When examined across cases of mass violence — genocide, slavery, war crimes, systematic oppression — a consistent pattern emerges. Dehumanization always precedes large-scale atrocity. It is not that atrocity produces dehumanization as a post-hoc justification (though that happens too). It is that dehumanization comes first, and atrocity follows.
This temporal ordering reveals something crucial about how human violence operates at scale. You cannot commit atrocity against a population you regard as fully human. The moral constraints are too strong. Something must first transform the out-group from human to non-human. Only then does systematic killing become possible.
The dehumanization is not incidental to atrocity. It is prerequisite. And understanding this ordering has implications for prevention: if dehumanization is the first step in the pathway to genocide, then preventing dehumanization is the first line of defense.
Rwanda, 1994: RTLM radio broadcasts began their dehumanization campaigns in 1991, calling Tutsis "cockroaches" and "snakes." The genocide did not begin until April 1994 — nearly three years later. The propaganda came first. The killing followed.1
Nazi Germany: The Nazi propaganda campaign against Jews began in the early 1930s, with newspapers, films, and political rhetoric depicting Jews as parasites, vermin, and subhuman. The killing did not reach industrial scale until 1941-42, after years of dehumanization propaganda. Hitler's early writings (Mein Kampf, 1925) already contained dehumanization metaphors. The Holocaust was the implementation of rhetoric that had been repeated for over a decade.
Slavery: The trans-Atlantic slave trade was justified through ideologies that depicted Africans as fundamentally different from Europeans — less intelligent, less moral, less fully human. These dehumanizing ideologies emerged and solidified in the 1600s-1700s, before slavery reached its industrial scale. The dehumanization provided the moral justification for the institution. The institution did not create the dehumanization; it followed it.
American genocide against Native Americans: Depictions of Native Americans as "savages," as "wild," as "uncivilized" and thus less than fully human, preceded European extermination campaigns. The rhetoric existed before the violence was systematized. The metaphors dehumanized the out-group, and then the violence became possible.
In each case, the temporal ordering is clear: propaganda and dehumanization metaphors come first. Atrocity comes second. The metaphor changes the moral logic. It opens the door. Then violence steps through.
The reason dehumanization must precede atrocity is neurobiological. The human brain has empathy systems (anterior cingulate, insula, ventromedial prefrontal cortex) that produce visceral opposition to harming other humans. When you see a human face in pain, these systems activate. They create an aversive response — a motivation to prevent harm, not cause it.
These empathy systems are ancient and powerful. They are present in all humans (except in cases of severe pathology or brain damage). You cannot consciously override them through reason alone. You can manage them through cognitive effort, but that effort is costly and depletes other cognitive resources.
The only way to overcome empathy systems is to change the input. If the brain does not process the target as human, the empathy systems do not activate. Dehumanization metaphors change the input. They make the brain process the out-group member as non-human — as insect, animal, or disease. The insula, which processes both disgust at contamination and moral revulsion at harm, gets activated by the metaphorical "disgusting" properties. Moral constraints weaken.
Once the dehumanization is established and habituated (through repetition in propaganda), the brain can process violence against the out-group as justified. Not because the person has no empathy but because the dehumanization has redirected empathy away from the target. The target is no longer seen as human, so empathy does not apply.
This is why dehumanization is so powerful and so necessary for atrocity. Without it, the human brain resists killing at scale. With it, killing becomes possible.
Dehumanization does not emerge spontaneously. It is constructed and distributed through propaganda. A government or political movement develops rhetoric that presents an out-group as non-human. That rhetoric is broadcast — through media, education, religious institutions, political speeches, art, entertainment. The metaphors become ubiquitous. Everyone hears them repeatedly. The dehumanization becomes normalized.
This ubiquity is crucial. If dehumanization only existed in fringe rhetoric, it would not reshape the population's moral intuitions. But when dehumanization becomes mainstream — when state media broadcasts it, when schools teach it, when religious leaders preach it — it rewires the brain's processing of the out-group. The metaphor becomes habitual. The disgusting properties become "real" through repetition and official sanction.
Rwanda's RTLM broadcasts illustrated this mechanism. The radio was constantly on. It broadcast for twelve hours daily. It reached people in their homes and workplaces. The dehumanization metaphors — cockroaches, snakes, tall trees to cut down — were hammered in repeatedly. The embodied cognition was relentless. By the time the genocide began, the dehumanization was thorough.
If dehumanization is prerequisite to atrocity, then preventing dehumanization is the first line of defense against genocide. This suggests a prevention strategy that differs from conventional approaches.
Conventional approaches focus on preventing conflict or managing conflict through negotiation and institutional design. These are valuable but limited. If dehumanization is already established, institutional design and negotiation cannot prevent genocide. Rwanda had institutions, legal frameworks, and even a multiparty system. None of these prevented genocide because the dehumanization was already deeply embedded in the population.
A prevention strategy focused on dehumanization would monitor political rhetoric and propaganda. It would identify when metaphors of contamination, disease, or non-humanity are being applied to an out-group. It would create counter-narratives that assert the target group's full humanity and shared identity with the broader population. It would ensure that educational and media systems resist dehumanization metaphors.
This is difficult politically. Politicians often use dehumanizing rhetoric to mobilize their base. They depict immigrants as invading hordes or disease. They depict religious minorities as contaminating influences. They depict political enemies as subhuman. This rhetoric is emotionally powerful. It activates threat response and in-group solidarity. Countering it requires sustained effort and political will.
But if the temporal ordering is correct — if dehumanization precedes atrocity — then preventing dehumanization is the most leverage point for preventing genocide. It is earlier in the causal chain than military intervention or humanitarian response. Intervening early, before dehumanization reaches critical mass, is more effective than intervening after atrocity begins.
One open question is whether dehumanization, once established, can be reversed. The case studies of peace-making suggest it can be — symbolic recognition of the other's humanity can reset the dehumanization. But peace-making typically occurs after violence has ended. Can dehumanization be reversed during an ongoing genocide? Or once the killing has begun, does the dehumanization become irreversible?
Historical evidence suggests that during active genocide, dehumanization becomes more entrenched, not less. The perpetrators' need to justify their actions intensifies the dehumanization. Attempts at cease-fires and negotiations fail because they require accepting the out-group's humanity — something the dehumanization explicitly rejects.
This implies that prevention must come before the genocide begins. Once dehumanization reaches critical mass and violence begins, reversing it becomes much harder. The window for prevention is narrow: before dehumanization becomes ubiquitous and before violence begins.
Dehumanization as Cause vs. Enabling Condition: Does dehumanization cause atrocity, or does it merely enable atrocity that would happen anyway for other reasons (resource competition, political struggle, threat)? The temporal ordering suggests causation, but causation in human affairs is rarely simple. Dehumanization may be necessary but not sufficient — additional conditions (threat, institutional capacity, leadership willing to deploy violence) may be required. The tension reveals that atrocity emerges from multiple causes aligning, not from any single cause.
Prevention vs. Inevitability: If dehumanization is preventable through counter-narratives and media management, are genocides preventable? Or once certain threshold conditions are met (threat activation, resource scarcity, political instability), is genocide inevitable regardless of prevention efforts? The tension reveals that genocide prevention requires sustained effort, and failure at any single intervention point may render prevention impossible.
Individual Responsibility vs. Systemic Dehumanization: If dehumanization is ubiquitous and normalized through propaganda, are individual perpetrators responsible for atrocity? Or is responsibility primarily systemic — located in the political leadership and propaganda apparatus that created the dehumanization? The tension reveals that responsibility operates at multiple levels. Individual perpetrators make choices, but those choices are made within systems that have already redefined moral constraints.
Behaviorally, dehumanization operates through specific mechanisms: metaphor confuses the insula's processing of physical disgust with metaphorical disgust at humans; repeated exposure habituates the brain to the metaphor; ubiquity through propaganda ensures that the dehumanization becomes a shared cultural norm.
Historically, examining cases of atrocity reveals that dehumanization is not an incidental feature but a systemic prerequisite. Atrocity requires removing moral constraints against harming humans. Dehumanization removes those constraints by redefining the target as non-human. Every major case of genocide shows dehumanization preceding atrocity.
The behavioral-mechanics insight is that dehumanization is not magical or mysterious. It is a specific rewiring of the brain's processing systems through propaganda and metaphor. It can be systematically created through deliberate rhetoric and media management. Understanding this reveals that dehumanization is not an inevitable human failing but a specific technique that can be deployed strategically.
The tension between domains reveals something critical: atrocity is not random or inevitable. It requires specific prerequisites to be put in place. If we understand those prerequisites (salient boundaries, threat activation, dehumanization propaganda), we can potentially prevent atrocity by disrupting any of them. Preventing dehumanization is one lever. But it is not the only lever. Preventing threat activation, or preventing the weaponization of boundaries, or preventing access to organizational machinery are also levers for prevention.
Psychologically, the brain's moral judgment systems (ventromedial prefrontal cortex, anterior cingulate) are highly sensitive to how out-group members are represented. If they are presented as fully human, empathy systems activate. If they are presented as non-human, empathy systems are bypassed.
Historically, propaganda campaigns reveal how systematically this reorganization can be accomplished. Nazi films showed Jews as parasites. Rwandan radio showed Tutsis as cockroaches. American ideology showed Native Americans as savage animals. In each case, propaganda reorganized the population's perception of the out-group before violence was deployed.
The psychological insight is that moral judgment is not fixed. It can be reorganized through perception and rhetoric. The historical insight is that this reorganization does not happen accidentally. It is deliberate strategy. Propaganda is a technology for rewriting moral judgment at population scale.
The tension between domains reveals that understanding propaganda is crucial for understanding atrocity prevention. If propaganda reorganizes moral judgment, then counter-propaganda and alternative narratives that assert the out-group's humanity can maintain moral constraints against atrocity. The brain can be reorganized toward dehumanization through propaganda. It can also be reorganized toward humanization through counter-propaganda.
The most consequential insight emerges when examining the complete pathway from dehumanization rhetoric to systematic atrocity.
A political leadership develops dehumanization metaphors that depict an out-group as non-human. The metaphors are distributed through propaganda (media, education, religious institutions). The brain's insula processes the metaphorical disgust and habituates to it through repetition. The moral constraints against harming the out-group are weakened.
Simultaneously, threat is activated — through actual conflict, economic crisis, or political rhetoric claiming existential danger. The population's threat-response systems are activated. They become more prone to accept violence as self-defense.
An organizational machinery is put in place — militia structures, military hierarchies, bureaucratic systems. The machinery provides the capacity to kill at scale and the diffusion of responsibility that makes participation possible.
Dehumanization continues through propaganda and media. The rhetoric intensifies as violence begins. The perpetrators need to justify their actions, so the dehumanization intensifies.
The complete pathway shows that atrocity is not random. It emerges from specific mechanisms activating in sequence. If we understand the sequence, we can potentially intervene at any stage. Prevent dehumanization, and the pathway is disrupted. Prevent threat activation, and the motivation for violence is reduced. Prevent the development of organizational machinery, and the capacity for systematic killing is limited.