A registered result in its source context

E8.ONEOBJECT.01

Original ledger classification: Exact [E] lattice/discriminant-form census results and an exact [E] group-theoretic kill, both behind explicit fences; the c_3 link stays typed INTERPRETATION (no P1 derivation); no marker move (Python-only per GATE.WOLFRAM.02).

Original claim text

The d = 4 forcing census + the 60-line clock synchronization kill, one module from two probes (21/21 + 23/23 checks, ~5 s; discovery probes one_object_reduction_probe.py ONE-OBJECT-PARTIAL and clock_fiber_product_probe.py CLOCK-FIBER-DEAD, both 2026-08-05). THE FORCING CENSUS (exact Fraction arithmetic, d = 2..12): the TFPT glue pattern (cyclic isotropic glue, evenness q = 0 mod 2Z, unimodularity |H|^2 = 4d) admits an even unimodular DIAGONAL glue of A_{d-1} (+) D_{d+1} iff d = 4 -- at d = 4 the census reproduces v92 EXACTLY (precisely two even Lagrangian glues, the graphs of the anti-isometries k -> k and k -> 3k), the explicit glued lattice is E8 (Gram det 1, exactly 240 norm-2 vectors, all integral over the glued basis), d = 9 (the only other perfect-square candidate from |H|^2 = 4d) dies at isotropy (0 even Lagrangians in Z9 (+) Z2 x Z2), and both relaxation controls fire (dropping evenness opens d != 4; dropping unimodularity opens d != 4). COROLLARIES AT THE FORCED d (verified arithmetic): g_car = d+1 = 5, N_fam = d-1 = 3, rank = 2d = 8, glue index sqrt(4d) = 4 = d = |mu4| -- the boundary object with four marked points reads P2 and the family count as the two functor readouts of ONE object. HONESTY GATE (binding): c_3 = 1/(2 pi d)|_{d=4} = 1/(8 pi) is a corpus-legitimate REWRITING (c_3 = 1/(2 e_1(a) pi) with e_1(a) = 4 = |mu4|, tfpt_1 anchor-first refinement / v23), NOT a corpus derivation of P1 -- ONE-OBJECT-FORCED was predeclared unreachable and P1 stays a primitive postulate. THE CLOCK KILL (part 2): the hypothesized fiber-product action L_60 = Z12 x_{Z6} Z30 of the seam clock (order 12, c = J o sigma, v629/v634) and the Coxeter clock (order 30, Kostant census {30: 8}, w^15 = -1) on the 60 Gaussian lines is DEAD at three levels -- the Coxeter clock does not preserve the line partition; the FULL line-compatible group (the 11520 line-reflection image extended by conjugation) has order 23040 with maximum element order 12, so line orders 15/30/60 are absent for EVERY realization of the clocks; and the fiber product itself is cyclic Z60 (exact SNF arithmetic; the full product Z12 x Z30 has invariant factors (6,60)) which cannot embed. All three census controls fire (the CC2 sampler-widening correction disclosed and carried). ROOTCLASS-MIXED (v775) binding in both parts: no line/code -> matter semantics. No marker move, NO RH claim. NOTE (2026-08-06, v807 -- marker UNCHANGED): the BOUNDARY.NULLSELECTOR addendum -- a second, independent d = 4 selector on the same invariant: the functor vector v_d = (d+1, -(d-1), d) (honest RR/Serre arithmetic) has |v_d|^2_L = det M(v_d) = d(4-d), null/rank-one iff d = 4, with M_4 = 2(1,2)^T(1,2) the certified spinor square and g^2 - N^2 = 4d = |H|^2 identically -- null <=> glue index = d; the glue census forces {4} via evenness/unimodularity, the null selector via 4d = d^2: independent inputs, same output; the missing premise (why the boundary reading carries the functor vector) is typed OPEN in v807. Scope note (Prime Front separation): v807 supplies only its original S1 RR/Serre/null-spinor algebra here; no prime-window hypothesis or additional physical selector follows.

Original classification detail

ONE-OBJECT-PARTIAL + CLOCK-FIBER-DEAD: even unimodular diagonal glue of A_{d-1}+D_{d+1} iff d = 4 (census 2..12, v92 reproduced, E8 certificate); g_car = 5 / N_fam = 3 / |mu4| = 4 as one-object corollaries; c3 = 1/(2 pi d)|_{d=4} typed corpus-legitimate rewriting, NOT a P1 derivation; the Z12 x_{Z6} Z30 clock action on the 60 lines dead at three levels (group order 23040, max element order 12); controls fire; no RH statement.

Dependencies as recorded

v1; v23; v53; v92; v623; v629; v633; v634; ARF.ROOTCLASS.01