Chapter 14 examined repair after knowledge intermediaries fail. An information chain that can still be traced may nonetheless end at correct expression: the agency acknowledges residents' experience, the report adopts cautious language, the meeting permits dissenting voices, while the risks to water, income, time, and health continue to be borne by the same people as before. The truthfulness of a public commitment cannot be judged by wording alone; it must also be judged by how the commitment enters budgets, authority, deadlines, and actual lives. The water station issues a Declaration of Joint Repair, promising to "listen to every voice, restore trust, and guarantee safety." The declaration need not be hypocritical, and the staff may sincerely hope for change. Yet downstream residents must still purchase drinking water themselves and take leave to collect samples; temporary water supply opens only on weekdays; the enterprise receives subsidies for the work stoppage, while gig workers, having no income records, cannot apply. Correct language and asymmetric losses exist at the same time. This chapter does not use a fictional scene to indict real institutions, nor does it treat all public expression as public relations. Commitments can coordinate expectations, establish an accountable addressee, and help a community articulate harms that the old institutions failed to acknowledge. The questions are when language acquires an interface to action, who is able to test it, and the harsher possibility: that power can absorb critical vocabulary, substituting recognition for the transfer of resources, so that opponents appear to have already been heard.
Commitments and Actual Losses
"We hope to restore the supply as soon as possible" expresses a wish, whereas "three water collection points will be set up downstream within ten days" specifies an object, a quantity, and a deadline. Good will has relational value, yet it cannot by itself sustain dependence; a commitment allows others to arrange their lives around it, and therefore carries a heightened duty of explanation. An agency should distinguish goals, plans, and decisions for which resources have already been secured. Demanding that every sentence read like a contract would make public communication rigid. An agency may retain exploratory expressions, provided it does not use visions to obtain the reputation of having already acted. The more specific the language, and the more it is used to ask the public to wait, the more precisely it should correspond to demonstrable capacity.
Telling residents "comprehensive protection" while applying to the budget authority only for a limited pilot creates a gap of authority between the two versions. Language may be adapted to different audiences, but the core objects, deadlines, and exceptions should be checkable against one another. Otherwise every audience receives the commitment best suited to placating it, while no version answers for the overall consequences. A public record of versions makes change visible. When new material appears, a commitment may be revised, but the agency should state what changed, why it changed, and who had already invested in reliance on the old commitment. Updating is not itself a breach of faith; it is silent contraction that leaves the costs of adaptation to those who depend.
"Respect," "inclusion," and "joint repair" can establish a direction of value, yet they cannot show how water supply, compensation, and decision-making power actually change. Abstract words should be connected to observable actions while retaining the part of them that cannot be fully quantified. Respect includes service outcomes, but also whether people can understand, refuse, and appeal. Looking only at indicators, in turn, induces formal compliance. That an agency has held three meetings does not mean residents' opinions entered the decision. Value words and objects of action must limit each other: language keeps indicators from losing their purpose, and material keeps language from being immune to failure.
If the water station for a long time does not contradict the claim that "expenses will be reimbursed," and allows residents to keep receipts accordingly, reasonable reliance may form. Formal documents are not the only source of relational obligation. Evaluation should consider who has authority to speak, whether the information persisted, whether the agency knew that others were relying on it, and whether clarification was feasible. Not everything overheard counts as an institutional commitment. Unauthorized personal opinions, clearly conditional predictions, and outside speculation must each be recorded as what they are. An institution can reduce ambiguity through a unified entry point, but it cannot use internal authorization rules to cancel the public impression it has long tolerated.
Actual Losses Must Be Tracked Across Subjects and Time
That the valley's average water-quality indicators recover does not mean that every tributary, every time period, and every household recovers. Aggregate figures have coordinating value, but the distribution still has to be checked by affected location. If the most severely affected sites have few samples, average improvement may even appear at the same time as their deterioration. Grouping cannot be refined indefinitely to the point where judgment becomes impossible. The scale should be chosen by risk, by rights, and by the use to which the decision will be put, and anomalous material should be allowed to trigger drill-down. Aggregate and local correct each other; neither automatically stands for the whole.
Residents queuing for water, learning how to apply, and repeatedly recounting their experiences all consume labor, care, and recovery. Agencies readily record the monetary value of goods distributed while treating the time spent obtaining them as zero. A complete account of costs should include provable direct expenditures, and should also estimate, by proportionate methods, the lives occupied by the institution. Not every inconvenience can or should be fully monetized. Some time is better repaired through extended service hours, home visits, and single submissions reusable across offices. The value lies in reducing losses, not in putting a price on every relational experience.
Setting up temporary water points requires vehicles and personnel; if outsourced drivers transport water at night without protection, residents' risk falls while labor's risk is displaced outward. Evaluation must not look only at the target group but must also trace the supply chain through which the goal is pursued. Repairing one harm does not license the creation of another, invisible one. This does not mean that any residual risk invalidates the action. The institution must compare necessity, alternatives, distribution, and compensation, and give those who bear the risk an informed position and a place to refuse. When risk cannot be eliminated, it must not be absorbed by default only by those with the least power.
Temporary inconvenience, long-term income interruption, damage to health, and the rupture of relationships differ in recoverability. When budgets are limited, prioritizing the prevention of irreversible harm is a value choice that must be publicly justified. The largest number of people need not always take precedence, and the most severe harm does not automatically acquire unlimited authority over other needs. Judgment must also consider the buffers of the subjects. The same week of stopped work has different consequences for an enterprise with savings and for a day laborer. Equal amounts may not amount to proportionate support, and differentiated support must rest on explainable conditions rather than impressions of identity.
Budgets, Harm, and Accountability
An agency may propose a plan for which funding has not yet been appropriated, but it cannot use a definitive tone to ask the public to rely on it. The funding source, the stage of approval, the minimum executable portion, and the fallback in case of failure should be clear. If a commitment has already earned trust in circulation while holding no place in internal resources, the organizational structure itself is manufacturing false expectations. Nor does an approved budget mean that capacity exists. Personnel, authority, supplies, and maintenance can each become constraints. Executability checks should enter before release, rather than waiting for failure so that responsibility can be pushed down to the front line.
That after the declaration is issued the first investment goes into image design while the water points still lack vehicles does not by itself prove malice, but it does reveal the organizational convenience and priority of different tasks. Evaluation should look at persistent patterns, amounts, substitutability, and who decides. Communication also requires resources, and not all spending on dissemination should be counted as waste. The question is whether reputational labor helps the affected understand and use the repair, or mainly helps the agency announce that repair has already occurred. The two can be distinguished by content, audience, timing, and the interfaces that follow.
Accepting only payrolls quickly verifies formal labor losses yet excludes gig work, caregiving, and non-monetary income. Abolishing documentary requirements altogether, in turn, invites misallocation. The institution should offer multiple forms of proof, spot checks, and appeals, so that the cost of verification remains proportionate to the scale of compensation. The application procedure itself also distributes costs. Language, digital entry points, deadlines, and repeated submissions select those most able to persist. A low application rate cannot be read directly as absence of loss or as public satisfaction.
Equipment purchases and project launches are easy to see; filters, repairs, data updates, and rest for personnel determine whether service continues. A one-time budget with no place for maintenance lets commitments decay once attention moves on. An agency should state who is responsible for daily operation, when review occurs, and how handover proceeds after exit. Nor are permanent projects always better. Once a temporary risk has passed, resources should return to other needs. The lifespan of a commitment should be determined by its object, not by the self-perpetuation of the organization.
Acknowledging Harm Must Not Occupy the Affected's Right of Interpretation
The water station can acknowledge delayed notification, missed samples, and absent appeals without prescribing for all residents what they have suffered. A collective apology establishes institutional responsibility, while individual experiences may still differ. Overly grand language sometimes lets concrete responsibility vanish once more into "collective trauma." Nor may an apology demand immediate forgiveness, participation in ceremonies, or the cessation of accountability. Whether to accept belongs to the affected; what the apologizing party can control is acknowledgment, action, and non-repetition. Relational repair may occur, or it may not.
When an agency invites residents to tell their experiences, omissions can enter the public record; but if the stories serve only publicity while budgets and rules remain decided by the original team, the subjects become material for the image of repair. Participation should state which questions it can influence and which it cannot, and provide boundaries for labor, privacy, and withdrawal. Not every resident wishes to shoulder the work of governance. Refusing to participate must not cost compensation or services, and representative participation does not automatically carry everyone's mandate. Turning rights into a standing duty to speak punishes, once again, the most exhausted.
That an organization learns to use language such as "structural," "subjectivity," and "co-production" may reflect real learning, or it may give old procedures a new moral appearance. One cannot judge authenticity from the origin of vocabulary; one must examine whether authority, resources, and the consequences of error have changed. The more correct the language, the more it must be prevented from becoming a signal to stop asking. Opponents, too, may label every institutional learning as co-optation, so that change can never obtain acknowledgment. Critique should allow partial progress to stand while preserving what remains unfinished. Otherwise the institution will only ever choose between denial and performance.
Downstream residents are not a single subject. Some need immediate water supply, some care more about long-term monitoring; landlords, tenants, shopkeepers, and workers bear different consequences. Letting the most visible representatives define all needs hides internal asymmetry beneath a shared identity. The institution should accommodate multiple entry points and contestable representation. Internal difference does not make the shared harm unreal, but it does determine that repair cannot take only one form.
Accountability Must Connect Failure to Change
An agency can explain how budget delays, technical limits, and false information combined, and this helps prevent repetition. But an explanation that proves only that "no one acted intentionally" still does not answer who repairs the residents' losses. Causal explanation, moral responsibility, and the duty of compensation are related, yet they are not the same question. A no-fault accident may still require institutional assumption, because the agency configured and managed the risk. Deliberate concealment adds disciplinary or legal liability, but punishing individuals cannot substitute for repairing the interface.
Front-line personnel who execute ambiguous rules bear a responsibility different from that of the managers who approved the rules; an enterprise that profits from stoppage subsidies may also bear obligations to submit truthful material. Those closest to the public must not absorb the burden for the whole system, nor may structural complexity be used to declare that no one is responsible. A responsibility map can list who knew, who could change, who decided, who benefited, and who bore the consequences. It does not seek one total enemy; it matches different repairs to different subjects.
If an agency defines "holding a meeting" as listening and "issuing subsidies" as restoration, the indicators will guarantee their own success. Evaluation should include whether subjects actually received, whether problems were resolved, unintended losses, and the outcomes of appeals, and it should be set with the participation of the affected. Indicators, too, can be adapted to strategically. Retaining sampled narratives, external review, and indicator updates prevents one set of numbers from permanently answering for reality. Nor can the limits of quantification justify retreating to moving stories alone.
When a commitment goes unfulfilled, the agency should state the reason, the losses already imposed on those who relied on it, whether the commitment continues, and who approves the extension. Writing every extension as "continued progress" drains time boundaries of meaning. Some complex repairs genuinely require longer, and an honest extension is better than a hasty declaration of completion. If continuation is hopeless, the commitment should be terminated, with alternatives, compensation, and archives arranged. Admitting failure may damage reputation, yet it is a necessary condition for restoring language that can be relied upon.
From Acknowledgment to Remedy
The Qingya Valley can publish budget categories, service capacity, the numbers who applied and who actually received, maintenance deadlines, and the reasons for non-coverage, while protecting personal privacy. A ledger is not truer for being finer; what matters is that resources can be traced from commitment into particular lives, so that losses along the way become visible. External auditors can examine the records, and residents can point to the costs in time and relationships that the indicators do not see. The two kinds of knowledge must meet: professional accounting must not cancel lived experience, and experience must not replace financial fact.
Some need cash compensation, others need continuous water supply, health checks, work arrangements, or public correction. Fully individualizing every remedy is costly, while a single scheme transfers the differences back onto households. A basic uniform guarantee can be set, with additional pathways, bounded by documentation, for situations of high loss. The choice of form should involve the subjects as far as possible, but an agency cannot use a menu of options to evade an insufficiency of totals. Where resources fall short, both the principle of distribution and the size of the gap should be public.
The water points close on schedule and the water-quality indicators recover, yet residents may still carry debts, unemployment, or distrust of the reports. The end of a project does not end all its consequences at once. The institution should distinguish completed objects, residuals requiring referral, and losses that cannot be compensated, rather than covering the whole of reality with one closing chapter. Indefinitely extending repair can also fix identities of victimhood and place the community under permanent institutional observation. Archives, services, and monitoring should carry exit and data-deletion conditions, with the affected participating in deciding what is worth keeping.
If the water station promises the water supply, compensation, and monitoring it can actually deliver, states its unknowns and budgetary limits, and bears the consequences when it fails, its language acquires the weight of reality. It need not promise the restoration of all trust, nor may it require residents to read a limited repair as total reconciliation. A public commitment is not more sincere for being grander. Its value lies in making future action something others can rely on, examine, and correct. Actual losses, in turn, remind the community: correct meaning must pass through material, time, and the capacities of subjects, and cannot be completed on paper alone.
Compensation, Restoration, and the Transfer of Power Are Not the Same Thing
That residents receive water and compensation for lost work can ease the material losses already incurred; but if the next round of sampling, publication, and water shutdowns is still decided unilaterally by the original agency, the same structure of knowledge has not changed. Compensation cannot therefore be called useless; it simply answers a limited question. Repair must also examine who in the future can propose samples, trigger review, and change measures. Conversely, using power reform as a reason to refuse compensation asks the affected to trade their present lives for long-term institutional goals. Different repairs can proceed in parallel; budgetary conflicts require public choices, and present losses should not be swallowed by grand change.
If gig workers lacked income documentation and the downstream lacked sampling points before the incident, a simple "return to normal" rebuilds the original invisibility. Repair should distinguish the new losses the event caused from the old gaps it exposed. The agency's responsibility for the two may arise from different sources, and the greater scale of historical problems does not cancel current responsibility. Nor should the thorough resolution of all historical injustice become the precondition for honoring current commitments. Safety and income can first be restored, and structural questions then taken up in subsequent decisions with subjects and deadlines. Phasing is not the same as deferral, provided each phase carries real capacity and conditions of failure.
Having a residents' committee take part in monitoring widens the positions of observation, yet it can turn sampling, translation, and conflict management into unpaid community labor. Decision-making power must come with budgets, technical access, rest, and exit. Otherwise the agency, having devolved responsibility, can describe its own withdrawal as "co-governance." Differences of standing also exist within the community. Rotation, public delegation, and multiple entry points reduce the permanent occupation of experience by a few representatives. Criticism of representatives should be directed at describable decisions, not used to cancel the entire position of participation on the grounds that "the community is unprofessional."
Missed care, the long-term rupture of relationships, and irreversible damage to health cannot be restored by a price. Institutions can still compensate part of the costs, preserve records, change rules, and acknowledge that the loss cannot be returned to zero. Calling compensation "complete restoration" demands that subjects certify, by their receipt of payment, that the matter is closed. But the impossibility of complete repair cannot become a reason for abandoning action. The honest boundaries of limited compensation protect public language better than grand promises of reconciliation. A community needs to allow certain losses to remain as memory, without letting them be endlessly used to mobilize new sacrifice.
Trust Cannot Be the Sole Indicator of a Fulfilled Commitment
That residents still choose to test the water themselves after quality has recovered does not necessarily signify a refusal of common life. Past delays in information raised their evidentiary standards. If the agency writes restored trust as a requirement, it may once again recast prudence as a relational problem. It can provide stable material and appealable procedures; it cannot command subjects to feel safe. Trust may also rise temporarily before repair, through fatigue, dependence, or publicity. Satisfaction and changes in tone can serve as reference, but they cannot replace the outcomes of water, compensation, and authority.
People can dislike the water station and still arrange their lives by clear deadlines; they can also credit the staff with good will and still be unable to predict the service. The institution should give priority to dependability: limited commitments, versions that can be checked, responses when things fail. Emotional relations may recover with time and need not be exacted as project performance. This does not deny the value of respectful communication. A cold but accurate system also raises the subject's costs. The difference is whether communication serves the actual relationship, rather than using warm language to cover what cannot be depended on.
Completing the project, releasing the budget, and protecting reputation give the agency and its outside consultants an incentive to end the evaluation. Interest does not prove the conclusion wrong, but it does require open methods, unresolved items, and residents' objections. When the same team promises, executes, and unilaterally declares success, the epistemic risk is especially high. Unlimited audits equally consume resources. Closure can rest on criteria fixed in advance, reviewed by different subjects, with a reopening path reserved for significant new material. The status of completion should have boundaries; it is not eternal truth.
Some residents may, after receiving compensation, still withdraw from the original cooperation, build their own testing, or leave the valley. The institution may regret this, but it cannot read departure as the failure of repair or as ingratitude. Repair first reduces harm and restores choice; it does not guarantee the continuation of the original relationship. The material of those who leave should still enter public learning, and their privacy and right of exit must be preserved. A subject's losses cannot be deleted from the historical ledger because he no longer attends meetings.
The Present Boundaries of Future Commitments
A five-year ecological restoration proposed for the valley may supply an intertemporal direction; but if immediate water, care, and income are all advanced by residents, the future public benefit is being built on the present losses of a few. A vision should list the resources of each phase, the earliest beneficiaries and the latest bearers, and provide minimum protection during the wait. The greater future good cannot be used to write today's subjects into abstract inputs. Conversely, attending only to immediate losses can exhaust the funds needed for long-term repair. Budgetary choices need an open time scale, so that residents know which present conveniences are being deferred and what verifiable capacities are gained in exchange. Intergenerational goals cannot become a reason to refuse present accounting.
The water station seconding personnel to monitoring may reduce other services; residents attending meetings likewise give up work and family time. Not every unchosen option can be precisely measured, but significant alternatives should be stated. Recording only a project's direct expenditures writes a collective choice as pure gain without sacrifice. Opportunity cost, likewise, cannot be expanded without limit to block action. Assessment should center on the most likely and most weighty alternatives, state the uncertainty, and keep observing during execution. Every plan leaves possibilities unrealized; the responsibility is to keep crucial losses visible and to allow the budget to be reallocated when new material appears. The promisor must not publish only the gains of the chosen path; it should also preserve why important alternatives were abandoned at the time.
An agency may, in the name of "restoring the valley for the next generation," ask current residents to accept long-term monitoring and data retention. Protecting the future environment carries legitimate value, yet the unborn cannot authorize an agency to acquire unlimited powers. Basic ecological capacity and the space of choice should be protected, while irreversible transfers of data and property are limited. Current residents, likewise, cannot, by virtue of a present majority, exhaust conditions that later people will be unable to recover. Intertemporal judgment requires testable predictions, distinct scenarios, and periodic revaluation, so that the future becomes neither a slogan of power nor the present the sole owner of all value. A public commitment possesses real boundaries only when responsibility can still be redistributed after adverse outcomes appear.
Public Commitments Must Preserve a Seat for Dissent
Residents can decline to attend reconciliation ceremonies, refuse to have private experiences written into publicity, and choose not to accept compensation that carries overly broad waiver clauses. Refusal must not cancel unconditional basic services. The agency needs to separate repair, communication, and legal arrangements, so that subjects are not forced to exchange one consent for all their rights. The agency, too, can refuse demands it cannot execute or that would create greater harm, but it should state its authority, resources, and alternatives rather than closing with "we have listened." Refusal is a public decision and must enter records and appeals. Accepting every request ambiguously and then letting it vanish silently in execution only enlarges the losses of reliance. Project personnel need to exit their roles, and community representatives grow weary. Handover must preserve unfinished commitments, access to materials, contacts, and continuing responsibilities; the departure of individuals must not become institutional forgetting. No individual should be permanently owned by one crisis, while public responsibility must persist through changes of personnel. Ultimately, the boundaries of a commitment are tested by two capacities together: whether others can arrange their lives by it, and whether the promisor can be required to change when it fails. Without the former, language is only vision; without the latter, language is only power's description of itself. Joined together, public meaning enters the actual world.
A public commitment also shapes the actions of other institutions. Hospitals, schools, employers, and families may adjust their services according to the water station's declaration; if changes to the declaration are notified only to direct applicants, the downstream dependents will keep using the old version. The promisor needs to identify the main chains of reliance and, when updating, actively notify the nodes able to change outcomes; other institutions, in turn, have their own duty to check sources and retain alternatives. Requiring the promisor to control every restatement is unrealistic, but the complexity of transmission does not justify correcting only the original web page. Cross-institutional versions, contacts, and clear scopes of force keep a sentence of public language from inflating without limit as it travels, and keep errors from being borne alone by those at the farthest end. A public commitment should also state what it does not promise, preventing the vision from being automatically expanded downstream. Defining boundaries is not a lowering of responsibility; it is what allows new requests to re-enter budget and authorization. The next chapter examines how polarization reduces the paths of negotiation: when the failure of commitments is swiftly translated into the essence of camps, which partial repairs and common objects disappear from public choice.