In April 1994, a small radio station in Rwanda began broadcasting a simple message: "Tutsies are cockroaches. The snakes are still in the grass. Cut down the tall trees." The radio station was RTLM — Télévision Libre des Mille Collines. The broadcasts were reaching hundreds of thousands of people. And they were organizing genocide.
Over one hundred days, eight hundred thousand people were murdered. Not soldiers killing soldiers. Civilians killing civilians. Neighbors killing neighbors. A genocide so efficient that at its peak, the killing rate exceeded that of the Holocaust. The perpetrators were not trained assassins or ideological fanatics. They were ordinary people. Farmers. Teachers. Priests. Ordinary Rwandans who, in the span of weeks, became murderers.
The question that haunts the Rwanda genocide is: how did ordinary people commit atrocity? How did a radio station, broadcasting metaphors about insects, turn a nation into a killing field?
The answer lies in understanding how all the mechanisms you have encountered in preceding pages — the minimal group paradigm, dehumanization, embodied cognition, sacred values — align into a single machinery of atrocity. Rwanda was not a unique horror. It was a textbook case of how the behavioral systems that evolved for tribal solidarity and group conflict can be systematically weaponized by a state apparatus.
The Hutu-Tutsi distinction was not ancient. It was a colonial invention. Before Belgian colonization, Hutu and Tutsi were fluid categories — social roles more than ethnic identities. A Hutu could become Tutsi through wealth accumulation. A Tutsi could become Hutu through poverty. The categories had some regional variation and some occupational clustering, but they were not the rigid ethnic divides that characterize modern identity.
Then Belgium arrived. Colonial administrators, needing to organize governance, required clear categories. They chose — or perhaps invented — Hutu and Tutsi as the fundamental distinction. They issued identity cards. They elevated Tutsi to administrative positions, believing them more sophisticated. They froze the categories. Hutu and Tutsi, which had been fluid and overlapping, became rigid and hereditary. The colonial category became identity. The boundary became real.
This is the minimal group paradigm operating at colonial scale. Belgium did not invent ethnic hatred. It drew a line. It made the line salient. It assigned administrative privilege based on the line. And in doing so, it activated the ancient in-group bias systems in the Rwandan population. On one side of the line were "us." On the other side were "them."
For decades, the category remained largely dormant. Hutus and Tutsis intermarried. They lived together. They shared language and culture. The boundary existed as an administrative fact, not a lived division. But the boundary was there, waiting to be activated.
In the late 1980s, Rwanda faced a cascade of crises. Economic collapse. Structural adjustment programs imposed by the IMF had decimated the economy. Coffee prices, Rwanda's primary export, crashed. Unemployment soared. Resources that had been distributed became scarce.
Simultaneously, a civil war erupted. The Rwandan Patriotic Front (RPF), led by Tutsis who had been exiled during earlier violence, invaded from Uganda. The Hutu government responded with military mobilization. Threat was activated. The Hutu population felt endangered by the approaching military force and the internal Tutsi population.
Into this context of economic crisis and military threat, Hutu-dominated political parties and media outlets began a propaganda campaign. The message was simple: Tutsis are the enemy. They are not like us. They are planning to enslave Hutus. They are invading our territory. And the most inflammatory rhetoric: the Tutsis are cockroaches. They are vermin. They are less than human.
This was dehumanization operating at state scale. RTLM was not independent journalism. It was a tool of the Hutu Power political movement. Its broadcasts were calculated propaganda. And they deployed the dehumanization metaphors with precision. Cockroaches. Snakes. Tall trees to be cut down. The language activated the insula's disgust system. It rewired the moral intuitions of the Rwandan population. Tutsis were no longer human. They were insects.
The radio station was continuously on the air. It broadcast for twelve hours daily. It reached people in their homes, in their workplaces, in public gatherings. It broadcast the same metaphors repeatedly. Hour after hour, day after day. The embodied cognition was relentless. The disgust metaphors were hammered into the brain through repetition and pervasive presence.
When the genocide began — triggered by the assassination of President Habyarimana — the propaganda had already done its work. The Hutu population was primed. The boundary was salient. The threat was activated. The out-group was dehumanized. The moral constraints on violence had been loosened.
What happened next was swift and organizational. Roadblocks were set up throughout the country. Young Hutu men, many of them mobilized through local administration and militia structures, were organized into killing squads. They were given machetes, clubs, and spears. They were given orders: kill Tutsis. Kill moderate Hutus who refuse to kill. Clear the roadblocks by killing everyone who tries to pass.
The perpetrators were not fanatics. They were ordinary people — many of them had never committed violence before. But they were embedded in a structure. They had political leadership telling them to kill. They had militia commanders organizing them. They had a population around them killing. They had a radio station broadcasting that they must kill to save their nation. They had dehumanization metaphors that made killing seem necessary and justified.
What emerged was a genocide of stunning efficiency. At peak periods, the killing rate exceeded any other genocide in recorded history. The machinery of death was not top-down coercion alone. It was a complex combination of propaganda, organizational structure, local mobilization, and psychological activation. It was the entire population being directed toward a single goal: the elimination of the Tutsi minority and Hutu opposition.
The violence was not impersonal. It was intimate. Neighbors killed neighbors. In some cases, husbands killed wives who had married out of their group. Teachers killed students. Priests killed refugees sheltering in churches. The intimacy made the killing even more horrific — and it revealed something essential: the dehumanization was not a consequence of distance but a prerequisite for intimacy combined with killing. You must first transform someone from human to non-human before you can kill them knowing their name, knowing their family.
The Rwanda genocide reveals how the mechanisms explored in preceding pages combine into a coherent machinery of atrocity.
The minimal group paradigm activated the boundary. Belgium had drawn the line between Hutu and Tutsi. It had made the distinction salient. And research on the minimal group paradigm shows that any salient boundary automatically activates in-group favoritism and out-group bias. Hutu and Tutsi were "us" and "them" — not because of deep historical hatred but because the boundary had been made administratively and politically salient.
Dehumanization removed the moral constraints. The propaganda campaign — particularly RTLM — deployed metaphors that transformed Tutsis from human to insect. Cockroaches. The insula — the brain region that processes both disgust and moral revulsion — was activated. The normal moral intuitions that prevent harming humans were loosened. If Tutsis were cockroaches, then killing them was not murder. It was pest control. It was hygiene.
Embodied cognition reinforced the dehumanization. The radio broadcasts were constant. They were visceral. They used metaphors of contamination and disgust. They associated Tutsis with parasites and disease. The brain's sensory systems — primed to respond to actual disgust — were activated by the metaphorical disgust language. What began as abstract propaganda became embodied in the nervous system. Tutsis did not just seem disgusting intellectually — they felt disgusting. The insula was ablaze.
Threat activation made violence seem necessary. The civil war provided real threat. The Hutu government was genuinely under military pressure from the RPF. This activated the amygdala and threat-response systems. The population was afraid. And fear makes people prone to accept violence as self-defense. When you are afraid, you accept narratives about preemption and necessity that you would reject if you were safe.
In-group solidarity made participation expected. Because the boundary had been made salient and the threat was real (or presented as real), in-group solidarity became powerful. Hutus felt they were defending their group against Tutsi invasion. Those who refused to kill were often killed themselves or shunned. Participation was not merely acceptable — it was required. The in-group bias system made participation feel morally mandatory.
Organizational structure diffused responsibility. The genocide was not carried out by lone killers. It was organized through local administration, militia structures, and military command. Each perpetrator had a narrow role. A roadblock operator did not decide who would be killed — he followed orders. A militia member did not decide the genocide's goals — he was told to kill. A radio broadcaster did not decide the propaganda strategy — he read scripts. The diffusion of responsibility across so many agents meant that no individual felt the full weight of participation in systematic murder.
One of the most chilling aspects of the Rwanda genocide was its apparent inexorability. Once it began, it continued. Cease-fires were violated. Negotiations failed. The killing accelerated. Why?
Because dehumanization is irreversible during conflict. Once Tutsis had been transformed into cockroaches, once the insula was ablaze with disgust, it became nearly impossible for the perpetrators to see their victims as human again. And once violence had been committed — once ordinary Hutus had killed their neighbors — the psychological need to justify that violence actually intensified the killing.
Research on cognitive dissonance shows that when people commit atrocities, they often increase their justifications and intensify their hatreds afterward. The killing created a need to believe the killing was justified. The dehumanization metaphors, already in place, provided that justification. But they had to be intensified. If you have killed someone, you must believe they deserved to be killed. You must believe they were not fully human. You must believe the killing was necessary. The propaganda intensified. The metaphors became more visceral. The killing accelerated.
Attempts at negotiation and cease-fires failed because they required reversing the dehumanization. They required seeing the out-group as human again. But at that point, the machinery of atrocity had created psychological and social conditions that made seeing the other as human nearly impossible. The perpetrators had invested themselves in the killing. Their identity had become defined by participation. Their in-group solidarity had been built on shared violence. Stopping would have meant admitting that the dehumanization was false, that the killing was murder, that they themselves were murderers.
The genocide continued not because perpetrators were inherently evil but because the machinery of atrocity had created conditions from which escape became psychologically impossible.
Colonial Invention vs. Ethnic Identity: Hutu-Tutsi were colonial constructs, not ancient ethnic divisions. Yet by 1994, they felt like ancient, real identities to the Rwandan population. The tension reveals that constructed identities can become psychologically and socially real through generations of reification. The distinction between "real" and "invented" identity may be meaningless — all identities are socially constructed; what matters is how salient and binding they become.
Individual Perpetrators vs. Structural Machinery: The genocide was carried out by hundreds of thousands of ordinary people, many of whom killed neighbors they had lived with for years. Yet the genocide was also organized through state institutions, militia structures, and propaganda. The tension reveals that the dichotomy between "individual evil" and "structural evil" is false. Individuals commit atrocity, but they do so within structures that make atrocity easier, more normalized, and more diffused in responsibility.
Dehumanization as Cause vs. Symptom: Did dehumanization propaganda cause the genocide, or did it emerge from pre-existing conflict and threat? The temporal ordering suggests propaganda came first — RTLM broadcast dehumanization metaphors before widespread killing began. But dehumanization may also have been a symptom of earlier, lower-intensity conflict. The tension reveals that causation in atrocity is circular — propaganda activates threat response, threat response makes dehumanization seem plausible, dehumanization removes moral constraints, violence validates the propaganda's claims.
The Rwanda genocide reveals how all the mechanisms of behavioral-mechanics — minimal group paradigm, dehumanization, embodied cognition, sacred values, in-group bias — align into a coherent machinery of atrocity.
Behaviorally, each mechanism operates according to specific rules. The minimal group paradigm shows that any salient boundary activates in-group bias automatically. Dehumanization shows that metaphors can remove moral constraints on violence. Embodied cognition shows that visceral imagery and metaphors can rewire moral judgment at the nervous-system level. In-group bias shows that solidarity with one's group makes violence toward out-groups seem justified.
Historically, Rwanda demonstrates how a state apparatus can activate all these mechanisms simultaneously and systematically. Belgium created the boundary through colonial administration. The Hutu government made the boundary salient through political identity. The radio station deployed dehumanization metaphors at massive scale. Militia organizations provided the structural machinery for carrying out violence. Each mechanism reinforced the others.
The insight from the cross-domain connection is that genocide does not emerge from any single cause. It is not propaganda alone that causes genocide — propaganda activates pre-existing mechanisms. It is not threat alone — threat activates defensive violence but not necessarily genocide. It is not dehumanization alone — metaphor can be ignored or resisted. It is the alignment of multiple mechanisms, each activated in context, that creates genocide.
Understanding this has implications for prevention. If genocide requires multiple mechanisms to align, then disrupting any single mechanism might prevent the outcome. If you can keep the boundary non-salient, the minimal group bias may not activate. If you can prevent dehumanization propaganda, the moral constraints might remain in place. If you can ensure that threat is not politically manipulated, fear-based violence may be contained. The mechanisms are not independent — they reinforce each other — but they are not inevitable.
Psychologically, the threat-response systems (amygdala, HPA axis) and identity systems (default mode network, self-concept) interact under specific conditions to create willingness to commit violence. When a person perceives their in-group under existential threat, and when out-group identity has been made salient through propaganda and dehumanization, the normal moral constraints against harming humans can be loosened.
Historically, Rwanda demonstrates this mechanism operating at population scale. The Hutu population had real military threat (the RPF invasion) and perceived existential threat (propaganda claims that Tutsis intended to enslave Hutus). The in-group identity had been made salient through colonial and post-colonial political emphasis. The out-group had been dehumanized through propaganda. Under these conditions, ordinary people became perpetrators.
The psychological insight is that humans are not inherently genocidal but conditionally genocidal under specific threat and identity conditions. The historical insight is that states can manufacture these conditions through propaganda and political messaging. The tension between domains reveals that individual psychology and historical circumstance are not separate. History operates through psychology. Atrocity emerges when individual threat-response systems are activated at population scale through institutional and propaganda machinery.
Examining Rwanda through both behavioral-mechanics and psychological lenses reveals a complete pathway from the minimal group paradigm to genocide:
Boundary creation and salience (behavioral-mechanics): Belgium created the Hutu-Tutsi distinction through colonial administration. The boundary existed but was dormant for decades.
Threat activation (psychology): Economic crisis and civil war created real threat. The population experienced genuine danger and resource scarcity.
Political amplification of threat (behavioral-mechanics + psychology): Political leadership amplified threat rhetoric, claiming that Tutsis intended genocide against Hutus. The propaganda weaponized pre-existing threat.
Dehumanization metaphors (behavioral-mechanics): RTLM broadcast dehumanization metaphors — cockroaches, snakes, contamination. The insula's disgust system was activated at population scale.
Embodied reinforcement (behavioral-mechanics): The radio broadcasts were constant and visceral, reinforcing embodied activation of disgust and fear.
Institutional mobilization (behavioral-mechanics): Militia structures, roadblocks, and military organization provided the machinery for carrying out killing at scale.
Diffusion of responsibility (psychology + behavioral-mechanics): The organizational structure meant no individual felt fully responsible. Killing became a role, not a conscious moral choice.
In-group solidarity (behavioral-mechanics + psychology): Participation became expected. Refusing to kill meant being killed or shunned. In-group identity was reinforced through shared violence.
Irreversibility (psychology): Once people had killed, cognitive dissonance required intensifying the dehumanization to justify the killing. The propaganda intensified. The violence escalated.
The complete pathway shows that genocide is not random or unexplainable. It is the outcome of specific mechanisms activating in a particular sequence, under particular conditions, within particular institutional structures. This is not reassuring — it means genocide is possible anywhere these conditions align. But it is informative: if we understand the mechanism, we can potentially interrupt it at any stage.