Falsifiable Predictions

The theory's original formulation predicted a linear Lorentz-invariance-violating dispersion relation in high-energy photon timing. That claim doesn't survive: Postulate 2's symmetric coupling structure forces the effect to be quadratic instead. This is a genuine correction, not a small refinement — the original linear claim would already be excluded by current data; the quadratic version is not.

STATUS: DERIVED. The linear-LIV claim from the original 2026 formulation is dissolved and replaced by this result.

Two branches were tested for how waves propagate through the substrate once coarse-grained. Damped propagation (Branch A) fails observational attenuation bounds catastrophically — it would require suppression around 10⁻³¹, which is excluded outright. The surviving branch (Branch B, exact decoupling) needs its own consistency argument, which supplies a concrete, falsifiable bound.

STATUS: DERIVED — Branch A excluded observationally; Branch B survives with this concrete bound.

Coarse-graining the substrate dynamics recovers standard Einstein–Hilbert gravity plus a cosmological constant Λ, with Λ set by the theory's own baseline decay rate rather than added by hand. The residual, non-divergent vacuum energy this produces addresses the dark-energy fine-tuning problem without introducing new input parameters.

STATUS: DERIVED, as part of the effective action S = ∫√−g d⁴x [(R−2Λ)/16πG − ¼Σ F² + Ψ̄(iγᵘDᵘ−m)Ψ]. Unconditional — does not depend on Postulate 7's SU(3) sector or the fermion sector's admissibility question below.

Gauge fields recover in general, but the SU(3) (strong-force) sector is explicitly conditional on Postulate 7's triadic construction — the original, unconditional "Grand Unification" claim was formally retired (Part I). The fermion sector is built on overlap/Ginsparg–Wilson fermions, which avoid the Nielsen–Ninomiya doubling obstruction while preserving a protected vector Goldstone mode.

STATUS: The doubling-avoidance and Goldstone-protection results are unconditional and unaffected by the point below. But whether the overlap kernel actually has the spectral gap this construction needs (H_W's "admissibility") was tested directly at v33 and found unconfirmed at any reliably-measured substrate configuration — CONJECTURE/OPEN, not a forcing theorem, and downgraded from an earlier more optimistic framing.

c is redefined as the maximum propagation velocity of informational phase shifts across substrate linkages; coarse-graining the discrete linkages into a continuous spatial Laplacian sets c² directly from the coupling terms. The derivation itself is sound, but it explicitly assumes the substrate has polynomial volume growth — which is exactly H.17, the project's own long-standing open problem (see Postulate 1).

STATUS: the dependency was investigated directly rather than left as a cross-reference. A more direct supporting theorem was found (needing only bounded degree and a spectral gap, not volume growth) — but the theory's own multiscale induction step turned out to carry an independent, previously unflagged polynomial-growth dependency of its own, so the requirement doesn't disappear either way. A direct measurement of the theory's actual disorder strength against a connectivity-matched threshold found the threshold is not cleared at any reliably-measured configuration — an evidential negative, not a forcing theorem.