The thrill of being violent as an antidote to posttraumatic stress disorder in Rwandese genocide perpetrators

Abstract

Background: The cumulative exposure to life-threatening events increases the risk for posttraumatic stress disorder (PTSD). However, over the course of evolutionary adaptation, intra-species killing may have also evolved as an inborn strategy leading to greater reproductive success. Assuming that homicide has evolved as a profitable strategy in humans, a protective mechanism must prevent the perpetrator from getting traumatised by self-initiated violent acts.

Objective: We thus postulate an inverse relation between a person's propensity toward violence and PTSD.

Method: We surveyed a sample of 269 Rwandan prisoners who were accused or convicted for crimes related to the 1994 genocide. In structured interviews we assessed traumatic event types, types of crimes committed, the person's appetitive violence experience with the Appetitive Aggression Scale (AAS) and PTSD symptom severity with the PSS-I.

Results: Using path-analysis, we found a dose-response effect between the exposure to traumatic events and the PTSD symptom severity (PSS-I). Moreover, participants who had reported that they committed more types of crimes demonstrated a higher AAS score. In turn, higher AAS scores predicted lower PTSD symptom severity scores.

Conclusions: This study provides first empirical support that the victim's struggling can be an essential rewarding cue for perpetrators. The results also suggest that an appetitive aggression can inhibit PTSD and trauma-related symptoms in perpetrators and prevent perpetrators from getting traumatised by their own atrocities.

Keywords: aggression; genocide; perpetrator; posttraumatic stress disorder.

Fig. 1

Path model of the relationship among the number of self-experienced as well as witnessed event types, number of different types of crimes, AAS sum score and PSS-I sum-score. Standardized regression weights are shown. Asterisks indicate statistical significance of the regression weights *p<.05, **p<.01, ***p<.001.