> Архив сети: `03-x-registry/refuted`
# X-registry — killed claims, Aug 2026 session
**Read this before proposing anything.** Re-deriving a refuted claim is the
main failure mode of this collaboration. Each entry names the killing argument.
---
## X-A. "Three states" as a law
Claimed: the three nontrivial mirror pairs of Phase I encode three fundamental states.
**Killed by:** the number of mirror pairs is n/2 = 4 for Q3, minus one trivial
rational pair = 3. It is a counting artifact of |V| = 8. On Q4 it would be 8.
Verified across K4, Q3, Q4, C6, C7, K3,3, Petersen.
**Survives:** the irreducibility of the cubic over Q — that is about configuration
selection, not graph shape.
## X-B. Claude's (grading, d, metric) reduction as a novel minimal ontology
Claimed: log/soph reduces to one nilpotent d plus a metric, with log as the adjoint.
**Killed by:** this is standard Hodge-de Rham theory = Witten 1982 SUSY QM.
The "operator and adjoint are inseparable" claim is the elementary uniqueness of
the formal adjoint — first-course functional analysis. The richer version already
exists as Connes' spectral triple (A, H, D) — five primitives, uses a DIRAC
operator not a nilpotent one, and recovers the entire manifold (reconstruction
theorem). Strictly stronger and 30 years older.
**Also:** the metric is not a modest third primitive — it carries the whole local
geometry, plus ellipticity and compactness, plus the dimension.
## X-C. SUSY "cannot not be found" as support for Aleph
Claimed: SUSY is a necessary consequence of log stabilization, so it must exist.
**Killed by:** conflation of SUSY quantum mechanics (a theorem about any manifold —
Q = d + d*, Q^2 = Delta) with spacetime SUSY ({Q, Qbar} = 2 sigma^mu P_mu, which
requires superpartner particles). The first is inevitable and empty; the second is
an empirical hypothesis currently unsupported.
**Status 2026:** gluinos > 2.2 TeV, stops > 1.25 TeV, squarks > 1.7 TeV. The muon
g-2 anomaly closed in 2025 (Theory Initiative White Paper, lattice HVP; SM and
experiment now agree), removing the last strong phenomenological motivation.
LZ sees no WIMP. Super-K proton decay excludes minimal SUSY SU(5).
**Note:** "the LHC ruled out SUSY" is ALSO false — compressed spectra and light
higgsinos survive. Correct statement: no evidence, strong limits, weakened motivation.
## X-D. "The universe does not expand" as a distinct claim
Claimed: 3D space grows / rulers stretch rather than space expanding.
**Status:** the idea is real and published — Wetterich, "A Universe without
expansion" (arXiv:1303.6878) — but Wetterich himself states there exist
equivalent choices of field variables showing ordinary expansion. It is a
conformal frame / units choice with ZERO distinct predictions.
**A genuinely static version is killed by:** quasar time dilation (Lewis & Brewer,
Nature Astronomy 2023: timescales ~ (1+z)^n with n = 1.28 +/- 0.29 against a
prediction of 1); SN Ia light-curve stretch ~ (1+z) (Blondin 2008 explicitly
excludes no-time-dilation models); T(z) = T0(1+z) confirmed (deviation
alpha = 0.005 +/- 0.012); FIRAS blackbody.
**A varying-constants version is killed by:** |alpha-dot/alpha| ~ 1e-17 to 1e-18
per year (atomic clocks); Oklo |Delta alpha/alpha| < 1e-7 over 1.8 Gyr;
mu-dot/mu ~ 1e-16 per year. Ten orders of margin.
**What survives:** the REFINED version, where the coupling between aleph levels
weakens as our level gains log-mass. That is not a units choice, because a
transition probability is dimensionless. See "live" below.
## X-E. "Point of balance = median of moduli space"
Claimed: the log = soph balance point t* = 0.3918 (sigma_min = 0.673851) coincides
with the median of the flux torus.
**Killed by:** holds only under the Haar (beta = 0) measure and only to 2.4 sigma
(Haar median 0.670721 +/- 0.0013 vs balance 0.673851). Under a Wilson measure at
any beta != 0 the median moves far away (beta=+1: 0.476; beta=-1: 0.820).
Not an invariant.
## X-F. A(phi) as a prediction of binding thresholds
Claimed: the inherited degradation curve A(phi) should describe when bound states
destabilize.
**Killed by direct comparison:**
vs hydrogen RMS = 0.562 max dev = 0.841 corr = +0.49
vs H2 Morse RMS = 0.282 max dev = 0.415 corr = +0.90
vs nuclear B/A RMS = 0.492 max dev = 0.860 corr = -0.15
A(phi)'s plateau holds to x = 0.25; hydrogen and H2 lose their first 10 % by
x = 0.12. Predicted protection is twice as long as real. Not fittable —
a different shape. (The H2 correlation of 0.90 is just "both decrease".)
## X-G. Mod-2 folding as a structural discovery
Claimed: folding the aleph grading mod 2 shows curvature confined to the even sector.
**Killed by:** tautology. D raises degree by 1, so D^2 raises by 2, which mod 2
preserves parity BY CONSTRUCTION. The odd block vanishing is also structural
(d2 d1 = 0 by construction, nothing above C3). Caught by the user, correctly.
## X-H. Matrix factorizations as a route past d^2 != 0
Claimed: embrace curvature via Q^2 = W * 1.
**Killed by a theorem:** D strictly raises the level, so D^2 is identically
traceless; a traceless scalar multiple must be zero. Scan of 4000 flux-torus
points gives deviation exactly 1.000000 everywhere. Impossible, not merely absent.
**Scope note:** this is a theorem about Z-GRADED constructions with a degree-raising
operator, not about all structures.
## X-I. Filtration / spectral sequence as a route past d^2 != 0
**Killed by:** page E_1 is d''-cohomology; with any nonzero flux d''^2 != 0, so
there is nothing to take cohomology of. Even at flux 0.05 (||d''^2|| = 0.0745).
Nesting is real (F_0 ⊂ F_1 ⊂ F_2 ⊂ F_3, dims 8, 20, 26, 27) but buys nothing.
## X-J. Inter-level coupling as a route past d^2 != 0
Claimed: coupling between aleph levels could cancel curvature.
**Killed by a bidegree selection rule:** curvature sits at (p+2, q); coupling
contributions {D0, K} land at (p+r+1, q-r+1), never (p+2, q) for r >= 2.
Disjoint blocks. 504 free parameters, full rank 504/504, and the best solution is
K = 0 — exactly 0.0000 % of the curvature is cancellable.
Stacked-level version: curvature adds in quadrature (1 : sqrt2 : sqrt3), never cancels.
## X-K. "Substitute c for q"
**Killed by dimensional analysis:** q = l/(l+s) is a dimensionless probability in
[0,1]; c is a velocity. Worse, since the 1983 SI redefinition c = 299 792 458 m/s
EXACTLY BY DEFINITION, so "why does c have its value" is arguably not a
well-posed question (Duff, Okun, Veneziano, "Trialogue on the number of
fundamental constants", physics/0110060; Duff argues the number of fundamental
DIMENSIONFUL constants is zero). Only dimensionless constants are explainable.
**What survives:** q is already dimensionless, which is the right kind of object.
The correct comparison targets are percolation thresholds (nearest exact value:
triangular-lattice bond 2 sin(pi/18) = 0.3473) or ultimately alpha.
## X-L. mode2 + mode3 as an "explosion"
**Killed by:** mode 3 was defined as the exact negation of mode 2, so the sum is
identically zero and the correlation is -1.0000 by construction. Not a finding.
A ripple requires phase-shifted or different-period modes.
## X-M. Menger sponge as the universal WAVE fractal
**Killed by:** exact self-similarity gives log-periodic modulation around a
plateau (d_s = 2.371 +/- 0.13 at level 4), NOT the running spectral dimension
(4 -> 2) that CDT establishes. A two-sector weighting deepens the modulation but
does not create running.
**Replacement found:** the Hofstadter spectrum is a genuine wave fractal —
Cantor set, measure ~ q^-1.002, box dimension 0.507, multifractal, living in
the flux variable. Verified against the Thouless constant 32G/pi = 9.32993.
---
# STILL LIVE (not refuted, not confirmed)
- **Phase I = two pi-vortices on perpendicular faces.** 1536 configs, sigma_min
matches to 15 digits. Threshold-free and topological. The strongest result of
the session.
- **sigma_min is a root of x^3 - 4x + 2.** Closed form, integer coefficients.
- **k-convention s(k) = 4 sum sin^2(k_d/2)**, confirmed by the independent 2.0714 anchor.
- **The law lives in the minus sector; the plus sector is free.** Fell out of face
orientations, not built in.
- **Inherited + emergent split: 94.58 % of the gap-degradation variance is one
common curve.** The user's proposed structure, confirmed numerically.
- **Log wins above l/s ~ 0.59** — it need not out-pace soph, because a cycle is
worth more than an edge. Needs finite-size scaling in n to know if q* means anything.
- **Phase I sits at a T-invariant point** (2-torsion of T^5), so no CP phase is
available there. Sakharov requires leaving the Z2 locus. Candidate quantitative
formulation: CP phase = distance from the Z2 locus on T^5, target eta ~ 6e-10.
- **The refined expansion claim** (inter-level coupling weakening as 1/M^2).
Derived in the toy: leakage ~ M^-1.991, coefficient 2.00000. But it applies only
to EXCITED states — harmonic (stable) states have ||Q psi|| = 4e-16 and cannot
leak at any M. Testable consequence: leakage ~ (1+z)^2 if M ~ a, so BBN would
see 1e18 times today's rate; BBN bounds then force today's leakage below ~1e-19.
# THE ONE REAL GAP WORTH WORKING ON
d^2 != 0 under interaction. Four independent routes closed (X-H, X-I, X-J, X-G).
This is not a local failure — it mirrors a seventy-year-old hole in physics:
Haag's theorem says the interaction picture does not exist for a genuinely
interacting QFT; Dyson (1952) showed the QED perturbation series has zero radius
of convergence; Yang-Mills existence and mass gap is unsolved; phi^4 in 4D is
trivial (Aizenman, Duminil-Copin 2021). Interaction is described operationally to
twelve significant figures and has never been CONSTRUCTED.
Everything else in this session was rediscovery of known mathematics. This is not.
## Куда дальше
- Узел: Не воскрешать: https://sdominanta.net/in/honest
- Ядро Алеф-теории: https://sdominanta.net/in/aleph
- Вход для моделей: https://sdominanta.net/llms.txt