2026-09-08 14:47:26
Anonymous:
Close story comparison with explicit inference limits and ship return scope
hold the flood/hospital story directions.md ..
@@ 28,19 28,19 @@
**The human misreading, and the failed intervention.** Humans found wounds closing and the dead getting up, and built a hospital. Research traced the effect to the core and **opened the seals** to strengthen it. Widening the field coupled it to the one inside, and the template began to include that anatomy — so healing now repairs toward something not entirely human, which is why healing machines turned harmful, the over-restored become creatures, and the core team changed without ever losing what they knew.
-
**Two legacies, and ship revival.** Builders' containment — isolator seals, pressure gates, field regulators in their cut script — against hospital machinery: restraint beds, dosing lines and triage stations under human labels, still working toward a drifted template. The intake scanner at the docking collar registered the astronaut at the airlock, so that is where the field returns them.
+
**Two legacies, and ship revival.** Builders' containment — isolator seals, pressure gates, field regulators in their cut script — against hospital machinery: restraint beds, dosing lines and triage stations under human labels, still working toward a drifted template. The docking intake registered the astronaut to a restoration point inside this ship, on the ship side of its airlock; the field returns them aboard here.
**Both alien languages, doing semantic work.** The builders' operational script covers seals, ports and regulator states, and the player uses it to work the containment.
-
The second is **the being's own**, used for utterances rather than labels. The observable distinction is not the address but a **terminal element that comes in two forms**. Inscriptions ending one way always sit beside a physical change made at that spot — a fitting pulled, a line cut, a panel left off. Inscriptions ending the other way never do. Compare a handful and the pattern is plain: one form is a request that somebody carried out, the other is a request nobody did, or a refusal. The player learns a modality by seeing which utterances the world acted on.
+
The second is **the being's own**, used for utterances rather than labels. The observable distinction is not the address but a **terminal element that comes in two forms**. Inscriptions ending one way always sit beside a physical change made at that spot — a fitting pulled, a line cut, a panel left off. Inscriptions ending the other way never do. One candidate reading is request versus refusal, but the same correlation also fits completed versus pending records. These cases provide a hypothesis, not a decoded grammatical distinction. A second context that distinguishes those readings is still needed before the player can know what the being is asking.
-
**Answering it, concretely.** The most recent inscription is in the acted-on form and names an act, and that act is closing the seals. So the deeper action is not the player's idea: it is the thing being asked for by the one it will shut in. Answering means performing it. Declining is equally available — seal anyway without answering, or leave it open and keep the healing dangerous — and the article does not settle what the being's reason is.
+
**Answering it, concretely.** The proposed later reading is that the most recent inscription asks for the seals to be closed. If further evidence establishes that reading, answering means performing that act for the one it will shut in. Leaving the seals open is the alternative, with healing still dangerous. A player who never decodes the inscription can still close them for practical reasons. The being's reason remains open, and the semantic evidence that would establish this request is not yet designed.
**A survivor with bounded evidence.** Yeva ran the hospital's plant and pressure systems: she knows which machines humans installed and which were already here, and that the core team came back different. She reads neither script. Her stake is Dorn, treated after the seals opened, changing in a way she cannot name.
**Three discoveries that change what the player can do.**
-
1. Marks on the creatures match those on the astronaut's healed injuries. *Changes:* certain builders' isolators read the field's pattern, so a marked astronaut passes where they could not — the same thing happening to Dorn.
-
2. Ward logs show healing was stable until the seals opened. *Changes:* the regulators are adjustable, so the field can be narrowed, not just avoided.
+
1. Marks on creatures reveal a pattern shared by certain builders' isolators. *Changes, if a player-attunement branch is selected:* recognised marks allow passage. An evidence-only branch instead makes this a clue to treatment and its bypasses; it grants no automatic player access power.
+
2. Ward logs show healing was stable until the seals opened. *Changes:* the regulators can isolate the contaminated repair template while preserving revival coverage.
3. The being's inscriptions address individual builders in sequence, the last addressing whoever comes now. *Changes:* it can be answered — an act the player performs, not a document they read.
**Consequential actions.** The apparent escape is undocking, which ends revival.
@@ 81,6 81,6 @@
## Recommendation
-
**Option A.** It keeps the image the owner leaned toward and turns it into something the player can act on: a final action that answers someone rather than operating a machine, and whose request is the same act the player would otherwise take on their own initiative.
+
**Option A.** It keeps the captive image the owner leaned toward and could let the same physical act mean either making treatment safe or answering a person. Its unresolved cost is that the second reading still needs convincing semantic evidence; the current inscription correlation does not establish it.
**Option B's advantage** is not directional shifting, which either option can have. It is that a machine with nothing inside is a colder explanation, and that its final act — supplying a template, possibly the astronaut's own — implicates the player rather than asking a favour. Its cost remains the prisoner: nothing left to answer, and the owner's strongest image gone.
