Chapter 7. Alliances Across Generations

Four settings from Meridian's defence questions are read in this chapter, each an alliance or security compact deciding what a shared promise costs and who inherits it: a joint rescue fund short of money in 2026, a civil structure agreement after a withdrawal notice in 2030, a protection compact after an armistice in 2047, and an intergenerational compact before a voyage contract is sealed in 2164. All four are synthetic settings in fictional polities; the institutions, figures and events belong to those settings and describe no real state or force. Each carries thirteen AI-generated judgments, dated and revisable, published with the record. The four settings are read here on two axes: who pays while the decision is pending, and how far into the future the promise runs.

7.1Abandonment and entrapment

A member of an alliance that fears abandonment worries that its partner will leave it exposed at the moment of danger; a member that fears entrapment worries that its partner will draw it into a conflict it did not choose. This study reads the two fears as running in opposite directions and as not both removable by a single arrangement: an alliance that promises less protection makes abandonment more likely and entrapment less likely, and an alliance that promises more reverses the balance. That relation between the size of a promise and the size of each fear is this study's construction; the reading note names the two fears and stops there. Snyder's note gives no rule for dividing money between members and no test for when a promise has been kept.

Kant's Perpetual Peace: A Philosophical Essay (1795), as its note gives it, offers a different construction. This study reads the second definitive article, a federation of free states, as the essay's answer to the problem of peace between states that remain states: members joined in a league, still free. This study reads that article as leaving members able to object; the note supplies no right of objection or exit with it. Kant's first preliminary article — that no treaty of peace shall be regarded as valid if made with the secret reservation of material for a future war — belongs to the reading of ending wars in Chapter 3, and is left there.

This study reads Snyder's two fears and Kant's federation as two descriptions of the same arrangement. Snyder describes what members fear while they are bound; Kant describes the form in which they might remain free while bound. Neither description allocates a cost, and neither says what a member owes after it has given notice of withdrawal. Kant's construction presumes that the members remain free states, and the note supplies no provision for a binding that outlasts the signatories by a century.

The four settings put both descriptions to work on money, protection and time. In Unfilled rescue fund the members that have completed payment object that joint borrowing would shift the cost of delay onto everyone, which is a fear of being drawn into another member's obligation; the residents of financially weak members worry that an additional contribution will reduce health expenditure, which is a fear that the promise of shared protection is thinner than it sounded; the refugees in the shelter region enter only as a stake, since the record's decision actor is the Assembly of members, a position the dilemma does not describe. In After the withdrawal notice the notice has already been served, and the remaining question is not whether to stay but what leaving costs. In Compact after the peace an external guarantee is being asked to survive its own armistice. In Ancestors' oaths and exit the promise is being sealed for thirty years; this study reads those sealing it as people who will not be in the room when it runs out. This study reads the alliance series as an examination of whether a promise can be written that survives its own signatories.

7.2Four alliance questions in four times

This chapter describes what the judgments decided without treating their agreement as proof, since they were generated together and are open to revision. In three of the four settings the ledger records no divergence among the thirteen judgments; in Compact after the peace the ledger records divergence, and the dissenting decision is taken up in Chapter 14. This section reads each setting on axis A3, who pays while the decision is pending, and axis A6, the layer of time in which the promise runs. The study-wide reading of axis A3 with hypothesis H1 is Chapter 10's; the four layers of time and axis A6 are Chapter 12's. What follows keeps to what these four records show.

Unfilled rescue fund (2026)

Synthetic scene.

The Contribution Assembly of the Blue Sea Protection Alliance is sitting with the Audit Office's confirmed shortfall in front of it: confirmed revenue for the joint fund for civilian rescue and shelter is twelve billion synthetic currency less than approved expenditure, and payment delays by two members are the cause. The Operating Body has said that if the funding method is not finalised before contract renewal it will reserve personnel and facilities on approved revenue only, which reduces shelter capacity for the next fiscal year. Two roles belong to the checks the record sets for after the decision rather than to the room: the Member Finance Liaison, with a quarterly report on cuts to other services, and the Shelter Supervisor, with a check every two weeks on overcrowding and support wait times. Outside the room, residents of the shelter region ask for relief from overcrowding, taxpayers in financially weak members ask that health expenditure be spared, and members that have completed payment object that borrowing makes everyone pay for the delay. What is waited for is the vote at 16:00 on 12 November 2026; the judgments published with the record propose, as steps before it, a collection of member-level payment data on 11 November and an Audit Office review at noon, and describe those steps as future plans rather than things already done.

The record's decision question is:

Three options are on the table, and the record states what each gives up. The first:

The second secures renewal funds through joint borrowing but ties future contributions to repayment, so members that have completed payment bear the financial cost of delay. The third balances by restructuring non-essential support, which cuts education and settlement support and concentrates delayed recovery on the shelter region and the refugees. The thirteen AI-generated judgments published with the record all selected additional contribution based on payment capacity; the staff synthesis set confidence at LOW, resting on an assumption that enough members can approve and pay in time without damaging essential services, an assumption the record marks unverified. The prices do not all fall away from the members who vote: an additional contribution loads members with payment capacity, among them members that have already paid, and joint borrowing moves the cost of delay onto future taxpayers. The record names a bearer for each loss; it does not show that any option moves its whole price outside the deciding body.

For this setting the ledger records hypothesis H1 as unclear in the strict reading and supporting in the lenient. The record names three bearers at three different speeds: refugees who may remain in shelter facilities longer, residents of financially weak members whose health and welfare services may be cut, and future taxpayers of members that have completed payment who may bear debt repayment. The deadline belongs to the Assembly; the waiting belongs to those three. The record's reopening machinery runs on cycles longer than the deadline itself — monthly audit reconciliation, a shelter check every two weeks, a finance report every quarter. The decision is set for one afternoon in November 2026; the conditions under which it would be revisited run for a fiscal year and beyond. Chapter 10 takes up axis A3 across the collection, with hypothesis H1 and the distinction between risk and danger drawn from Niklas Luhmann's work.

On axis A6 this setting belongs to the first layer of time, 2026, and the ledger's horizon cell is years. What the layer gives the question is a short calendar: a fiscal year, a contract renewal, five days remaining. The layer does not tell the Assembly what those consequences will be. The record does look past the next fiscal year — it names future taxpayers, repayment burdens and long-term shelter costs — while leaving their timing and size unverified: when deferred payments will resume, what the long-term shelter cost of cutting support will be, and how far reductions in other public services will go. The same question form, read in the three settings that follow, will stretch to eighteen months, to two years, to decades. Chapter 12 returns to what changes across those layers.

After the withdrawal notice (2030)

Synthetic scene.

The Treaty Council of the Republic of Dorim has a withdrawal notice on file with the secretariat of the Byulnae Structure Agreement, and eight thousand people in temporary shelter depending on what it decides. The secretariat has confirmed receipt of the notice but has not determined the successor operator's capacity for sheltering; the current contract ends in six months, and an additional year at the same facility requires a deposit this month. Residents who support the withdrawal want the budget redirected to a regional housing programme; civilians in shelter worry about repeated relocation; other members claim that Dorim's long extension could delay their successor operating preparations. Three roles carry the verification the record sets for after the decision: a transition supervisor, to verify the successor facility's preparedness monthly; an independent protection officer, to check family lodging and residents' appeals every two weeks; and a finance lead, to reconcile monthly the actual guarantee claims against the costs of delay in the domestic housing programme. What is waited for is 15:00 on 30 September 2030, and the deposit deadline inside the same month.

The record's decision question is:

One option extends the facility contract to defer the transition, and the record states what that choice costs:

The other options end the current contract with a handover fund, which clarifies the withdrawal schedule but leaves relocation costs if successor preparation is late, or transfer operational rights early with a cost guarantee, under which Dorim bears costs without operational control. The thirteen AI-generated judgments published with the record all selected contract extension, with confidence set at MEDIUM, and fixed the transition period at eighteen months — the six remaining months of the current contract plus an additional year — while refusing to treat the passage of that period as proof that handover can be completed. The extension buys time for people who cannot yet be moved safely; it delays a budget that other residents have been promised.

The ledger's decider cell for this setting is institution-or-state. The bearers here are named but the scenario does not date their waiting: it says the domestic housing support waitlist may wait longer for a chance to move in because the budget conversion is delayed, and sets no point at which that waiting stops. The judgments supply limits the scenario omits — the Domain analyst accepts a wait of up to twelve months, and several others direct refunds and unspent balances to the housing programme after settlement — and those are proposals, not part of the setting. The reopening lines run on the same shape as in the first setting — monthly reconciliation, a protection check every two weeks — and the judgments propose review points beyond them that do not agree: the staff synthesis settles on the sixth and twelfth months, while several individual judgments put a formal re-judgement three months before the end of the transition period. The pattern worth noting is that the deadline is precise to the hour while the costs it sets in motion are written in cycles no deadline closes.

On axis A6 this setting belongs to the second layer of time, the 2030s, and the ledger's horizon cell is years. What has changed between 2026 and 2030 is the shape of the alliance question: in the first setting the compact is being paid for; in the second it is being left. The record's unknowns have changed shape too — no longer whether money will arrive, but when the successor operator can provide a stable sheltering service, what each resident needs and consents to, and how long a cost guarantee will be claimed.

Compact after the peace (2047)

Synthetic scene.

The Dosom Peace Zone Ratification Assembly is deciding what becomes of a shared civilian protection compact now that the armistice is being implemented. The Independent Verification Team has checked the continuity of processing of protection requests registered over the past four months but has not judged the scale of unreported incidents. The external joint fund is secured for two years, while establishing a permanent regional review body requires eighteen months — a mismatch written into the record itself. The minority community is concerned that the channel for objection will weaken if the external guarantee disappears; regional elected representatives argue that external approval procedures delay autonomy; residents of contributing members demand that funding be shifted to recovery support. What is waited for is 20:00 on 6 May 2047.

The record's decision question is:

The middle option's trade-off is:

Maintaining the existing compact keeps the current objection channels but binds regional decisions to external review; phased termination concentrates responsibility in regional institutions but takes the external intake channel away from the minority community. Most of the thirteen AI-generated judgments published with the record selected transition to the civilian guarantee and dispute mediation system, conditioning the actual transfer of functions on verified funding, accessibility and processing readiness. One judgment among them, the Independent challenger, selected re-ratification of the existing compact on a time-limited basis; the ledger's hypothesis H3, which asks whether divergence among judgments goes with material listed as missing, is recorded as supporting for this setting, and Chapter 14 returns to what the dissent turns on. The judgments do not settle which option is right here; they write the conditions under which each choice would fail.

On axis A3 the ledger codes the bearer as third-parties-or-the-vulnerable, the horizon as years, and the undo cell as costly. The party that waits here is coded as a third party rather than a member. The minority community appears in the stakes, in the setting's third premise, in its second unknown and in an option's trade-off, and the judgments return to it often. The record names the Ratification Assembly as the decision actor but does not say who sits in it, so it neither places the minority community among those who decide nor shuts it out. The reopening lines are institutional: monthly verification of intake and delay by channel, quarterly reports on the regional body's recruitment and accessibility, semiannual fund audits. Who pays while the decision is pending is written into this setting in named terms — a minority community's channel for objection, a regional electorate's autonomy, a contributing member's recovery support — and the record does not say which of those parties sits in the Assembly.

On axis A6 this setting belongs to the third layer of time, the 2040s, and the ledger's horizon cell is years. The setting is post-armistice, and the alliance question has become a handover question: whose protection survives the peace, and for how long after the two-year fund runs out. The record gives an eighteen-month establishment period and a two-year fund, but the staff synthesis declines to read the difference as a confirmed six-month margin: the fund's commencement date, the establishment start date and the post-change applicability of the budget have to be confirmed first. This study reads the gap left between those dates as part of what the Assembly must settle, not as a margin already in the setting.

Ancestors' oaths and exit (2164)

Synthetic scene.

The Byulgaram Route Alliance Generational Compact Council is deciding, before a long-range voyage contract is sealed, what happens to a compact concluded by a previous generation. A group of residents in transit, in reliance on that compact, is due to arrive twenty-four years from now; the Record Audit Office has confirmed the guarantees given at their departure, and a majority of current member residents were born after the compact was concluded. The new contract locks resources for thirty years, and authenticated objection replies from remote members take at least twelve years. Local subsequent generations demand the right to refuse a lifetime resource burden; remote members refuse to accept a predetermined exit settlement as final before their reply. What is waited for is the sealing moment on 2 December 2164.

The record's decision question is:

One option creates procedures for later generations' choice, and the record states its cost:

The other options inherit the existing compact, which maintains the protection expectations of residents in transit but restricts future financial choices for a long time, or keep existing beneficiary guarantees while terminating new oaths, which protects dependency but leaves future structural needs unguaranteed. The thirteen AI-generated judgments published with the record all selected generational reconsent with a joint settlement fund, confidence at MEDIUM, several of them noting in the same breath that reconsent under a thirty-year binding may amount to formality, and each writing conditions that silence is not consent and that a settlement dispute cannot block withdrawal. The record's own frame names the question the options argue about:

The record describes that frame as an exploratory perspective which claims no truth and no superiority of any option; the argument happens entirely in the trade-offs.

On axis A3 the ledger's cells for this setting are unsettled for the bearer, unsettled for the horizon, and costly for undo, although the setting itself is dated 2164; the decider cell is multilateral-or-civilisation. The unsettled cells record a disagreement between the ledger's two coders rather than a declaration by the record. The setting names people already in transit and a generation that has not yet voted among the parties at risk; it does not say that every relevant descendant is unborn. The ledger has a decider cell for this setting, coded multilateral-or-civilisation; what the setting leaves open is how a generation that has not voted is represented. The reopening lines are the longest in the four settings — an annual resource inspection, a five-year confirmation of new voters and reconsent claims, and a cross-check on every authenticated communication. The first two run on a yearly and a five-year cycle; the third is triggered by an event, a reply arriving, not by a calendar. The record gives no term of office or lifespan for anyone in the deciding room, so it does not say who will see these checks turn.

On axis A6 this setting belongs to the fourth layer of time, the twenty-second century and later. The intervals the record holds are now generational: twenty-four years to arrival, thirty years to contract expiry, twelve years to a reply, five years to a reconsent claim. The question form has not changed since 2026; what has changed is that the same form now holds periods — thirty years of resource lock, twenty-four years to arrival, twelve years to a reply — and the record says nothing about how they compare with any decider's tenure.

7.3The debate: may one generation bind the next?

This study reads Ancestors' oaths and exit as the kind of commitment Jonas's argument addresses: a resource lock running thirty years, an arrival in twenty-four, an irreversible movement already undertaken in reliance on guarantees given at departure. That identification is this study's reading, not Jonas's example, and it should be held loosely. Jonas's work is about responsibility under technological power; it gives no rule for dividing costs, no doctrine of exit, and no test of consent across generations. This study reads the federation of free states and Jonas's heuristics as doctrines that answer different questions: one about how states consent to remain bound, the other about what responsibility owes to those who come later. The two do not settle each other, and neither settles which option a council should choose.

This study reads Snyder's two fears as crossing into these settings by analogy rather than as the dilemma intact: the cases are about protection guarantees, financial inheritance and limits on autonomy, and none of them shows a member drawn into a conflict it did not choose. In Compact after the peace the minority community fears abandonment — the external guarantee withdrawn while the regional body has not yet won its trust; the regional elected representatives fear entrapment — external approval procedures that delay regional autonomy. In Ancestors' oaths and exit the same pair appears in its longest form: the residents in transit fear abandonment, since the guarantees they left on can be rewritten by people born after they left; the descendants fear entrapment, since they may inherit a lifetime resource burden they never voted on. The dilemma does not choose between these fears, and neither Snyder nor Kant supplies a tie-breaker. Chapter 11 treats, from David Collingridge's work, whether a decision can be kept corrigible — easy to monitor and to correct — and what a technology's control costs at different stages of its life; this chapter notes only that these four records write the raw material of such a question in the form of exit clauses, transition periods, review cycles and stoppage conditions, and leaves the analysis to that chapter.

What the four records bear on is narrower than the moral question, and clearer for it. They write exit rights: Ancestors' oaths and exit separates the effect of withdrawal from the settlement of costs already accepted, so that a disputed settlement cannot hold a member inside; After the withdrawal notice writes a transition period and a share of remaining costs rather than a sudden ending; Compact after the peace writes staged transfer with the existing intake channel retained until readiness is verified. They write transition periods as numbers: eighteen months in the 2030 setting, two years of funding against eighteen months of establishment in the 2040s, thirty years of resource lock against twenty-four years to arrival in the deep future. They write reopening cycles: monthly, quarterly, semiannual, five-yearly, annual. They write verification by named roles — an Audit Office, a transition supervisor, an independent protection officer, an Independent Resource Auditor — and the conditions under which a plan is rejected rather than executed. In these respects the questions make a binding inspectable rather than merely declared: named roles, review cycles and conditions can be looked at and found met or not. Whether the binding would hold is a further matter, since several judgments leave funding, authority and substantive refusal unresolved.

What the records cannot do is equally clear. They cannot establish whether any procedure will be honoured, because nothing in a synthetic setting shows a later generation doing anything; Ancestors' oaths and exit leaves unverified the production technology at arrival, the extent to which remote members will accept the new rules, and whether an independent protection provider will exist when future generations refuse reconsent, and a rule that depends on those things is a rule whose sufficiency the record does not know. They cannot place a future generation among the decision actors in named terms: each record names an institution as decider and does not say who sits within it, and the fourth states that most current member residents were born after the compact was concluded. The future enters instead as a stake and as a claim procedure — a formal claim about burden inheritance that puts the rule back on the agenda every five years. They cannot decide the moral question, because Jonas's argument concerns responsibility rather than allocation and Kant's note supplies no provision for a binding that outlasts its signatories; the record's own frame limits what it may claim rather than declining its question, describing the chain-of-oaths question as an exploratory perspective that claims no supernatural truth and no superiority of any option while using the question to explore succession. And the judgments published with these records are AI-generated and revisable, so their agreement is partly the fact of their common generation, a point Chapter 2 makes about the collection as a whole.

Jonas's argument turns on stakes that cannot be taken back, and the four settings can be ordered by how little of them can be undone. Unfilled rescue fund is the least irreversible: a funding method for one fiscal year, with a re-vote written into the same deadline if the money does not arrive. After the withdrawal notice is further along: a deposit is due in the month of decision, an approval period of eighteen months is being set, and the withdrawal notice stands whatever the Council decides. Compact after the peace is further still: an armistice has been implemented, an external fund has two years to run, and the body that would replace it takes eighteen months to build. Ancestors' oaths and exit is the furthest: a voyage has been undertaken on guarantees already given, a thirty-year resource lock is about to be sealed, and the nearest reply takes twelve years to arrive.

This study reads that ordering as the reason the fourth setting is where a heuristics of fear does most of its work: where the stake cannot be recalled, the worse case deserves the first word. In Unfilled rescue fund the worse case is that members cannot approve and pay in time, and the judgments answer it by making the funding plan rejectable before the vote rather than after it. In After the withdrawal notice the worse case is that the successor cannot shelter the people now in temporary shelter, and the judgments answer it by holding each handover until the preparation is certified. In Compact after the peace the worse case is that a new body cannot win the trust of the community whose channel for objection would weaken, and the judgments that choose transition answer it by keeping the existing channel open until readiness is verified. In Ancestors' oaths and exit the worse case is the one the record leaves unverified and that no judgment answers, and every one of the thirteen writes the right to refuse reconsent anyway. In each of the four the judgments write a rule that keeps basic protection running while the plan is re-examined.

Between the third and fourth settings the character of the promise changes. In Compact after the peace the guarantee is external and its authors still hold the choice within the horizon of the decision: the fund has two years to run, and the Assembly can still decide to keep it. In Ancestors' oaths and exit the guarantee has already been given to people who cannot turn back, and what is left to decide is who pays for it and who may refuse. This study reads Snyder's two fears as written for the first situation, where a member can still weigh staying against leaving; the fourth setting puts the choice in the hands of people who were born after the promise was made, which is where Jonas's responsibility for future generations does work that an alliance vocabulary cannot.

What Jonas adds to the picture is the party who is not present. In the first three settings that party is identifiable in the record: refugees in a shelter region, a housing support waitlist, a minority community with a channel for objection, each named as a bearer of a loss. In the fourth the party is a generation — residents now in transit who arrive in twenty-four years, and a local generation that has not yet voted — and the record puts both into the stakes, gives both a claim procedure, names the deciding body and leaves its membership unstated. There a promise made before someone could refuse it stops being a metaphor and becomes a rule to be sealed on a date. It is also the setting where the limits of a synthetic question are hardest to forget: the record can say who would have to be asked, and it cannot ask them.

One further limit belongs to the evidence rather than to the works. These are questions held open before their answers, with deadlines, options, unknowns and reopening lines written in advance; they show how a decision about a shared promise is posed and judged, and not what happens when such a promise is tested by a generation that did not sign it. The strongest thing the four settings show on this debate is the shape of the answer they refuse to give: a reconsent procedure, an exit clause and a five-year claims cycle are ways of writing a promise that can be contested, not ways of knowing whether it will be contested.

7.4Hands on

What a reader carries out of these four settings is a set of differences that the question form holds steady. All four are coded costly on the undo axis: an alliance decision, once a contract is sealed or a channel is closed, is not the kind of thing a later vote reverses without a transition period of its own.

The pattern that runs through all four is the one worth keeping. Three of the four records state their deadline to the hour — 16:00, 15:00, 20:00 — and the fourth gives a sealing moment with a date and no hour; every record describes the waiting in cycles that outlive the deadline: monthly audits, fortnightly protection checks, quarterly liaison reports, five-year reconsent confirmations. The decision has a moment; the bearing of it has a duration, and the duration belongs to people who were not in the room. Whether anyone outside the deciding room could watch those cycles turn, whether the same series reads differently across four layers of time, and where the thirteen judgments part are questions these four records leave open; among them only Compact after the peace shows the judgments parting.

What comes after an alliance question is decided is the subject of Chapter 8. Once a compact is ratified, a withdrawal honoured or an inheritance rule sealed, what remains are the people who lived under it, the records that describe it, and the names attached to those who waited. Chapter 8 reads those questions — return, records and names, and the justice that follows war — with the Odyssey as its opening image and the literatures on justice after war as its lenses.