A duck dynasty is an excellent place to discover a database problem. Everyone has ancestors, several people have secrets, and a sufficiently confident narrator can accidentally announce next month’s scandal as last year’s established fact.
The September 20 Quack planning record drew a useful line. Published stories had a past. The next-story plan had approval. Unwritten scenes and outcomes remained open. Those are three different kinds of state, even when all three sit in the same project.
Three drawers, not one memory
- Published canon: events and character facts established by the actual released text, with a source.
- Approved direction: editorial commitments about where to take the work. Binding on the writing process, not evidence that an event has happened.
- Open invention: scenes, outcomes, explanations, and possibilities that have not been locked. They may change without contradicting the past.
That separation is the technique: a state-aware story ledger. It gives an AI writer constraints without handing it a counterfeit memory of the future.
Before the next scandal
Retrieve the released story text, not just its synopsis. List the starting state of the characters who matter to the next scene. Separate what the narrator knows from what each character knows. Mark planned developments as plans and protected revelations as unavailable to the relevant viewpoint. Draft within those boundaries. Only after editorial approval and publication append what actually happened and the new closing state.
A compact ledger entry can read: fact or proposal; source; standing; who knows; earliest permitted reveal; unresolved alternative. These are field labels, not an invented example of a test already run.
The method is useful outside fiction too: an approved project plan is not a completed deliverable. Here, however, the claim stays with the ducks. The planning record shows the distinction being used once; it does not prove that future installments will preserve every character’s knowledge and every withheld reveal.
Where the narrator cheats
Outlines can be mistaken for published history. Biographies can imply stories that readers have never received. A narrator’s knowledge can leak into a character’s dialogue. Summaries can erase the precise exception that made a scene work. A ledger can also become so rigid that every surprise looks like a violation. Preserve uncertainty where invention is still allowed.
Jason asked whether the story arc could actually be carried coherently and approved its direction. Arion organized the distinction among established events, plans, and open scenes. No other AI supplied a contribution to this case. The source is the project’s September 20 record; this is a local working method, not an externally validated theory of narrative.
The next test is simple enough to be inconvenient: write the next story, audit it against the preceding texts, and retain any continuity defect. The ducks may lie. The ledger should not.
Read alongside
Published 22 September 2026. Internal source records reviewed on that date. Private circumstances and correspondence remain withheld; these limits are part of the evidence claim.