The river was a border. Shells crossed it, then infantry. The United City-States of Ends ordered its forces to repel the attack. Months later, the lost ground had been recovered and the troops stood on the eastern bank. A different decision now presented itself: should they cross?

In this series’ fictional case, the western neighbor, the Democratic Republic of Purposes, still has artillery five to eighteen kilometers beyond the river, within range of Reedtown. The Office of Defense says the town cannot escape that range without a crossing. The defense chief disagrees in writing: distant fire could suppress the batteries; a ground advance would lead to prolonged occupation. The Executive for External Affairs also objects, fearing the loss of a chance for mediation.

The Wartime Council of Three votes two to one to cross. The next day, the defense chief is removed in what the office calls a normal command adjustment. Troops advance about twenty kilometers and establish a buffer zone containing eleven settlements and roughly ninety thousand people. The First Executive explains that this is not occupation: it is what must be done so Reedtown’s children can sleep.

Those children have indeed endured shelling. But can that reason also authorize everything that follows: staying across the river, detaining residents, requisitioning grain, and eventually replacing school lessons?

The next decision is a new decision

This essay does not choose between the military assessments. Whether remote fire would suffice, and whether crossing is unavoidable, require military judgment. It does distinguish three decisions: repelling an attack, declaring a wartime legal status, and crossing months later. The first cannot silently authorize the rest.

The argument for beginning defensive war asks whether nonmilitary routes are exhausted and whether, without resistance, the people being defended would lose the conditions for continuing their lives. Expansion asks anew about the present: without this expansion, would that defended position be lost? Changes in range, intensity, participants, objectives, duration, or means need their own justification. ‘We are already fighting’ supplies none. A possible future attack is not the attack happening now.

Who determines necessity matters too. If only the three people choosing to cross may establish that crossing is unavoidable, necessity becomes a stamp on their own decision. Confirmation must not be monopolized by the commanding side; it must have a time limit and remain challengeable. Continuing requires renewed reasons. Otherwise the burden shifts: power continues until somebody else proves that its justification has expired.

The legislature previously voted to declare wartime status, before this crossing was proposed. Six hours before the crossing, a joint committee received a closed briefing. The record establishes neither its confirmation nor authority to confirm or refuse. Receiving information is not independent authorization. On the available record the required conditions cannot be certified; missing documentation should not, however, be turned into a finding that every unrecorded safeguard certainly did not exist.

The chief’s removal after dissent deserves investigation, not an automatic finding of retaliation. Nor should a two-to-one vote become three unanimous voices. Responsibility requires distinctions among proposing, objecting, authorizing, and carrying out. The country answers for action in its name without erasing what each officeholder actually did.

These are institutional obligations and responsibilities of office, not a moral command imposed on every soldier and civilian. A declaration of war does not write everyone’s convictions for them. An institution must give an executor a route to raise an objection or bring an unacceptable demand into review, rather than leaving only obedience or solitary exposure to consequences.

A second account in the waiting hall

Across the river, four checkpoints screen men aged sixteen to sixty. About 1,900 people are held overnight or longer; 230 are transferred to UCSE detention facilities, without charges, for as long as eleven months. The Wheat Plain buffer zone also experiences curfews, grain requisitions, and electricity cuts. Most residents are DRP citizens without a UCSE vote.

Farther west, Junction Town carries both military and passenger trains. An authorized strike at 7:40 in the morning hits the railway junction and the station’s waiting hall. The DRP reports sixty-three dead, including forty-one civilians; the UCSE reports twenty-two civilian deaths. These conflicting totals are not resolved here. One entry in the targeting memorandum is clear: passenger timetables were not verified.

The passengers were not the intended target. Nevertheless, the strike reached them. The buffer-zone residents are foreigners, but UCSE troops established their checkpoints, issued their requisition receipts, and confined them. They cannot opt out of these actions. For these actions the UCSE must answer to them. That does not confer UCSE citizenship or authority over the remainder of their lives.

Combatants who have laid down arms, been captured, or become wounded or ill and are no longer fighting are not outside protection either. Capturing, confining, or treating them carries responsibilities. They and local civilians need independent, reachable complaints procedures they can use without retaliation. Force against combatants still fighting is a distinct question, governed by constraints on striking present coercive capacity without destroying the conditions for rebuilding lives after war. This is not an account that makes every use of force equivalent.

War requires two accounts that cannot cancel each other. One asks whether a military action was necessary in its circumstances. The other asks what it did to particular people. Even an action genuinely unavoidable in military terms can violate the civilians it harms; calling the harm a cost does not clear that account. Conversely, recording the violation does not by itself establish that the action should never have been undertaken. Both questions require judgment. Neither settles the other in advance.

This does not equate defending against invasion with invading. The distinction remains in the reasons for beginning and expanding war, and in whether each step goes beyond removing the threat. What it refuses is a further move: converting a justified defense into permission for every person harmed to be treated merely as a side effect of doing the right thing.

Suppose the waiting hall had happened to be empty. There would be no deaths to record there, but the failure to verify the timetable would remain. Luck can change whether someone is harmed. It cannot perform an omitted duty for the officials preparing and approving a strike. Actual harm and failure of office need separate examination.

Expansion requires fresh confirmation beforehand. Individual strikes and checks cannot necessarily wait for a meeting about each one. They must follow rules made public in advance and face independent review afterward, at a stated time. Detainees must be able to ask while detained, not first when the war ends. Routing their complaints into the same defense office raises a question about independence, but does not alone decide it: who can control acceptance, closure, or continuation of the inquiry must be established.

Repair is not permission to rule

The buffer zone acquires a temporary administration. It selects eleven local residents for a self-government committee and replaces the schools’ National Purposes lessons with civics prepared in the UCSE. Thirty-seven teachers who refuse to teach the new course are dismissed.

During its earlier occupation of Reedtown, the DRP also replaced lessons and took away two doctors who remain missing. Those facts matter. They do not authorize the UCSE to decide what children across the river should learn or become. The enemy’s crossing of a boundary gives a reason to resist, not a right to cross every boundary oneself.

Three questions help distinguish defense from its drift: does this step directly remove the present threat; does it instead prescribe how the other population should reorganize its life; and can the control name a limit and a route to ending? Repairing a school is not objectionable merely because it takes place abroad. Deciding its curriculum and dismissing teachers who refuse is a different act.

At the ceasefire, the original attackers have withdrawn to their side and hostilities stop. On these facts the original ground for resisting this attack has ended. A possible return does not prolong the same justification indefinitely. Remaining across the river becomes a fresh expansion, requiring its own determination.

Three proposals follow: fund rebuilding but retain the administration until elected institutions satisfy UCSE standards; establish long-term trusteeship; or withdraw within ninety days. The first two leave the key to ending control with the occupier. ‘Elected’ does not alter who sets the standard. These are proposals, not accomplished acts; the criticism concerns what their arrangements would do if adopted.

Some residents ask the UCSE to stay; others demonstrate against its administration. Both need to be heard. Approval from some cannot retrospectively authorize an occupier’s earlier decisions for everyone. When withdrawal allows the old curriculum to return, that result does not prove an entitlement to have stayed. Returning control includes accepting that others may rebuild differently from one’s hopes.

Nine committee members leave with UCSE troops and receive asylum; the two who remain face prosecution by the DRP. Their predicament begins with the UCSE’s selection of them. Ending the administration does not end responsibility for that action and its consequences.

The withdrawal agreement also permits two thirty-person observation posts for three years, with NUT observers alongside. Remaining alert to fresh threats is not the same as installing facilities on foreign land. The posts still need an account of authority, entry, powers, and withdrawal. Agreement and a deadline matter without settling everything; neither does the mere fact of foreign troops establish colonization in advance.

The promise was that Reedtown’s children could sleep. Once troops cross, the children they encounter are not scenery behind that promise. Defense need not therefore stop. It must keep asking whether it is removing a threat or beginning to prescribe another population’s life. A war protects people who can go on living; it must not put them behind a state’s objective and postpone their questions until everything else is over.