By Mira Fey
Lee Ann Fujii,
Killing neighbors: Webs of violence in Rwanda
Cornell University Press 2011
Previous research on driving forces of the genocide
Questions of intimate mass violence in ethnic killings are especially puzzling for researchers investigating political violence. Various approaches examine reasons for popular participation in the relatively recent genocide in Rwanda which exhibited especially brutal killings of acquaintances such as neighbors by regular villagers.
Some of these approaches link ethnic violence with structural violence by looking at cultural and historical particularities that allegedly led to distinct ethnic identities and a society divided along ethnic lines that can be equated with class divisions (Mamdani, 2001). Others focus on the role of the state and the hierarchical organization of Rwandan society as facilitating conditions for mass participation in the genocide (Prunier, 1995).
According to Straus who conducted one of the most extensive field studies in post-genocidal Rwanda, these perspectives portrayed the genocide as “a state-organized, planned extermination campaign that served elite interests and drew on constructed ethnic categories” (2004, 86), thereby shifting the analytical focus on macro-level examinations and elite actions away from the original puzzle of popular participation. Consequently, Straus’ own research focuses on local-level dynamics of the Rwandan genocide.
He conducts surveys, case studies, and interviews with 210 randomly selected sentenced inmates to answer the narrow question how many Rwandans participated in the genocide (Straus, 2004, 90). Instead of concentrating on his estimate (200,000 perpetrators), he stresses the need to disaggregate the category of “perpetrators” according to their degree of participation, and to investigate reasons for their participation in future research (Straus, 2004, 95).
Taking a local-level approach
This is the starting point of Lee Ann Fujii’s PhD research which was later published as the book Killing Neighbors – Webs of Violence in Rwanda. Her overall focus is on understanding how violence becomes “ethnic”, and how these two categories are intertwined (Fujii, 2011, 10). She distinguishes her approach from existing macro-level research, stressing the need to go beyond ethnicity-based approaches.
These include both the “ethnic hatred” thesis, aforementioned distinct ethnic group identities that include a strong collective hatred passed on over generations through myth or memory, and the “ethnic fear” thesis, which is described as “elite ambitions or moves” perpetuated through extremist media, state enticed violence such as organized riots and arbitrary arrests, and other techniques (Fujii, 2011, 4). Both approaches perceive ethnicity as some kind of underlying divisive force in society which might erupt as violent acts under specific conditions, especially those of threat and insecurity.
In order to surpass these macro-level approaches, Fujii proposes, following Straus’ suggestion, a disaggregation of macro-level concepts and categories. This includes the concept of “masses” as the perpetrators of the genocide and as “an undifferentiated hole” (Fujii, 2011, 9), and the perception of ethnic groups as unitary actors sharing the same interests. Additionally, Fujii underlines the need to break up the concept of perpetrators into smaller groups of individual actors and group actors, namely “local leaders, their collaborators, and a group called ’Joiners’ ”.
These are defined as regular or “ordinary” members of their community who had the least to gain from the genocide, but who were the ones who carried it out at large (Fujii, 2011, 15-16). These subdivisions of the overall concept of perpetrator are very useful when analyzing what motivated the majority of the participants who did not kill enthusiastically, but reluctantly or under force as stressed (Bhavnani, 2006, 651). Instead of focusing on the killers who clearly exhibited hatred towards Tutsi from the beginning or on the instigators who gained most from the genocide, concentrating on the ones who were not driven by hatred and who gained the least is the only way to understand which forces drive individuals to participate in mass slaughter of neighbors and other acquaintances.
Ethnicity as state-sponsored script
Additionally, Fujii views ethnicity not as external force, but as a state-sponsored script for violence that people are supposed to act out like a play. She refers to it as a “dramaturgical blueprint for an imagined world”, created by threatened elites which see genocide conducted by large parts of the population as the best way to protect their positions of power (Fujii, 2011, 12).
This is one of the cornerstones of Fujii’s thesis, but also one of the aspects that is easily questionable. While she wants to distinguish her approach from the “ethnic fear” thesis, both view ethnicity as prescribed views by the elite and the state in order to promote violent actions. The divisional line between Fujii’s state-sponsored script for violence and the “ethnic fear” thesis remains blurry. One can only distinguish these by their different foci, the first on the actions as being acted out like in a play, the second on the motivations behind these actions as fear resulting from state-propaganda. Unfortunately, Fujii does not resolve this tension throughout her analysis.
Genocide as a process
With regard to further terminologies, Fujii stresses her definition of “genocide” as a process rather than a clearly bounded event. This leads to different expectations on actors and their actions as not necessarily linear, but rather fluid and extending into various categories. Consequently, the same actor might have been victim, perpetrator, bystander, and rescuer at some point during the genocide. Instead of focusing on fixed roles, responses to the genocide should be seen as various parts within a spectrum (Fujii, 2011, 30).
Again, this breaking up of a fixed term gives greater flexibility to the analysis as it allows for a greater range of possible actions and responses to the genocide and does not confine actors to specific roles. It follows Fujii’s constructivist approach and is useful within this framework. However, it impedes easy comparability of her thesis with rationalist approaches such as Bhavnani’s model on the development of ethnic norms and their explanatory role in mass participation in the genocide (1).
As both authors try to solve the puzzle of what drove public participation in the genocide, a comparison of these approaches could have been useful. The ambiguity resulting from Fujii’s definition of “genocide” and her view on actors as occupying multiple, sometimes contradictory roles hinders an easy juxtaposition with Bhavnani’s approach even though he does not takes actors as static or unitary, but also acknowledges change and adjustment to evolving norms. Specific obstacles to the comparability of both approaces can be found in Fujii’s research design which led to an inability of generalizing her findings; these will be described in the following paragraph.
Fujii’s research design
When conducting her research, Fujii focused on interviews conducted on several research sites, two rural communities and central prisons in the provinces Ruhengeri (North Rwanda) and Gitarama (Central Rwanda). The communities in Ruhengeri consisted mostly of Hutu, but were subject to attacks from the Rwandan Patriotic Front (RPF) since the beginning of the 1990s which led to mass killings of Tutsi before the start of the genocide in 1994.
Conversely, communities in Gitarama had larger proportions of Tutsi, but did not experience the war or mass killings of Tutsi before the beginning of the genocide (Fujii, 2011, 27). As Fujii states, she selected these two very different provinces to capture these distinctions, but also other given historical and political differences present in pre-genocidal Rwanda. This method of selection is not immune to selection bias and therefore impedes an easy generalization of her findings (which, according to Fujii, is not necessarily an aim of her research). The selected rural communities are not representative for the whole of Rwanda.
Moreover, given the sensitive topic of her research and her intention to cover the spectrum of different responses to the genocide, Fujii relied on snowball sampling. This is another aspect that contradicts the scientific ideal of randomly selecting respondents to avoid biases, to which Straus adhered to as much as possible when randomly selecting interviewees from the prisons.
Fujii also used repeated visits for a specific core subset of people to go more in-depth for especially sensitive issues without justifying the criteria used for selecting which subjects where questioned repeatedly. All in all, her research design does not allow for a generalization of her findings, but restricts these to the communities from which she selected her respondents. Thus, her thesis cannot really be compared to a rationalist approach such as Bhavnani’s model.
Groups as constitutive element for the killings
In the remainder of her book, Fujii first gives a short overview of Rwandan history before covering local narratives and explanations for the genocide given by her interviewees. She then describes and analyses views on ethnicity and its complex relation in Rwandan society. In the following chapter, she presents her understanding of the mechanisms which led to popular participation in the genocide, namely local ties to power, friends, and most importantly groups. Consequently, she describes how groups worked as constitutive element for the execution of the killing by creating a particular social actor with a particular identity, the Interahamwe (Fujii, 2011, 174-78).
She stresses the form of the killing as a crucially important feature: it was conducted in groups as public acts in broad daylight, was physically intimate through the use of clubs, hoes, axes, machetes, and other similar tools, and contained certain “theatrical elements” (such as chants, banging, or dressing up). Moreover, Fujii lists specific acts that were performed by groups as part of the killing, but never outside of this context (such as spotting and circling the victims, threatening these, forcing others to join in, raping women and girls, torturing victims, pillaging, and mutilating dead bodies) (Fujii, 2011, 173).
Fujii’s descriptions of the group actions and the consecutive analysis compellingly make use of constructivist theory to understand mass participation. Nevertheless, the focus on group action and the constitution of a new, collective identity that is resumed instantly when the individuals assemble and as easily forgotten for the night when the group dissolves leaves several questions unanswered. The step from being individual Hutu who had no particularly hateful feelings towards their Tutsi neighbors to a murderous group of Interahamwe remains unclear; while the idea of a specific group identity seems reasonable, one is left to wonder what motivated individuals to assume this identity.
This is where Bhavnani’s thesis of punishment, coercion, and reversed ethnic norms might have provided a useful addition to Fujii’s argument. Moreover, one needs to keep in mind the aforementioned aspects about the lack of generalizability due to Fujii’s research design. In addition to the obstacles described in the previous paragraph, her method of quoting selected interviewees on specific topics as evidence for her thesis is also questionable. Again, she does not justify which criteria were used to select these specific persons from the 81 she interviewee. Instead, the reader is left wondering if the interview parts were particularly generalizable or very peculiar aspects.
In conclusion, Fujii’s Killing Neighbors – Webs of Violence in Rwanda presents a different perspective on the genocide by analyzing local-level group behavior in order to understand which factors drove mass participation in the killing, especially of acquaintances and neighbors. The findings from her interviews conducted in two rural communities and prisons lead Fujii to conclude that social ties and group dynamics played a key role in motivating “ordinary” people to join in the violence in the first place and to continue killing for weeks. Accordingly, not increased hatreds or fears propelled Joiners to continue their participation, but the constitutive power of killing in groups. These killings constituted the group as a particular kind of social actor with the particular identity of the Interahamwe (Fujii, 2011, 187).
Fujii’s constructivist approach allows her to break up fixed terms and categories such as “ethnicity”, “genocide”, and “perpetrator” which proves useful for her analysis. However, as stressed throughout this review, this dissolution of terms into fluid categories impedes comparability of her thesis with other research focusing on very similar questions such as Bhavnani’s model on the development and reversion of ethnic norms during the genocide.
Moreover, due to aforementioned flaws in her research design, Fujii’s findings are restricted to the communities she conducted her field study in and not easily generalizable to the whole of Rwanda, let alone other genocides and mass killings linked to ethnic violence.
Nevertheless, I would still recommend her book to graduate students as it contains a detailed description of Fujii’s research and can help students identify and avoid flaws that will devalue the findings of future research. Additionally, despite the shortcomings, one needs to keep in mind the wide cognitive divide between the extreme brutality and gruesomeness of the genocide on the one hand and the participants as ordinary members of their communities on the other hand.
To bridge this gap and to understand what drove mass participation is extremely difficult, more so years after the event when former participants spent much time rationalizing their behavior. Thus, Fujii’s book can be seen as a piece to the puzzle, but it should be combined with other parts such as Bhavnani’s model when one tries to recreate the whole image.
1. Like Fujii and Straus, Bhavnani also wants to go beyond conventional explanations of the genocide which focus on the macro-level and elite actions. Nevertheless, instead of conducting field research and trying to understand the actors’ motivations through interviews and the like, he creates a model analyzing the likelihood of the development and potential reversion of ethnic norms. Accordingly, through the use of punishment, mass participation in the genocide was enforced; “conformity was cultivated through compulsion over the course of the genocide” (Bhavnani, 2006, 657). His model supports this initial hypothesis and shows that social norms might be reversed under specific circumstances; thus, “ordinary people can engage in violence en masse (Bhavnani, 2006, 667).
Bhavnani, R. (2006). “Ethnic norms and interethnic violence: Accounting for mass
participation in the Rwandan genocide.” Journal of Peace Research, 43(6):651–669.
Fujii, L. A. (2011). Killing neighbors: webs of violence in Rwanda. Cornell University
Mamdani, M. (2001). When victims become killers: Colonialism, nativism, and the
genocide in Rwanda. Princeton University Press.
Prunier, G. (1995). The Rwanda crisis: History of a genocide. Columbia University Press.
Straus, S. (2004). “How many perpetrators were there in the Rwandan genocide? An
Estimate.” Journal of Genocide Research, 6(1):85–98.