In the Name of Protection

In 2017, the Japanese cultural critic Uno Tsunehiro published Maternal Dystopia (母性のディストピア), a 512-page argument that postwar Japanese animation was trapped. The trap, as he described it, was structural: male protagonists could not act — could not fly, could not fight, could not achieve agency — without a female figure to protect or a maternal presence to grant approval. Protection was not one theme among many. It was the deep imaginative grammar of the medium itself.

Uno traced this grammar through three canonical directors — Miyazaki Hayao, Tomino Yoshiyuki, Oshii Mamoru — and concluded that it was a symptom of postwar Japan's peculiar condition: a nation whose democratic sovereignty was structurally fictive, whose political agency was externally constrained, and whose male subjects compensated for this impotence by retreating into narratives of maternal shelter and protective devotion. In an earlier work, Imagination in the Zero Decade (ゼロ年代の想像力, 2008), he had argued that the collapse of grand narratives in the 1990s — the burst of the economic bubble, the Aum Shinrikyo attacks, the endless Heisei recession — left a legitimacy vacuum that "protection" conveniently filled. You no longer needed to believe in justice, progress, or national destiny to justify violence. You only needed someone specific to protect.

Uno was not wrong. But he was looking at a local outcrop of something much larger.

This essay argues that what Uno diagnosed in Japanese animation is not a cultural pathology specific to postwar Japan, nor a narrative quirk of the anime industry, but an instance of a far more pervasive phenomenon: protection as an organisational logic — a way of structuring relationships, institutions, narratives, and polities that is so deeply naturalised it is no longer recognised as a logic at all, but mistaken for nature itself.

What follows begins as a concept history (Begriffsgeschichte) — a study of what a single word has been made to carry — and then leaves the vocabulary behind for the institutions, because the linguistic evidence can raise the question but cannot settle it. The argument proceeds through a series of increasingly broad claims. First, that the Japanese word mamoru (守る) — which means simultaneously "to protect," "to obey," "to keep a promise," and "to maintain" — offers a heuristic entry point for noticing a relationship between protection and order that contemporary English vocabulary tends to distribute across several terms. Second, that protection operates through three specific mechanisms: it constitutes the protector's moral identity, it reorganises the relationship between protector and protected into a durable and difficult-to-reverse asymmetry, and it constrains the grammar through which the protected might refuse. Third, that these mechanisms are not confined to intimate relationships or fictional narratives but operate across a spectrum from religion to insurance contracts, from Confucian governance to operating-system updates, from heroic narrative to trade policy. And fourth, that in a series of cases since the turn of the millennium protection has taken over justificatory work that other vocabularies used to do — a substitution that cannot be explained by increased danger alone and may also register a crisis of legitimacy, as institutions lean more heavily on protection-based justification when other legitimating vocabularies weaken.

A word about what this essay does not claim. It does not claim that protection is inherently harmful. To keep the analysis precise, it is necessary to distinguish three levels at which protection operates:

A protective act is a temporally bounded intervention that prevents or mitigates harm: a firefighter pulling someone from a burning building, a vaccine conferring immunity, a bystander stepping between an aggressor and a target. The act ends when the danger passes. It does not, by itself, install a durable hierarchy or generate an expectation of ongoing obedience.
A protective relationship is a more stable configuration in which one party regularly provides protection to another: a parent and child, a bodyguard and client, a patron and dependent. The roles are relatively fixed, but the relationship remains negotiable — the protected party can, in principle, renegotiate the terms, seek alternative protection, or declare the relationship concluded. Gratitude may be expected, but exit is not coded as betrayal.
A protection regime is the point at which protection becomes the organising principle of the relationship, the source of the protector's authority, and the moral framework within which the protected party is expected to operate. In a protection regime, the protector's identity depends on continuing to protect; the relationship has been reorganised by an economy of gratitude and obligation that runs in one direction; and refusal — the protected party's attempt to exit or challenge the arrangement — is coded not as a legitimate choice but as ingratitude, impiety, treason, or naivety. Protection has ceased to be something someone does and has become the structure within which both parties live.

This essay analyses the third level. Its central question — can the protected say no? — is the diagnostic that distinguishes a protection regime from a protective act or relationship. Where the answer is "yes, without penalty," protection is an act or a relationship. Where the answer is "no, or only at the cost of being labelled ungrateful, disloyal, or dangerous," protection has become a regime. The essay's further question — under what conditions does a protective relationship slide into a protection regime? — is what the three mechanisms in Section II are designed to answer.

One further orientation. In a previous essay, "The Prepared Soil," I traced how competitive logic — the formula of no ceiling, no floor, and no mercy for losers — functions as an organisational principle so naturalised that it appears to be a description of reality rather than a design. This essay examines the other side of that coin. In the modern institutions examined here, competition and protection frequently reinforce each other: competition eliminates losers, and protection legitimises the authority of those who claim to shield the survivors. The two are not logically inseparable in all conceivable arrangements, but in the institutional contexts this essay examines — from Confucian governance to security states to narrative industries — they have historically operated as complementary registers of the same authority.

I. What "Protection" Contains

Begin with a word.

The Japanese verb mamoru (守る) is conventionally translated as "to protect." This is not wrong, but it is drastically incomplete. Mamoru means, simultaneously:

To protect or defend (人を守る — to protect a person)
To obey or observe (規則を守る — to obey rules)
To keep a promise (約束を守る — to keep a promise)
To maintain or preserve (伝統を守る — to preserve a tradition; 沈黙を守る — to maintain silence)
To guard or watch over (門を守る — to guard a gate)

One traditional etymological account derives mamoru from 目 (ma, eye/sight, archaic form) + もる (moru, to watch over) — "to watch with one's eyes." The kanji 守 has been analysed as combining 宀 (roof/shelter) and 寸 (hand/measure). These component analyses should be treated as heuristic rather than definitive historical etymology, but they point in the same direction as the polysemy itself: toward a semantic core in which protection is understood as sustained watchful maintenance — keeping something in its proper state through continuous attention.

What unites all five meanings is this orientation: maintaining something in its proper state through watchful presence — whether that "something" is a person's safety, a rule's integrity, a promise's fulfilment, or a tradition's continuity.

The polysemy of mamoru does not, by itself, prove that Japanese speakers experience protection and obedience as "the same activity." A word can have multiple meanings without its users consciously perceiving the connection between them. But the polysemy does offer a useful heuristic — an analytical entry point for noticing a relationship that is difficult to notice in English, where the relevant meanings are distributed across separate words. When the same verb covers both "protect a person" and "obey a rule," the analyst is prompted to ask: is there a structural relationship between protecting and obeying that the English vocabulary obscures? The argument of this essay is that there is — but the evidence for that relationship comes from institutional analysis and political theory (Sections II–V), not from the linguistic fact alone.

This is why the phrase 守りたい (mamoritai, "I want to protect") carries such weight in Japanese popular culture — particularly in anime, manga, and games, where it functions less as a declaration of intent than as a totalising claim on a relationship. When a character says 守りたい, the phrase can draw on a wider semantic field of watchfulness, maintenance, and commitment: I want to watch over you, maintain your world, keep my promises to you, preserve what matters to you. It is not a one-time act but a pledge of continuous existential custodianship. This is substantially heavier than the English "I want to protect you," which denotes a specific defensive action without the connotations of obedience, maintenance, and preservation.

The word also differentiates internally. Three different kanji can be read as mamoru, each with distinct connotations: 守る (the general form, covering all senses including "obey rules"), 護る (emphasising care for the vulnerable — sheltering, nursing), and 衛る (military perimeter defence — patrolling, guarding). The general form — the one that includes "obey" — is the default. The specifically protective and specifically military forms are marked specialisations. The unmarked, everyday form of "protection" already contains obedience.

The English fracture

English speakers may find it difficult to perceive the relationship that mamoru's polysemy makes visible, because contemporary English tends to distribute the relevant meanings across several lexical items, each housed in a different etymological tradition.

Table 1

Historical layer	Source	Words for mamoru
Old English (Germanic)	Native vocabulary	hold (healdan), ward, watch, keep
Post-1066 (Norman French)	Conquest import	guard, protect, defend, observe
Renaissance (Latin/scholarly)	Written borrowing	preserve, maintain, conserve

The Old English verb that came closest to mamoru was not keep but healdan ("hold"), whose recorded senses run "hold; keep; preserve; protect; maintain," "observe," and "rule; govern" — shelter, maintenance, compliance and rule inside a single verb. (Keep, from Old English cēpan, is a poor exhibit for the opposite reason: its "guard, defend" sense is a twelfth-century development and its "protect from harm" sense later still, so keep acquired its protective range only after the Conquest, not before it.) As French and Latin words flooded in, they created synonymic competition with the native stock. The result was not the victory of one layer over another but semantic specialisation: each word retreated into a narrower ecological niche. Protect became exclusively "to shield from harm." Obey became exclusively "to follow commands." Observe split into "to watch" and "to comply with." Preserve became "to keep unchanged." The unity at issue is lexical, not psychological: guarding, obeying and maintaining were once available inside a single verb, and lexical history distributed them across several. Whether speakers ever felt them as one gesture is not something the vocabulary can tell us — the caution this essay applies to mamoru above applies to healdan equally.

The same pattern governs adjacent concepts. English has duty (Old French deu, what is owed), obligation (Latin obligare, to bind), and responsibility (Latin respondere, to answer) where Chinese and Japanese use two words (義務, 責任) to cover roughly the same territory.

This matters for analysis because the English lexical distribution makes it harder to notice a relationship that mamoru's polysemy makes harder to miss. When "protect" and "obey" are separate words with separate etymologies, the analyst must do theoretical work to reconnect them. Iris Marion Young wrote an entire essay (2003) demonstrating that "protection" contains "domination." Wendy Brown, in States of Injury (1995), argued at length that the protection the state offers is purchased with dependence. Foucault traced the pastoral lineage back to shepherds to reconnect "care" with "governance." These are genuine intellectual achievements. The polysemy of mamoru does not make them unnecessary — one can use a word for centuries without reflecting on what its semantic range implies — but it does offer a more economical starting point for the same inquiry.

It is worth noting that this is not a uniquely Japanese phenomenon. English retains traces of the same unity in hold (hold a position, hold to a promise, hold an office), and Latin observare (to watch / to obey) preserves the watching-obeying link. The polysemy of mamoru is heuristically useful not because Japanese is unique but because it is the language in which the cultural phenomenon this essay examines — the centrality of 守りたい in popular narrative — is most visible.

The starting hypothesis, then, is this: protection and order are more closely related than the English vocabulary suggests, and the polysemy of mamoru is a useful entry point for investigating that relationship. Whether the relationship holds — and how it operates — is the business of the sections that follow.

One further distinction: "protection" is not synonymous with "care," "aid," "rescue," or "defence." Care can be mutual; aid can be temporary; rescue ends when the danger passes; defence can be collective and reciprocal. What makes protection specific is the asymmetric, durable relationship it tends to create — the protection regime described in the Introduction. It is this relational structure, not the act of shielding someone from harm, that the sections below examine.

II. How Protection Works: Three Mechanisms

Protection's power does not lie in any single operation but in the convergence of three mechanisms, each reinforcing the others. The first constitutes the protector's identity. The second reorganises the relationship between protector and protected. The third eliminates the grammar through which the protected might refuse.

Mechanism 1: Identity constitution

Most concepts analysed in ideology critique function through concealment. The word 天灾 ("natural disaster") conceals human responsibility behind a meteorological label. The "hidden rule tax" conceals the gap between formal and actual rules. These are operations of disguise: the reality is one thing, the name says another, and the critic's task is to expose the discrepancy.

Protection does not work this way. It does not conceal the protector's power — it constitutes the protector's moral identity. When someone says "I am protecting you," they are not hiding what they are doing. They are, in many cases, genuinely shielding someone from harm. The operation is not deceptive. It is something more durable than deception: it is self-confirming. Each act of protection reinforces the protector's sense of being the kind of person who protects. The identity becomes load-bearing — it is no longer something the protector does but something the protector is. And once an identity is load-bearing, abandoning the behaviour that sustains it feels not like a policy change but like a loss of self.

This is why challenging protection provokes a different kind of resistance than challenging, say, 天灾. If you demonstrate that a "natural disaster" was actually a policy failure, the person who used the term can retreat to "I was just describing the situation." The word was instrumental; the user's identity is not at stake. But if you demonstrate that someone's "protection" was actually control, the protector does not experience this as a factual correction. They experience it as a denial of their love. The response is not "you are wrong about the facts" but "you are saying I don't care" — a response from the level of identity, not argument.

This is what makes protection harder to analyse than most ideological concepts. The standard critical move — show that the word disguises a different reality — does not apply, because protection is not primarily a disguise. It is a subject-position. The protector is not lying about what they are doing. They are constituting themselves through what they are doing. And because the constitution is genuine — it really does feel like love, it really does produce care, it really does involve sacrifice — the critical analyst faces a distinctive challenge: they must explain how something that is genuinely felt as love can simultaneously function as power, without claiming that the love is fake.

The psychoanalyst Saito Tamaki offered a partial framework for this in Beautiful Fighting Girl (戦闘美少女の精神分析, 2000). Analysing the male viewer's desire to "protect" the beautiful fighting girl — a character who manifestly does not need protection, since she is the one doing the fighting — Saito argued that the desire to protect is a strategy for constituting the viewer's own subject-position. It is not about the character's needs. It is about the viewer's need to establish a relationship with the fictional object. The "protection" posture is the bridge through which the subject constitutes itself as a subject-who-cares. Whether the object actually needs protection is, structurally, irrelevant.

Mechanism 2: Relational alchemy

Protection does not merely add a behaviour to an existing relationship. It transforms the relationship itself. Before the act of protection, two parties may stand in any configuration — strangers, equals, rivals, collaborators. After one party protects the other, the relationship has been reorganised into an asymmetry that is difficult to reverse:

Before: neither party owes the other anything. After: the protected owes gratitude.
Before: either party can leave. After: the protected who leaves is ungrateful.
Before: the protector's actions require justification. After: the protector's actions are pre-justified ("I am protecting you").
Before: either party can criticise the other. After: the protected who criticises is ungrateful, disloyal, or — in the vocabulary of many traditions — impious.

This is the relational alchemy of protection: the transmutation of a neutral or symmetric relationship into one structured by obligation, gratitude, loyalty, and subordination — a transmutation that appears entirely natural, because "of course you should be grateful to someone who protected you."

Marcel Mauss's analysis of gift exchange in The Gift (1925) provides the structural key. Mauss demonstrated that gifts are never free: they create obligation. The recipient of a gift is bound to reciprocate, and the failure to reciprocate is a social offence. Protection can be understood as a particular kind of gift — an unreturnable gift. You cannot give back equivalent protection to the person who protected you (especially if they are stronger, more powerful, or more institutionally positioned than you). The debt is therefore difficult to discharge. This does not mean it is always permanent — debts can be forgiven, renegotiated, mutualised, or simply ignored. But in the absence of active renegotiation, the default trajectory is toward a durable asymmetry: the protector remains the one who gave, and the protected remains the one who owes.

Shu-Shan Lee's analysis of Confucian political obligation as "paternalistic gratitude" (2021) traces this mechanism into institutional form. In the Confucian framework, the ruler protects the people; the people owe the ruler gratitude; gratitude generates a duty of obedience. This is not presented as a coercive arrangement but as a moral one — the natural response of virtuous subjects to benevolent governance. The obligation is conditional: it depends on the ruler actually governing well. But — and this is the critical structural point — the conditionality is evaluated from the ruler's performance, not from the subject's right to refuse. The subject can say "you are not governing well enough" (a judgement about the quality of protection) but cannot say "I do not wish to be governed" (a refusal of the protective relationship itself). Even resistance, when it occurs, must be framed in the ruler's own vocabulary: not "I have a right to refuse your protection" but "your protection is insufficiently protective." The grammar of refusal is locked inside the grammar of protection.

Sungmoon Kim's concept of "passive virtues" (2025) identifies the subjective complement to this institutional structure. Kim's target is a normative position rather than a historical description: contemporary Confucian meritocrats hold that moral education should not aim at making people politically active, and should instead cultivate deference, dependence and paternalistic gratitude — the cluster Kim names "passive virtues" and argues against. What matters here is what the prescription reveals. It is a fully worked-out account of good citizenship in which the citizen's excellence consists in receiving protection well: the protected party is not merely required to tolerate the asymmetry, but to find their virtue in it. Gratitude is not the price of protection — it is, on this account, the moral achievement of the protected. Whether any actual polity produced such dispositions is a separate question, and not one this essay settles.

What the prescription describes is the most complete form the relational alchemy could take: an arrangement in which the asymmetry would not be experienced as asymmetry at all, but as mutual moral fulfilment — the protector fulfilled by protecting, the protected fulfilled by being grateful. Naming the ideal is worth doing whether or not it was ever achieved, because it specifies what a fully realised protection regime would look like from the inside: a hierarchy that had disappeared from view not by being hidden, but because both parties were invested in experiencing it as love.

Mechanism 3: The elimination of refusal grammar

The third mechanism is the most consequential and the most culturally variable: the systematic prevention of a grammar through which the protected might refuse protection.

The claim here is not that no one has ever refused protection — people refuse it all the time, and pay the social costs. The claim is comparative and about degree. Every tradition examined here gives the protected some language for complaint, and in every one of them that language is issued by the protector's own institutions. What differs is how much of it survives contact with an actual refusal: whether the tradition also developed vocabulary that can be turned against the protective relationship itself — rights discourse, contractarian theory, individual autonomy — or whether the only available idiom is corrective ("your protection is insufficient"), so that refusal must be spoken in the protector's words or not at all.

Compare three traditions:

In the Hobbesian framework, the protection-obedience exchange is contractual. The sovereign protects; the subject obeys. But the contract is conditional: "if protection ceases, the state too ceases, and every obligation to obey ceases" (Carl Schmitt's summary of Hobbes). The subject has, at least in theory, a recognisable grammar for saying "you have failed to protect me, therefore I no longer owe you obedience." The grammar is thin — it does not extend to "I refuse your protection even if it works" — but it exists.

In the Foucauldian pastoral model, the shepherd tends the flock. The relationship is not contractual but caring. There is a third party — God — above both shepherd and sheep, and the sheep can in principle appeal to God against a bad shepherd. But this check is weaker than it looks, and for exactly the reason that will constrain Heaven below: the shepherd's own institution is what interprets God's will. What the pastorate supplies is not an independent court but a vocabulary of complaint — and the vocabulary belongs to the protector.

The Confucian familial model differs from these two by degree rather than kind. The difference is not whether mediating authorities exist but how completely they are absorbed into the protector's own vocabulary. This is not to say that no third parties exist: Heaven, the ancestral lineage, ritual propriety (禮), the censorate, and local community norms all function, in different periods and contexts, as potential courts of appeal beyond the immediate protector. The imperial Chinese system was not a simple dyad of ruler and ruled. But these mediating authorities share a structural feature that distinguishes them from the Hobbesian contract and the pastoral God: they operate within the same moral vocabulary as the protector, and they tend to reinforce rather than check the protective relationship.

Heaven (天) is invoked to judge the ruler — but who interprets Heaven's signs? Predominantly, the officials who are the ruler's appointees, operating within an interpretation loop that "The Prepared Soil" described as closed. That description was too flat: the loop is strongly self-reinforcing rather than sealed — individual officials did sometimes risk their careers to deliver unwelcome interpretations — but they did so from inside the appointment structure they were judging. Ritual propriety constrains the ruler — but the content of ritual is defined by the scholarly tradition that the state sponsors. The censorate monitors officials — but it monitors them for failure to fulfil their protective duties, not for the existence of those duties in the first place. At every level, the available language of challenge is a language of corrective remonstrance within the protective framework, not a language of refusal of the framework itself.

The relationship is structured by the isomorphism of family and state (家国同构): the emperor is the father of his people, and subjects owe the obedience that children owe parents. 雷霆雨露，俱是天恩 — thunder and rain are both Heaven's grace. The child cannot sue the father. To question the father's discipline is not dissent but impiety. The claim here is not that the grammar of refusal was "never" present in any form — it is that imperial Confucian institutions repeatedly subordinated available languages of resistance to a paternal vocabulary of corrective remonstrance, channelling challenge into a form that reinforced rather than threatened the protective structure.

The point is not confined to the ruler-subject case. Hagop Sarkissian has examined the passages standardly cited to show that Confucian ethics licenses remonstrance against wrongdoing parents ("How Remonstration Fails: Filial Piety and Reprehensible Parents," Journal of Confucian Philosophy and Culture, 2023). He finds that all of them require remonstrating without opposing — gently, and only so far as it does not produce ill feeling or estrangement — and that if the remonstrance is ignored, the son is to remain reverent rather than become disobedient. The conditions are severe enough that in exactly the cases where a protector is genuinely harmful, the permission is functionally equivalent to its absence.

This is not to say that no Confucian thinker ever articulated a challenge to authority. Mencius famously declared: 闻诛一夫纣矣，未闻弑君也 — "I have heard of the execution of a mere fellow named Zhou, but not of the assassination of a ruler." This appears to provide exactly the grammar of refusal that the analysis claims is absent. But the appearance is misleading, and the history of this passage is itself the strongest evidence for the claim.

In 1394, the Ming dynasty founder Zhu Yuanzhang — a former beggar and rebel who had ridden heterodox religious movements to the throne — ordered the scholar Liu Sanwu to produce an abridged edition of the Mencius. Eighty-five of the Mencius's 260 chapters were deleted — about a third of the chapters, and by character count roughly 47 per cent of the text. The deletions targeted precisely the politically dangerous content: 民为贵，社稷次之，君为轻 ("The people are most important, the altars of grain and soil come next, the ruler is least important"); the passage on executing tyrants; every sentence concerning the right of rebellion. Only the abridged edition was permitted for the civil service examinations.

Zhu Yuanzhang did not misunderstand Mencius. He understood him precisely — and responded with deletion.

The censorship lasted only seventeen years before the full text was restored (1411). But the restoration did not empower Mencian radicalism. It returned the text to a framework that had already domesticated it.

That framework predated both Zhu Yuanzhang's censorship and its reversal. In the late twelfth century — two hundred years before the Ming dynasty — Zhu Xi (朱熹, 1130–1200) had compiled the Mencius into the Four Books (四书) alongside the Analects, the Great Learning, and the Doctrine of the Mean. He framed the text within Neo-Confucian commentaries that emphasised metaphysics and moral self-cultivation, not political resistance. When the Yuan dynasty adopted the Four Books as the core curriculum of the civil service examinations in 1313, the institutional domestication entered a new phase: the Mencius was now mandatory reading for every aspiring official, to be read through Zhu Xi's commentarial lens. The Eight-Legged Essay format (八股文), which further constrained interpretive freedom by requiring formulaic exposition of classical passages, was fully codified somewhat later (its mature form is generally associated with the Ming dynasty examination system), but the trajectory was consistent: the Mencius was to be expounded within parameters that rewarded moral-philosophical exegesis, not independent political reasoning. The examination system would not have rewarded a candidate who used 民贵君轻 to argue for limiting imperial authority.

Zhu Yuanzhang's censorship (1394) and its reversal (1411) are therefore best understood as a brief episode within a longer process: the Mencius was canonical before the censorship, during it, and after it. What changed was not the text's status but a single emperor's willingness to tolerate even the decorative presence of radical vocabulary. The domesticating framework began before the censorship — Zhu Xi's commentaries and the examination system were already in place — and was subsequently reinforced by the mature Eight-Legged Essay format. For nearly six hundred years, generations of educated elites in China could quote Mencius on the people's importance. They learned and deployed those quotations within an institutional context designed to produce service to the emperor. The text was canonical. Its radical implications were neutralised by the interpretive and institutional framework surrounding it.

As the political philosopher Justin Tiwald has argued (2008), Mencius's permissions fall short of a proper right of rebellion in two crucial respects: the people are sometimes permitted to participate in rebellion but are not permitted to decide for themselves when rebellion is warranted; and what matters is not the people's judgement about justifiability but "certain affections and behaviours that closely track their life satisfaction." The people are the medium through which Heaven expresses displeasure — not agents exercising a right. The grammar remains: "the protector should be better" (您不够仁), not "I have the right to refuse protection."

Yuri Pines has identified the structural root of this limitation within Mencius's own thought: Mencius insisted on the commoners' primary importance for the polity while equally strongly opposing their participation in political life. The people matter, but the people do not act. This internal tension — popular importance combined with popular exclusion — made 民贵君轻 structurally toothless: it tells the ruler to care about the people but gives the people no agency. Pines argues in The Everlasting Empire (2012) that this unresolved tension was an asset for the empire's longevity: the ideology persisted precisely because it was never operationally tested.

The specific contribution of imperial Confucian institutions to the architecture of protection, then, is not the invention of "protection as virtue" — that is cross-cultural. It is the repeated subordination of available languages of resistance to a paternal vocabulary of corrective remonstrance — a vocabulary that channels challenge into forms that strengthen, rather than threaten, the protective structure. To cite Mencius as evidence that a robust grammar of refusal exists within this tradition is to mistake the presence of radical vocabulary for the presence of institutional mechanisms that could operationalise it. The vocabulary existed. The mechanisms did not — and the history of how the vocabulary was handled (elevated, censored, restored, domesticated) is itself the strongest evidence for the claim.

One might read all of this as history. It is not only that. In 2024 the philosopher Andrew T. W. Hung published a defence of hard paternalism grounded explicitly in Confucian familism ("Hard Paternalism and Confucian Familism," Fudan Journal of the Humanities and Social Sciences 17:3, 455–478). Hung argues that Joel Feinberg's soft paternalism — which licenses intervention only where a person's choice is insufficiently voluntary — is untenable, because voluntariness cannot be reliably assessed; and that Confucian familism supplies what soft paternalism lacks. Because a person's life and identity are inseparable from their family, actions conventionally classed as self-regarding are not in fact self-regarding: they affect family members too, so protecting a person from self-harm also protects their kin, which brings paternalistic intervention inside Mill's harm principle rather than into conflict with it.

The argument is worth stating precisely, because its force is easy to misplace. Hung is not claiming that the protected should be overruled when they object. He is claiming something prior: that for a relational self, the category of a purely self-regarding choice does not exist. The consequence — Hung does not put it this way — is that "leave me alone" is not a demand the framework refuses but a demand it cannot parse. That is the elimination of refusal grammar stated as a philosophical position rather than inferred from an institution, and it is the most direct evidence in this section, because unlike the imperial material it cannot be set aside as belonging to a superseded order.

III. Genealogy, and a Comparison Across Domains

The three mechanisms described above — identity constitution, relational alchemy, and the elimination of refusal grammar — are not confined to any single cultural tradition or institutional domain. They operate wherever protection is invoked as the organising principle of a relationship. What follows has two parts, and they do different work. The first three subsections are genealogical: they trace a descent, from the Christian pastorate through the sovereign bargain to Tilly's protection racket. The fourth is not genealogical at all but synchronic — a comparison of six domains that are not descended from one another, set side by side along shared dimensions to test whether the same structural feature recurs where no line of transmission can explain it.

The structural feature is this: in each of these domains, protection has been arranged so that it naturalises asymmetry and forecloses refusal. The domains differ in everything else — theology, enforcement mechanisms, cultural context, historical period. What recurs is the double operation: hierarchy made to appear natural, refusal made to appear ungrateful or incoherent. Whether that arrangement is a necessary property of protection or a recurring design choice is a question Section VI returns to with a case that cuts the other way.

The pastoral prototype

Foucault identified the oldest institutional form of protection-as-governance in the Christian pastorate, analysed in his 1977–78 lectures at the Collège de France (Security, Territory, Population). The shepherd tends the flock. He watches over each individual sheep and the flock as a whole — simultaneously individualising and totalising. He is willing to sacrifice himself for the flock. His power is exercised for the benefit of those he governs, not against them. It is, in Foucault's phrase, "a power of care."

The pastoral model is the prototype of all subsequent protection-as-governance because it establishes the core paradox: the more benevolent the power, the more total it becomes. A tyrannical ruler can be resisted because his cruelty provides the grounds for resistance. A caring shepherd cannot be resisted on the same grounds — he is not cruel, he is attentive. To resist him is not to resist oppression but to refuse care. The pastorate inaugurates a form of power that disarms resistance by making resistance feel like ingratitude.

Foucault argued that the secularisation of pastoral power — the transfer of the shepherd function from church to state — was the genealogical origin of modern governmentality: the state that manages, optimises, and cares for its population. The secular state inherited from the pastorate a dual operation that Foucault named with a Latin phrase in his 1979 Tanner Lectures at Stanford: omnes et singulatim — "each and all." Modern political rationality, on that account, is a permanent integration of individuals into a totality, governing each person individually (through data, records, case files) and the population as a whole (through statistics, policy, regulation). This is pastoral power with paperwork.

The sovereign bargain

Carl Schmitt sharpened the political logic to a formula: protego ergo obligo — I protect, therefore I obligate. This is, Schmitt claimed, the cogito ergo sum of the state. The state's fundamental claim to existence rests on its capacity to protect; the citizen's fundamental obligation to obey rests on the state's provision of that protection. "If protection ceases, the state too ceases, and every obligation to obey ceases."

The formula is elegant and apparently reciprocal — a mutual exchange of protection for obedience. But the reciprocity is illusory, for a reason that Quentin Skinner has identified in Hobbes and Republican Liberty (2008): Hobbes redefined freedom so that surrendering autonomy for protection does not register as a loss. Freedom, in Hobbes's framework, is the absence of physical impediment — not the absence of domination. As long as no one physically stops you, you are free, even if your obedience is coerced by the threat of unprotection. This is the intellectual infrastructure for modern security-state rhetoric: the cost of protection is redefined so that it is no longer experienced as a cost.

The protection racket

Charles Tilly, in "War Making and State Making as Organized Crime" (1985), made the latent violence of the sovereign bargain explicit. States, Tilly argued, are "quintessential protection rackets with the advantage of legitimacy." The word "protection" has two faces: the shelter provided by a powerful friend, and the racket in which a strongman forces merchants to pay tribute to avoid damage that he himself threatens. "When the threats against which a given government protects its citizens are imaginary or are consequences of its own activities, the government has organized a protection racket." Banditry, piracy, policing, and war-making, Tilly insisted, all belong on the same continuum.

Tilly's contribution was to collapse the distinction between legitimate protection and criminal extortion — not by claiming they are morally equivalent, but by demonstrating that they are structurally identical. The mechanism is the same: create or exaggerate a threat, offer protection from it, extract compliance as payment. What differs is only the scale and the legitimacy granted by institutional recognition.

Cross-domain comparison

The pastoral prototype, the sovereign bargain, and the protection racket establish the theoretical architecture. But the argument of this essay is that protection regimes operate not only in the domains where political theorists have looked for them — state power, sovereignty, international relations — but across a far wider spectrum of human organisation. To make this claim precise rather than merely suggestive, it is necessary to compare domains along shared analytical dimensions rather than simply listing them.

The following table applies the essay's central diagnostic — can the protected say no? — across six domains, using five dimensions that operationalise the question:

Table 2

Dimension	Religion (pastoral)	Family (Confucian)	Domestic politics (security state)	Technology (security software)	Commerce (insurance)	Colonialism (protectorate)
Who defines the threat?	God / clergy	Father / elders	State security apparatus	Software vendor	Actuarial models	Colonial power
Can the protected consent in advance?	Baptism as infant; conversion as adult (but apostasy is punished)	Born into the family; no prior consent possible	Citizenship mostly by birth; naturalisation involves oath	Bundled with purchase; consent via EULA few read	Contract signed, but often legally mandatory (e.g. car insurance)	Imposed by treaty or conquest; no consent of the population
Can the protected exit, and at what cost?	Apostasy: social death, sometimes physical death (varies enormously by period and denomination)	Filial impiety: gravest moral failure in classical ethics	Dissent: ranges from electoral opposition to criminalisation as sedition	Uninstall: possible but system degrades; OS updates: increasingly difficult to defer or disable	Cancel: possible for voluntary policies; legally penalised for mandatory types (e.g. car insurance)	Independence: often required war, framed as ingratitude
To whom is the protector accountable?	God (interpreted by protector's own institution in most traditions)	Heaven (interpreted by ruler's own officials, though censorate provides partial check)	Electorate in democracies; much weaker mechanisms in authoritarian states	Market competition; regulators (weakly)	Regulators; contract law	International opinion (weakly); usually no binding mechanism
Can the protected participate in defining the threat and the protective response?	Varies: from none (infant baptism) to significant (adult conversion, congregational governance)	Remonstrance exists but constrained within paternal vocabulary	Varies by regime type: democratic participation vs. authoritarian exclusion	Minimal (users can configure settings but not define threat models)	Minimal (policyholder cannot define what counts as a covered risk)	None (the protected population was not consulted)
Does protection end when the threat ends?	Generally no (salvation is permanent; the soul is always at risk)	No (filial obligation persists until death; ancestor veneration extends it further)	Tends toward permanence (emergency powers rarely expire — Agamben)	No (updates are continuous; "protection" is effectively a subscription)	Yes, nominally (policy has a term); but renewal is expected or mandatory	Tended toward permanence in practice, though the formal category did not specify duration

This table does not claim that these six domains are "the same." Insurance regulation and divine grace operate through entirely different mechanisms, in entirely different institutional contexts, with entirely different histories. The claim is narrower: when examined along shared dimensions, these domains exhibit a family resemblance — not identity but a recognisable pattern. Each domain exhibits several elements of the pattern, though no single element is present uniformly across all six. In most cases, the threat is defined primarily by the protector, consent is constrained or absent, exit is costly, accountability flows more readily upward than downward, participation of the protected in defining the protective response is limited, and the protective arrangement tends to outlast the threat that ostensibly justified it. The exceptions — democratic accountability, voluntary insurance cancellation, adult religious conversion — are real and important; they mark the points at which specific institutions have partially succeeded in building checks on protective power.

Three observations emerge from the comparison.

First, religious protection is the genealogical prototype from which other forms descend. Foucault's genealogy traces pastoral power from the Christian church to the modern state. Talal Asad, in Formations of the Secular (2003), argues that the secular state absorbed the protective function previously claimed by religious authority, transferring the protection-obedience exchange from God to sovereign. The family sits between religion and state as the connective tissue: the Confucian concept of 父母官 (fumu guan, "parent-official") makes the bridge explicit. Officials are parents; subjects are children; governance is nurture; punishment is discipline. The eight characters 雷霆雨露，俱是天恩 — "thunder and rain are both Heaven's grace" — simultaneously invoke the religious layer (Heaven), the familial layer (grace as parental beneficence), and the political layer (thunder as punishment, rain as reward), and their function is to deny the protected the right to distinguish between protection and harm. Both are grace.

Second, the "sunset" row reveals a characteristic tendency of protection regimes: they persist beyond their justifying conditions. Emergency powers become permanent (Agamben). Software updates never stop. Insurance must be continuously renewed. Filial obligation extends beyond death into ancestor worship. The protective arrangement, once installed, generates institutional and affective momentum that sustains it independently of the threat it was created to address. This is not inevitable — some protective arrangements do end — but the tendency toward permanence is structurally built in, because the protector's identity (Mechanism 1) and the relationship's reorganised asymmetry (Mechanism 2) both create incentives to continue protecting whether or not the original threat persists.

Third, the colonial "protectorate" is revealing precisely because of its name. A "protectorate" is a political entity whose official designation contains the word "protection." The naming performs the entire operation analysed in this essay: it frames conquest as care, forecloses the word "invasion," installs an asymmetric relationship (the "protecting" power controls foreign affairs and defence; the "protected" entity retains nominal sovereignty), and makes the independence movement — the refusal of protection — appear as ingratitude or betrayal. The protectorate is protection-as-organisational-logic at its most linguistically transparent.

IV. Protection as Operating System

In 2001, the cultural theorist Azuma Hiroki published Otaku: Japan's Database Animals (動物化するポストモダン), arguing that in the postmodern condition, consumers no longer consume stories (grand narratives) but extract affective elements from a database — cat ears, twin-tails, tsundere personality, maid outfits — and respond to them as stimuli, without the mediation of meaning. This was the "animalisation" of the postmodern subject: stimulus-response without interpretation.

Azuma's database model is powerful for analysing the surface layer of popular culture — the differentiating attributes that distinguish one character from another, one product from its competitors. But it cannot account for what lies beneath those attributes: the non-differentiating infrastructure that all products in a genre share and that no consumer selects because no alternative is offered.

In otome games (乙女ゲーム) — the Japanese dating simulation genre where a female protagonist navigates romantic relationships with multiple male characters — a recurrent convention is that pursuable male characters undertake, in some form, to protect the heroine. Hakuouki: Shinsengumi Kitan (薄桜鬼〜新選組奇譚〜, PS2, 2008) organises all six of its routes — Hijikata, Okita, Saitō, Tōdō, Harada, Kazama — around the male character's guardianship of the heroine Yukimura Chizuru. She carries a sword throughout, and its function in the plot is not defence: it is one of the Yukimura clan head's tokens, two centuries old, of which she holds one and her brother the other, and what it does is identify her — marking her as oni rather than human, and therefore as the rare blood that both sides want. The object that looks like her capability turns out to be the certificate of her value to other people. Amnesia (PSP, 2011; released in English as Amnesia: Memories in 2015) runs the convention through five routes at once, and its most-discussed route is the limit case: Toma confines the amnesiac heroine to a cage in his apartment and justifies it, in so many words, as keeping her safe. Code: Realize ~~Guardian of Rebirth~~ (2014) puts the word in the title and builds its premise on a heroine whose body is lethal to anyone who touches her — protection as the precondition of the romance rather than an episode inside it.

This is not a character trait that differentiates one route from another; it is closer to an entry ticket to the category of "pursuable character." A player does not choose a route because "this character will protect me" — most of them will. The vocabulary is there at the genre's origin, though not yet pointed at the heroine. In Angelique (1994), the game generally credited with founding the genre, the nine love interests are 守護聖 — "guardian saints," rendered in the English materials simply as Guardians — but what they guard is the territory the heroine is cultivating in her contest to become queen, not her person. Protection was the job description from the beginning; what the genre added later was the transfer of its object from the world to the woman. Departures in the other direction do exist: Love and Deepspace (恋与深空, 2024) gives its heroine professional combat abilities and positions her as a fighter alongside her love interests rather than a charge in their care.

Nor is the naming confined to games marketed to women. In Katekyo Hitman Reborn! (家庭教師ヒットマンREBORN!, Weekly Shōnen Jump, 2004–2012), the innermost circle of the Vongola crime family are titled 守護者 — Guardians — one for each of seven ring-borne attributes, and the boss is not the figure they surround but one of them, holding the Sky. The rival organisation is built the same way: the Millefiore executives who correspond to the Vongola Guardians hold the parallel Mare Rings across the same attributes, six of them under a boss who holds the Sky. That symmetry rules out the reading that protection is the protagonists' particular value. It is the setting's only available word for the position of closest ally, and both sides use it.

It also complicates the asymmetry, and the complication is worth keeping rather than smoothing away. These Guardians fight beside the person they guard, who outranks them; the protective vocabulary here attaches to something much closer to a peer relation than the otome pattern does. The word travels further than the hierarchy does — which is a possibility Section VI takes up directly.

This distinction matters theoretically. In Azuma's framework, a moe element is a database entry: optional, differentiating, selected by the consumer from a menu of alternatives. "Protection" does not function as a moe element in this sense. It is not in the database — it is in the database's operating system: the infrastructure that determines what kinds of entries the database can contain. You can choose between a tsundere protector and a gentle protector, but the option "no protector" is rarely available.

The hero's structural dependency on protection

The operating-system quality of protection extends far beyond otome games. It is embedded in the deep structure of heroic narrative itself.

Joseph Campbell's monomyth — separation, initiation, return — is often summarised as a story of individual transformation: the hero leaves the ordinary world, undergoes trials, and returns transformed. But the return stage is structurally a protection function: the hero brings back a boon (a gift, a power, a piece of knowledge) that transforms or saves the ordinary world. The hero exists, in the monomyth's own logic, in order to protect the community. Without something to protect, the hero often loses a central narrative function.

Robert Jewett and John Shelton Lawrence identified a specifically American variant in The American Monomyth (1977): "A community in a harmonious paradise is threatened by evil; normal institutions fail to contend with this threat; a selfless superhero emerges to renounce temptations and carry out the redemptive task; aided by fate, his decisive victory restores the community to its paradisiacal condition; the superhero then recedes into obscurity." They argued that this pattern is anti-democratic: it presupposes that ordinary institutions are inadequate and that salvation requires an extraordinary individual. In the successor volume, The Myth of the American Superhero (2002), Lawrence and Jewett pressed the point further, describing the mythology's addressees as "crusading loners, attracted by guns, bombs, and the call to destroy evil." The community cannot save itself. It must be saved — which is to say, it must be protected.

The structural dependency runs in both directions. Protection needs heroes (someone must wield the power that "protecting" requires), and heroes need protection (without something to protect, the hero has no purpose). This mutual dependency helps explain why protection functions as an operating system rather than an optional feature: removing it destabilises much of the narrative architecture that heroic genres depend on.

What happens to heroes who refuse to protect

If protection functions as operating system in heroic narrative, then heroes who refuse the protective function — who rebel against the order rather than defending it — should be difficult for narratives to sustain. In practice, they frequently are.

Rebellious heroes exist: Prometheus steals fire from the gods; Sun Wukong wreaks havoc in Heaven; Robin Hood robs the rich; V (in V for Vendetta) destroys Parliament. Their animating principle is not protection but transgression — action directed against the prevailing order rather than in its defence. But their narrative fates follow a remarkably consistent pattern:

Punishment: Prometheus is chained to a rock in the Caucasus, his liver devoured daily by an eagle.
Co-optation: Sun Wukong is imprisoned for five hundred years under a mountain, then released on the condition that he serve as protector of the monk Tripitaka on the pilgrimage to the West. The rebel becomes a guardian. The Tightening Headband Spell (紧箍咒) — which causes agonising pain when the monk chants the sutra — is the disciplinary technology that converts rebellion into service. The trajectory is the novel's own: rebellion → imprisonment → domestication → protection.
Death: Robin Hood, in most versions, is reconciled to the returning legitimate king — or dies. V for Vendetta complicates the pattern rather than completing it: V is killed before the final act of destruction, which Evey carries out after taking the mask. The individual rebel is eliminated, but the rebel position is inherited — one of the places where the trichotomy leaks.

The pattern — punishment, co-optation, or death — is not a law, but the tendency is pronounced: what occurs less often is the sustained existence of a hero who refuses the protective function and is allowed to continue refusing it within the narrative's own terms. Permanent rebellion is difficult for narratives to sustain because it challenges the necessity of the order that the narrative itself typically presupposes.

The pattern is legible without any single critic's framework. The narrative performs a service for the order that the rebel threatened: it eliminates the rebel so that the protector — the order — can continue. The rebellion is permitted, even celebrated, as long as it is temporary. The system absorbs the transgression, processes it, and returns to equilibrium — an equilibrium in which someone, once again, is protecting someone else.

"I don't need your protection!" — resistance as diagnosis

Within the operating system of protection, apparent resistance does occur. Across shōnen and shōjo manga and in otome games alike, female characters refuse protection in so many words — I don't need you to protect me. This appears to be a direct challenge to the protective structure. But a closer examination reveals that it is more accurately described as what Max Gluckman called a ritual of rebellion (1954) — a socially sanctioned performance of protest that ultimately reinforces the system it appears to challenge.

Gluckman observed that in the Swazi Incwala ceremony, subjects performed ritualised acts of rebellion against the king — acts that "obviously include a protest against the established order, yet are intended to preserve and even to strengthen the established order." The rebellion is licensed: it occurs within a frame that contains it, and its completion returns the system to stability. Slavoj Žižek's concept of "inherent transgression" extends this logic: every power system comes with an "obscene supplement" — a permitted zone of rule-breaking that is not external to the system but secretly sustained by it.

The female character's declaration that she does not need protection functions in the same way. She protests — and then the narrative resolves her protest through one of several characteristic moves. In Naruto, Sakura Haruno resolves to stop relying on others, trains for two and a half years under Tsunade, and becomes one of the story's most accomplished medical ninja — and the narrative goes on positioning her as support and healer rather than as an equal combatant in its decisive fights. In Bleach, Orihime Inoue's Hueco Mundo arc turns on her being taken by Ulquiorra and recovered by Ichigo. In Naruto's Pain arc, Hinata Hyūga reverses the protective dynamic by throwing herself between Pain and Naruto — and is struck down within moments, to be healed afterwards by Sakura. In Ouran High School Host Club (episode 8), Haruhi confronts a group of harassers on her own, is thrown from a rock into the sea, and is pulled out by Tamaki, who then upbraids her for having acted as though she were not vulnerable. In each case the heroine's bid for autonomy is permitted — even dramatised — and then answered by a demonstration that the protective structure was necessary after all. The protest is licensed. The conclusion is not in doubt.

What the narrative typically does not permit is sustaining the refusal — allowing the character to remain a protagonist while persistently declining the protective relationship. Yona of the Dawn (暁のヨナ) is the nearest thing to an exception here, and it is instructive that it is only a partial one. The heroine does move from sheltered princess to fighter and keeps that agency across the long run of the series — and acquires, in the process, four dragon warriors bound by prophecy to protect her. The refusal is granted and the protective structure is rebuilt around her in the same motion. More commonly the heroine's resistance functions as a safety valve: it releases the pressure of the asymmetry without altering the asymmetry itself.

Lila Abu-Lughod's methodological insight is useful here: resistance should be used not as evidence of freedom but as a "diagnostic of power" (1990). It tells us about the structure of domination, not about escape from it. The "I don't need your protection" moment diagnoses the protection structure — it reveals that the character is aware of the asymmetry — but it does not dismantle it. The diagnostic is permitted. The dismantling is not.

David Lyon's framework of "surveillance as care" (Surveillance Studies: An Overview, 2007) captures the structural tendency: surveillance exists on a spectrum from care (a lifeguard at a beach) to control (a police stakeout), and institutions routinely slide from one to the other while maintaining the rhetoric of protection. In the domain of security software, effective protection often requires privileged visibility and intervention — the capacity to scan files, monitor network traffic, and override user preferences — creating a structural tendency for protective capacity to become supervisory capacity. This is not an identity (security systems can be designed with privacy-preserving architectures, minimal-privilege principles, and local processing), but it is a tendency that the rhetoric of protection makes difficult to challenge: "we need access in order to protect you" is a claim that carries its own moral warrant.

V. Why Protection, When Other Justifications Fail

If protection is an organisational logic as old as the pastorate, what is specific about the present? A common answer is that protection discourse has intensified since the turn of the millennium. This essay cannot show that, and the obvious rival case cuts against it: Cold War civil defence — shelters, drills, the protection of the free world — saturated public life more thoroughly than anything that has followed. No measurement is offered here, and the reader should not accept one by implication.

The narrower claim the following cases can carry is not about volume but about substitution: in each of them, protection took over justificatory work that some other vocabulary had been doing, and took it over as that other vocabulary lost its force. That is a claim about what protection replaced, not about how much of it there is.

The hypothesis — offered as such, not as a demonstrated fact — is that this substitution tracks a change in the conditions of legitimation. Specifically: where progress-based legitimation weakens, institutions have stronger incentives to shift toward protection-based legitimation.

The legitimacy-crisis hypothesis

Stuart Hall, Chas Critcher, Tony Jefferson, John Clarke and Brian Roberts's Policing the Crisis (1978) provides the template. They showed that the British state's crisis of legitimacy in the 1970s — economic stagnation, industrial unrest, the breakdown of the postwar social contract — was managed not by addressing its structural causes but by displacing it onto a moral panic about mugging: the interacting economic, social and political conflicts of the decade were relocated onto the problem of crime. The security apparatus expanded not because crime had increased but because legitimacy had decreased. In this case — and it is one of the few where the expansion was documented rather than inferred — the turn to protection was a symptom of the crisis, not a response to its cause.

The post-2000 global landscape is consistent with the same pattern at a larger scale. The collapse or discrediting of grand narratives — the end of the Cold War, the failures of liberal interventionism, the 2008 financial crisis, the erosion of democratic norms, the climate emergency — has produced legitimacy deficits across political systems. Institutions that can no longer plausibly claim to be delivering progress, justice, freedom, or prosperity have, in several visible cases, shifted their rhetorical weight toward the language of protection.

Four cases:

Table 3

Domain	Legitimacy pressure	Protection-rhetoric shift
Japanese popular culture	Bubble collapse + Aum Shinrikyo + Heisei recession	セカイ系 / 守りたい as narrative default
American politics	Cold War's end + 9/11	"War on Terror" / "protecting our freedom"
International law	Rwanda + Kosovo + failures of liberal order	Responsibility to Protect / R2P (2005)
European politics	Refugee crisis + terrorist attacks	Right-wing "protecting our culture / borders"

This is not a claim of strict causal determination — genuine increases in threat, institutional path dependence, and other factors are always at work alongside legitimacy pressures. The hypothesis is more modest: where other legitimating vocabularies weaken, protection-based legitimation tends to become more prominent, because protection has properties that make it especially resilient in conditions of justificatory scarcity.

Protection after deconstruction

There is a deeper layer to this substitution, one that connects it to the broader intellectual history of the late twentieth century. Postmodern thought — particularly in its deconstructive variants — registered, analysed, and in some contexts accelerated the loss of self-evidence surrounding modernity's grand narratives: progress, reason, universal values, objective truth. These had served as the legitimating vocabulary of modernity. As they came to be understood as historically contingent, power-laden, and culturally specific — a process to which deconstruction contributed but which it did not single-handedly cause — they became less effective as self-evident justifications for action.

But the need for action did not disappear. If anything, it intensified — under the pressure of neoliberal competition (which Uno correctly identified as a driver), ecological crisis, and geopolitical instability. The gap between the necessity of action and the impossibility of justifying it is the structural condition of the post-grand-narrative era.

"Protection" fills this gap — but the reason it does so, rather than other possible responses to the same gap, requires specification. After all, protection is not the only available posture in a world without grand narratives. Solidarity, resistance, mutual aid, flight — these are also available, and none of them depend on grand-narrative justification. Why does protection tend to become an increasingly prominent grammar rather than, say, solidarity?

Three properties distinguish protection from its alternatives and explain its competitive advantage in the post-grand-narrative landscape.

First, protection creates asymmetry, which is what order requires. Solidarity presupposes equality — it is a horizontal relation between peers. Resistance challenges existing hierarchies. Mutual aid is reciprocal. None of these can serve as the organising principle of a system that requires someone to be above and someone to be below. Protection is especially well suited to this function because it simultaneously feels like virtue and installs a hierarchy. An order that has lost its other justifications still needs to reproduce its own structure — and protection is one of the few moral vocabularies that produces structure rather than flattening it.

Second, protection operates as a gesture rather than a proposition. Progress, justice, freedom, and truth are propositions — claims that can be argued, challenged, and refuted. "I am protecting you" is not a proposition in the same sense. It is a bodily orientation, an affective stance, a relational posture. You can deconstruct the concept of justice (what is justice? whose justice? justice for whom?). You cannot, in the same way, deconstruct the act of someone placing their body between you and danger. Solidarity is also a gesture — but solidarity does not create the asymmetry that power needs. Resistance is also a gesture — but resistance threatens order rather than sustaining it. Protection is uniquely positioned: it is a gesture (and therefore resistant to deconstruction) that produces hierarchy (and therefore useful to order).

Third, protection is more self-legitimating than most alternative postures. Solidarity requires an account of who counts as a peer. Resistance requires an account of what is being resisted and why. Flight requires an account of what one is fleeing from. Each of these, in other words, still requires justification — a "because" that can be challenged. Protection requires less: "I am protecting you" carries its own moral warrant more immediately than "I stand in solidarity with you" (which invites the question: on what basis?). This is not to say that protection requires no justification — it can be questioned, and this essay is an extended exercise in questioning it. But its justificatory threshold is lower than that of competing postures, which gives it a competitive advantage in conditions of justificatory scarcity.

Together, these three properties help explain why protection tends to become more prominent when other legitimating vocabularies weaken. It is not more true than the alternatives. It is more structurally useful: it reproduces hierarchy, it operates as gesture rather than proposition, and its justificatory demands are minimal.

Protection as symptom

If this analysis is on the right track, then protection's growing share of the justificatory work is evidence not that the world has become more dangerous — though it may also have — but that prevailing orders have become less confident in their ability to justify themselves through other vocabularies. This does not invert the common-sense reading so much as complicate it: protection may be doing more of the work both because threats have grown and because other forms of justification have weakened, and disentangling the two is an empirical task that varies by domain.

Wendy Brown's analysis of border walls in Walled States, Waning Sovereignty (2010) illustrates the complication. Walls proliferate, Brown argues, not because sovereignty is strong but because it is waning. The walls' actual function is "theatrical" — they project "power and efficaciousness that they do not and cannot actually exercise." They do not simply respond to xenophobia but "generate and intensify it." In Brown's analysis, protection infrastructure does not merely respond to threats — it participates in producing the perception of threats that justifies its existence.

The Copenhagen School of security studies formalises a related insight as "securitisation theory" (Buzan, Wæver, and de Wilde, 1998): security is a performative speech act, not merely a description of an objective condition. Any issue can be "securitised" — transformed into an existential threat requiring extraordinary measures — through the speech act of declaring it a security concern. The declaration does not only describe a pre-existing reality; it participates in constituting one. Securitisation theory was built for states, and the scale it was built for matters: a government declaring a war on terror commands instruments an anime protagonist declaring 守りたい does not. What carries across is narrower than the theory itself — only the observation that declaring protection participates in shaping the threat-perception that justifies it.

VI. Limits of the Concept

A framework that can explain everything explains nothing. If protection-as-organisational-logic is visible in religion, family, intimate relationships, narrative, domestic politics, technology, commerce, international economy, colonialism, and international politics, a reasonable reader will ask: is there anything it cannot explain? If not, the concept is unfalsifiable — and unfalsifiable concepts are not analytical tools but articles of faith.

Three kinds of relationships fall outside the scope of this analysis, and specifying them sharpens the concept by contrast. A fourth item follows them, and it is of a different sort: not a category the framework excludes but a case the framework may get wrong.

Genuinely reciprocal relationships. When two parties protect each other in roughly symmetric ways — where roles rotate, where either party can withdraw without moral penalty, where no durable hierarchy is installed — the framework does not apply. The framework targets structural asymmetry: a fixed protector and a fixed protected, bound by one-directional obligation. Recognising this distinction matters because the conflation of reciprocal care with asymmetric protection is one of protection's most effective rhetorical strategies.

Exit-able contractual relationships. When a relationship can be terminated by either party without moral penalty — a commercial transaction, a freely cancelable subscription — the framework applies weakly or not at all. What it targets is specifically the irrefusability of protection: the point at which exit is coded as ingratitude, impiety, or treason. Many relationships that appear contractual are in practice asymmetric (employment "at will" in a market with no alternatives), and the framework can help identify that gap — but it does not claim that all relationships contain hidden protection.

Protector-initiated dissolution. The framework analyses refusal from below but has less to say about dissolution from above — a parent stepping back, a state dismantling its security apparatus, an institution transferring power to those it formerly "protected." The three mechanisms create pressures against such dissolution (the protector's identity depends on continuing to protect), but the pressures are not deterministic. Why some protectors let go is a question the framework raises but does not answer.

A case, not a category. The three exclusions above are definitional: they say what the framework does not cover. A framework should also be put against a case it might simply get wrong, and there is one. Henrik Berggren and Lars Trägårdh have described the Swedish social contract as a statist individualism whose organising purpose is close to the inverse of what Mechanism 2 predicts. Welfare provision and family law were built to release the individual from dependence on family, employer, church and charity, on the premise — their "Swedish theory of love" — that equality and autonomy are preconditions of genuine attachment rather than obstacles to it (Är svensken människa?, 2006; in English as The Swedish Theory of Love, trans. Stephen Donovan, University of Washington Press, 2022). Here a protective apparatus is engineered to dissolve the asymmetry that this essay has been treating as protection's characteristic product.

The obvious rejoinder is that the dependence was relocated rather than removed: a citizen who no longer needs a family needs a state, and a state is the larger protector. That rejoinder is available and it is not obviously wrong. But it is worth seeing what accepting it costs. If every redistribution of dependence counts as confirmation, the framework can no longer fail, and the worry this section opened with returns in full. The more useful reading is narrower: the Swedish case shows that the asymmetry described in Mechanism 2 is a design choice rather than a necessary property of protection as such. Protective institutions can be built to reduce the protected party's dependence. The question is whether they are — and most, including most of those examined here, are not. That is a weaker thesis than "protection produces asymmetry." It is also one that can be checked against a case instead of illustrated by one.

Finally, a caveat about scope versus depth. The cross-domain comparison in Section III surveys six domains along shared dimensions. This breadth is the argument's strength (it demonstrates that the structural feature recurs across radically different contexts) and its vulnerability (it cannot offer the same depth of analysis in each domain that a specialist study would provide). The essay's purpose is not to replace specialised studies of pastoral power, Confucian political thought, or surveillance technology. It is to identify a structural feature that those specialised studies, read together, reveal — a feature that no single specialist domain can see because it crosses all of them. Each domain-specific claim is offered as a hypothesis to be tested by those with deeper expertise in that domain.

VII. Conclusion: Return to Uno

We began with Uno Tsunehiro's diagnosis: postwar Japanese animation is trapped in a "maternal dystopia" where male protagonists cannot achieve agency without a female figure to protect. Uno treated this as a Japanese problem — rooted in Japan's postwar condition, manifested in Japan's popular culture, addressable through Japan's cultural evolution.

The analysis offered here suggests that Uno was simultaneously more right and more wrong than he knew.

He was more right because the structure he identified — the dependency of the subject's agency on the act of protection — is not a local peculiarity of Japanese animation but a recurrent organisational logic visible across domains, cultures, and centuries. From the Christian pastorate to the Confucian family-state, from the Hobbesian social contract to the colonial protectorate, from the otome game to the operating-system update, a recognisable structural feature recurs: protection naturalises asymmetry, constitutes the protector's identity, reorganises relationships through durable obligation, and constrains the grammar through which the protected might refuse. Uno found the local outcrop. The vein runs deeper than he dug.

He was more wrong because his causal framework — the social-reflection model in which cultural products mirror social conditions — is both too direct and too narrow. Too direct, because the relationship between social conditions and cultural production is not one of reflection but of mutual constitution: 守りたい is not a symptom of postwar Japanese society but a participant in the reproduction of the structures it appears merely to represent. Too narrow, because the turn to protection in 2000s Japanese popular culture is not a uniquely Japanese response to uniquely Japanese conditions but part of a broader, uneven tendency in which legitimacy pressures across political systems, cultural industries, and international institutions make protection-based justification increasingly attractive.

His critics, too, were partially right. Maternal Dystopia has been faulted for treating individual works as material for a general theory rather than reading them on their own terms, and the charge has force. So does a second one, which this essay's own opening makes visible: a book that takes motherhood as its central structural category builds its argument on three male directors — Miyazaki, Tomino, Oshii — and derives what maternity means in postwar Japanese animation from how men imagined it. Its evidentiary spine is male. A study of maternal structure that draws its evidence that way reproduces, at the level of method, something close to the erasure it sets out to diagnose.

But these corrections do not diminish the force of Uno's central intuition: that protection, in the world he was examining, had ceased to be one narrative option among many and had become the condition of possibility for narrative itself. Protection is not a theme. It is the operating system. What this essay has tried to show is that a similar dynamic operates far beyond the world Uno was examining — that across the institutional contexts surveyed here, protection functions as a recurrent infrastructure, the default justification that remains available when others have been exhausted.

And in every case, the same question arises — the question that the logic of protection is structurally designed to prevent:

Can the protected say no?

The Hobbesian subject can, in theory, withdraw consent when protection fails — but Hobbes redefined freedom so that the withdrawal costs nothing conceptually but everything practically. The pastoral flock can, in theory, appeal to God against a bad shepherd — but the shepherd is the one who interprets God's will. The Confucian child can, in theory, remonstrate with the father — but as Sarkissian has shown, every passage cited to establish a right of remonstrance in fact requires remonstrating without opposing: the son speaks gently, and if he is ignored he is to remain reverent rather than disobedient. The duty to remonstrate is not a right to refuse. The otome-game heroine can declare "I don't need your protection!" — but the narrative will resolve her protest into romantic attachment. The citizen can oppose the security state — but will be told they are endangering the very safety that the state provides. The user can try to decline the software update — but the dialogue box will return tomorrow.

The goal of this analysis is not to abolish protection. People protect each other, and much of that protection is genuine, necessary, and good. The goal is to restore the grammar of refusal — to make it possible, in each of the domains examined here, for the protected to say "no" without that refusal being coded as ingratitude, impiety, treason, naivety, or system error. The question is not whether someone is protecting you. The question is whether you can decline — and what it costs you if you try.

雷霆雨露，俱是天恩。

Thunder and rain are both Heaven's grace. The formula has survived for centuries because it is, in eight characters, a perfect compression of everything this essay has tried to decompress: the fusion of punishment and care, the foreclosure of the distinction between harm and help, and the positioning of the subject as a recipient of grace who has no standing to ask whether the thunder was necessary — because even the thunder is a gift.

The task, then, is not to deny that rain falls or that some thunder clears the air. It is to insist, against the weight of the formula, that the person standing in the storm has the right to say: this is not grace. This is rain. And I did not ask for it.