Justice

Repair had an object, and choosing it was already a judgment.

The institution I had followed had delivered something it owed. I preserved the force of that accomplishment while asking what determined the content of the obligation. A response directed toward one loss did not acquire responsibility for every loss merely by succeeding. Yet neither did its chosen boundary justify itself. I was examining what repair should restore and what it justifiably demanded of others. Punishing a wrong required its own grounds; the severity of a response did not settle those questions.

I approached that work through the catalog's arrangements for remedy. Their limits were easy to name and harder to evaluate. Some boundaries concerned an institution's proper responsibility or another person's claim; others acknowledged the impossibility of reversing a loss. Each needed examining before I decided whether it left an established obligation unmet. A tidy heading was a considerable convenience to the person doing the filing.

What interested me was the relation between the loss recognized and the action authorized. Recognition gave a response its object. A demand to return something wrongfully taken differed from a demand to compensate for its absence; both differed from an acknowledgment of the wrong itself. More of one response would not necessarily supply what was missing from another. An apology could recognize what a payment did not express, while remaining entirely inadequate to replace a payment that was due.

Earth had given some of these distinctions formal expression. Its principles concerning remedies for gross human-rights violations and serious violations of humanitarian law distinguished restitution, compensation, rehabilitation, satisfaction and guarantees of non-repetition. The distinctions described purposes and obligations in that field, not a record of their universal fulfillment. I valued the refusal to make a single act of response answer for every task.¹

The differences also constrained criticism. If compensation had been owed and paid, the continued existence of a memory did not prove that the payment was fictitious. If acknowledgment was independently owed, a payment did not silently perform it. I needed the grounds of each demand before judging the response sufficient or insufficient. A person's importance did not depend on whether their losses fitted the available remedy.

A boundary around a remedy was therefore a place to examine reasons. An institution's limited jurisdiction explained the reach of its powers, without establishing that losses outside its reach deserved no response from anyone. Conversely, an unaddressed loss did not identify which institution or person owed what. The boundary identified the next question; it did not supply the answer.

Restoration made the difficulty especially visible because it seemed to promise a destination. Return the condition to what it had been. But the earlier condition needed identifying, and time complicated the return. Restoring a resource after an interval did not restore its use during that interval. Replacing it elsewhere raised a further question about whose access to it had been restored.

One European environmental-liability framework distinguished work returning damaged resources toward their baseline, complementary provision where that return was incomplete, and provision for interim losses of resources or services. In this setting, compensatory remediation did not mean cash paid to the public. The framework specified different tasks within restoration itself.²

I followed the distinction into the justification of a baseline. Estimating the condition without the damage gave an account of the particular injury. It did not make that earlier condition an ideal distribution. Repairing a wrong and correcting an earlier injustice might require different obligations. Neither task became dispensable because the other existed. I kept responsibility for the identified injury distinct from a demand to perfect the whole arrangement; that larger demand needed grounds of its own.

Equivalence required equal care. A replacement of similar size or value did not settle whether the same people had access to it or were able to resume the activity it supported. Those differences needed reasons, even when substitution was the best available response. Literal return was no solution where return was impossible. But accepting any substitute merely because it was available let availability decide the obligation.

The difficulty did not make repair futile. It gave achieved repair a more exact meaning.

I turned to a separate restoration programme in Louisiana. The Rabbit Island plan placed bird-habitat work in an area open to commercial and recreational fishing and containing public oyster seed ground. After the Deepwater Horizon spill, the work raised nesting ground; trustee reports described birds using it after inundation of older, lower areas. That was reported use, not proof of complete ecological recovery.³

I was following the habitat strand of the response. Its achievement did not settle what was owed to people, including those using the surrounding waters. It did establish a narrower reported consequence worth preserving. A bird using higher ground was not a verdict on the whole response to an oil spill. It was also not an administrative fiction.

The distinction changed how I assessed completion. Finishing the assigned work mattered. Whether the work restored the protected condition required another kind of attention, and whether that condition adequately answered the loss required judgment. I kept these questions connected without treating them as the same question asked three times.

Partial success therefore required judgment about what remained. Where a further duty was established, the value of work already done did not discharge it. Where the response reached a justified limit, continued loss was not enough to overturn that judgment. I needed the grounds of the remaining demand, rather than a rule that converted every incomplete reversal into failure.

I had begun with the content of fulfillment. The means of achieving it now drew other claims into view.

What repair asks of others

The restoration design had to contend with more than the intended nesting habitat. An access route was eliminated because of adjacent oyster resources. The plan also anticipated temporary restrictions on recreational fishing near the construction work. The selected alternative reflected considerations about construction, usable habitat and retained marsh. These were stated design reasons and projected burdens, not findings that fishing restrictions had occurred or that collateral harm had been avoided.⁴

The anticipated restriction on fishing made the reach of the decision visible. Work intended to restore one condition also called for limits on people's use of another. Its restorative purpose did not decide, in advance, how those limits should be justified.

Across the catalog, I had reasons to be suspicious of arrangements that moved a burden beyond the account in which a benefit was celebrated. Here the useful comparison concerned the boundary of the account. Assessing repair only by its benefit to the designated object left out the claims affected by its means. The intention to restore did not make those claims less real.

Nor did their existence require abandoning the work. If every competing interest could veto a response merely by being affected, repair would become impossible in precisely the circumstances requiring difficult judgment. What mattered was how those interests entered the choice, what weight their differences deserved and why the chosen burden could be justified to those carrying it.

This was where procedure acquired a substantive place in my inquiry. I had already examined what made a claim practically usable. Now I was concerned with the position of someone whose interests a justified claim might affect. The fact that one party had grounds for action could not settle everything another party was entitled to contest.

In the European remediation framework, a court had required authorities changing remedial measures to hear operators, apart from urgent situations, invite affected landowners' observations and explain their choices. Environmental protection remained a legitimate objective; the powers used to pursue it still had limits. This was an adjudicated requirement, not a measured finding that the hearings improved every result.⁵

The distinction mattered even without such a finding. A person facing an imposed burden had a reason to know the grounds and an opportunity to contest their application. Their standing did not depend entirely on whether hearing them would help the authority reach its preferred outcome. They were among those to whom the decision needed justification.

I found an unexpected source of discipline here for an observer attracted to consequences. A desirable result did not entitle me to treat the objections to its means as mere friction. Some objections would fail. Some evidence would be unpersuasive. Giving a claim a hearing did not promise its success. It required the decision to encounter the claim before dismissing it, subject to the reasons that justified any urgent departure.

There was a corresponding limit on procedural satisfaction. A well-conducted hearing could establish that people had been heard. It could not, by that fact alone, establish that the chosen distribution was just. The reasons for the outcome still needed examination. Procedure protected a relationship among parties and authority; it did not relieve that authority of the need to justify what it imposed.

The same issue reached the extent of a response. A demand justified by restoration had to be assessed in relation to restoration, including its feasibility and effects on others. Punishment concerned a distinct justification for imposing a burden. Carrying its severity into a remedial judgment changed the task without explaining the change.

Proportionality gave me a way to keep the burden connected to its grounds. I needed to examine a proposed means in relation to the legitimate end, compare less burdensome alternatives and account for the disadvantages remaining. The word did not supply a universal calculation. Where two responses protected different things well, estimating costs did not settle their relative importance.

I resisted making the cheapest option the moral winner by default. Cost mattered because resources committed to one response were unavailable for another use. But a saving achieved by abandoning the very interest that justified action was not an equivalent means of protecting it. The comparison needed to preserve the objective while examining reasonable means of pursuing it.

These considerations made judgment harder without making it arbitrary. Reasons mattered in ruling out an option, exposing an overlooked burden or explaining a departure from literal restoration. A decision did not have to become uncontestable to be justified. It had to account for the demands it placed on others, beyond the satisfaction of those empowered to make it.

The grounds of a burden

Those demands included the work and resources needed to carry out repair. I needed to distinguish the reason someone should contribute from the fact that they had been made to contribute. The financial flow showed who paid directly; it did not vindicate the allocation.

Responsibility for causing a wrong supplied one ground. Control over the activity supplied another question about the power to prevent or correct it. Benefit, capacity and need introduced further considerations. These grounds sometimes pointed toward the same actor. I could not assume that they always would, or that naming one made the others irrelevant.

The distinction restrained an easy reading of public expenditure. If a public body financed necessary repair, the payment did not itself prove that private responsibility had been excused. Nor did its public purpose settle the fairness of the burden. Funding an urgent response and determining how its costs should finally be allocated were related decisions with potentially different timing.

The United States' Superfund financing made a bounded part of that relationship visible. General Fund transfers had been the trust fund's principal revenue source over a long period; reinstated chemical and petroleum taxes later changed the mix. This existed alongside cleanup work carried out by responsible parties. The fiscal record described direct sources of funding, not the final distribution of costs among consumers, workers or owners.⁶

That last distinction stopped my account where the evidence stopped. A legal payer was not a complete map of economic burden. Possible changes in prices or returns required evidence about how much moved, to whom and under what conditions. An intuitive account of who must eventually pay supplied none of that. The stated allocation remained available for examination; its further consequences remained unresolved.

The unanswered incidence question did not empty the known arrangement of meaning. Direct public funding committed resources to repair. A rule assigning payment to a responsible actor made a different demand. I examined both at that level. What I could not do was turn either description into a total account of who had borne the loss.

There were also grounds for acting before every allocation was resolved. Where a damaging condition continues during a dispute over responsibility, postponement imposes a burden on those still exposed. This does not establish a general right to charge any convenient party. It gives a reason to distinguish immediate provision from subsequent recovery of costs.

I found it useful to keep the distinct questions in view long enough for their tension to appear. The actor best placed to carry out repair might not be the actor who ought finally to finance it. The person with the strongest claim to assistance might have contributed nothing to the wrong. Capacity explained who could act; it required another argument to explain who should bear the expense.

The catalog's language of transferred losses remained useful, provided I did not let it decide the judgment before the transfer was understood. Sharing a loss could protect people from an otherwise unbearable burden. It could also shelter a responsible party at others' expense. The existence of sharing alone did not distinguish the two. The grounds, alternatives and consequences had to do that work.

This brought me back to the human importance of the inquiry. A person seeking repair did not cease to matter because responsibility was difficult to allocate. Neither did a person asked to fund or endure the response become expendable because its purpose was good. Justice had to hold those claims together long enough to decide between them where deciding was necessary.

I now saw why the adequacy of a remedy was not contained entirely in what its recipient obtained. A response took a position on other people's claims through its means, its demands and the losses it left to be borne. Even a valuable achievement needed reasons for that ordering.

There was still a cost that no later allocation could undo: having suffered the injury in the first place. That drew my attention toward the work required while harm could still be prevented.

Before there is a claim

Prevention needed grounds of its own. The existence of an accomplished repair did not prove that another duty had been neglected. But where an actor controlled a dangerous activity and had reasonable means to reduce its hazards, the availability of compensation was a poor reason to postpone examining what that actor should do.

An arrangement offering a response after injury addressed a different condition from one requiring care before it. The difference concerned more than timing. Supporting a person living with a loss and reducing the occurrence of that loss were distinct purposes. Measuring either task solely through the other's achievements obscured what it was meant to do.

Ordinary duties gave this earlier work an institutional form. California's workplace rules assigned employers responsibilities for safe conditions and for implemented prevention programs, including the identification and correction of hazards. Such requirements established work to be done before an injured person brought a claim. Their existence did not demonstrate that every employer performed it.⁷

To assess protective effect, I followed an intervention into an observed consequence. A study of California's random workplace-inspection program during 1996–2006 compared eligible establishments in high-hazard industries with matched controls. It estimated fewer injury claims after inspection and lower workers' compensation injury costs. The study found no detectable adverse effect on the business outcomes it examined. That last finding did not mean compliance was costless, and the injury-claim measure did not count every possible injury.⁸

The consequence changed my account. Inspection had evidence of protective effect in the setting studied. Filing it as a ceremony because it was a procedure, or because injuries had not disappeared, misdescribed the finding. Attributing that result to each separate requirement in the prevention rules overstated it. The evaluated intervention and the legal account of duty supported different parts of the inquiry.

What mattered was the condition the action sought to protect. The work was not made meaningful only when it generated an enforceable claim after someone had been hurt. It had a purpose before that event, including for people who would never need the remedy. The evidence gave that purpose a bounded practical consequence.

This also changed how I considered its burdens. An inspection used resources and imposed demands. Those demands required justification in relation to the hazards and the means of reducing them. Counting their cost while giving no weight to avoided injury predetermined the judgment. Treating any safety objective as permission for unlimited restriction predetermined it in the other direction.

The existence of an effective intervention did not settle every choice about its reach. A more intrusive measure might add little protection; a less burdensome one might preserve much of the benefit. Those comparisons required evidence about the alternatives. Where it was absent, I retained the demonstrated effect without inventing a precise account of the best possible policy.

There was a difference between allowing uncertainty to limit a claim and using it to cancel what had been learned. The inspection study did not show how every workplace should be governed. It did support a particular protective achievement. That achievement did not have to answer every question to count within justice.

Prevention also clarified the smaller question of maintaining an achieved repair. Keeping a restored condition usable was work directed at the durability of that achievement. Preventing a repetition of the original wrong concerned the activity that had produced the injury. Protecting nesting ground from erosion, for example, would not establish that the conduct leading to an oil spill had changed. The two aims could both matter without borrowing evidence from each other.

I had no reason to demand that a single remedial institution undertake every task. Distinct capacities and responsibilities gave reasons to examine how the tasks were assigned. I wanted the description of success to match the task accomplished. A maintained repair deserved credit for remaining useful. A prevented injury deserved attention even though it left no recipient of compensation.

That was a different view of success from one organized entirely around the visible conclusion of a dispute. It required attention to conditions that protection preserved, while leaving the person seeking a remedy fully within the promise. There was no need to diminish recovery in order to value not having suffered the loss.

I returned to the fulfilled obligation with which I had begun. Its achievement survived the inquiry. What had changed was the range of judgment needed to place it: the loss it addressed, the claims affected by its means and the grounds for protection before another loss occurred. Each question had acquired a more definite object. Declaring the response complete answered none of them by itself.

These differences mattered to my comparative survey. A reasoned limit and an established failure required different explanations, even when both left a loss unrepaired. An effective intervention required attention to the condition protected, including where there was no injured person collecting an award. The catalog needed to preserve those differences if its account of justice was to describe more than the occasions on which justice was invoked.

The standards used in those judgments now required another kind of attention. Someone had authority to establish the terms, delegate decisions and revise the arrangement. Being heard about a particular burden did not necessarily give a person continuing power over those terms. Equally, the wish to change them did not make every proposed change just.

I recognized a fulfilled obligation and a justified decision while still asking who had the power to shape the next one. Justice had given me reasons for assessing the standards. I turned toward the people living under them, and toward what authority they possessed to make those standards their continuing concern.