Published June 3, 2026 | Version v1

Public-Data Falsification Tests for the Entropy-Response Gravity Framework: Methodology, Tiered Test Suite, and Explicit Postulate-Dependence Caveats

  • 1. Independent researcher

Description

We present a public-data test suite for the Entropy-Response Gravity (ERG) framework, accompanying companion Colab notebook ERG_Framework_Tests.ipynb archived alongside this paper. The suite is organised into two explicit tiers: Tier-1 tests probe derived framework predictions and constitute genuine falsifiers; Tier-2 tests probe downstream postulates whose structural derivations are not yet established (specifically the bipartite modular spectral triple with Pell-Lucas eigenvalues, the resonance scale a_res = a_0*sqrt((4*pi^2-1)/2), and the cosmological feature at z ~ 0.617 where H(z) = sqrt(2)*H_0). Null results from Tier-2 tests constrain the postulates without falsifying the framework's core predictions; we make this distinction operational. The suite uses NuFIT-6.0 oscillation parameters, the SPARC galactic database, ACCEPT cluster data, Pantheon+ supernovae, and DESI BAO measurements, all publicly available. We provide explicit reproduction instructions for each test.

The intent is to expose the framework's empirical commitments transparently, including the speculative commitments, so that observational programmes over the next decade (Euclid DR1, JUNO, DESI-Simons Observatory, CMB-S4 / LiteBIRD, nEXO) can constrain or rule out specific framework components without ambiguity about what is being tested.

Tier-1 tests (six derived predictions, framework falsifiers):
- T1.1 SPARC galaxy rotation curves: standard mu(x) at 6.4 sigma across 173 galaxies
- T1.2 CMB spectral index: n_s = 0.9741 from N_e = 77.23 (ACT DR6 agreement at -0.06 sigma)
- T1.3 Tensor-to-scalar ratio r = 0 exactly (CMB-S4/LiteBIRD test)
- T1.4 Gravitational slip eta = 1.000 (Euclid DR1 October 2026)
- T1.5 Cluster M-T scaling at 12.4 sigma (eFEDS)
- T1.6 Sum of neutrino masses 58.72 +/- 0.30 meV (DESI-SO ~2028)

Tier-2 tests (four postulate-dependent constraints, NOT framework falsifiers):
- T2.1 Silver-ratio mass ratio m_3^2/m_2^2 = (1+sqrt(2))^4 at 0.08 sigma (NuFIT updates)
- T2.2 Cosmological feature at z = 0.617 (DESI BAO future-data test)
- T2.3 Galactic transition scale at r ~ 3 kpc (SPARC stacked residuals)
- T2.4 Cluster temperature feature at g = a_res (ACCEPT)

Plus verified mathematical identity: zeta_{Q(sqrt(2))}(-1) = 1/12 (sympy-rigorous).

The framework operates an explicit discipline rule (10 retractions, 168 caught near-misses, 3 verified derivations, 0 framework-breaking findings). This paper is the operational expression of that rule in publication strategy: rather than publish unverified Tier-2 conjectures as if derived, we publish them as conjectures with explicit caveats and explicit falsification pathways.

Notes

Version 1 (June 2026, first deposit).

Methodology / artifact paper for the ERG programme. The companion 
Colab notebook ERG_Framework_Tests.ipynb implements all ten tests 
(six Tier-1 + four Tier-2) with public-data interfaces.

Operational rule, stated explicitly:
- A Tier-1 null result at indicated significance falsifies the 
  framework (no parameters to adjust).
- A Tier-2 null result rules out the relevant structural postulate, 
  NOT the framework.

The Tier-2 separation is essential because the framework has both 
derived predictions (Tier-1) and structural conjectures whose 
derivations are open specialist problems (Tier-2). Without the 
explicit tiering, any null result on a postulate test would 
ambiguously implicate the entire framework; this paper makes the 
dependence chains explicit so the interpretation is decided in 
advance, not after the result.

This paper does NOT claim:
- That any Tier-2 test has detected a signal (none have).
- That the bipartite modular spectral triple has been constructed 
  (open specialist problem in operator-algebraic NCG).
- That the identity zeta_{Q(sqrt(2))}(-1) = 1/12 derives Newton's 
  constant (the connection is heuristic and labelled as such).

Pre-registration: Tier-1 predictions are pre-registered at OSF 
10.17605/OSF.IO/8Z27K (master, 3 predictions) and 10.17605/OSF.IO/
5HCGF (Paper XII specific, 6 predictions including v2 silver-ratio 
extension). Tier-2 predictions T2.2-T2.4 are pre-registered as v1 
entries in the present deposit; they will be added to OSF in 
conjunction with this paper.

Discipline-rule track record (June 2026): 10 retractions, 168 caught 
near-misses, 3 verified derivations (beta-function from foundations, 
J_{02} = K_R, alpha_Dirac = 1/8), 0 framework-breaking findings.

Files

T4_METHODS_v1.pdf

Files (341.5 kB)

Name Size Download all
md5:7a665488f71da0f11a2393ba5240c288
312.6 kB Preview Download
md5:6f3592baf4fe0a722a3073b9d881f5a9
28.9 kB Download

Additional details