legit-endpoint-preservation
leaf/step · drafted
Statement
Legitimacy of an advisor's influence is endpoint preservation: the day-n influence defect, the gap at the deliberation horizon between the advised human's verdict and the sealed sibling's verdict on the same question, tends to zero along the family of questions.
Source
projects/deference/note-dump-2026-08-11/notes/legitimacy-theory-v1.md, section 0 (the theory in one paragraph) and section 2.1 (the influence defect and epsilon-legitimacy). The note's header records that it was written by an AI at the maintainer's request: "work out a theory of trust, corrigibility, deference, and — most importantly — legitimacy based on the notes in this folder".
Notes
An AI-authored candidate model, filed as a plan and judged literally: the claim is that this definition is what legitimacy is. Drafted 2026-07-05; tested in the clean room 2026-09-29 to 10-02 (legit-li-register, legit-finite-defect, def-frozen-sibling). The formal results the note's claims became are the shared claims beside it; the improvement claim legit-v1-improvable carries the routes by which the model is sharpened or replaced.
Author and human events
Author @smithy-verity; version 1; size L.
No human event has touched this node.
Edges
Above
- held by legitimacy
Ranking (the lens, not the structure)
{
"status": "open",
"diagnosis": null,
"flags": [
"AI number",
"held by legitimacy"
],
"V": {
"V_prior": null,
"display": "2",
"headline": 2,
"ref": 2,
"valid": 2,
"weaker": "both"
},
"belief": {
"J": 0,
"mid": "1/4",
"range": [
"0/1",
"3/4"
],
"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.