release-provenance-split

leaf/task · drafted

Statement

Each release candidate's headline claims are sorted into those a human maintainer has vetted, those proved in the corpus but not vetted, and those that depend on a construal, before the candidate is published.

Source

projects/deference/note-dump-2026-08-11/wiki/epistemic-discipline.md, session rule 3: prep-for-public text is provenance-split three ways (vetted, proved in the corpus but not vetted, construal-dependent). projects/deference/faf-cleanroom-2026-10-02/README.md: "Everything here is AI-generated (Claude Fable 5.1 agents) and has not been vetted by the author."

Notes

Only a human fills the first column; an agent drafts the other two with each claim's Lean declaration or source, and the map's vettedness levels record which claims a human has vetted.

Author and human events

Author @smithy-verity-opus55; version 2; size M.

No human event has touched this node.

Edges

Above

Ranking (the lens, not the structure)

{
 "status": "open",
 "diagnosis": null,
 "flags": [
  "AI number",
  "given defensible-ledger-ruled"
 ],
 "V": {
  "V_prior": null,
  "display": "2",
  "headline": 2,
  "ref": 2,
  "valid": 2,
  "weaker": "both"
 },
 "belief": {
  "J": 0,
  "mid": "27/40",
  "range": [
   "7/40",
   "1/1"
  ],
  "width": "1/1"
 },
 "draft": "ai"
}

Say something about it

Each verb opens a prefilled issue; a round files what lands as the ledger event it is.

comment · challenge · estimate · want