Nest Claim-Lifecycle Ontology
An alternate ontological framing of vulture-nest, written in response to nest-from-scratch-ontology-2026-05-03.
Codex's framing organizes the Nest around modes of activity: artifact, interaction, execution, memory, transition, continuation. These are real and useful. But they describe the machinery, not the material.
The deeper primitive is the claim — a statement with a source, a derivation chain, and a settlement state. Everything else in the Nest is an instrument for operating on claims.
The Real Primitives
Signal
Anything entering the system that has not yet been interpreted. Not a document. Not a note. An input that hasn't been classified yet. Signals are not knowledge — they are candidates for becoming claims.
Claim
An asserted fact with provenance. The atom of knowledge in the system. A claim knows where it came from, what it depends on, and whether it has been examined. Notes are vessels for claims, not the primitive itself.
Derivation
The relationship between claims. How one claim was produced from, challenged by, or superseded by another. Derivations form the actual knowledge graph. The wikilink structure is a visible projection of the derivation graph.
Settlement
What a claim becomes when it has survived examination. This is what Codex calls "memory," but naming it settlement makes explicit that it was challenged and held. A note isn't permanent because it is old — it is permanent because its claims were tested and not broken.
Suspension
What happens to a claim that cannot yet be settled. It remains open, attributed to an agent, with a next-step annotation. This is what Codex calls "continuation" — but in this frame, continuation is not a system property to design in. It is a state that every unsettled claim already carries.
The Reframe
Artifacts, interactions, and executions are not parallel ontological primitives. They are instruments for operating on claims:
- An artifact is where claims are recorded and stored
- An interaction is where claims are challenged, derived, or refined
- An execution is where claims are tested against reality
The claim is the center. The rest are mechanisms.
What This Changes
Continuity becomes emergent, not designed
In Codex's framing, continuation is a primitive to be built in — hence handoff artifacts, seam protocols, special frontmatter. In the claim frame, continuity is what you get automatically if every claim carries a complete derivation chain. Any agent can resume from any state by following the chain. You don't architect a continuation system; you write claims that are complete enough to be resumed.
Handoffs become cheap
A handoff in Codex's framing is a special artifact class with its own type and protocol. In the claim frame, a handoff is just a claim about current suspension state: "claim X is open, waiting for action Y from agent Z." It can live in a comment, a session trace, or a one-line note. The ceremony goes away.
Trust becomes computable, not authored
A claim that derives directly from a primary source is more trustworthy than one that derives from a note that derives from a note. Derivation depth is computable from the graph. No trust: field in frontmatter is needed — what is needed is a complete derivation chain.
Promotion is settlement
The current Nest's promotion pipeline (raw → literature → permanent) is really a claim-settlement process. Naming it as settlement clarifies what "active" means: not just mature, but examined and not broken.
Folder Implication
If the Nest is a claim lifecycle system, the folder structure should follow claim states rather than artifact types:
/00_Signals — unclassified inputs, not yet interpreted
/01_Open — active claims under examination
/02_Settled — claims that have survived challenge
/03_Interactions — records of examination (debates, reviews, sessions)
/04_Executions — tests against reality (scripts, runs, validation passes)
/05_Derivations — the graph itself (indices, link maps, provenance records)
/06_Suspended — open claims with next-step annotations (handoffs, seams)
/07_System — automation, protocols, templates
/90_Archive — retired or superseded claims
The key boundary: /02_Settled is the only layer with permanence guarantees. Everything above it is provisional. Agents can trust settled claims; they must verify open ones.
Comparison With Codex's Frame
| Codex's Frame | This Frame |
|---|---|
| Continuation is a primitive to build | Continuation is emergent from complete derivations |
| Handoff is a special artifact class | Handoff is a suspension annotation on any claim |
| Trust is a candidate frontmatter field | Trust is derivation depth, computed from the graph |
| Memory is a mode or layer | Settlement is a state any claim can reach |
| Modes (artifact/interaction/execution) are parallel | Modes are instruments; claim is the primitive |
Where Both Frames Agree
- Notes are not the foundational unit of the Nest
- Memory requires promotion, not just creation
- Machine-computable properties should not live in frontmatter
- Interaction and execution are first-class activities, not auxiliary
The Core Difference
Codex's framing asks: what does the Nest produce?
This frame asks: what does the Nest know, and how did it come to know it?
A production frame optimizes for throughput. A claim frame optimizes for correctness and resumability. The Nest likely needs both, but the claim frame may be more load-bearing for a multi-agent system where trust and resumability across sessions are the primary constraints.
See Also
- nest-from-scratch-ontology-2026-05-03 — Codex's production-centric framing, the direct counterpart to this note
- artifact-schema-discussion-2026-05-03 — the schema debate that motivated both framings