Peace Science Digest

Violence, Intergroup Bias, and Dehumanization in South Sudan

Photo Credit: UNMISS via flickr

This analysis summarizes and reflects on the following research: Calissendorff, L., Brosché, J., & Sundberg, R. (2019). Dehumanization amidst massacres: An examination of Dinka-Nuer intergroup attitudes in South Sudan. Peace and Conflict: The Journal of Peace Psychology, 25(1), 37-48.

Talking Points 

In the context of recent violence in South Sudan’s civil war:

  • There was evidence of only “mechanistic dehumanization” among Dinka respondents towards Nuer and no evidence of any form of dehumanization among Nuer respondents towards Dinka.
  • Both Dinka and Nuer groups exhibited bias against the other group by assigning more positive characteristics to their own group than to the other group, as well as more emotional capacity overall (both positive and negative) to their own group, while largely denying the other group this same emotional range.
  • Recent violence played a central role in structuring and motivating what bias and dehumanization was found among Nuer and Dinka communities.


Two years after gaining independence from Sudan in 2011, South Sudan became embroiled in a civil war. Violence began in December 2013 after an alleged coup attempt against President Salva Kiir (a member of the Dinka community) on the part of Vice President Riek Machar (a member of the Nuer community), with government forces massacring hundreds of Nuer civilians, followed by revenge killings against Dinka. The violence escalated into a civil war that has ultimately killed anywhere from 50,000 to 400,000 people and displaced 4 million. A peace agreement reached in 2018 is in the process of being implemented, though progress is slow and halting. With a special interest in the relationship between dehumanization and violence, the authors pose two questions in the context of this recent civil war in South Sudan: 1) “what levels of dehumanization exist between the Dinka and Nuer communities,” and 2) “how did the violent events of December 2013 structure and inform the expression of inter-group dehumanization”?

Dehumanization is a widely recognized dimension of intergroup conflict. According to previous research, it can take the form of either animalistic dehumanization or mechanistic dehumanization. Dehumanization is known to interact in particular ways with violence, both as a condition enabling violence against the out-group and, to some extent, as a product of violence. There are few existing studies, however, that empirically examine how “episodes of mass violence may affect dehumanization attitudes” or how dehumanization processes unfold in non-western contexts—two ways in which the authors hope to contribute to the literature on dehumanization.

Animalistic dehumanization

casting an out-group as animal-like, lacking distinctly human traits like self-control, rationality, or moral sensibility, as well as “secondary” emotions that are considered to be uniquely human.

Mechanistic dehumanization

casting an out-group as akin to automatons, lacking traits associated with human nature like warmth or emotional receptiveness.

Source: Loughnan, S., & Haslam, N. (2007). Animals and androids: Implicit associations between social categories and nonhumans. Association for Psychological Science: Research Report, 18(2).

The authors carried out their research in summer 2014, expecting to find high levels of dehumanization between Dinka and Nuer communities. In addition, they expected to hear that the December 2013 violence was important in shaping people’s attitudes towards the other group. To find out, they conducted interviews (47 individuals) and focus groups (18 individuals) with roughly equal numbers of people from the Dinka and Nuer communities in the South Sudan capital of Juba, close to where the December 2013 violence took place. Respondents were all men, ages 22-71, most of whom had directly experienced or witnessed the recent violence. The interviews asked respondents to assign each of 19 attributes and 12 emotions (measuring animalistic and mechanistic dehumanization) exclusively to either their own group or the other group. Responses were then statistically analyzed to see if the particular distribution of attributes and emotions constituted dehumanization of the out-group—or, at the very least, bias against the out-group (demonstrated through the distribution of “positive” versus “negative” attributes and emotions). The focus groups narrowed in on the motivations for and structure of whatever dehumanization may have existed. The researchers asked focus group participants to assign the same 12 emotions to their own group or the other group and then to discuss why they assigned the emotions where they did. The researchers then looked for common themes emerging through these discussions about why participants felt the way they did about members of the other group.

The findings of the study both challenged and confirmed the authors’ preconceptions. Quantitative analysis revealed only mechanistic dehumanization among Dinka respondents towards Nuer and no dehumanization among Nuer respondents towards Dinka. Both groups did, however, exhibit a bias against the other group by assigning more positive characteristics to their own group than to the other group, as well as more emotional capacity overall (both positive and negative) to their own group, while largely denying the other group this same emotional range. The fact that stronger dehumanization between the two groups was not found presents a puzzle in the context of such extreme violence. The authors suggest that one explanation, beyond the possibility that dehumanization may not necessarily accompany violence to the extent expected, might lie in the poor reliability of their measurements for dehumanization—in particular, that indicators developed out of western theoretical constructs may not mean the same thing and therefore may not accurately measure dehumanization in non-western contexts.

The qualitative analysis of the focus group discussions confirmed the hypothesis that recent violence was central in structuring and motivating what bias and dehumanization was found. When participants discussed their perceptions of the other group—and, in particular, the emotions and other characteristics they did or did not attribute to that group—the violent events of December 2013 “were frequently referred to for justification.”

In short, while dehumanization was not as pronounced as expected between Nuer and Dinka in South Sudan, widespread negative intergroup perceptions and biases were present and could largely be explained with reference to recent episodes of violence.

Informing Practice

There are two main take-aways from this research, one substantive and one methodological. First, it is crucial to note the role that violence plays in structuring negative attitudes towards the “other” side in conflicts. Violence is never simply an instrument for pursuing political or security goals. It inevitably feeds negative conflict dynamics through the harm that it causes, solidifying hostile perceptions of the “other” and contributing to the list of grievances motivating either side to continue fighting. In fact, these broader effects of violence largely account for why it does not operate in the straightforward way its proponents might hope it to: rather than simply deterring or cowing the other side into submission and therefore bringing the desired political objective with clean-cut means/end precision, violence instead usually creates the justification necessary for continued fighting, as well as deep-rooted legacies of hate and mistrust that take generations to overcome.

Second, this study draws attention to the implications of a chosen research framework, especially how particular concepts are measured. The findings on dehumanization are only as accurate as the indicators used to measure it are. For instance, if the attributes of being trusting and sociable are considered uniquely human in a western context but other attributes are understood to be uniquely human in other contexts, then using the absence of the former to measure “animalistic dehumanization” in a context where they are not meaningful in the same way will render an inaccurate determination of dehumanization. As the authors note, researchers should proceed by studying the “value structures” of a particular cultural context in order to identify which traits are associated with dehumanization and therefore would more accurately measure its presence or absence in that context. Such studies could entail close readings of local discourses—found in texts like popular stories, songs, films, fables, and so on—to determine the traits associated with humans versus animals or automatons. More fundamentally, the methodological difficulties here should caution researchers—whether scholars or those working with peacebuilding, development, or humanitarian organizations—from simply importing theoretical and methodological models from elsewhere and assuming that the intended meaning will carry over. Although it is only through studying non-western contexts that theoretical frameworks will broaden and change to incorporate a range of human experience, the initial research of these non-western contexts should be more inductive in nature, generating meaning from the ground up, rather than from the top down. Findings from such studies can then not only transform existing theoretical frameworks but also inform work at the community level to rebuild relationships torn apart by violence.

Continued Reading 

Haslam, N. (2006). Dehumanization: An integrative review. Personality and Social Psychology Review, 10(3), 252-264.

Staub, E. (1999). The origins and prevention of genocide, mass killing, and other collective violence. Peace and Conflict: Journal of Peace Psychology, 5(4), 303-336.

International Crisis Group. (2019, November 4). Déjà vu: Preventing another collapse in South Sudan. Retrieved January 9, 2020, from

Keywords: dehumanization, bias, South Sudan, civil war, violence

The following analysis appears in Volume 4, Issue 6 of the Peace Science Digest.