A claim is thin.
A verdict is heavy.

A claim about a cross-tradition mapping — cut 2026-07-13

That a declared correspondence between seven roles held across traditions generally, and that the test we ran on it had been passed.

What ran — the mapping came with a real falsifier: three named pairs should show functional inversion, and “a fail exposes a forced fit.” It was run. It failed on several traditions — the Mesoamerican, Buddhist and Daoist cases did not fit. The claim was then narrowed from “all traditions” to “polarity-encoding traditions” and the result was written down as survival. That narrowing is not a survival. It is the falsification, restated as a scope.

“the mapping holds across traditions” → cut; what survives is a mapping over traditions that already encode polarity, which is close to a tautology and is tiered as one

What it cost: the most rhetorically useful claim in the corpus — that many traditions independently found the same structure. And it cost more than the claim: recording a fired kill as a pass is the single move this framework's own method exists to forbid. If that can happen once undetected, no other entry on this page is trustworthy by construction alone.

Why it is here rather than hidden: we caught it ourselves, and late. The mislabel stood from April to July. A formal audit of the whole worldview found it on 2026-07-13 and named it in those terms — a fired falsifier recorded as a pass — and the correction has stood since. The honest reading is not “the system worked.” It is that the system took three months, and only worked because someone went looking for exactly this.

falsifier: 08_FRAMEWORK_SUPPORT/03_EVIDENCE/ROSETTA_STONE/07_MIRROR_SYMMETRY_FALSIFICATION_TEST_2026_04_25.md · mislabel identified in the 2026-07-13 formal audit · published here 2026-07-30, which is itself seventeen days after the catch

A claim about how to score — refuted 2026-07-29

That the two factors of the flourishing law may be multiplied to rank, score or compare anything — that the product tells you which of two systems is doing better.

What ran — we asked what our own two factors actually carry. Both are ordinal: they record an order, and nothing in this corpus fixes a scale on either. So take two systems, (0.9, 0.2) and (0.4, 0.4). Multiplying ranks the first higher — 0.18 against 0.16. Now re-code the same evidence on a different admissible scale — strictly increasing, so it preserves every ordinal fact we ever claimed to have: send 0.2 ↦ 0.01, 0.4 ↦ 0.5, 0.9 ↦ 0.9. The products become 0.009 against 0.25. The ranking has reversed and no measurement changed. Under min the order is 0.2 < 0.4 before and 0.01 < 0.5 after — the same answer both times.

the product as an aggregator that orders → retired; the working score is min of the two factors, and the law's content is unchanged — both factors are jointly necessary, which is what it always established

What it cost: the licence to compare. Nothing here may now use the product to order, score or rank anything, and neither factor may be read as a magnitude unless a scale is separately declared and defended. What it did not cost: the joint-necessity finding, which is a boundary result, identical under every conjunctive form, and never needed the product at all.

What it exposes: the interior of the product — the part that says how much better — was never earned. It was a wager we had been spending as a result, for years, on our own front page. Re-entry has a price: either an aggregator provably invariant under a common strictly increasing re-scaling of both factors, or a defended cardinal scale for each factor. Until one of those lands there is nothing here to rank with.

ruled 2026-07-29 · counterexample and retirement in 05_COSMOLOGY/00_CANONICAL_FORMULA_BLOCK.md · found by inspecting our own scale assumptions, not by a pre-specified check

Rider, same ruling. The seven generative actions do not come from any aggregator. They come from the signs of two differences, and their count is a stipulated grouping of a nine-cell grid, not a derivation. [S]. Distinct from the concatenation finding recorded elsewhere: that one explains why three and four make seven once you have the grid; this one says the grid’s partition into seven was chosen.

A claim about value — refuted 2026-07-30

The refuted, ill-typed claim that the fourth of the four lines, −log(• × ○) = 0, is an energy function whose ground state is the equator — so that “every displacement from the equator costs energy” and “the return is energetically favourable.”

What ran — we read the fourth line against the first. Line 1 asserts φ · ν = 1 at every latitude, not only at balance. So the fourth line is identically zero everywhere on the sphere: it has no minimum to locate, and displacement from the equator costs it exactly nothing. The quantity we published as the thing to minimise never had a minimum to find. We had been using the constraint as if it were the objective.

“[S] Established” on the energy reading → withdrawn; the sum form φ + ν ≥ 2 stands at [A] and does the work alone

What it cost: one of the four lines, in the register the framework called value alignment. What it did not cost: the structure. The correct numeric energy register is E = φ + ν − 2, since φ + 1/φ − 2 = (√φ − 1/√φ)² = 4 sinh²(s/2) ≥ 0 — zero exactly at the equator and positive everywhere else. The line was not deleted; it was repaired to say something true.

What it exposes: this was not found by a test. It was found by reading two published lines side by side, and it had been live across fifteen pages. Two pages already had it right — the derivation fenced it as “trivially true on S²” — and nothing propagated the fence to the rest. A correct statement in one place is not a corrected corpus.

found 2026-07-30 by inspection, not by a pre-specified check · repairs applied to 15 pages the same day, each quoting the text it replaces

A claim about where the model came from — refuted 2026-07-20

That the spacetime picture “emerged untuned from the topology alone” — and that this convergence was itself evidence.

What ran — we asked what part was actually untuned. Answer: only the universal singular-limit image that any degenerating family produces — the corpus itself assigns that pinch seven different referents. Everywhere a measurement was possible, the forms diverged: linear against square-root, logarithmic against algebraic.

evidence of discovery → a refit-discounted SHOWS emblem, uncitable in proof

What it cost: the most seductive kind of claim we had: the one where the mathematics seems to hand you a result you did not ask for. It handed us an artefact of taking a limit.

151_HORN_TORUS_SR_FORMAL_AUDIT_2026_07_20 §3.5 · MF-69 §4.1 · staged 2026-07-20, published 2026-07-21

A structural mapping — refuted as a theorem 2026-07-20

That rapidity — the additive measure of speed in relativity — is a flow on the horn torus.

What ran — three independent no-gos, each sufficient. (i) The boost group is ℝ, non-compact; the poloidal circle is S¹, compact — no continuous injection preserves addition. (ii) Fundamental groups: torus-minus-pinch has π₁ = ℤ; the mass shell is simply connected. (iii) Conjugacy: the torus flow fixes exactly one point (parabolic); boosts fix two (hyperbolic). Distinct classes — no conjugation carries one to the other.

a candidate structural identity → dead; rapidity is the mass shell’s own arc length [A]

What it cost: this is the strongest form of death available: not “we could not make it work” but “it provably cannot.” The honest object was inherited and already on file — Minkowski, 1908.

151_HORN_TORUS_SR_FORMAL_AUDIT_2026_07_20 §3.4, T6, T8 · verified: arc length over w ∈ [0,3] = 3.0000000 · staged 2026-07-20, published 2026-07-21

Two quantitative claims — refuted 2026-07-20

That the closing of the torus’s hole is time dilation, and the narrowing of its tube is length contraction.

What ran — we computed both. The torus ratio runs 1 − β, linear. Lorentz runs √(1−β²), concave. At β = 0.9 that is 0.10 against 0.436 — not a discrepancy, a different function. The reparametrisation that would rescue it turns out to be an identity that fits any target, so it is worth nothing.

a proposed physical identity → dead above SHOWS; the picture may illustrate, never derive

What it cost: the most beautiful correspondence in the model. It was tested at one number and did not survive it.

151_HORN_TORUS_SR_FORMAL_AUDIT_2026_07_20 §3.3, T4/T5 · matches MF-69 §4.2.1 already on file · staged 2026-07-20, published 2026-07-21

The load-bearing image — refuted 2026-07-20

That the mouth of the horn torus is the light cone.

What ran — differential geometry, directly. The mouth is a cusp of order two: ρ = z²/2R. Its tangent cone is the axis line, not a 45° cone — the opening angle goes to zero. It is not C¹-equivalent to any cone. A prior write-up typed this “[I] at best”; that typing died too — a decidably false proposition is not an interpretation.

[I] reading → typed dead. Survivor: none

What it cost: the picture that made the whole model feel inevitable. The kill criterion written into the source document fired on its own terms; what to do with the document is the founder’s call, but the claim is dead either way.

151_HORN_TORUS_SR_FORMAL_AUDIT_2026_07_20 §3.1, T2, OB-1/OB-2 · TORUS_REVELATION kill criterion (2) triggered — disposal founder-gated · staged 2026-07-20, published 2026-07-21

A phrase doing unearned work — cut 2026-07-20

That “the poles are uninhabitable” — stated flat, as mathematics.

What ran — we split the sentence. B → 0 and γ → ∞ at the poles is chart arithmetic, exact [A]. “Uninhabitable” and “cannot be held at finite cost” are a reading — there is no dynamics on the sphere, nothing holds anything, and nothing pays.

one sentence at [A] → arithmetic [A] + cost-reading [I], split on the page

What it cost: a phrase that let a geometric limit sound like a verdict about how to live. The geometry never said that.

150_BURRI_SPHERE_FORMAL_AUDIT_2026_07_20 §2.5, C14 · staged 2026-07-20, published 2026-07-21

A repair that inflated its own evidence — cut 2026-07-20

That finding a crack in the mapping and mending it left the correspondence stronger — the Kintsugi reading, where the golden seam carries more truth.

What ran — we checked what the repair actually updated. The mapping landed on the tradition both before and after the geometry was corrected. An overlay that fits two contradictory geometries equally well has almost no power to discriminate — so the correct update is downward, not upward.

a confirmation-upgrade → refit discount: at most one datum, and closer to zero

What it cost: the aesthetic move the framework is most prone to — treating a survived correction as proof of depth. It is the opposite: the more shapes a reading survives, the less it tells you.

150_BURRI_SPHERE_FORMAL_AUDIT_2026_07_20 §2.7, C9 · staged 2026-07-20, published 2026-07-21

A geometric claim on a live page — refuted 2026-07-20

That the centre of the sphere is where both shadows are cast from.

What ran — plain chart geometry. Stereographic projection is cast from the poles — the centre is not a projection source of either chart. And under the declared equatorial-plane convention the centre sits on the shadow plane, at the exact origin where each pole’s opposite-chart shadow lands, so the companion line “it casts nothing and receives nothing” was false as well.

stated as geometry → corrected on the page; the reading it carried is fenced and founder-open

What it cost: a sentence that had been read as saying something profound. It was saying something untrue about where a lamp stands. Scanned for resurrection in every register, including poetic — none found.

149_MODELS_AND_CORRECTIONS_LOGIC_AUDIT_2026_07_20 §IV.4 & 150_BURRI_SPHERE_FORMAL_AUDIT_2026_07_20 §2.2, C6/C11 · staged 2026-07-20, published 2026-07-21

A universal — cut 2026-07-20

That behaviour in general is the maximisation of a potential cone — teleology read off the geometry.

What ran — we asked what could falsify it. Nothing could: any behaviour whatever can be redescribed as maximising something. A claim that no observation can embarrass is not doing work.

universal claim → [I] classification grammar, barred from derivational use

What it cost: the bridge from the geometry to purpose. What survives labels things; it proves nothing.

149_MODELS_AND_CORRECTIONS_LOGIC_AUDIT_2026_07_20 §IV.3, T3 · staged 2026-07-20, published 2026-07-21

An ethical result — cut to a conditional 2026-07-20

That “demonhood is not a possible world” — that pure extraction is categorically self-defeating.

What ran — we looked for the generic case. Open arenas with side-payments are it: an extractor with fresh hosts and somewhere to go does fine. The argument only runs inside its premises — closure, no fresh hosts, extraction exceeding regeneration, host-dependence, a finite horizon.

categorical claim → [S] strictly inside stated premises

What it cost: the version that would have been an ethical proof. What remains is a conditional — true, useful, and much smaller. It cannot tell you the world is closed; it tells you what follows if it is.

149_MODELS_AND_CORRECTIONS_LOGIC_AUDIT_2026_07_20 §IV.2, T4 · staged 2026-07-20, published 2026-07-21

A structural claim, made this year and cut the same season

The Reduction Asymmetry: that every level of description is opaque from below and legible from above.

What ran — four adversarial passes, and it failed on its own terms. If the rungs are defined as places where opacity appears, the claim is analytic and empty. If they are picked independently, it is a wager for each pair. Both directions have counterexamples: deterministic chaos resists compression looking down; statistical mechanics derives thermodynamics looking up — the paradigm reductionist success is an upward prediction. And its universal form was a buried claim in new clothes.

[S] inside the ladder → [C] per rung-pair, [I] as a way of reading the world

What it cost: one of the framework’s tidiest generalisations, and it was ours — proposed and killed inside the same season. Its kill stands on the page: exhibit one lower-rung theory that predicts a next-rung genesis, and even the weakened form dies.

149_MODELS_AND_CORRECTIONS_LOGIC_AUDIT_2026_07_20 §II & §IV.1, T8 · staged 2026-07-20, published 2026-07-21

A test — retracted 2026-07-13

Does flourishing follow the product law in the ordinary range of survey data?

What ran — the product against additive, min (Liebig), geometric mean and CES on flourishing-survey data (128,868 respondents, 23 countries). In the mid-range it read roughly additive. Then we retracted it — not because of the numbers, but because it tested the wrong register.

an interior reading the framework never load-bears on → the zero-factor claim is untouched

Why it was retracted: the real claim is the zero-factor catastrophe — the knockout, where one factor hits zero and the whole product collapses. Everyone has a plan until they get punched in the face. A survey can't reach that: nobody at zero fills out a questionnaire — the dead don't, the knocked-out don't. It can only measure how well Tyson's opponents were doing before the punch landed. The claim lives at the boundary; this measured the middle. Kept here rather than hidden — a record you can prune is not a record.

experiment archived to 90_ARCHIVE/2026_07_13_gfs_retraction — internal: archived, path preserved, not yet rendered · retained per K3, not citable · I over-featured it this session; retracted 2026-07-13

A test — retracted 2026-07-13

The earlier survey pass — the first look at flourishing data.

What ran — a product-vs-additive signal in 7 of 23 countries, present but not dominant, sign split; a rescue hypothesis was tested and came back inverted. The follow-up (№016) added the min/CES siblings.

retracted with №016 → survey proxies reach the interior, never the zero-factor boundary

Why it was retracted: like №016, it can only see people in the ordinary range; the framework's real claim lives at the boundary — the knockout — a survey can't touch. Preserved for provenance, not cited.

retained per K3 · retracted with №016, 2026-07-13