PUH Theorem 290 (Recovery and Correction) — The T169 Exclusion Rule Was Filed in April: the Second Addendum's Centralizer Criterion Closes T284's Named Gap Before It Was Named; T284 Section 6 Corrected; and the Null-Channel Theorem [E_(±θ), E7] = 0 M

Photonic Universe Hypothesis (PUH) — Recovery and Correction. BACKGROUND. T284 (10.5281/zenodo.21316357) corrected a real error: the branching 248 = (133,0)+(1,0)+(56,+1)+(56,−1) sums to 246, in five places across T167, T168 and T169; the correct decomposition restores the singlets 1_{±2}. THAT CORRECTION STANDS AND IS NOT REOPENED. But T284's Section 6 went further, asserting that T169 "supplies no rule excluding them from the photon's environment — because T169 did not know they were there," and entered "derive the exclusion rule" as a standing open item. THE RECOVERY. That sentence is false, and the archive itself falsifies it. T169's SECOND ADDENDUM (10.5281/zenodo.19431882) — filed the SAME DAY as the main paper, April 5, 2026 — contains: (a) the CORRECT charge decomposition (Lemma 1): 134 states at charge 0 (126 E7 roots + 8 Cartans), 112 at ±1, and 2 at ±2, EXPLICITLY IDENTIFIED AS THE HIGHEST ROOT ±theta, summing to 248; and (b) THE RULE: N_algebra = dim(ker ad T_gamma) − 1 = 134 − 1 = 133 — the centralizer of the photon minus the photon itself, with E7 forced as the centralizer ("It is the definition of E7 in this context"). One criterion; the charge-±2 states fail it for exactly the same reason the charge-±1 states do. The gap T284 named in July was closed in April by T169's own family. The archive contradicted itself within a single day's filings — the wrong equation in the main paper, the correct table in its own addendum — and carried both, unremarked, for three months. THE CORRECTION. T284 Section 6's "no rule / did not know" claims are WITHDRAWN. Retained in full: the five-occurrence equation corrigendum, the T167/T168 impact analysis, the survival of 137, the "exact and canonical" retraction. How the miss happened, honestly: T284 hunted the erroneous EQUATION and verified every paper containing it; the Second Addendum never quotes the equation — its Lemma 1 works directly from the charge partition — so it fell outside the corrigendum's search pattern. The lesson is the archive's own first rule: read ALL relevant papers in full, not merely those matching the pattern of the error. VERIFICATION. Re-derived this session on the explicit 240-root construction of E8 (photon = H_theta, charge q(alpha) = (alpha, theta)): charge table 1 / 56 / 126 / 56 / 1 plus 8 Cartans = 248 exactly; the only charge-±2 states are ±theta (direct enumeration); centralizer = 134; N_algebra = 133 from the single criterion — the 133 in, the photon out, the 56s out, the ±2 out — no exceptions. THE NEW RESULT — THEOREM 290.1 (NULL CHANNEL): [E_{±theta}, E7] = 0, IDENTICALLY. Proof in two lines: (i) for any E7 root alpha (charge 0), theta + alpha carries charge +2; the unique charge-+2 root is theta itself; theta + alpha = theta forces alpha = 0, not a root — so theta + alpha is never a root and the bracket vanishes; (ii) the E7 Cartan is the orthogonal complement of theta, so theta(h) = 0 on all of it. Verified: zero violations across all 126 charge-0 roots. The complete bracket inventory of the ±2 states: the doublet action 56_{−1} → 56_{+1} (all 56 of 56 — matter mapped to matter) and [E_theta, E_{−theta}] = H_theta — back to the photon itself. Nothing else. By contrast the 56 channel is LIVE: its brackets reach ALL 126 of the 126 E7 roots — the entire environment. THEOREM 290.2 (CONNECTIVITY-FORCED EXCLUSION): defining the photon's algebraic environment, per T169's own coupling narrative, as the states the photon's channels can reach: the environment is exactly E7 — all 133, reached in full by the live 56 channel — and the 1_{±2} contribute nothing: as environment they fail the charge criterion; as channels they open onto nothing outside {photon, themselves, matter-to-matter maps}. With the 1_0 they close the su(2)_theta — the same internal su(2) T283 Section 6 identified and rejected as spin's. THEY ARE THE PHOTON'S OWN LADDER, NOT ITS NEIGHBOURHOOD. One cannot count a window that opens onto the room one is standing in. The centralizer rule and the connectivity rule agree exactly; the exclusion is doubly grounded — by membership and by reach — and 1/alpha_0 = 133 + 4 = 137 is FORCED. THE RULE, ONCE: Environment_algebra = ker(ad T_gamma) \ {T_gamma} = E7, as a subalgebra. Four verdicts, no exceptions: 133_0 IN; 1_0 OUT (the photon does not count itself); 56_{±1} OUT as environment, IN as the live channel that renders Sector A countable; 1_{±2} OUT twice over — charged and null. Sector B's +4 (the graviton Cartans; the addendum's orthogonality argument for additivity) is untouched in either direction: it stands as filed. NOT CLAIMED: any new physics beyond Theorem 290.1; any change to T284's equation corrigendum; any reopening of Sector B's additivity; anything regarding the anti-screening / Interpretation-B thread (T209 is context only); or that T284's method was unsound — the lesson is only that a corrigendum's search pattern must include the papers that DON'T contain the error. PROCESS NOTE, in the archive's own spirit: this is a correction of a corrigendum — the immune system pointed at its own antibody. T282 withdrew predictions; T284 corrected equations; T287 v1 was withdrawn after upload; T288's draft errors were caught before build; T290 now corrects T284. Each correction has been caught earlier, and by the archive's own machinery, than the last.