Quantum-Gravitational Decoherence as the Origin of Primordial Perturbations: A Cyclic Cosmology from Exposed Singularities

We present Version 6.0 of a cosmological framework in which primordial perturbations originate from quantum-gravitational decoherence at Planck-scale Kerr remnants. This version introduces a previously absent consistency check: two rigorously independent calculations of the Planck-scale initial condition for inflation are compared for the first time. The first, established in V5.0, derives the Euclidean instanton action B = 0.74^(+0.23)_(−0.09) purely from horizon geometry — the LQG area gap, the Kerr/Schwarzschild horizon area ratio, and the perturbative rotational correction δ_rot ≈ 0.15. The second, introduced here, applies the Euclidean Friedmann constraint at the bounce surface using only the independently fixed nucleation radius a_i ≈ 2.2 ℓ_P, yielding (1/2)φ̇²_i ≃ 0.620 M_P⁴ — a result parametrization-independent by construction, since the Starobinsky potential contributes ≤ 10^(−10) M_P⁴ in the relevant range. The two calculations, derived by entirely different routes and with no parameter tuned to force agreement, converge to within 16%. We argue that this convergence constitutes a non-trivial structural consistency check on the framework's overall architecture, while emphasizing that it is not claimed as an exact identity. A previously asserted Hamilton-Jacobi derivation of B = (1/2)φ̇²_i is explicitly retracted after source verification. Three open problems are documented in full. First, the residual 16% gap and the associated shear/rotation partition cannot be independently calculated without first establishing whether the Euclidean section of Kerr with a/M = √3/2 constitutes an admissible Riemannian geometry under Witten's (2021) criterion — a question of independent interest that remains unresolved in the literature. Second, contrary to a previous internal claim, numerical integration reveals that the framework does not produce N_eff ≈ 60 e-folds of slow-roll inflation regardless of the initial field value φ_i; only φ_i ≳ 15 M_P succeeds, while smaller values overshoot the Starobinsky plateau and abort inflation within a few e-folds. This φ_i is not currently fixed by any element of the Kerr/LQG nucleation geometry. Third, the resulting kination phase N_kin ≈ 3.73–3.76 e-folds is logarithmically insensitive to the 16% discrepancy. The framework's central falsifiable predictions — r ≈ 0.0033, n_s ≈ 0.9678 — depend only on N_eff, not on B or N_kin, and remain unchanged whenever N_eff ≈ 60 is realized. LiteBIRD (2028+) measurement of r > 0.0083 or r < 0.0013 at 5σ excludes the model without parameter adjustment.