Chapter 16 distinguished open exchange, external influence, and covert control. Judging whether a society can resist control also often relies on another grand word: atomization. It is used to describe individuals isolated from one another, the disappearance of shared values, and the weakening of intermediary organizations, and it does point to genuine risks. But if we search only for lonely individuals standing before the system, we will overlook the fact that family care, supply-chain credit, platform groups, professional ethics, and improvised mutual aid continue to connect people, and we will fail to see who maintains these connections and who controls them. Days of continuous heavy rain briefly paralyzed the transportation and payment networks. The official map showed that several community organizations had deregistered, and the report accordingly concluded that residents lacked social ties. Actual recovery, however, relied on building-level care groups, cross-company messaging among repair workers, vegetable vendors extending credit, remittances from relatives living elsewhere, and hospital professional networks. At the same time, these connections depended heavily on a few administrators, commercial platforms, and the women caregivers who always filled the gaps; they refute the claim that society has become fully atomized, while also exposing new concentrated fragility.
The "fluidized society" depicts a disturbing scene: once the ties of family, rules, and identity are dissolved, individuals are left only as replaceable flows. This image preserves the alarm over the disappearance of shared bearing capacity, yet it misdraws a complex society as a single state of matter, and it easily treats migration, weak ties, and identity change themselves as decline. It can therefore serve only as a narrative under examination, not as a conclusion. What actually needs to be checked is whether relationships can transmit resources and responsibility, whether key nodes are replaceable, whether subjects can enter and exit, and who bears the cost of maintenance. Formal fluidity does not mean that connections have disappeared, and formal stability does not prove that connections are still working.
This chapter romanticizes neither families, neighborhoods, nor traditional organizations, nor does it pronounce online relationships false. Connections can support subjects, but they can also exclude newcomers, propagate surveillance, and demand unlimited sacrifice. The harsher possibility must be stated: power need not destroy all relationships; it need only route key connections through a few interfaces, exhaust the caregivers, and keep groups from checking one another horizontally, and in a crisis compliant concentration or even a strongman takeover may emerge. Analyzing this risk requires evidence about relationships and permissions, rather than writing urban change directly as a unified conspiracy.
External relationships can serve as a fulcrum by which individuals verify reality, and horizontal relationships can serve as checks within organizations. This chapter then widens the scale to the infrastructure of living: the family care, informal credit, peer messaging, and improvised fill-ins ordinarily obscured by "personal life" — how exactly do they allow a city to keep going when formal systems fail, and on whom does the cost of maintaining them fall?
Atomization and Actual Relational Networks
A person who joins many groups may still have no one who can deliver medicine during a blackout; another person who rarely socializes may have a stable family, colleagues, and public services. The experience of loneliness deserves acknowledgment and cannot be negated by a count of contacts; institutional capacity, meanwhile, must be judged by commitment, responsiveness, resources, and substitutes. The two observations answer different questions. Treating emotional statistics directly as social structure misses concrete connections; drawing only network diagrams overlooks whether relationships can bear weight. Research and governance should state which layer they are measuring, rather than replacing a whole society with a single index.
The 2025 report of the World Health Organization's Commission on Social Connection likewise distinguishes loneliness from social isolation: the former concerns the subject's painful experience of a gap in relationships, the latter a deficiency in the objective quantity or frequency of social contact; the report also stresses that the two are distributed unevenly across age, region, and living conditions. This chapter does not use the report's global estimates to prove that Chengjiang City is already "atomized," nor does it reason backward from health outcomes to a single social cause. The calibration it offers is that loneliness, solitude, number of contacts, relationship quality, and public bearing capacity are not one variable. Any judgment that declares society intact or collapsed on the basis of a single headcount, group count, or loneliness score has first confused its object of observation.
The vegetable vendor and the residents are not close, yet repeated transactions allow short-term credit; repair workers exchange fault information across companies and may never have met. Limited, low-affect relationships can transmit knowledge and resources without first forming a shared identity. If the atomization narrative recognizes only thick community, it cannot see the vast support that underpins modern collaboration. Weak ties are also easily severed when costs rise. Institutions can use contracts, insurance, public credit, and multiple points of entry to improve continuity, without requiring every relationship to be converted into kinship.
The family that cares for the elderly during the rainstorm provides support that public systems cannot immediately cover; the same family may also control members' communications, finances, and ability to leave. Relational density is not the same as subject capacity. Evaluation should look at refusal, privacy, the distribution of responsibility, and external help-seeking, rather than taking family integrity as proof of social health. Criticizing family power must not hand members back to an isolated market either. Public services and independent relationships allow people to retain voluntary intimacy, without making intimacy the sole point of entry for survival.
Mutual aid may be active within each building while no one knows that the neighboring district is enduring the same cutoffs; the platform keeps groups apart, and the administrative department receives only individual tickets. Members are not lonely, yet public problems are still sliced into private difficulties. Horizontal contact and a common master record determine whether local relationships can form a public subject. Group connection does not require merging all identities. It need only allow similar objects, sources, and responsibilities to encounter one another, while preserving difference and exit.
Family Ties Conceal a Large Amount of Institutional Labor
After transportation shut down, hospitals discharged patients early into the care of family members, and children out of school were looked after by relatives. Institutional statistics record services completed, yet never record the time and income of those receiving them. The family is not a costless buffer; its capacity is determined by housing, health, work, and the historical division of labor. Short-term mutual aid can arise from genuine love and commitment, and should not be dismissed wholesale as exploitation. The questions are whether institutions normalize the exception, whether they provide refusal and alternatives, and whether the benefiting institutions participate in maintaining the conditions of care.
Relatives living elsewhere help pay for medicine and expand local options; remittances may also come with demands concerning occupation, marriage, or residence. Dependence itself is not domination — what matters is whether the conditions are relevant, whether the subject can refuse, and whether basic life has only one path. Public provision reduces the chance that private help acquires total explanatory authority. Cross-territorial families also refute the equation of outmigration with the disappearance of relationships. Migration reorganizes connections and produces new maintenance labor and time-zone costs. Cities should observe how relationships change rather than merely counting the resident population.
A relief registry that accepts only the views of the household head may improve speed, yet it renders invisible those being cared for, tenants, and people whose family form does not meet the standard. The family can serve as a unit of service, but members still need individual points of entry for high-consequence matters. Emergency proxies should have scope, records, and after-the-fact confirmation. Requiring everyone to apply separately would also add burden. Limited proxy combined with independent complaint is more workable than either totalizing the family or fully individualizing it.
Younger members willing to shoulder recovery because they were once cared for make commitments of genuine moral weight; if elders, institutions, or the community use gratitude to demand a permanent renunciation of life, the historical relationship becomes a debt one cannot exit. Concrete care, capacities, substitutes, and the needs of all parties should be discussed openly. The freedom of the subject does not require denying the help received. Gratitude and boundaries can coexist; the costs of continuing tradition will be treated further in the next chapter.
Supply Chains Connect Strangers into Mutual Dependence
Vegetable vendors extend credit, drivers prioritize transporting medicine, and repair contractors defer settlement, relying on repeated reputation and professional judgment. The market does not consist of instantaneous prices alone; crisis capacity comes from long-term relationships, inventories, standards, and dispute handling. Treating market actors as isolated egoists makes this shared maintenance invisible. These commitments may also exclude newcomers without history. Public emergency procurement, basic rationing, and transparent eligibility can supplement acquaintance credit, so that relational capital does not decide survival.
The city distributed repairs among multiple tiers of contractors, yet the workers still coordinated through informal groups. The organizational chart looks dispersed, but operations depend on a single dispatcher and the same warehouse. When failure occurs, each firm can claim responsibility for only one segment of the contract. The existence of relationships does not mean responsibility is clear. A responsibility map should track information, materials, decisions, and benefits, and state who can change the conditions. Complex supply chains require a division of labor; complexity cannot be used to declare that no one bears the overall interface.
Selecting only the lowest bid year after year, standby suppliers and local repair capacity gradually withdraw. Ordinary costs fall, and in a crisis the connections snap simultaneously. Redundancy is not a matter of the more the better; it must be configured against irreversible loss and recovery time. The price ledger should include continuity capacity rather than treating it as a free background. Protecting the local can also become a means of excluding competition and entrenching interests. The object of audit is function, substitution, and responsibility, not something automatically decided by territorial identity.
The party that controls payment, logistics, and certification need not directly command residents; it need only change the conditions of access for multiple domains of life to contract at once. In severe cases it can bind eligibility for resources to political or cultural expressions. Judging this kind of control requires evidence of rules, permissions, selective consequences, and alternatives; one cannot presume malice merely because infrastructure is concentrated. Governance can separate critical eligibility, provide a public floor, and establish independent complaint. The more concentrated the connections, the greater the duty of explanation and the stricter the exit conditions required.
Platforms Have Not Abolished Relationships but Rearranged Them
Building groups post requests for medicine, check on the status of the elderly, and arrange pickups; commitments formed through digital media are not false for being online. Evaluation looks at response, responsibility, and consequence, not at whether bodies share a location. Platforms also let professionals far away enter into local support. Online relationships also depend on electricity, accounts, and ranking. Exporting contacts, retaining offline entry points, and redundancy across media ensure that platform failure does not mean the simultaneous disappearance of relationships.
Administrators decide who joins the group, which message is pinned, and how conflicts are handled; they perform necessary labor and also hold the power of visibility. If a community treats the role as purely technical, it overlooks questions of representation and privacy; calling all administration arbitrary makes it impossible to maintain order. Rules, rotation, logs, and appeal make the permissions limited. The administrator's private life likewise needs boundaries, and cannot be required to provide round-the-clock service merely because a public role has been assumed. Organizations should share the labor rather than letting relational continuity press on a single person.
Platforms help similar needs find one another and reduce isolation; if ranking only reinforces similar sentiment, the group becomes unable to check other scenes. Active search, chronological order, shared public channels, and cross-group intermediaries can provide comparison. The effect of algorithms must be judged from rules and data evidence, not inferred directly from usage outcomes. Public institutions cannot take platform popularity as the whole of opinion. Those who are offline, the silent, and demoted material need other points of entry.
A screenshot of one argument recirculates in a new crisis, and the subject is defined by an old position. Shared memory requires decisions, corrections, and accountability, and need not permanently preserve every private conflict. Purpose, retention period, and context protect people's capacity to change, and also prevent administrators from rewriting history at will. Deletion and accountability are not an either-or choice. Public decisions and records of serious harm can be preserved under control, while ordinary social material enjoys a stronger boundary of forgetting.
Professions, Intermediary Organizations, and Crisis Connections
Hospital staff discover through professional networks that similar drug shortages have appeared across several regions; reports from individual institutions have not yet been aggregated. Horizontal professional contact provides early signals, and can also propagate unverified shared impressions. Information should be labeled with source and status, with early warning and formal determination layered separately. Banning all informal communication turns the hierarchy into a single point; allowing all materials to leak harms privacy and security. Controlled professional channels require clear objects and responsibilities.
Repair workers prioritize life-safety equipment rather than merely obeying their employer's immediate returns. Professional standards allow strangers to expect certain behavior, forming social connection. Standards may still maintain industry thresholds or ignore differences on the ground, and need joint revision by the public, users, and practitioners. Professional autonomy is not freedom from accountability, but the assurance that judgments of knowledge do not change entirely with the command of a single organization. An independent ethics channel can protect those who refuse overreach under internal pressure.
Institutions that do not recognize one another's training leave capable people unable to enter during a crisis; abolishing qualification altogether may increase risk. Mutual recognition, temporary authorization, and after-the-fact review can widen connections. They require clear responsibility, and must not permanentize the emergency exception. If qualification is monopolized by a single platform, professional networks come under the control of its ratings. Multiple forms of certification and appeal reduce the single gate.
Technical shorthand speeds up communication among peers, yet leaves residents unable to judge decisions. Intermediary translation should preserve the raw material, its scope, and its uncertainty, rather than seizing total explanatory authority through simplification. The public need not master every professional detail, but should still know what bears on rights, risks, and choices. Anti-professional language also severs critical connections. Demanding that experts explain, rather than demanding that they renounce knowledge, better forms shared judgment.
Intermediary Organizations Do Not Automatically Represent the Public Interest
Residents' committees, unions, trade associations, and mutual-aid networks can aggregate experience, preserve timelines, and negotiate with government. Individuals need not understand the entire system alone. Part of the risk of atomization lies precisely in the insufficiency of this translational capacity. The number of organizations that exist does not prove function. One must ask whether members can grant authority, whether dissent is preserved, whether materials return to their sources, and whether the organization actually changes decisions.
Long-standing spokespersons master policy language and external resources, and easily mistake their own judgment for the membership as a whole. Term limits, feedback, multiple entry points, and disclosure of interests can restrain this. That representatives hold discretion is required for efficiency; decisions that far exceed their mandate must still return to the members. When government receives only a few "mature partners," it also shapes the public through its choice of representatives. A relationship of convenience cannot substitute for authorization.
Mutual-aid organizations depend on government contracts or foundations; funding may support real services, or may steer them away from unfavorable issues. A funding relationship does not prove control — one should examine conditions, the power to withdraw funds, content permissions, and actual change. Multiple sources and portable results reduce the capacity of any single party to shut things down. Demanding that organizations have no dependencies at all would allow only the wealthy to participate. The goal is dependence that is visible, limited, and subject to member oversight.
Old neighborhoods coordinate quickly through acquaintance trust, yet exclude tenants and newcomers; trade associations protect members, and may also push costs onto consumers and informal workers. Strong internal connection can coexist with public atomization, because acknowledgment between groups is missing. Evaluation must examine who bears the consequences outside the boundary. Open membership need not abolish all requirements. Conditions that are task-relevant, accessible, and appealable differ from inherited identity and secret loyalty.
Crisis-Time Concentration Often Builds on Existing Connections
When platforms, supply chains, and intermediary organizations all depend on a few interfaces, a single coordinator can rapidly concentrate information and resources in a crisis. Concentration may solve genuinely urgent problems, or may extend temporary capacity into long-term obedience. Analysis needs to trace the preexisting networks and authorizations, rather than writing the public as atoms suddenly surrendering their freedom. If the strongman narrative blames only individual psychology, it misses the alternative paths that institutions have already deleted; if it blames only structure, it cancels the responsibility of those who actively expanded their powers and of their supporters. Both layers exist simultaneously.
When multiple agencies shift blame to one another and versions of the information conflict, people may choose the party whose promises are clear. That choice can be rational, especially when danger is imminent. The questions are whether guarantees can be verified, whether permissions are temporary, and whether errors have an external point of entry. Clarity does not equal truth, but ambiguous institutions do genuinely create the demand. Repair cannot merely criticize the public for believing simple answers; it must also restore polycentric coordination and intelligible responsibility.
The residents' committee still hangs its sign but only forwards official notices; the trade association still meets but has lost independent funding and the right of appeal. The number of organizations is stable, while actual horizontal capacity has already contracted. An audit should ask whether an organization can raise unfavorable material, protect members, and refuse demands beyond its mandate. Absorption need not come from conspiracy either. Contractual convenience, personnel flows, and resource squeeze can produce it gradually. The absence of unified intent does not remove the need to repair the structure.
A crisis committee issues orders rapidly through families, enterprises, and platforms; the connections speed execution, and may also require every intermediary to surveil its members. Once caring relationships carry political qualification review, people hide information and long-term trust declines. Security reports should target specific risks, not fold ordinary dissent into relational obligation. Where serious threats exist, limited information sharing may be necessary. Object, duration, and review prevent one mobilization from permanently rewriting all relationships.
The Quality of Connections and the Risk of Propagation
A group that only forwards central commands has connections of transmission, but lacks connections of feedback. Whether workers can report supply risks, whether residents can correct the roster, and whether administrators can admit the unknown determine whether the network touches reality. Unfavorable information need not always be adopted, but it must have somewhere for a response to go.
The same caregiver, dispatcher, or volunteer continually filling in appears relationally strong, while actually being sustained through exhaustion. Rotation, compensation, substitution, and refusal make connections sustainable. Commitments may involve asymmetric contribution, but that cannot be naturalized into permanent identity.
If internal cohesion is conditioned on cutting off external contact, errors reinforce themselves. Controlled professional exchange, shared public material, and multi-organization coalitions provide comparison while protecting privacy and security. Those who cross boundaries are not naturally neutral, and also need review of interests and permissions.
Leaving a family, a group, a platform, or an association should not simultaneously cost one all services and professional credentials. Specific commitments still have to be honored; relationships cannot acquire a master switch over survival. A public floor and multiple entry points make voluntary connection possible.
Latent Connections Require Conditions to Be Activated in Crisis
Contact lists and organizational rosters record relationships, but do not say whether the other party knows of the need, possesses resources, or is willing to bear responsibility. Preparedness for recovery should check channels of contact, response duties, substitutes, and failure referrals, while avoiding sustained surveillance of private relationships in the name of preparation. Public capacity requires limited commitment, not everyone online around the clock.
An institution that appears only at the moment of disaster can hardly ask residents to trust it immediately. Everyday service, honest promises, correction of errors, and low-cost appeal accumulate deployable credibility. Crisis propaganda cannot manufacture trust in an instant, nor can public hesitation be presumed to be manipulation.
A group knows where medicine is lacking, but without transport permission, payment, and receivers, information does not become help. A relationship map should connect knowledge to interfaces of action. Conversely, possessing resources without knowing what the scene needs also produces mismatch. Intermediary labor joins the two ends, and should receive budget and a position of responsibility.
Regular drills test contacts, roles, and substitutes; participants who know the procedures may then improve real capacity. If the pursuit is only attendance and elapsed time, caregivers, the offline, and non-standard families remain excluded. Drill results must allow failure to change resources; a success report cannot substitute for differences on the ground.
Overly Dense Connections Can Also Propagate Risk in Synchrony
Multiple groups appear dispersed, yet all depend on the same account and cloud service, and vanish simultaneously when these are suspended. External contact, offline lists, and different media provide redundancy. The quantity of redundancy should be proportional to the risk; one does not establish unlimited copies of data merely to prevent failure.
When all members cite the same professional group, an error can spread quickly across communities. A shared source is not necessarily unreliable, but it should be recognized as a single strand of evidence. Different professions, scenes, and public records correct one another, so that the speed of relationship does not overwhelm verification.
Members who have received help are expected to join every action, take public positions, or share locations, and connection turns from support into qualification. Specific reciprocity can be legitimate; total loyalty cannot be inferred from one instance of aid. Those who refuse additional demands should still retain basic services and exit from relationships.
A unified administrator can allocate resources quickly; if it also controls membership, information, and appeal at the same time, a single bias acts across domains. Splitting permissions, rotation, and external review limit the master switch. A coordination center should not be abolished; it should be correctable and replaceable.
Regrowing Actual Bearing Capacity
Demanding that everyone return to acquaintance neighborhoods may place newcomers, minority ways of life, and those who have left old relationships back outside the door. Recovery should retain the effective support of old networks, while also establishing public and professional entry points. Continuity of relationship is not conditional on purity of identity.
Paid care, online peers, and cross-territorial families can bear sustained responsibility, and are not false for being modern or commercial. Contracts, promises, exit, and actual response should be examined. Old forms require the same audit; age confers no moral exemption.
Leaving a violent family, a closed organization, or an exploitative supply chain reduces some connections while enlarging the possibilities of life. A network metric that rewards only the number of connections writes necessary exits as social loss. Evaluation should look at subject capacity, the handover of responsibility, and substitution, rather than pursuing maximum density.
A crisis mutual-aid group can become a standing organization, or can end once its task is complete. Continuation should be decided by member authorization, resources, and public need. Making crisis relationships permanent accumulates the interests of administrators and funders; immediate dissolution loses the experience. Archives can be preserved, with permissions exiting along with the task.
Returning from Network Illusions to Actual Bearing Capacity
Chengjiang redraws its recovery relationship map: beyond counting organizations and accounts, it records who transmits medicine, information, care, payment, and repairs, which nodes are replaceable, and who continually bears invisible labor. Family proxies have individual appeal, platform groups can export their data and rotate administration, supply chains disclose overall responsibility, professional networks retain controlled early warning, and intermediary organizations must demonstrate authorization and feedback. This map does not yield a perfect community. Relationships may oppress, weak ties may break, organizations will contend for resources, and platforms can hardly become fully transparent. What it prevents is using the claim that "people are already atomized" as a reason not to see the society still working, and crediting all recovery to a new center; it also prevents using the continued abundance of relationships to ignore critical interfaces, care depletion, and group isolation. The city should also observe whether the recovery of connection merely restarts old burdens. Groups becoming active again, supplies restored, and families resuming care do not necessarily indicate improved institutional capacity; if public institutions still hand care, information verification, and final delivery over to unpaid relationships, the post-crisis "social resilience" may be nothing but private depletion. Assessment should track who responds continually, whether rotation is possible, whether costs are formally acknowledged, and whether the next failure can still only summon the same people.
Connection also has a part that institutions cannot fully design. If friendship, trust, and mutual aid were all converted into indicators and tasks, subjects would perform relationships to satisfy recovery evaluation, and private space would shrink. Public policy can protect encounter, time, basic services, and conditions of exit, but should not prescribe which community each person must join. Social capacity comes from the conditions under which relationships can form, and from people's right to choose the concrete form of their relationships. The judgment of atomization itself therefore needs conditions of failure. If people can still transmit unfavorable information across institutions, help one another when the center fails, turn local experience into public issues, and leave oppressive relationships, one cannot declare society atomized merely because traditional organizations have declined. Conversely, when contacts are many yet none can bear weight, groups cannot check one another, and all entry points are controlled by the same platform, a high connection count cannot be used to declare the problem nonexistent. The harshest institutional risk thereby becomes concrete: when shared values and intermediary organizations fail, a crisis may indeed let a single coordinator concentrate power; but this process usually passes through existing families, supply chains, platforms, and professional networks, rather than occurring in a social vacuum. Protecting the capacity for public self-defense requires making connections plural in source, capable of feedback, open to exit, and bounded in what they bear — not re-bundling everyone into one total community.
Evaluating connections should not pursue a single aggregate score of social density. Care relationships, transmission of fact, delivery of resources, shared decision, and appeal each need different networks; the same relationship may be reliable for one task and overreach on another. By separating functions, a city can add the connections it lacks, restrain overly dense control, and acknowledge that some ruptures are the subject's protection. The critique of atomization thus turns from a nostalgic judgment into a testable audit of relationships and institutions. The value of a connection is ultimately tested by whether it increases the capacity of concrete subjects for access, feedback, sharing of burden, and exit. The next chapter discusses who bears the costs of continuing tradition: when the preservation of language, ritual, architecture, and relationships is regarded as a shared responsibility, how to prevent the labor of maintenance from falling once again on those with the fewest choices.