PUH Theorem 321 (Vacuum Birefringence Requires Anharmonicity) — The Harmonic Bond Term Contributes EXACTLY ZERO; the Quartic Term Is UNIQUE Because E8 Has No Degree-4 Casimir; and QED's Parameter-Free Prediction Determines Its Coefficient

Photonic Universe Hypothesis (PUH) — Limitation and Constraint. THE OBSERVATION. Polarimetry of a magnetar over more than 140 hours has reported polarization nearly three times that of comparable sources, interpreted as evidence for vacuum birefringence — the prediction that sufficiently strong magnetic fields make empty space an anisotropic optical medium, the two polarization states propagating at different speeds. THE REFRAMING, WHICH RUNS OPPOSITE TO THE OBVIOUS ONE. It is tempting to read a substrate-modified vacuum as support for a framework built on a substrate. THAT READING IS WRONG, for two reasons. FIRST, the effect is already predicted quantitatively with NO free parameters: n∥ − n⊥ = (α/30π)·(B/B_c)²·sin²θ with B_c = m²c³/eℏ = 4.41×10⁹ T, prefactor α/30π = 7.74×10⁻⁵ — both the coupling constant and the critical field measured elsewhere. Magnitudes: 8.1×10⁻²¹ in the strongest laboratory magnet (45 T), 4.0×10⁻⁸ for a typical neutron star (10⁸ T), 4.0×10⁻² for a magnetar (10¹¹ T, though above the critical field the perturbative form is only indicative). SECOND, this framework does not replace electrodynamics — T186 DERIVES the field equations from lattice phase modes, so the vacuum polarization producing birefringence would itself be a lattice effect. THE FRAMEWORK MUST THEREFORE REPRODUCE THE ESTABLISHED NUMBER EXACTLY — not exceed it, since any excess is already excluded by the measurement, and not fall short. A harder test than accounting for an anomaly, and a more useful one. THEOREM 321.1 (the harmonic term contributes exactly zero). T320's coupling is E₂ = (J/2)Σ_x Σ_α ⟨φ(x)−φ(x+α), φ(x)−φ(x+α)⟩ with J constant — a quadratic energy, whose Euler-Lagrange equations are therefore LINEAR. Linear equations obey superposition: each solution evolves as though the others were absent. A background field, however strong, cannot alter the propagation of a probe wave through it. ∎ Three consequences, none of them small: no scattering of light by light; no intensity-dependent refractive index; NO FIELD-INDUCED BIREFRINGENCE OF ANY MAGNITUDE. A purely harmonic lattice has no nonlinear optics at all. SCOPE STATED CAREFULLY: the full Lagrangian is not harmonic — the potential and the Casimir constraint are nonlinear — but the constraint is active at the core rather than in vacuum, and the potential expanded to quadratic order about the vacuum is harmonic too. So the required nonlinearity must come either from the potential's quartic term or from an anharmonic bond term. T320 supplies neither, and claimed only that its coupling was quadratic AT LEADING ORDER — a hedge the observation now converts into a requirement. THEOREM 321.2 (the quartic term is unique). The anharmonic single-bond term is unique up to a coefficient: E₄ = (K/4)Σ_x Σ_α ⟨Δφ, Δφ⟩². PROOF: a quartic single-bond term must be a degree-4 invariant polynomial in Δφ, and invariant polynomials of degree d are spanned by products of basic invariants whose degrees sum to d. E8's invariant degrees are 2, 8, 12, 14, 18, 20, 24, 30 — degree 4 can be reached ONLY as 2 + 2, the square of the quadratic invariant, since THERE IS NO INDEPENDENT DEGREE-4 INVARIANT. ∎ AND IT IS THE SAME ABSENCE TWICE: the missing degree-4 invariant is precisely what made the lattice isotropic to sixth order in T320, since with no degree-4 or degree-6 invariant the only rotational invariant at those orders is a power of |k|². ONE STRUCTURAL GAP, TWO UNRELATED CONSEQUENCES — an unusually isotropic medium, and an anharmonic response with no adjustable form. NOT GENERIC: su(4) has degrees 2,3,4; so(7) 2,4,6; so(8) 2,4,4,6 — all with degree 4. G2 (2,6), F4 (2,6,8,12), E6 (2,5,6,8,9,12) and E7 (2,6,8,10,12,14,18) lack degree 4 but all have degree 6. ONLY E8 LACKS BOTH. RESULT 321.3 (the coefficient is determined, not free). Because the established prediction carries no free parameters, matching it does not merely permit the new coefficient — it FIXES it. The pair: the quadratic term (J/2)Σ⟨Δφ,Δφ⟩ has J fixed by the wave speed, c² = 60J/ρ; the quartic term (K/4)Σ⟨Δφ,Δφ⟩² has K fixed by the birefringence, Δn = (α/30π)(B/B_c)². TWO COEFFICIENTS, TWO INDEPENDENT OBSERVATIONS, neither a fitted parameter in the ordinary sense — each determined by a measurement the framework did not choose and cannot adjust. A framework that acquires a new term and a new observation to fix it on the same day has not gained a free parameter; it has gained a constraint. WHAT IS NOT DONE. The index difference cannot yet be computed from K: that requires knowing how the electromagnetic field amplitude maps onto lattice displacement — the field normalisation — which the archive does not fix. THAT IS THE SAME OBSTRUCTION LEAVING J UNDETERMINED, and it now blocks two coefficients rather than one, raising its priority accordingly. This note establishes three things and not a fourth: that the harmonic term contributes nothing, that the anharmonic term replacing it is unique, and that the observation determines its coefficient. IT DOES NOT PRODUCE THE NUMBER. Whether the framework reproduces, exceeds or falls short of the established prediction remains open, and is the sharpest quantitative test now available to it. KILL-CONDITIONS: (i) if the quartic coefficient yields an index difference exceeding the established prediction, the excess is already excluded by measurement and the anharmonic term is wrong; (ii) if it yields substantially less, the framework fails to reproduce an established effect; (iii) if a degree-4 invariant of E8 is exhibited, Theorem 321.2's uniqueness fails; (iv) if the nonlinearity is supplied instead by the potential's quartic term, the bond term may remain harmonic and this note addresses a requirement the framework does not have — though the potential's expansion about the vacuum would then need computing, which the archive has not done. NOT CLAIMED: that the framework reproduces the established birefringence, since the number is not computed; that the observation supports the framework, which Section 1 argues it does not — it constrains it; that the anharmonic term is derived rather than required; that T320 is in error, since it stated its coupling was quadratic at leading order and made no claim about nonlinear optics; or that the field normalisation has been addressed.