The Category as Load-Bearing Structure
On the Psychological Conditions That Make Mass Subjugation Administratively Possible
Argument in Brief
Mass subjugation, whatever its scale, tends to follow the same underlying sequence. A population is first sorted into a category rather than seen as individuals; that category then numbs the surrounding society against the moral weight of what follows, through bureaucratic distance, administrative scale, or the deliberate suppression of prior relationship; participants and bystanders produce justification sentences that function as alibis rather than arguments, since they require no knowledge of the specific person involved; and the categorization apparatus, once built, persists as standing capacity available for reuse against the same population or another. This essay traces that sequence across Salem, Japanese American incarceration, the Holocaust, the Bosnian genocide, and the persecution of the Karen and Rohingya in Myanmar, before examining specific practices within current U.S. immigration enforcement as a live instance of the same mechanism. The claim throughout is one of recurring structure, not equivalent magnitude: these cases differ enormously in scale and consequence, and the essay is explicit about that difference at every point where the two could be confused.
A Note on Method
This essay examines a psychological mechanism, not a hierarchy of historical crimes. The cases it draws on differ enormously in scale and consequence: some describe the deliberate murder of millions, others the forced removal of a population, others the separation of children from parents in government custody. They are not equivalent, and nothing here should be read as claiming they are. What recurs across them is not magnitude but structure, the conditions under which a society can act against a categorized population, or permit that action, without most participants experiencing themselves as doing something monstrous. Naming that structure does not level the outcomes. It identifies what made each of them, at whatever scale it reached, sustainable. This mechanism operates independently of, but alongside, the structural account of individual psychology developed in Psychological Architecture.
Three further boundaries. This is not an essay about racism, about where prejudice originates or how in-group hostility forms; those are real questions, but different ones. It asks instead what has to happen for a society to act on hostility, fear, or expedience at scale once those already exist. It is not an essay about immigration law or legal status; where current enforcement practices appear, they appear as evidence of what happens to people once they have been sorted into a category, not as an argument about who should be permitted to enter or remain. And it treats the settled and the ongoing differently: most of what follows is historically adjudicated, but one case is live and contested, and I have tried to mark that difference wherever it matters rather than smuggling a verdict in under the cover of history.
The Categorization Moment
Subjugation at scale requires a precondition that is easy to overlook because it happens quickly and rarely announces itself as significant. A population stops being perceived as a set of individuals and becomes perceived as a type. This is not the same as disliking a group, or fearing it, or believing false things about it. Those are inputs. Categorization is the structural event that converts those inputs into something actionable at scale, because it is far more difficult, psychologically and administratively, to act against a person you recognize as singular than against a category you have already decided you understand.
In Salem in 1692, the category was witch, and it was constructed with remarkable speed. Once a person was named, the accusation itself did the work that evidence, character, or years of known relationship would ordinarily do. A woman who had lived in the community for decades, who had been a neighbor, a midwife, a churchgoer, ceased to be evaluated as that person once the label attached. The label preceded and superseded everything the community actually knew about her. This is the essential feature of categorization: it does not require the category to be accurate. It requires only that the category be believed, and that belief in it license treatment that individual knowledge would have prevented.
Executive Order 9066, signed in February 1942, categorized an entire ethnicity, including tens of thousands of American citizens, as a security threat based entirely on ancestry.¹ No individualized evidence of disloyalty was required, or in most cases even sought. Japanese ancestry was treated as sufficient information to determine how a person would be treated by their own government. Family histories, military service records, church membership, decades of residence, none of it mattered once the category had been applied. The category did not describe a finding. It replaced the need for one.
What both cases share, and what recurs throughout this essay, is that categorization is not a conclusion reached after examining a person. It is a substitute for examining a person. That substitution is what makes everything that follows possible. It is worth pausing on why the substitution is so effective. Individual attention is expensive; it demands time, contact, and a willingness to have one's assumptions complicated by particulars. A category is cheap, and more than cheap, it is stable. Once a person has been filed under it, no further information is required, and, crucially, contradictory information can be discarded without the discarding feeling like dishonesty. The category is experienced not as a decision the observer made but as a fact about the person observed. This is the quiet inversion at the heart of the mechanism: a judgment the categorizer performed comes to feel like a property the categorized possesses. This is the same distortion examined at the level of individual cognition in schema-constrained default attribution.
The Numbing Function
Once a category exists, it performs a second function beyond identifying a target. It anesthetizes the surrounding population against the moral weight of what happens next. This is the mechanism that most needs naming, because it explains something that otherwise seems inexplicable: how ordinary people, not sadists, not ideologues, but administrators, clerks, neighbors, and bystanders, participate in or permit atrocity without experiencing themselves as doing wrong.
The bureaucratic architecture of the Nazi camp system illustrates this function starkly, because it was in part designed to produce it. The system divided the work of genocide into thousands of narrow, discrete tasks: scheduling trains, processing paperwork, maintaining records, operating machinery. Each task, isolated from the whole, could be performed by someone who understood themselves as doing a job rather than participating in mass murder. The division of labor was not merely a matter of efficiency; it distributed moral responsibility so thinly across so many hands that no single participant had to hold the whole of it. The category, Jew, undesirable, enemy of the state, did the moral work that the individual task-performer was never required to do, and the administrative structure ensured that the task-performer rarely had to see the connection between the form they processed and the death it enabled.
A related mechanism, at incomparably smaller scale but recognizable in form, operated in the American camps at Manzanar, Tule Lake, and elsewhere. Guards, administrators, and local officials carried out an incarceration built entirely on ancestry, and the administrative framing, security measure, wartime necessity, "relocation" rather than imprisonment, allowed most of the people executing it to understand themselves as following procedure rather than participating in the mass punishment of citizens for a crime none of them had committed. The euphemism did real work here. Language that describes incarceration as relocation, and prisoners as evacuees, is not merely dishonest; it is functional. It gives the participant a vocabulary in which their own actions can be described without moral residue, and a vocabulary in which cruelty can be discussed calmly is a vocabulary that permits cruelty to continue.
Bosnia offers a version of the numbing function that is in some ways more disturbing, because it operated not through bureaucratic distance but through the deliberate collapse of personal relationship. In Srebrenica and elsewhere, Bosnian Serb neighbors who had lived beside Bosniak families for generations, who had attended the same schools and shared the same streets, participated in or permitted the expulsion and killing of those families once wartime propaganda had successfully recategorized "neighbor" as "Muslim," and "Muslim" as existential threat.² The personal knowledge that should have made cruelty impossible was overridden by a category that was, in psychological terms, more recently and more urgently activated than decades of lived relationship. This is the numbing function operating at its most direct: not distance from the victim, but the deliberate suppression of a proximity that had already been established. It reveals something the bureaucratic cases can obscure, that prior relationship is not, by itself, protection. A sufficiently activated category can override years of it, which means the intuition that atrocity only happens between strangers is not merely wrong but dangerously reassuring.
Across these cases the mechanism recurs even where the method differs. Bureaucratic distance and the erasure of intimacy are different techniques for the same psychological outcome: allowing a person to participate in, or fail to resist, something that direct confrontation with a recognized individual victim would have made unbearable. What they share is the interposition of the category between the participant and the person, so that the participant never quite has to meet the person at all. This is identity organized through opposition in its most consequential form, the self stabilized against a target rather than through internal coherence.
The Justification Sentence
Once categorization and numbing have done their work, a further and quieter event tends to occur. People who participate in, benefit from, or simply witness subjugation without resisting it often produce a sentence that explains their acceptance to themselves and others. These sentences have the surface form of reasoning. They are not reasoning. They are the residue of a judgment the category already made, retrieved after the fact to lend the appearance of deliberation to something that was never actually deliberated.
During the incarceration of Japanese Americans, the justification took a familiar shape: they could be saboteurs, their loyalty cannot be verified, better safe than sorry. Note what this sentence does. It does not engage with the specific person being incarcerated. It restates the category as though the category were itself a form of evidence. It sounds like reasoning about risk; it is actually a restatement of the ancestry-based classification, dressed in the language of prudence. The tell is that the sentence would apply identically to every person in the category, which means it is not doing the work reasoning does, distinguishing this case from that one, but the opposite work, dissolving all cases into the one undifferentiated type.
Propaganda preceding and accompanying the Bosnian genocide and the persecution of the Rohingya performed the same function at greater volume and with greater deliberate intent. Populations were described, in state and quasi-state media, in language that framed their continued presence as a threat to the nation's survival, purity, or future.³ This is the justification sentence built at industrial scale: not a private rationalization reached after the fact, but a manufactured public narrative designed to produce that private rationalization in millions of people at once, so that participation or acquiescence would feel like self-defense rather than aggression. The individual who later says I had no choice, it was them or us is not lying about their experience. The experience was engineered.
In the contemporary American context, a common justification takes the form: they should have followed the proper legal channels. This sentence deserves particular attention because its surface reasonableness so effectively obscures its function. It is, formally, a statement about legal process. Structurally it performs the same work as every justification examined above: it lets the speaker feel they have reached a considered conclusion about a specific situation, when the category, illegal, has already answered the only question the sentence pretends to ask, which is whether this particular person, in this particular circumstance, deserves what is happening to them. The sentence requires no knowledge of the individual. It requires only that the category be invoked. This is the same work examined at the level of everyday social control in indirect power.
This is not to say legal process is an illegitimate subject of debate. It is to say that when a sentence about legal process is deployed not to argue a policy position but to dissolve discomfort at the sight of a child separated from a parent, it is functioning as an alibi, not an argument. The test is simple and worth stating plainly, because it is the same test that exposes every justification sentence in this essay: does the sentence require any knowledge of the specific person and circumstance, or does it work identically regardless of who the person is? A judgment about a person depends on facts about that person. An alibi works no matter who is standing there. That interchangeability is the signature of the category speaking through the person who believes they are reasoning. It is one of the enabling conditions and common misrecognitions through which indirect power operates without announcing itself.
Escalation and Reuse
Categorization and its administrative apparatus, once built, do not dissolve when their first use ends. They persist as capacity, and states that have built them tend to redeploy them, often against a different population, sometimes many years later.
Myanmar is the clearest illustration. The military apparatus responsible for decades of campaigns against the Karen and other ethnic minorities was, in its essential features, the same apparatus later turned against the Rohingya.⁴ The specific target changed; the ethnic and religious justification changed; the underlying machinery did not need to be rebuilt. The willingness to categorize an entire population as a threat to national identity, the administrative and military infrastructure for displacement, the propaganda apparatus for manufacturing justification at scale, had already been constructed, tested, and normalized against earlier populations, and it was available to be pointed in a new direction. In structural terms, persecution of the Karen can be read not as wholly separate from persecution of the Rohingya, but as an earlier use of the same state instrument.
This is among the most important structural findings for anyone trying to understand how subjugation recurs. The danger of a categorization apparatus is never confined to the population it currently targets. Once built and normalized, it is standing capacity, available for use against whichever group a state or society next decides to categorize. This is why the precedent set by the American incarceration of Japanese Americans mattered well beyond the 1940s: the legal and administrative reasoning that a population could be detained on the basis of ancestry rather than individual conduct did not vanish when the camps closed. It entered the repertoire of available state action, cited and argued over in subsequent decades whenever the question of detaining a categorized population returned. A mechanism used once is easier to use again, not only because the infrastructure exists but because the prior use has already answered, for many people, the question of whether such a thing is possible here. It is.
The Live Case
A brief restatement of the boundary before proceeding, because it matters most precisely here. Nothing in this section claims that current immigration enforcement is equivalent in scale or consequence to any historical case above. It examines documented practices as evidence of a mechanism, not as a verdict, and it makes no argument about legal status or the proper scope of enforcement. What it observes is narrower: that the structural conditions described throughout this essay are recognizably present in a live, ongoing system, and that a mechanism is more useful to name while it is operating than only in retrospect.
Family separation as a practice did not end with the policy reversal of 2018. Under enforcement operations beginning in 2025, researchers estimated, as of spring 2026, roughly 400,000 interior detentions since January 2025, and calculated that approximately 205,000 children, including roughly 145,000 United States citizen children, had experienced the detention of a parent as a result; more than a third of the affected citizen children were under the age of six.⁵ These separations are not incidental byproducts. They are the foreseeable consequence of an enforcement structure that does not require a person's parental status to be accounted for before detention proceeds, that treats the category as the operative unit and the family as a detail to be discovered later, if at all.
Detention itself scaled sharply, with an average of more than 60,000 people held on a given day as of late 2025, the highest figure recorded to that point, and roughly $45 billion allocated toward further expansion of detention infrastructure.⁶ Scale of this kind requires, and produces, exactly the administrative distance described in The Numbing Function. A system processing tens of thousands of people does not, and at that scale is structurally disinclined to, evaluate each person as an individual before enacting the consequence the category assigns. The category, arrived illegally, overstayed, absconded, becomes the operative unit, and the individual recedes into it in precisely the way this essay has traced.
Reporting on implementation documented what internal accounts described as administrative disorganization severe enough that separated family members could not reliably be tracked, echoing the documented failures of 2018, when investigators found that children had effectively been erased from government databases.⁷ Whether such confusion is a design feature or a failure of coordination is a separate question from the one examined here. What matters structurally is that the confusion functions the same way regardless of intent: it prevents individual accountability from attaching to any specific decision, and it lets each participant experience their own role as narrow and procedural. That is the same distribution of responsibility described in The Numbing Function, arrived at by a different route.
I want to be exact about the limits of this comparison, because the distance between the cases is as important as the resemblance. The resemblance is at the level of mechanism: categorization prior to individual evaluation, numbing through scale and administrative distance, justification sentences that operate as alibis. The distance is at the level of everything else, intent, scale, outcome, and the presence, so far, of functioning courts, journalists, and oversight bodies contesting the practice in real time. To notice a mechanism operating is not to predict its furthest possible development, still less to assert that its furthest development has arrived. It is to say that the early structure is visible, and that the value of recognizing it early is precisely that the later stages are not inevitable. A mechanism named while it is still contestable can still be contested.
The Qualifier, Restated
These cases are not equivalent, and the differences are not incidental to the argument; they are part of it. Salem produced roughly twenty deaths. The Holocaust was the deliberate murder of six million Jews and millions of others. Srebrenica was a judicially determined genocide. What was done to the Karen and the Rohingya has been described by international bodies in comparably severe terms. Current American immigration enforcement, whatever judgment history renders, has not been credibly likened by serious observers to industrial genocide, and this essay does not liken it to one. To collapse these outcomes into a single moral category would be its own version of the error the essay has spent its length describing: the substitution of a label for careful attention to what specifically happened, to whom, and at what scale.
What is equivalent, and what the essay has tried to isolate, is the mechanism. A population is sorted into a type rather than seen as individuals. That sorting anesthetizes the surrounding population against the moral weight of what follows, through bureaucratic distance, through the suppression of prior relationship, or through sheer administrative scale. Justification sentences emerge that lend the appearance of reasoning to the residue of a judgment the category already made. And the apparatus, once built, remains available for reuse, against the same population or another, indefinitely.
Recognizing the mechanism does not, by itself, prevent it. But a society that can only recognize it in retrospect is a society that can only ever name what happened after naming it can no longer change anything. The reason to examine these structures together, across eras, cultures, and vastly different outcomes, is not to deliver a verdict on any single case. It is to make the mechanism legible enough, early enough, that recognizing it might occasionally matter more than remembering it. The category is load-bearing. It holds up everything that follows. And the first act of resistance available to anyone is also the simplest: to insist, against the considerable convenience of the category, on seeing the person it was built to replace.
Notes
The figures in The Live Case concerning current United States immigration enforcement reflect reporting available as of mid-2026 and describe conditions at the time of writing. They are point-in-time measurements, not permanent facts, and readers consulting this essay later should treat them accordingly.
Executive Order 9066 (February 19, 1942) authorized the exclusion of persons of Japanese ancestry from designated military areas, leading to the incarceration of approximately 120,000 people, roughly two-thirds of them United States citizens. See the U.S. National Archives records on Japanese American incarceration and Executive Order 9066.
The July 1995 killing of more than 8,000 Bosniak men and boys at Srebrenica was determined to constitute genocide by the International Criminal Tribunal for the former Yugoslavia (Prosecutor v. Krstić, 2001) and affirmed by the International Court of Justice (Bosnia and Herzegovina v. Serbia and Montenegro, 2007).
The role of state and quasi-state media in preparing populations for mass violence is documented across the scholarship on both the Balkan conflicts and the Myanmar military's campaigns, including the findings of the Independent International Fact-Finding Mission on Myanmar (United Nations Human Rights Council, 2018). The structural point made here concerns the function such narratives perform rather than any single broadcast or publication.
United Nations reporting on Myanmar has documented the military's successive campaigns against ethnic minority populations, including the Karen and other groups over several decades and the Rohingya most severely from 2016 onward, carried out by substantially the same institutional apparatus. See the Report of the Independent International Fact-Finding Mission on Myanmar (UN Human Rights Council, A/HRC/39/64, 2018).
Maria Cancian, Nissi Cantu, Lanikque Howard, and Tara Watson, "The administration has detained 400,000 immigrants: What do we know about their children?" The Brookings Institution, May 18, 2026. The analysis estimates approximately 400,000 interior-arrest detentions between January 20, 2025 and April 9, 2026, and, on its baseline assumptions, roughly 205,000 children affected by parental detention, including approximately 145,000 United States citizen children, of whom more than a third of the affected citizen children are under age six. https://www.brookings.edu/articles/the-administration-has-detained-400000-immigrants-what-do-we-know-about-their-children/
Detention population figures (an average exceeding 60,000 people held per day as of late 2025) and the $45 billion appropriation for detention expansion under the 2025 reconciliation legislation are reported in the Brookings analysis cited in Note 5 and in contemporaneous reporting, including PBS NewsHour, "What happens to children when immigrant parents are detained by ICE," May 19–20, 2026.
Raul Pinto et al., American Immigration Council transparency project on the implementation of the 2018 family separation policy (October 30, 2025), documenting that separated children were effectively erased from government tracking systems; and ProPublica, "Over 100,000 American Kids Had Parents Detained in Immigration Sweeps, Report Estimates" (May 18, 2026), documenting parallel tracking failures and the absence of any systematic government accounting of separated families under current enforcement.