Skip to content

design(research-overhaul): the register + the foundations — lexicon, claim ladder, deskcheck papers, laws ledger, Lean ladder (draft) - #47

Open
ascalva wants to merge 2 commits into
mainfrom
design/dn-research-register
Open

design(research-overhaul): the register + the foundations — lexicon, claim ladder, deskcheck papers, laws ledger, Lean ladder (draft)#47
ascalva wants to merge 2 commits into
mainfrom
design/dn-research-register

Conversation

@ascalva

@ascalva ascalva commented Aug 9, 2026

Copy link
Copy Markdown
Owner

What — an overhaul, two notes and a lane, one review

Per the owner's directive this PR now carries the full design treatment, ratified as a unit:

  1. docs/design-notes/dn-research-register.md (draft, track: workflow) — the research register: the lexicon (docs/lexicon.md, alias-and-deepen, five-clause teaching contract), the claim ladder (agent ceiling: Proposition-with-checkable-inline-proof; "Theorem" honest by absence), the deskcheck evolved in place into an artifact-evaluation paper (carrier byte-compatible, board.py zero lines, book LaTeX factored, PDFs via the exhaust lane), and the empty owner-only Question slot. Licenses RR-1/2/3 on merge.
  2. docs/design-notes/dn-mathematical-foundations.md (draft, track: mathematical-foundations — ⚑ manifest minted in this PR, owner adjudicates the coordinate at the merge) — the frame (structure-instantiation, the implemented bridges tabled), the laws ledger (docs/LAWS.md: seven mandatory columns, ~30 seed rows extracted from real code and audited line-by-line; memberships resolved as a coordinate-keyed relation with multiset projection — the capsule's "Boolean algebra" demoted to a named-falsifier Conjecture about the support image), the Lean ladder L0–L4 (from the Lean deep-dive: stress-test the proof-assistant park in dn-research-register — coverage, bridge, cost, and the agent-era premise #48 investigation: formal/ zero-coupling sibling, Plausible attacks, sorry-ratchet with the axiom-escape guard, entry-gated L2 discharge), and float-ε as substance (three arithmetic tiers). Licenses MF-1..6 on merge.
  3. Eight surgical edits to the register note so the siblings point at each other: its parked Lean row now reads default L0 — ladder in the foundations note §2.4; re-entry for L1: the ledger's first laws exist to state; §1.3 names the real file. The foundations note's MF-4..6 licenses are coherent only with these edits — the two notes and the edits are one ratification unit, not severable.
  4. docs/tracks/mathematical-foundations.md — the new track manifest (scored-beliefs/erratum-relation precedent).

Rulings PRESENTED to the reviewer (not resolved)

  • The track coordinatemathematical-foundations is minted ⚑; rename or reject at the merge.
  • Register edit E5 upgrades the register's non-goal 9 tag [INFERENCE] → [ESTABLISHED] on the warrant of owner seed 3 (a chat directive quoted in the note). Merging ratifies the upgrade; strike the edit if the tag should stay humble.
  • dc verdict mechanics post-regime: does the hand-set verdict field survive as the one owner-hand value? #43 (dc verdict mechanics) still stands as the register's presented default.
  • MF-1 is entry-gated on the staged memberships store landing (the bp-152 lineage files are staged in your checkout, on no branch — the note's Sequencing names this third dependency).

Sequencing

Merge #42 first (the warrant capsule); then this PR; the bp-152 store's own PR gates MF-1 whenever it lands.

Issues

Raised alongside: #43 (owner ruling), #44 (deskcheck README de-stale), #45 (dormant gate machinery, parked), #46 (ACM re-verification), #49 (χ_s windowed-law divergence — found by this PR's audit). Closes #48 — the Lean investigation's proposal is adopted wholesale by the foundations note; the issue's content is ratified into design by this merge.

Verification

Docs-only surface; all six CI verdicts green on the branch. Process: two ultracode treatments (11 agents each — 5 grounding readers, 1 drafter, adversarial audit, revision). The register note: 4 audit lenses, 4 must-fixes applied. The foundations note: 5 audit verdicts across two runs (a session-limit outage killed one run's reviser; the orchestrating session performed the revision inline), 6 must-fixes + ~25 should-fixes applied, including: a fabricated code quotation replaced with the real text (conductance.py:159-160); the χ_s windowed-call counterexample recorded as a stated hypothesis (→ #49); χ_s demoted out of the exact tier (it is one float64 division); GAP-capped rows brought under the register's own ceiling (Conjecture until the falsifier lands); ledger anchor semantics fixed to symbol-at-HEAD with the repair model stated (the refactoring PR carries the row repair); the SimpleGraph-vs-weighted Mathlib transfer priced honestly (the one-line proofs are for binarized-support statements; the weighted reduction lemma is house work). Audits verified all ~34 seed-row anchors and falsifier pointers against HEAD + the staged store, all 8 register-edit old_strings unique, all table rows ≤190 chars; the orchestrating session re-verified the contested pins independently before landing.

🤖 Generated with Claude Code

https://claude.ai/code/session_01HQcmECo4hNxNnQuRPxWo5K

…e without inflating its claims

Three mechanisms, one surfaced absence: a lexicon that teaches (alias and
deepen, never replace — ids/commands/parsed strings frozen), a claim ladder
that prices every label (agent ceiling: Proposition-with-checkable-proof;
"Theorem" honest by absence), and the deskcheck evolved in place into an
artifact-evaluation paper (same dc- lifecycle, board.py zero lines, book
LaTeX factored not forked, PDFs to the exhaust lane never git). Plus the
empty slot: the research Question, owner-only, surfaced not filled.

Warrant: owner seed 2 in docs/brainstorms/mathematical-foundations.md
(PR #42, sequenced first). Graduation licenses RR-1/2/3 on merge.
… Lean ladder — the overhaul's second note

Structure-instantiation as the foundation (each component names the structure
it models, states its hypotheses, imports theorems only under them; the
implemented correspondences tabled as the working bridges). The laws ledger:
docs/LAWS.md, seven mandatory columns, ~30 seed rows extracted from the real
code (memberships first — resolved as coordinate-keyed relation with multiset
projection, NOT the capsule's Boolean algebra; Boolean structure survives only
in the support image, stronger readings demoted to Conjecture with a named
falsifier). The Lean ladder L0-L4 adopted from issue #48: formal/ as a
zero-coupling sibling, L1 statements with Plausible attacks and a sorry
ratchet guarded against the axiom escape, L2 entry-gated discharge, L3/L4
parked. Float-epsilon as substance: three arithmetic tiers named and bound.

Adversarially audited: 5 lenses (house rules, math-correctness vs code,
regime/consumed-interfaces, feasibility x2), 6 must-fixes + ~25 should-fixes
applied in revision — among them a fabricated code quotation replaced with
the real text, the chi_s windowed-call counterexample recorded as a stated
hypothesis, GAP-capped labels brought under the register's ceiling, anchor
semantics fixed to symbol-at-HEAD with the repair model stated, and the
SimpleGraph-vs-weighted Mathlib transfer priced honestly.

Rides with: eight surgical edits to dn-research-register (its parked Lean row
now points at the ladder; sect 1.3 names the real file) and the new track
manifest docs/tracks/mathematical-foundations.md (owner adjudicates the
coordinate at the merge).

Warrant: owner seeds 1+3 (docs/brainstorms/mathematical-foundations.md, PR #42
+ this session's directive); the Lean investigation is issue #48.
@ascalva ascalva changed the title design(research-register): the palace adopts the literature's language without inflating its claims (draft) design(research-overhaul): the register + the foundations — lexicon, claim ladder, deskcheck papers, laws ledger, Lean ladder (draft) Aug 10, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Lean deep-dive: stress-test the proof-assistant park in dn-research-register — coverage, bridge, cost, and the agent-era premise

1 participant