distinct, if for nothing else, because it has genocidal intent and may displace men who are not its direct victims. To the extent Copelon normatively argues we ought not recognize genocidal rape, we also disagree. Because the world newly noticed wartime rape following Bosnia's genocide, she worries, “the terrible war-time rape of women in the former Yugoslavia will [either] disappear into history” alongside rape in every conflict, or “survive but be viewed as an exceptional case” (1995:198). We too lament under-emphasis on mass sexual violence, but ignoring the distinctiveness of genocidal rape exacerbates misunderstandings of interrelated forms of violence.⁹

Arguments that atrocities like Darfur's are nongenocidal are likely produced by and threaten to reproduce the misunderstandings this article aims to correct. For instance, perpetrators can intend rape to socially destroy without intending physical group damage and with unexpected results. Similarly, much debate over Darfur's genocide involves the surviving refugees, from which many conclude the attacks are “ethnic cleansing” and therefore nongenocidal (e.g., Schabas 2000; Mann 2005). Others suggest gender-selective killing proves the atrocities are no more than overzealous counterinsurgency (e.g., International Commission of Inquiry 2005; de Waal 2007a). It can be all three. Genocide requires deliberate acts with intent to destroy a group; why perpetrators wish to socially destroy (their motive) is entirely irrelevant to classifying the crime. Ethnic cleansing by genocide and counterinsurgency by genocide are both genocide, whether successful or not.

It is unsurprising that many find “genocide” conceptually confusing. We too often uncritically accept one group's experiences as representative of an entire phenomenon, and the defining group is too often the dominant, hegemonic one. Sexual violence, for instance, more often targets women than men, so its presence in wartime and genocide is consistently ignored (Seifert 1994; Copelon 1995). Even recognizing genocidal crimes that more often target women, children, and the elderly is not enough, however.

Understanding Darfur's and other collective violence requires critically analyzing social processes that differentially target various group members and, through their experiences, affect the entire group. We must recognize indirect effects crimes have on all subgroups, through their interactional interpretations of victimization. During widespread, systematic attacks on a population, even rape and torture victims are not the only group members who recognize destruction of their livelihoods and the consequent need to find a new home; the exclusive focus on extermination has obscured even the effects of mass killing on those who “merely” witness it. Killing itself functions not only as extermination but also as another form of elimination, terrorizing survivors and forcing them to abandon home for starvation, disease, and death along paths of exile.

Together, there are the forgotten victims, left to live on without their homes, histories, or cultures. If we are to learn from their experiences and understand collective violence, we must remember that they, too, are victims of crimes that target their social existence for destruction. They, too, are victims of genocide.