A Galaxy Law Tied to Dark Energy, and a Map of What It Leaves Unexplained (Working Note)
Description
Working note, version 1.2 (2026-11-04). AI-assisted research programme; not peer reviewed; not reproduced by any outside group. kappa = 1/2 is fitted; the theory is not closed; nothing here says the data favour this framework over any other.
A programme report after its own final grade. The galaxy law g_N = g^2/(g + a0), with a0 = kappa c sqrt(G rho_DE), is MOND's simple form with a proposed origin for its scale, proposed earlier by other authors. It describes isolated disc galaxies well on the canonical footing, and its tie to dark energy has no decisive test yet. It fails IC 1613, several Local Group classes, cluster shear, X-COP (factor 2.93) and the unscreened Solar System, its Milky Way result depends on a vertical form it doesn't supply, and it has no account of the CMB; its one action, v3s, is adopted only in a limited scope. The cold sector built to cover those failed on measured data and is retired. The note gives the scorecard, the two walls, the covariant attempts on record, twelve requirements for any successor and the dated tests that can still kill the law. It is numerical evidence for the families tried, not a no-go theorem.
Version 1.1 adds lanes run after the retirement; none was blind, all were run by AI subagents, and no outside group has reproduced them. AR1 landed, MOVES AND FAILS, conditional on one missed check. A gas term fitted on X-COP reaches 14 of 15 keys on the canonical footing (fragile, never 15; alt 10), and tied to present gas it is dead on the Bullet. Published AeST, action searches, a one-component requirement calculation and a 96-row harness are reported; all fail or are empty for what was listed, and none of this is a no-go theorem. A repaired test gives no statistical support that the a0 formula is special (5.51 bits, FDR-DEAD; 92.7 % of scrambled scalars decompose as simply); that doesn't refute it, and the test lacks power at 16.6 % precision. Correction to v1.0: v9's alpha_1 = -2(K_B+2) is a static-branch ratio; a moving source has alpha_1 = alpha_2 = 0 at G = G~, and v9 stays non-viable because it loses the law for moving systems. Added: MUSE-DARK III reports a0 rising with redshift, which counts against the flat band at face value. Alcock (MNRAS 2026) puts Lambda-tied frameworks under the same pressure. We read the rise as mass-route dependent and not decisive; on the survey's masses the band fails.
Lanes through 2026-11-04 add these, under the same caveats. No unscreened action gives the law and passes Cassini's Q2: classes whose static limit is AQUAL or QUMOND miss it by 1.69-5.18 times the bound. Only the frozen screened theory's screening length passes the Solar System, cleanly once a high-precision rerun cleared its one failed check; that theory still fails cosmology, it adopts nothing, and a0 enters by hand in every action class. The filter line, a phantom mass smoothed over a length, has no surviving form once the full 3D field is used: the local rule fails Cassini by 13.3 times (alt 18.1). Inside 0.3 Mpc the KiDS galaxy-lensing data sit about 0.09 dex above the law under every estimator on record, and the hot gas measured through kSZ doesn't close that gap. A light thermal fermion beside the law, run under the owner's lift of the no-particle rule for that lane only, fails the early universe at 5 of 6 masses and adopts nothing; the rule stands elsewhere. The derivation of kappa = 1/2 reduces to one missing length, c/sqrt(G rho), that no field equation on record supplies, and telling 1/2 from 0.461 needs a0 to 2.8 % against a 9.47 % floor. Correction: v1.0 and the first v1.1 draft said nobody had computed what published AeST's dust does in nonlinear wells; reduced 1D and 3D runs exist, and both read NO VERDICT. A dimensioned search's level L2, 7.4e11 expressions per footing, also reads NO HIT, and its scrambled control comes about as close.
Version 1.2 records the owner's decision of 2026-11-04, which superseded a no-action call made the same day: v3s, a screened Khronon theory, is adopted as the action for the a0 law, scoped to isolated galaxies, the Solar System, galaxy lensing and gravitational waves. The gates are AX1 (only v3s's screening length passes; unscreened classes fail Cassini's Q2 by 1.69-5.18 times), C6P (Q2 margin at least 4.8e7, alt 5.7e7), V3W (gamma = beta = 1 derived; KiDS isolated lensing A 1.129 +- 0.206, alt 1.035 +- 0.187; c_T = c) and V3X (Solar-System health items (a) and (c) pass). Carried weak spots: V3X's W2 and W4c(b) are undecided on numerics, the constraints act instantaneously, four items are asserted rather than shown, the lensing gate has little power, and a0 is put in by hand. v3s supplies no cluster or CMB mass, and its 2026-10-24 failures in clusters, growth and the CMB stand; those stay open. Also added: a power-law screening length opens a window at p 0.30-0.45 (FN6), a description, not adopted; ten fresh selector ideas find no action (FA1); every high-z sample on disk is blocked under the a0(z) rule (AT2).
Files
PAPER67_galaxy_law_and_what_it_leaves_2026.pdf
Additional details
Related works
- Continues
- Working paper: 10.5281/zenodo.23868979 (DOI)