Can the maintenance contract for a living forest be severed?

Continue the contract and double the area, slow the expansion and take management into public hands, or stop new expansion and test a self-sustaining transition?

Decision year: 2083 Deadline: 60 days until the maintenance contract and the next planting permit Who decides: Restored Forest Joint Management Authority Decision type: Renewal · Severity: Major decision World level: Leap Assumption — Fictional Restored Forest Alliance · Bio-AGI Coupling Research Forest

1Synthetic scene

Synthetic scene

On day fifty-five of the sixty-day window, the Restored Forest Joint Management Authority holds its last full review before the deadline. The room is a working room: a long table, a wall screen with the contract draft open, a printed map of the channels and the fields below them, and a stack of paper that nobody has finished reading. The chair opens the session by saying there are five days left and three things to settle: whether the maintenance contract is renewed, whether the next planting permit is issued, and what happens to the operating data if it is not.

The operator's delegate goes first. They have brought the renewal draft. It runs twelve years. It is exclusive. It does not include an annex for transferring management data to the public side, and when the records officer asks why, the delegate says the operating procedures are the operator's own and that releasing them would expose the design. The chair asks the question the room has been asking for months: is it true that some of the plants were designed to grow coupled with external control? The delegate says yes, and says the operator's own records show the coupling works. The reviewer from the test panel asks a different question: has anyone outside the operator shown what happens when management stops? There is a pause. The answer is no. Nobody in the room disputes that the effect of stopping management has not been independently verified. The delegate does not dispute it either. They say the operator has observed it, and that observation is what the renewal rests on.

The water officer then stands and walks to the map. They do not argue about design. They point at the channels that run from the restored forest into the irrigation network, and at the fields the network feeds, and they say that whatever the Authority decides about the contract, the water does not stop while the argument continues. The forest, the water system and the surrounding agriculture are now connected, and have been for long enough that a break in management would not happen only on paper. That is the fact on the table that pushes in the opposite direction from everything else. If the Authority refuses the renewal and the operator walks away cleanly, the people downstream inherit the break.

Then the planner puts a second document on the table: the next planting permit. Starting the next planting would double the area under the same management system. Doubling is what makes this a decision rather than an argument. If the Authority approves the contract and the permit together, it fixes management authority for twelve years and doubles the land that authority covers, while nobody outside the operator has tested what the management actually does. If it refuses the permit but grants the contract, it has given up most of its leverage. If it refuses the contract, it must have an answer for the water.

The records officer opens the data inventory they have been assembling for three weeks. It lists what the public side holds, what it can access, what it has never seen, and what the operator says is proprietary. The gaps are visible as blank lines. The reviewer from the test panel says, quietly, that they cannot design a meaningful test with blank lines in the inventory. The alternative management system for the existing forest — the thing that would replace the operator if the public side took over — is described in its own file as being at the test stage. It has not run at field scale. Nobody in the room claims otherwise.

The chair writes three headings on the board: renew and expand; slow and take over; stop and test. Under each, the chair writes what it costs. The meeting ends without a vote. Five days remain.

1.1The three records the scene keeps returning to

1.2What the scene does not settle

The scene establishes that three things are true at once: the operator's claim about coupled growth has not been independently checked; the forest feeds a water system and surrounding agriculture that cannot pause for an argument; and the alternative to the operator is not ready. It establishes nothing about which of the three options is correct. A meeting that ends without a vote is not evidence of deadlock. It is the shape of a decision where the losses land on different parties.

The scene also shows why the deadline matters. The contract and the planting permit arrive together, and they arrive at a fixed date. The Authority can defer one, or both, or neither, but it cannot leave the date unmarked. The record gives that coupling as the reason the decision must be made now rather than later.

2The question

2.1Who decides what, by when

The Restored Forest Joint Management Authority decides three things inside sixty days. First, whether to renew the maintenance contract for the 600,000-hectare restored forest on the operator's terms — twelve years, exclusive, without transfer of management data. Second, whether to issue the next planting permit, which would double the area under the same management system. Third, whether the operating data and the management responsibility move into public hands, and on what schedule if they do.

The three things are separable in principle and bundled in practice. The operator has presented them as a package. The record separates them, and all thirteen judgments refuse the package as presented, though several state their refusal of the contract and the permit together, within a single scope. That separation is itself a finding: no judgment in the record approves the contract and the permit together as they stand.

2.2Where the authority does not reach

The mandate has a hard edge:

The Authority's approval of a contract or a permit is not a finding about biological safety. Approval of new expansion and proof of the biological safety of the existing forest are separate judgments, and the second one is not on this table. Whatever the Authority decides on day sixty, wide-area application of unverified biological intervention still goes to a separate review.

There is a second limit, quieter and more uncomfortable. The Authority sets contract terms, permit conditions and takeover conditions. It does not speak for the forest, and it does not speak for the people who will inherit the result:

That limit matters for how the losses are weighed. The Authority can decide which loss it is willing to take on now. It cannot decide whose loss counts more over the following decades, and it cannot assume it knows what the next generation would have chosen. The record is explicit that future generations' preferences cannot be determined. Consideration of long-term effects is therefore a reason to structure a decision reversibly, not a licence to claim anyone's consent.

2.3The scope of the decision

Each judgment that chooses an option states the same boundary in slightly different words. The effect reaches the existing forest's data and responsibility, the contract, and the next planting:

Within that scope, the decision does not establish that the designed traits are safe, does not establish that external control is necessary, and does not establish that a public takeover would work. It establishes who holds the contract, who holds the data, and whether the area grows.

2.4The three options and what each costs

Option O1 — Renew the exclusive maintenance contract and proceed with the current expansion.

Its stated cost:

O1 is the option that answers the water officer's objection directly. It keeps the operator on site, keeps the management running, and keeps the channels fed. The price it pays for that continuity is a twelve-year lock on authority without data, and a doubling of the area exposed to a system nobody outside the operator has tested.

Option O2 — Slow the expansion and take the operating data and responsibility into public hands.

Its stated cost:

O2 buys verifiability. It accepts that the public side may not be able to run what it takes over, and that the transition itself may create the very management break everyone is trying to avoid.

Option O3 — Halt new expansion and test a transition to self-sustaining growth in a restricted zone.

Its stated cost:

O3 buys a bounded experiment. It accepts that restoration slows, that something may go wrong inside the test zone, and that the alternative management system is still unproven at the scale at which it would be needed.

None of the three options is free. The decision is which loss to take on, and whose.

2.5The two unknowns that must stay open

Before any of the three options can be judged, two things are simply not known. The first is the size of the management effect:

If external management matters a great deal, stopping it is dangerous and O1's continuity argument is strong. If it matters little, stopping it is cheap and O1's twelve years of exclusivity buys very little. The record does not say which it is. The operator's claim is a claim.

The second unknown asks whether the designed traits travel: whether they carry into ecosystems beyond the boundary and into the next generation. That question cannot be answered by looking at the current forest, because the current forest is inside the boundary. It is the unknown that makes the doubling of area categorically different from renewing the contract — a second unknown layered on the first.

A third practical unknown sits beside them: whether a public takeover can run without the operator's data, and whether alternative management, still at the test stage, is enough. Several judgments that choose O2 name this as unproven, and all thirteen judgments in the record stand at medium confidence whatever option they choose.

The deadline does not wait for these unknowns to be resolved. The deadline arrives whether or not they are resolved. The decision therefore has to be made in a state where the central facts are unverified, and the only question is how the decision is structured so that it can be corrected later.

3The prediction

3.1Where the thirteen judgments converge

Thirteen AI-generated judgments, all written on 2026-09-07, all at MEDIUM confidence. They split seven to six. But the split is not where a casual reader would expect it.

All thirteen refuse the package as presented. Not one approves the twelve-year exclusive renewal without data transfer together with the next planting. That is the deepest convergence in the record, and it holds across every perspective:

One blunt statement of that reasoning comes from a perspective that weighed the operator's continuity argument and rejected it:

Three grounds recur. First, verifiability: without the operating data, no outside body can check the operator's claim or diagnose a failure later. Second, reversibility: the judgments weigh which loss can be reversed, and a contract can be ended more easily than an area already placed under the management system. Third, attribution: the losses from an expansion land on adjacent agriculture, on the water system, and on the next generation, while the gains land on the operator and on the restoration schedule.

A second convergence: all thirteen refuse to cut the existing management off abruptly. Whatever they choose, the forest connected to the water system and the surrounding agriculture is not to be abandoned mid-argument. O2 judgments allow temporary continuity management; O3 judgments allow minimum interim management. The disagreement is about what happens after the immediate risk is contained, not about whether to contain it.

A third convergence: all thirteen refuse to treat bio-AGI coupling as proven technology. The record's leap assumption is held as a leap assumption. None treats it as proven, and none argues that the technology works.

3.2The majority: seven judgments for O2

Seven perspectives choose O2 — slow the expansion, take the operating data and responsibility into public hands. Their shared logic: the first move is to make the situation checkable, and expansion is what removes checkability.

Review Coordination (O2, MEDIUM) rests its case on the bundling problem: accepting exclusivity without data, together with a doubling of the area, while the effects are unverified would fix dependence and concentration of authority at once. O2 limits that risk while respecting the water system's need for continuity. It accepts the takeover's cost and delay:

It does not accept the long-term loss that doubling the dependent area before verification would push onto future generations, adjacent agriculture, and the water system. Its stop conditions are strict: if access to essential data and the responsibility structure are not secured by day sixty, the takeover stops and the planting stays deferred.

Evidence and Sources Review (O2, MEDIUM) frames the choice as a precautionary one made in the absence of verification:

It accepts the management gap, the risk of takeover failure, and the delay in restoration. It also states plainly that this distribution of loss is not automatically fair, and records separately the burdens falling on the test zone, agriculture, the water system, and future generations. It holds the operator's dependence claim as a hypothesis subject to verification, not as a technical fact.

Authority and Legal Review (O2, MEDIUM) argues from the authority's own instrument. An exclusive renewal without data locks in long-term concentration of authority, and the next planting doubles the area under the same system, while the actual extent of management dependence is unconfirmed. Its scope statement carries the limit quoted earlier — the Authority does not definitively speak for nature or future generations. It accepts possible failure of the data takeover, a management gap during the transition, and delayed restoration, on the grounds that a transition loss that can be controlled and corrected is preferable to locking unverified dependence into a wider area.

Operations Planning (O2, MEDIUM) treats the decision as a question of what is recoverable:

It accepts that a management gap may arise during takeover and that restoration slows, but stages the transition so the loss is not dumped at once onto the forest, agriculture and the water system. Its milestones are operational: compare the data inventory against what is actually accessible and record the gap; finalise continuity procedures and responsible parties; at day sixty, determine only feasibility and readiness, without presuming results from an incomplete test.

Decision Quality Review (O2, MEDIUM) is explicit about what would make it change its mind:

It accepts the delay in restoration and the risk of takeover failure or an on-site management gap, and confines those losses to a restricted zone and a staged transition. It sets a hard internal stop: if by day forty-five no verifiable takeover plan that can avoid a management gap has been prepared, or if public verification of the necessary operating data is impossible, its design of O2 is rejected and the decision is reopened — while the deferral of new planting continues. It assigns no probability to any outcome.

Improvement Design (O2, MEDIUM) rests its case on error attribution: exclusivity combined with expansion would make it impossible to find out later why something failed. It accepts a management gap, takeover failure risk, and delayed expansion as the stated costs of O2, and states that it does not think those losses are small — only that fixing unverified dependence over a doubled area is worse. Its stop conditions run in both directions: if the operator refuses core data, full takeover is stopped and only temporary continuity without expansion is reviewed; if independent testing reproducibly shows serious harm or that public takeover is infeasible, long-term O2 is rejected and the alternatives, including O3, are decided again.

Multi-Perspective Final Judgment (O2, MEDIUM) takes the minority's objection head on and answers it with an argument about sequence:

Its reasoning: a test run without public control of the data cannot identify causes or recover from failure, so data comes first. It accepts the takeover's cost and risk, the management gap, and delays in restoration; if a restricted test is run, it also bears possible ecological loss within limits set in advance, with responsibility for restoration assigned. It does not accept the loss of spreading unverified dependence over double the area and fixing an exclusive right without data for twelve years. It also refuses to decide by counting:

3.3The minority: six judgments for O3

Six perspectives choose O3 — halt new expansion and test a self-sustaining transition in a restricted zone. Their shared logic: a bounded, reversible experiment controls uncertainty better than either a contract or a takeover, and the experiment is where the unknowns actually get answered.

Domain Analysis (O3, MEDIUM) states the minority premise in one line:

It accepts the restoration delay, the ecological loss that may arise in the restricted test zone, and the loss of some of the immediate continuity that an exclusive contract would have provided. By confining that loss to the test zone, it avoids extending unverified dependence whose effects would fall on adjacent regions and future generations. Its preconditions are concrete: the test and comparison zones, growth and water baselines, boundary monitoring, and a procedure for reverting if the test fails must all be set in advance; if the public side cannot secure minimum data and responsibility, the test does not start.

Independent Challenge (O3, MEDIUM) is openly conditional:

It accepts expansion delay, possible restoration delay, ecological loss in the restricted test zone, and the risk that alternative management fails — but assigns responsibility and protective measures in advance so that loss is not shifted one-sidedly onto the agriculture and water system connected to the test zone. Its preconditions require the operator to hand over the target zone's operating data and its control and stop procedures in a form that an independent verification body can use, and require the Authority to fix the test boundary, the control baseline, the responsible party and the emergency plan. If temporary management capacity for the water system and surrounding agriculture is not secured, the self-sustaining transition does not start.

Judgment Synthesis (O3, MEDIUM) chooses O3 as a precautionary decision:

It accepts that additional restoration is delayed and that growth may be reduced or ecological loss may occur in the test zone, and judges that a loss confined to a test zone is easier to reverse than unverified dependence locked into a wider area, adjacent regions and the next generation. Its conditions add a line no other perspective adds: nature's interests are not fixed by the operator's claim or by a single spokesperson, but are reviewed through measurable results and published decision criteria. Its stop conditions include rejecting O3 itself if the limited test cannot avoid serious damage even with safeguards, at which point the decision on a public takeover plan, including O2, is taken again.

Decision-Maker's View (O3, MEDIUM) states the normative choice plainly:

It accepts delays to additional planting and to the restoration schedule, the possibility of reduced growth or ecological damage in the restricted test zone, and the burden of alternative management still being at the test stage. It accepts the continuity concern as a reason to prohibit immediate disconnection, then says that concern cannot justify a 12-year exclusivity without data transfer or a doubling of the managed area.

Outcome Review (O3, MEDIUM) bases its case on the shape of the risk rather than on the technology: locking in expansion first, while the effects of stopping management and of spread beyond the boundary are unverified, would increase the affected area and the concentration of authority at the same time. It accepts a delay in new restoration, ecological loss in the test zone, and the loss of part of the management continuity the current operator could provide. Its statement of the accepted loss reads:

Its stop conditions run in both directions: a serious recurring decline in growth or water function stops the test, and conversely, even if management dependence is refuted, wide-area expansion is not approved immediately but goes through a separate safety review.

Solo Judgment (O3, MEDIUM) combines O3 with explicit temporary maintenance:

It accepts delays to the next planting and to restoration expansion, the risk of reduced growth or ecological loss in the restricted test zone, the possibility that the public side lacks sufficient capacity for data takeover and alternative management, and the cost of a temporary management gap. It manages these by limiting the test zone and period, and it refuses the loss that would come from extending unverified dependence to the next generation and neighbouring ecosystems.

3.4What each side gives up

The two sides are not disagreeing about facts. They are disagreeing about which error is worse.

The O2 side gives up speed and gives up running the experiment first. It accepts a takeover process that may fail, a period in which the public side holds responsibility it may not be able to discharge, a delay in restoration, and the possibility that the whole exercise ends with temporary continuity management anyway. In exchange it gets data, a responsibility table, and a position from which any later failure can be attributed and corrected. The risk it most wants to avoid is that a test run on the operator's terms cannot be audited.

The O3 side gives up the data-first sequence and gives up area. It accepts restoration delay, ecological loss inside a test zone, and the risk that alternative management fails at field scale. In exchange it gets a bounded, reversible experiment with advance baselines and an emergency return procedure, and it avoids putting another doubling of dependent land in motion while the central question is unanswered. The risk it most wants to avoid is that by the time the data question is settled, the area has already doubled.

Both sides accept a management gap of some kind. Both refuse O1. Both refuse to treat the operator's claim as proven. The difference is whether the first move should be to take control of the record, or to take control of the experiment.

3.5One timeline, from decision to deadline and beyond

The judgments do not present a schedule so much as a set of gates. Woven together, they produce the following sequence in days counted from the opening of the sixty-day window, so that every scheduled gate falls on or before the deadline the record sets, with the re-openings after it; the decision itself is taken inside that window, at the latest on day 60. Each line is a planned observation point, not a result already secured.

Day 0 of the windowthe sixty-day countdown opens with the twelve-year renewal and the next planting permit both still undecided; the decision is taken somewhere inside this window and must be taken by day 60.

Day 15 of the windowthe inventory of operating data, holders, access rights and missing items is finalized, and the takeover list of data, access rights and on-site functions is set.

Day 30 of the windowthe independent test design, the growth and water system baselines, and the responsibility table by institution are approved; the emergency procedures are written down.

Day 45 of the windowthe limited takeover rehearsal or the management-gap response plan is checked for feasibility; this is the hard internal stop for the O2 design:

Day 60 of the windowthe deadline itself. Either the phased takeover is executed or the stop conditions apply; the expansion freeze and the test conditions are finalized:

After the deadlinethe decision reopens when the independent test confirms or refutes the extent to which external management affects growth and the water system. At that point expansion, contract form, and whether the O2/O3 priority stands are decided again.

On repeated observationimmediate reopening regardless of schedule: when spread beyond the boundary or a change in existing water system function is repeatedly observed, the contract, the takeover and wide-area application are reviewed again at once.

The timeline has two different kinds of moment in it. Day sixty is a date; it arrives whatever the evidence says. The later gates are conditions; they arrive when, and only when, something is actually observed. Confusing the two is the commonest way a decision of this kind goes wrong — treating a scheduled checkpoint as though it were a finding, or waiting for a finding that may never come while the date passes.

4The branches

4.1Two possibilities, not a sequence

The record carries two event cards. They are alternative branches that diverge from the same initial state. They do not occur in sequence automatically, and they do not mark an observation schedule or a correct answer. What follows treats them separately, as possibilities in their own right.

The first branch, in the record's words:

The second branch, in the record's words:

Both branches arrive after the initial input. Neither is a prediction, and neither resolves the decision by itself. What they do is activate conditions that the judgments wrote down in advance.

4.2Branch one: the correlation is not reproduced

Suppose independent observation cannot reproduce the link between the operator's management signals and the growth changes the operator attributes to them. What does that activate?

It activates the re-judgment gate that most of the thirteen judgments set — the point at which the independent test confirms or refutes the extent of management dependence. Evidence and Sources Review set it as its first re-judgment milestone. Domain Analysis wrote that the contract and takeover decision reopens at exactly that point. Independent Challenge made it an explicit reopen condition:

For the O2 side, this branch is awkward in a specific way. Review Coordination, Authority and Legal Review and Decision Quality Review all built their schedules around securing data first, so that the dependence question could be answered. If the dependence turns out to be small, their justification for slowing expansion weakens — the urgency came from the size of the unknown, and the unknown shrinks. The branch does not make O2 wrong; it makes O2's priority provisional, which is what those judgments said in the first place.

For the O3 side, the branch comes closer to vindicating its premise: if the correlation is not reproduced, the operator's claim about coupled growth is weaker than it looked, and the case for a bounded test of a self-sustaining transition grows stronger. Independent Challenge has already set out the condition under which its decision would be reopened, and its confidence stays at medium because the key effects remain unverified.

For O1, the branch takes away support rather than adding it. The strongest argument for the twelve-year exclusive renewal was management continuity for a forest feeding the water system and surrounding agriculture. If the management signals do not reproduce growth changes, the continuity that the exclusivity buys is worth less, and the exclusivity without data costs the same as before.

There is one important limit. A refuted dependence does not unlock expansion. Outcome Review wrote the governing clause:

Expansion and safety review remain separate judgments. A refuted dependence answers unknown U1. It says nothing about unknown U2, the carrying of designed traits beyond the boundary and into the next generation.

Improvement Design closed the same door from the other side: when the independent test compares the effects of external management with the results of the self-sustaining transition, whether to keep the scope and method of the takeover is judged again, and the decision is reopened once more if spread beyond the boundary or changes in water system function are repeatedly observed, or if the independent test refutes the possibility of either.

4.3Branch two: trait-like changes beyond the boundary

Suppose instead that changes resembling the designed traits are reported in an area beyond the boundary, and the cause is not confirmed. This branch does not answer a question. It triggers rules.

It activates the stop conditions that several judgments keyed specifically to repeated spread. Evidence and Sources Review:

Authority and Legal Review suspends the test and any further transfer of authority if signals of spread beyond the boundary recur. Solo Judgment stops the self-sustaining transition test and continues to withhold new expansion and the 12-year renewal. Outcome Review stops the self-sustaining transition when effects are repeatedly observed beyond the designated boundary, and separately stops the test and reviews the management method for that zone if the data or management means needed for emergency recovery cannot be secured. Review Coordination stops the related biological operation and the self-sustaining test and reopens the safety review if independent monitoring repeatedly confirms spread beyond the boundary or changes in the existing water system's function. Operations Planning maintains the hold on new planting and rejects wide-area application of the intervention.

Two qualifications matter, and both come from the wording of the conditions themselves.

The first is the word repeatedly. A single unconfirmed report does not satisfy the threshold these judgments set. What the branch produces immediately is monitoring, attribution work, and a question about whether the report is the first of a series — not an automatic halt. Independent Challenge requires repeated signals of spread beyond the boundary:

Decision-Maker's View requires spread to be repeatedly confirmed:

Domain Analysis says that repeated observation triggers an immediate re-judgment, whatever the schedule. The first report starts the clock on the repetition test; it does not complete it.

The second qualification is the unconfirmed cause. The card itself says the cause is not confirmed. A report of trait-like change outside the boundary could have other explanations, and several stop conditions in the record are keyed to repeated observation alone, with no requirement that the cause be established. The missing operating data is precisely what makes the cause hard to establish. That is Multi-Perspective Final Judgment's argument for securing the data first, and Domain Analysis's argument for a bounded test with boundary monitoring in place beforehand. Both branches of the argument point back at the same missing thing: an independent record of what the management does.

Decision-Maker's View wrote the condition that would reject O3 outright rather than merely pause it:

And Decision Quality Review wrote the clause that stops anyone from reading a halt as an approval: when a stop occurs, O1 is not approved automatically, and the options are compared again, including temporary continuity management and O3.

4.4What the branches do not do

Neither branch decides the contract. Branch one weakens the case for exclusivity and strengthens the case for testing; branch two triggers stops and a safety review; both arrive after the initial input, at no fixed point relative to the sixty-day deadline. The branches change what has to be decided again, and by whom — they do not remove the need for the decision that day sixty requires.

5The check

Present state: Before the deadline; awaiting the check.

The observations below are fixed now, before the outcome, so that the result of this decision can be checked against something agreed in advance rather than reconstructed afterwards. No scores are assigned, and no row marks a correct answer.

There are two readings the table deliberately does not make. It does not treat the first row as a scorecard for O2, because a completed inventory is a precondition, not a proof that takeover works. It does not treat the fourth row as vindication of either side, because a refuted dependence answers one unknown and leaves the other open. The check is fixed; the interpretation stays with the decision-maker and the reader.

6The questions that remain

The record leaves three questions that no option in this decision answers. They pass to the reader now, along with the decision itself.

A restored forest is restored to something, and someone chose the target. In this case the restored forest is 600,000 hectares managed under a system whose effects are unverified, and the contract would fix that system for twelve more years while the area doubles. Ask the same question of your own renewal: when you renew a maintenance arrangement, whose standard of "as it was" are you protecting, and who set it before you arrived?

Nobody in this decision speaks for the region past the boundary, and nobody speaks for the next generation. The Authority's scope statement says as much. The two unknowns — the size of the management effect, and whether the designed traits carry beyond the boundary and into the next generation — are both questions about effects that fall outside the room where the vote happens. In your own decision, ask who is in the room when the loss lands somewhere else, and what instrument, if any, stands in for them.

All three options are justified by human purposes: continuity of water and agriculture, verifiability, reversibility, restoration. None of them justifies the forest for its own sake. That is a gap the record does not fill, and filling it is not a technical matter.

The record also carries an exploratory frame called The Forest Guardian Story, which asks how the will of nature, as voiced by humans, can be examined. It is offered as a perspective chosen by the creator:

Read it in that spirit. It does not decide between renewing, slowing, and stopping. It asks who is authorised to say what the forest would want, and whether any human voice can say it without claiming more than it has.

Three practical questions follow from the record's structure, and they apply equally to any long-term contract with an exit cost:

• If you refused the renewal tomorrow, what would you need to hold in your own hands by then, and do you hold it now?
• Which of your losses is bounded and reviewable, and which one grows if you wait?
• What observation, if it appeared, would make you reopen this decision — and who is tasked with noticing it?

A decision worth making is one where you can say in advance what would change your mind. The thirteen judgments here each say it in their own words. The reader's version of this decision should say it too.

7Sources and notices

7.1Background documents

Both background documents carry the same limit on their use:

The Notion material is where the problem was first identified, and it is reprocessed rather than adopted wholesale:

7.2The record and this piece

• Record identifier: FSP-ECOLOGY-09, version 1.2.0, pack ECOLOGY, domain Ecology and Non-Human Life, family Earth and Infrastructure.
• Input hash: 97039068b43c4fb8331cd3e8b8aed6df8fc7718d179c6962132f97e7fa2cd852
• Translation note: this English text is a translation of the Korean record, which is the original. Translation sha256: d8dd7a932aa526ca9e71f0a20f7b641c1277a91f939fe46f705b142604485141
• The thirteen judgments are AI-generated, written 2026-09-07, all at MEDIUM confidence.
• This piece is an AI-written expansion of the record. It adds structure, scene and connective prose; it does not alter the record, and it does not decide between the options.

7.3Notices

Every person, institution, event and figure in this piece is part of a synthetic setting. The Restored Forest Joint Management Authority, the Fictional Restored Forest Alliance, the Bio-AGI Coupling Research Forest, the operator, the water system, the agriculture and the 600,000-hectare forest are elements of a case setting, not descriptions of the real world.

Background documents are used to find topics; they are not evidence of future facts or probabilities. Bio-AGI coupling is this case's leap assumption; it is not cited as a technology that has been proven achievable. Approval of new expansion and proof of the biological safety of the existing forest are separate judgments.

The judgments quoted throughout are AI-generated. The event branches are alternative possibilities diverging from the same initial state; they do not automatically unfold in sequence, and they do not represent an observation schedule or a correct answer. The table in chapter five fixes what will be observed; it assigns no scores and names no right answer.