The Neuroscience Of Dehumanization — What Happens In The Brain

To understand what dehumanization does in the brain, you first need the baseline — what happens when we perceive other people normally.

Human beings have an extraordinary cognitive capacity that took millions of years to develop: Theory of Mind. The ability to attribute mental states — beliefs, desires, intentions, emotions — to other beings and to use those attributions to predict and interpret behavior. This capacity distinguishes humans from virtually all other species in its sophistication, and it is the computational foundation of everything that makes human social life possible: language, cooperation, culture, morality, politics, love.

The neural architecture of Theory of Mind is not singular but distributed across several regions that together form what researchers call the "social brain network" or "mentalizing network." Key nodes include:

The medial prefrontal cortex (mPFC), particularly its ventromedial subdivision, which is involved in thinking about the mental states of ourselves and close others. The temporoparietal junction (TPJ), at the intersection of temporal and parietal cortex, which is critical for taking the perspective of another person — for the cognitive act of shifting out of your own viewpoint into someone else's. The superior temporal sulcus (STS), involved in the perception of biological motion and the interpretation of social signals. The anterior cingulate cortex (ACC), which plays a role in empathy — specifically in the affective component, the felt response to another person's pain or distress.

When these systems are active — when you are genuinely engaged with another person as a minded being — a specific and remarkable thing occurs: you automatically track their mental states. You model their beliefs. You predict their intentions. You feel something when they suffer. This happens fast, below conscious deliberation, as a default response to perceiving another person.

This automatic extension of mental-state attribution to perceived persons is the neural foundation of basic human morality. You cannot easily harm what you automatically perceive as a mind. The perception of mind generates moral consideration. This is not a philosophical inference that human minds arrive at by deliberation — it is a rapid, automatic product of the social cognition systems. Damage to the mPFC or TPJ reliably produces deficits in moral judgment. The neural substrate for empathy and the neural substrate for moral concern are not separate systems that can be independently manipulated — they share hardware.

Lasana Harris and Susan Fiske's landmark 2006 paper in Psychological Science introduced the concept of "neural dehumanization" and produced findings that should be among the most discussed scientific results of the 21st century.

Harris and Fiske showed study participants photographs of people from different social groups while scanning their brains with fMRI. The social groups were selected using what they called the Stereotype Content Model — a framework that categorizes social groups along two dimensions: warmth (perceived friendliness and benevolence) and competence (perceived capability and status). Most groups, whether liked or disliked, occupy either high-warmth/high-competence, high-warmth/low-competence, or low-warmth/high-competence positions — and for all of these, the mPFC activated normally when participants viewed group members.

The exception was the low-warmth/low-competence quadrant: groups perceived as both cold and incompetent. In American society at the time of the study, the groups that were consistently placed in this quadrant were homeless people and drug addicts. When participants viewed photographs of these groups, the mPFC — the region associated with social cognition, mental-state attribution, and considering the minds of others — failed to activate at the levels seen for other groups. Instead, the pattern of activation resembled the pattern produced by viewing objects rather than people.

This was the neural signature of dehumanization: not hatred, not active hostility, but the failure of the system that generates humanization in the first place. The participants were not choosing not to empathize with homeless people. Their brains were simply not generating the social cognition that normally precedes empathy, because the social categories assigned to these groups had suppressed the automatic humanization response.

Subsequent research has extended and refined these findings across multiple dimensions. Nick Haslam's theoretical framework distinguishes two forms of dehumanization: "animalistic dehumanization," in which the target group is denied uniquely human traits (civility, morality, refinement) and implicitly compared to animals; and "mechanistic dehumanization," in which the target is denied human nature traits (warmth, emotion, openness, agency) and implicitly compared to objects or machines. Both forms produce measurable changes in how the brain processes targets, but through somewhat different mechanisms.

Nour Kteily and colleagues at Northwestern developed the "Blatant Dehumanization Scale" — a direct measure in which participants rate how "evolved" various human groups are on a scale that runs from primitive hominid to modern human. Results using this scale have been disturbing: significant percentages of participants in politically polarized groups in the United States are willing to explicitly rate outgroup members as less than fully human on this scale. And blatant dehumanization scores — not subtle, unconscious bias scores, but explicit, blatant dehumanization — predict support for aggressive policies against those groups, willingness to endorse violence against them, and reduced moral concern for their welfare.

The mechanism connecting neural dehumanization to behavioral outcome is the suppression of what Jean Decety and colleagues have called "empathic concern" — the affective, motivational component of empathy that actually drives helping behavior and moral action. You can understand, cognitively, that someone is suffering without being motivated to do anything about it. What drives the motivation is the felt response — the emotional resonance with the other person's pain. And that felt response depends on the brain processing the other person as a being like oneself — a minded, suffering creature whose inner life registers as real.

When the mPFC goes quiet and social cognition fails, the empathic concern circuit has nothing to work with. The suffering is visible but does not register as meaningful. And in that gap between visible suffering and absent meaning, the conditions for atrocity are established.

Individual dehumanization — the spontaneous failure to extend full humanity to members of certain social groups — is disturbing enough. But individual dehumanization is maintained and amplified by social architectures specifically designed to prevent its correction.

Philip Zimbardo, in his analysis of the conditions that produce atrocity among ordinary people, identifies situational forces as at least as important as dispositional ones. Most accounts of historical atrocity are what Zimbardo calls "dispositional" — they focus on the specific evil of specific individuals, the madness of individual perpetrators, the uniqueness of particular circumstances. Dispositional accounts are reassuring because they suggest that atrocity is an aberration, the product of bad people who most of us are not like.

The situational analysis is less reassuring but more accurate: the conditions that produce ordinary people committing atrocities are more specific than rare. They include the deindividuation of both perpetrators and victims (uniforms, numbers, categories replacing names), the diffusion of responsibility across chains of command, the gradual escalation of harm that makes each step seem small even as the cumulative trajectory is enormous, the social proof that atrocity is normal (everyone else is doing it), and — critically — the structural prevention of humanizing contact between perpetrators and victims.

The prevention of humanizing contact is the key architectural feature. Every sustained system of mass harm has included mechanisms to prevent the people inflicting harm from having genuine encounters with the people experiencing it. The plantation system included prohibitions on certain forms of social contact between enslaved and enslaving populations. Nazi concentration camp administrators were deliberately rotated and insulated from extended contact with individual prisoners, because contact produced the humanization that disrupted the operational logic of the system. Modern supply chains are engineered to ensure that the people consuming products made under exploitative conditions never encounter the people making them — not because of a conscious conspiracy to maintain dehumanization, but because the opacity is structurally convenient.

This is the social architecture of maintained dehumanization: the engineering of conditions in which the humanizing information — the names, the faces, the specific grief and specific hope of specific people — never reaches the people whose behavior toward those people matters.

David Livingstone Smith, in Less Than Human, provides the most comprehensive historical and philosophical account of the relationship between dehumanization and violence. Smith's central argument: dehumanization is not a byproduct of violence; it is a prerequisite. The moral architecture of normal human psychology makes direct harm to perceived humans difficult and psychologically costly. Dehumanization is the psychological technology that removes that difficulty — that makes it possible to harm without the harm registering as harm to the human brain doing the harming.

Smith documents the consistent rhetorical and symbolic patterns by which populations have been prepared for mass violence: the prior use of animal metaphors (rats, cockroaches, vermin, beasts) that shift the neural categorization of the target group from person to something less than person; the physical separation and distinctive marking that reinforces categorical difference; the propaganda that attributes subhuman characteristics to the target; the suppression of humanizing information about individual members of the target group. These patterns appear across cultures and centuries with such consistency that Smith argues they reflect something about the deep structure of human moral psychology — specifically, the neural requirement for dehumanization as a prerequisite for sustained violence.

The most important question the neuroscience raises is also the most hopeful one: is the neural dehumanization response reversible? Can the mPFC be switched back on?

The answer, based on the research, is yes — and the mechanism is simpler and more available than most people expect.

Harris and Fiske's follow-up research found that the neural dehumanization of the low-status/low-warmth groups could be reversed by a simple contextual manipulation: asking participants to imagine what kind of vegetable the person in the photograph might like. This apparently trivial instruction — which directed participants to mentally attribute a preference to the person, to imagine them as a subject with wants — reliably activated the mPFC and restored the social cognition response. The person was registered as a mind again. The dehumanization was neurally reversible through a single act of mind-attribution.

This finding is more significant than it appears. It means that the neural basis of dehumanization is not a hard categorical distinction between human and not-human, but a failure of active mental simulation — a failure that can be corrected by the act of imagining the other person's inner life. The circuit is not broken; it is simply not being engaged. Engage it with any genuine act of imaginative perspective-taking and it fires normally.

Adam Galinsky's research on perspective-taking extends this finding into more complex social contexts. Galinsky has consistently found that perspective-taking — the deliberate cognitive act of imagining the situation and experience of another person — reduces stereotyping, reduces discrimination, increases cooperation across group lines, and in contexts of negotiation and conflict, produces better outcomes for both parties than either self-advocacy or compromise strategies alone.

The mechanism appears to work through the same neural pathway: deliberately activating the mentalizing network through perspective-taking restores the individuation of the other — makes their specific experience available — and this individuation is what blocks the categorical, dehumanized response.

Contact theory research, as previously discussed, provides the behavioral complement to these neural findings. The conditions under which intergroup contact reduces prejudice — equal status, common goals, cooperation, institutional support — are also the conditions that maximize the quality and depth of interpersonal engagement, which means the conditions that maximize the likelihood of genuine perspective-taking and mutual mind-attribution. The structural conditions for good contact are the structural conditions for sustained neurological rehumanization.

Here is what the neuroscience ultimately establishes: "We Are Human" is not an ethical claim requiring philosophical justification. It is a description of a neural capability that all normal human brains possess and that social architecture can suppress or support.

Every human brain with intact social cognition systems has the capacity for full humanization of every other human being it encounters. The mPFC can fire. The TPJ can take the perspective. The ACC can resonate with the suffering. This is not a gift some people have and others don't — it is a standard feature of the human brain, present in virtually everyone.

What varies is not the capacity but the activation. And the activation is shaped by:

Language — the words used to describe groups of people prime specific neural responses. Dehumanizing language ("illegals," "animals," "vermin") primes the non-social cognition response. Humanizing language — proper names, specific attributes, personal histories — primes the mPFC response.

Contact — genuine encounter with specific individuals from outgroups activates individuation processes that override categorical dehumanization. Not all contact is equally effective; shallow, status-unequal, competitive contact can reinforce stereotypes. Deep contact, which involves genuine mutual disclosure and collaboration, reliably activates the humanizing response.

Narrative — research on narrative transportation, the experience of being immersed in a story, shows that fiction produces genuine empathic responses and perspective-taking that can reduce prejudice and increase prosocial behavior. This works because the brain simulates the experiences of characters using the same neural systems it uses to understand real people. Reading a novel that gives you access to the inner life of a person from a despised outgroup is neurologically comparable to actually meeting that person. Paul Bloom's critique of empathy as a moral guide — that it is biased toward the salient, the similar, and the nearby — is specifically answered by narrative, which extends the reach of the empathic simulation system to people who are not salient, not similar, and not nearby.

Structural opportunity — the social design of institutions, cities, schools, and organizations either creates or prevents the conditions for genuine encounter. Segregation — whether formal or informal, by race, class, nationality, or any other marker — maintains dehumanization by preventing the contact that would correct it. Integration creates the structural opportunity for humanization. The research is consistent that where integration is genuine rather than superficial — where different groups work together toward common goals under conditions of rough equality — humanization follows.

What would it take for the neural dehumanization response to become the exception rather than the rule? It would require a social architecture that, at every level, favored humanizing contact over segregating distance. Schools that mixed across every available demographic line. Media that systematically provided access to the inner lives of people from every social position. Cities designed for encounter rather than avoidance. Institutions that named and individuated rather than categorized and averaged. Language that preserved specificity rather than collapsing into slur.

None of this is impossible. Some of it is already being done in places. The barrier is not human nature — human nature, at the neural level, runs toward humanization by default. The barrier is the social architecture that suppresses the default.

That architecture can be changed. And if it were changed universally — if every person on this planet lived in conditions where the humanizing information about every other person was structurally available and the neural dehumanization response was consistently corrected by genuine encounter — the conditions for mass violence and mass indifference would dissolve. Not because humans became different. Because they were finally allowed to operate as themselves.

The Name Practice: When you encounter a person in a role that makes them easy to categorize rather than see — a service worker, a homeless person, a person behind a desk — find out or use their name. Say it. The act of using someone's name is a micro-intervention in the dehumanization process; it activates individuation, which is the first step in the humanizing response.

The Inner Life Question: For any group you notice yourself treating categorically — thinking about in terms of what "they" are like rather than what specific individuals are like — practice the Harris-Fiske manipulation deliberately. Pick a single person from that group (in your mind, or in your actual life) and ask: what do they care about? What's hard for them right now? What do they want that they might not get? This is not sentimentality — it is the specific cognitive act that switches the mPFC back on.

The Dehumanizing Language Audit: Pay attention, for one week, to the language you use and consume around groups of people. Specifically: language that replaces people with categories ("the homeless," "illegals," "those people"), language that attributes non-human characteristics to groups (animal metaphors, machine metaphors), language that strips individual variation from groups. This language is everywhere. Simply noticing it is the beginning of understanding how the neural ground is being prepared.

The Counter-Narrative Seek: The research on narrative and perspective-taking is clear: reading, watching, and listening to first-person accounts from people whose humanity you haven't fully extended is neurologically comparable to meeting them. Find one book, documentary, or long-form piece this month that gives you genuine access to the inner life of a person from a group you haven't thought much about as individuals. Go deep enough for the perspective to land. Notice what it does.

The Encounter Design: Look at your actual daily life and map who you genuinely encounter versus who you see but don't meet. Most of us live in conditions of de facto segregation — by class, by neighborhood, by professional world — that minimize genuine cross-group encounter. This is not usually a personal moral failure; it is structural. But it has personal neural consequences. Find one change you can make to your actual daily context that creates more genuine encounter with people whose life experience differs substantially from yours.

The neuroscience of dehumanization does not produce despair if you follow it far enough. It produces something more useful: a precise account of the mechanism that allows ordinary humans to participate in extraordinary harm, alongside an equally precise account of the mechanism that corrects it. The correcting mechanism is not complicated. It is the thing your brain already knows how to do, when it isn't being prevented: see the person in front of you as real. Extend the same cognitive courtesy you extend to your friends — the assumption that they have a rich inner life that matters — to everyone else.

Your brain can do this. It's built for it. The question is what you're going to do with that fact.