# TWIST-J Public Canon v52 **Release identity.** Public Canon v52. Normative authority and activation state are declared exclusively by [STATUS.md](../STATUS.md). An identical tree on any other ref is a release candidate, not a second authority. **What TWIST-J is.** TWIST-J tests one risky hypothesis: physical reality is a closed, exact, deterministic integer system; continuum, geometry, probability, and fields are readings of it. Its single algebraic axiom is J. Public Canon v52 also declares the discrete architecture used to read that axiom. Those architectural definitions are inventoried below and are not claimed to be uniquely derived from J. No fitted dimensionless parameter is introduced in the stated forms; the single SI calibration anchor is the electron mass m_e. **Axiom (A0).** Reality is the closed integer J-Cayley plenum. Its algebraic generator is J = 1 + zeta_5^2. The public model has no external boundary and no external clock: after the architecture below is declared, one state determines its successor by one map U. J is the seed of the two algebraic projections. Public Canon v52 does not claim that the checkpoint space, the five kernel generators, the selector, or the decoder interface are uniquely forced by J or M_J. CZ: A0. Skutecnost je uzavrene celociselne J-plenum. Jeho algebraickym generatorem je J = 1 + zeta_5^2; diskretni architektura je v aktualnim verejnem Canonu deklarovana zvlast a neni vydavana za jednoznacny dusledek J. System nema vnejsi hranici ani vnejsi hodiny. **Reading.** Time is a counter. Space is a commutator. J is the verb; phi and pi are projections of J, not primitives. The assignment of the modulus to gravity and scale and of the argument to electromagnetism and phase is the public dictionary AXIOM-PROJECTION-DICTIONARY [D], not part of the algebraic theorem. Plenum, not vacuum. **Conventions.** No em dashes. No decimals unless justified: integers and ratios are primary; decimals appear only as computed or measured witnesses, engineering readouts, or measured comparisons, never in a conclusion, and are labeled as such. If it cannot be calculated in integers, it is not physics. Falsification is first class progress. Assertions use exact arithmetic; preregistration precedes computation; a computation only theorem requires byte identical output on two architectures; the action layer protocol names six layers (L1 state, L2 manifold, L3 boundary, L4 support, L5 stream, L6 measure) and any lift between layers requires its own named gate. For GATE-L1-L5-TM-SYM2-SELECTOR-STREAM only, gate kind FIRED_NEGATIVE records that the preregistered scientific negative route has fired. It is terminal and grants no execution, reopening, definition swap, or threshold repair. **Statuses.** Rigid order T-LOCK > T > D > C > H > O > F. T-LOCK immutable theorem; T theorem; D derived; C computed at finite range; H hypothesis with an explicit falsifier; O open obligation; F falsified. Every claim carries a label; no summary is stronger than the label it summarizes. The machine readable registry is canon/REGISTRY.tsv; its claim identifiers are status neutral and stable. The registry is the authoritative public ledger: a bracketed status in the narrative without a registered claim identifier is not thereby a registered public claim, and this sentence does not legalize it; every such label is reconciliation debt, registered with evidence, rewritten, or removed before the synthesis pull request opens. `canon/FRONTIER_PROGRAMS.tsv` is a validated scheduler snapshot for the live frontier. Its program, queue, work-state, and work-mode labels create no claim, status, scope, dependency, layer, gate, evidence, or permission to run a verifier. Those authorities remain in their dedicated ledgers. **Notation.** j = zeta_5, j^5 = 1. J = 1 + j^2. phi = (1 + sqrt5)/2 = 1/|J|. pi = -5 i Li_1(J), with Li_1(J) = i pi/5 exactly. tau = sqrt(J) in F_25; m_8 = zeta_8 = sqrt(i) at prime 2. The Dirac mass variable is m_D. The clock counter is n, its Thue-Morse bit is theta_n, and the sixth finite-kernel coordinate is r. T_pl denotes the plenum point; M_J denotes multiplication by J. The finite phase character is z_6 = Tr_6 and the piston character is Tr_4. The ramified chord is s_J and the kernel transport offset is s_c. p = 5 and d = 3 throughout. The tensor-to-scalar ratio is r_T(k), distinct from the finite-kernel coordinate r and the radial coordinate used in continuum formulas. K_chi5 = 1/(864 pi) is the conformal mode prefactor. script-Q is the exact bridge scalar of BRIDGE-DEFECT [T], fixed by script-Q phi^2 = 2 pi; it is derived notation, not a new primitive. **Primitive and architecture inventory.** ``` algebraic axiom J = 1 + zeta_5^2 derived algebra j, phi, pi, M_J, L, and T_pl declared architecture Omega, U, F_5^6, a,b,c,d,e, and the selector decoder interface D_matter, D_geom, D_clock (typed partial interface) calibration anchor m_e only ``` This is a definition boundary, not an omitted reduction theorem. Every downstream statement is conditional on the declared architecture. Restoring a stronger compression slogan requires a public theorem deriving the architecture from J; Public Canon v52 contains no such theorem. --- ## 1. The axiom and the two projections J = 1 + zeta_5^2 is a cyclotomic unit: N(J) = 1 and Tr(J) = 3 in Q(zeta_5), with Galois conjugate moduli (phi^-1, phi, phi, phi^-1) (J-UNIT [T], reproduce/kernel). Its two archimedean projections are: ``` |J| = 1/phi the modulus projection arg(J) = 2 pi / 5 the argument projection ``` Derivation (J-PROJECTIONS [T]): with j = e^(2 pi i/5), J = 1 + j^2 = 2 cos(2 pi/5) e^(2 pi i/5); 2 cos(2 pi/5) = phi - 1 = 1/phi, so |J| = 1/phi and arg J = 2 pi/5. The algebraic half is exact in the kernel witness as J Jbar = 2 - phi = phi^-2 (J-MODULUS-CHORD [T], reproduce/kernel). AXIOM-PROJECTION-DICTIONARY [D] reads the modulus projection as gravity and scale and the argument projection as electromagnetism and phase. It also reads the CRT factors of T_pl as the prime 5 write and prime 2 read and the unit and ramified chords as the gravity and space channels. These assignments rest on the exact rows above and below; they are neither algebraic consequences nor uniqueness claims. pi enters from the argument side (PI-FROM-J [T]): 1 - J = -j^2 is a primitive tenth root of unity (J-TENTH-ROOT [T], reproduce/kernel), 1 - J = e^(-i pi/5), so Li_1(J) = -log(1 - J) = i pi/5 exactly and pi = -5 i Li_1(J). A third transcendental enters from the modulus side: ln phi, the boost rapidity. The two logarithmic axes pi and ln phi are linearly independent over the algebraic numbers (LOG-AXES-INDEPENDENCE [T]): i pi = Log(-1) and ln phi = Log(phi) are logarithms of algebraic numbers, Q-linearly independent, hence independent over the algebraic numbers by Baker. No algebraic independence claim is made. The golden bridge (J-GOLDEN-BRIDGE [T], reproduce/kernel): J phi = j, (J - 1)^3 = j, and J^5 phi^5 = 1. ### The harmonic seam Put `K = Q(j)`, `O_K = Z[j]`, `psi = 1-phi = -phi^-1`, and let `F_0=0`, `F_1=1`, `F_(n+1)=F_n+F_(n-1)`. For `n>=1` define ```text u_n = F_n phi - F_(n+1), H(x) = sum_(n>=1) u_n x^n/n, x in mu_10. ``` Since `|psi|=phi^-1<1`, the series converges absolutely on `mu_10` and, on the principal archimedean branch, ```text u_n = -psi^n = -F_(n+1)-F_n j^2-F_n j^3, H(x) = Log(1-psi x). ``` The first identity holds for every `n`, not just a finite range. Its induction starts at `u_1=-psi`; the identities `psi phi=-1` and `psi(-1)=phi-1` carry the coefficient pair `(F_n,F_(n+1))` to `(F_(n+1),F_(n+2))`. The complete torsion-axis classification is ```text H(x) is real iff x in {1,-1}, Re H(x) = 0 iff x in {-j,-j^-1}. ``` Indeed, `1-psi x` is real exactly when `x=conj(x)`, and the only real elements of `mu_10` are `1,-1`. Their positive landings are `1-psi=phi` and `1+psi=phi^-2`. For the imaginary axis, ```text |1-psi x|^2 = 1+psi^2-psi(x+x^-1), ``` so unit modulus is equivalent to `x+x^-1=psi`, whose two roots in the complete ten-element class are `-j,-j^-1`. The four landing values are ```text 1-psi = phi, 1+psi = phi^-2, 1-psi(-j) = -j^2, 1-psi(-j^-1)= -j^3. ``` The last two have exact order ten. The full arithmetic unit group is ```text O_K^x = mu_10 x . ``` For a unit `e`, the quotient `e/conj(e)` has every conjugate on the unit circle and is therefore a root of unity. Reduction modulo `1-j` places it in `mu_5`; after division by its square root in `mu_5`, the remaining unit is real. The elementary Pell-unit classification `Z[phi]^x={+/- phi^m:m in Z}` then gives the displayed product, with trivial intersection between `mu_10` and ``. Consequently the distinguished values and reconstruction are ```text H(1) = log phi, H(-j) = -i pi/5, Log J = -H(1)-2H(-j) = -log phi+2 pi i/5. ``` These statements form J-HARMONIC-SEAM [T] at L1, evidenced by `probes/P-J-HARMONIC-SEAM-1`. They classify one exact contracting integer ladder and its four torsion-axis landings. They do not promote AXIOM-PROJECTION-DICTIONARY [D] or TWO-PLACE-PHYSICS [D], identify the order-two sign in `mu_10` with the separate `zeta_8` read place, strengthen BOOST-COUNT-LADDER [D], or add a decoder, measure, observer, force, spacetime, SI bridge, new constant, or lift to L2-L6. Multiplication by J on the integer lattice Z^4 in the basis {1, j, j^2, j^3} is the canonical step (J-STEP [T], reproduce/kernel): ``` (a, b, c, d) -> (a - c + d, b - c, a, b - c + d), det = N(J) = 1, trace = 3 ``` C20-TEICHMULLER-SPLIT [T], evidenced by probes/P-C20-TEICHMULLER-SPLIT-2, is an L1 arithmetic theorem. Put O = Z[zeta_5], lambda = 1 - zeta_5, A_m = O/(lambda^m), and R = A_4 = O/(5). The last equality uses (5) = (lambda)^4 as ideals in O; it does not assert 5 = lambda^4 as elements. Each A_m is local with residue field F_5, its units are exactly the elements of nonzero residue, |A_m| = 5^m, and |A_m^*| = 4 x 5^(m-1). In R, lambda^4 = 0 != lambda^3. The characteristic-five identity gives J^5 = 2, hence J^10 = -1, J^15 = 3, and J has exact order 20. With t = J^5 = 2 and u = J^16 = 3J, the factors have orders 4 and 5, J = tu, and = x isomorphic to C_4 x C_5. Moreover u - 1 = lambda(-1 + 3 lambda), whose second factor is a unit, so (u - 1)^4 = 0 != (u - 1)^3. At depth four, the kernel of R^* -> F_5^* is a 5-group. Therefore the four nonzero scalar constants are exactly both root sets: mu_8(R) = mu_4(R) = F_5^*. For every m >= 1 the kernel of A_m^* -> F_5^* is again a 5-group, so the Sylow 2-subgroup of A_m^* is C_4 and no A_m contains an element of order 8. The literal scalar-root description is asserted only for R. For the operator leg of C20-TEICHMULLER-SPLIT [T], reducing the four public J-STEP columns modulo 5 gives M_R, which agrees with multiplication by J. It has exact order 20 and satisfies M_R^5 = 2I, M_R^10 = -I, and M_R^16 = 3M_R. Since M_R - 2I is multiplication by -lambda(j + 1), (M_R - 2I)^4 = 0 != (M_R - 2I)^3. For U = M_R^16, U - I = 3(M_R - 2I), so (U - I)^4 = 0 != (U - I)^3. The scope of C20-TEICHMULLER-SPLIT [T] is exact L1 arithmetic only. The all-m unit-group result is not an all-k order theorem for M_J modulo 5^k and does not strengthen TIME-QUANTUM-TOWER [C]. It supplies no time, tick, clock, decoder, physical carrier, metrology scale, unique interpretation, or L2-L6 lift. The plenum point (PLENUM-POINT [T]): T_pl = s_J + i phi = 2i(1 - J) with s_J^2 = 1 + J Jbar = 3 - phi = sqrt5/phi. Derivation: J Jbar = 2 - phi gives s_J^2 = 3 - phi, and (3 - phi) phi = 2 phi - 1 = sqrt5. |T_pl| = 2; arg T_pl = 3 pi/10; zeta_5 T_pl^2 + 4 = 0, since T_pl^2 = -4 j^4; T_pl^10 = -2^10; T_pl/2 = zeta_20^3, and the CRT split is T_pl/2 = zeta_4^-1 zeta_5^2. The ramified chord s_J = |1 - zeta_5| obeys N(1 - zeta_5) = 5 (J-RAMIFIED-CHORD [T], reproduce/kernel), while the modulus chord is a unit chord. Their physical channel assignment belongs only to AXIOM-PROJECTION-DICTIONARY [D]. At the magic prime the axiom has a square root: sqrt(J) = tau in F_25 (section 11). ## 2. Time, space, and the decoder The autonomous state is not the finite checkpoint alone. It is ``` Omega = N_0 x F_5^6, omega = (n, psi), theta_n = s_2(n) mod 2, z_6(psi) = sum_k psi_k mod 5, sigma(omega) = z_6(psi) + 2 theta_n mod 5, U(n, psi) = (n + 1, g_{sigma(omega)}(psi)), (g_0, g_1, g_2, g_3, g_4) = (a, b, c, d, e). ``` Here s_2(n) is the finite binary digit sum. The clock coordinate N_0 is the distinguished forward orbit 0, 1, 2, ... of the 2-adic odometer, embedded in Z_2. No Thue-Morse parity function on all of Z_2 is asserted. The finite checkpoint projection is pr_checkpoint(n, psi) = psi; psi by itself is not claimed to be an autonomous state. The driver word and every registered event log are derived orbit records, not additional state variables. For a binary observable lambda supplied by a registered decoder claim, its Log is (lambda(U^k omega_0))_{k >= 0}. The counter closure (ODOMETER-INTERNALIZED [D], reproduce/foundations-places) is exact on this forward orbit: theta_{2n} = theta_n, theta_{2n+1} = 1 - theta_n, and the local carry law updates theta without consulting an external list. Therefore U has no external step parameter. This is an autonomous skew product by definition; no extension of theta to every 2-adic integer is used. On the sheet z5 in {1,4}, use the local probe aliases z5 = z_6(psi) and t = theta_n in {0,1}; t is the binary drive bit, not the sixth checkpoint coordinate r. The selector ``` i = (z5 + 2t) mod 5 ``` takes only the values {1,3,4}, so only b, d, and e ever fire. Their fired pair commutators are the pure fiber translations ``` [d,e] = T_(0,0,0,0,3,0), [b,d] = T_(0,0,0,0,3,3), [b,e] = T_(0,0,0,0,1,3), ``` all with zero piston component. The linear parts of b, d, and e lie in the abelian Klein group {I, -I, B, -B} of exponent two. Hence every group commutator in is a pure translation, and the derived subgroup D() is exactly the 25 fiber translations. This is FIRED-COMMUTATOR-NOGO [T] at L1, evidenced by probes/P-FIRED-COMMUTATOR-NOGO-1: the fired dynamics is spatially abelian and piston-block noncommutativity cannot arise from the fired steps. The silent control [a,c] is not a translation. No curvature operator selection, value, or physical reading is claimed. The phrase "space is a commutator" remains a dictionary reading, not a unique curvature construction. One historical construction is now typed exactly. Let X = F_5^6, F = Q^X, H = , let R_H be Reynolds averaging over H, and let P_0 remove the constant function. With ``` V = F^H intersect 1_X^perp, dim V = 818, C_0 = T_a T_c - T_c T_a, K_hist = (P_0 R_H C_0 R_H P_0)|_V, ``` two complete exact routes give ``` Tr_V(K_hist^2) = -881/8. ``` This is CURVATURE-HISTORICAL-TRACE [T], evidenced by probes/P-CURVATURE-TRACE-VALUE-1. The registered proposal -21/8 is therefore false for this operator (CURVATURE-TRACE-VALUE [F]); the separately printed ten-mode historical checksum is -22 and is not asserted as the spectrum of K_hist. The historical compression also has an exact intrinsic/exterior split. Put `P = P_0 R_H`, `Q = I - P`, `A = T_a`, and `C = T_c`, and restrict every displayed endomorphism to `V`. Then ``` K_amb = P(AC - CA)P = K_hist, K_int = [PAP,PCP], K_ext = PAQCP - PCQAP, K_amb = K_int + K_ext. ``` Two complete exact routes and byte-identical aarch64 and x86_64 executions give `K_ext = 0` entrywise and hence `K_amb = K_int`. The ambient and intrinsic operators both have rank 292 and nullity 526; the exterior operator has rank 0 and nullity 818. Their exact trace split is ``` Tr_V(K_int^2) = -881/8, Tr_V(K_ext^2) = 0, 2 Tr_V(K_int K_ext) = 0. ``` The leakage maps `QAP|_V` and `QCP|_V` are both nonzero, so the vanishing exterior term is exact cancellation, not invariance of `V` under `A` and `C`. This is CURVATURE-HISTORICAL-GAUSS-SPLIT [T], evidenced by probes/P-CURVATURE-GAUSS-SPLIT-1. It is an algebraic identity for the frozen historical operator, not a differential-geometric Gauss equation, embedding theorem, spectrum, or physical interpretation. None of these facts selects a canonical spatial-curvature operator. CURVATURE-OPERATOR-CANONICAL [O] asks whether the public architecture determines exactly one equivalence class after the carrier, measure, projection group, and ambient versus intrinsic commutator choice are fixed publicly. No golden spectrum or continuum-curvature reading is asserted. The decoder is a typed partial interface, not a completed total map. Let K be the set of forward U-orbits. Let MatterData, GeometryData, and ObservableHistory denote records whose fields exist only where a registered claim defines them. Then ``` D_matter : dom(D_matter) subset K -> MatterData D_geom : dom(D_geom) subset K x MatterData -> GeometryData D_clock : dom(D_clock) subset K x MatterData x GeometryData -> ObservableHistory D(kappa) = D_clock(kappa, m, g), m = D_matter(kappa), g = D_geom(kappa, m). ``` Functional order is matter, then geometry, then clock. D_matter reads only the orbit and the registered quadratic/Born and matter maps; D_geom reads the orbit plus MatterData and the registered linear, boundary, wedge, and chain maps; D_clock reads the counter projection plus the accumulated records and is terminal. None of these outputs feeds U, so the declared dependency graph is acyclic. Totality, uniqueness, and completeness of D are not claimed. The typed quadratic/Born `D_matter` action, including its exact factorization through the declared quadratic pair, remains in QUADRATIC-DECODER-DATA [O]. No umbrella full-decoder completeness claim is registered. The preceding read-only clause defines only the declared partial interface. No registered nonempty complete admissible class of future nontrivial observer-output-to-`U` writeback extensions exists, so the clause is not a theorem about such extensions. Any such extension requires its own newly typed architecture, output schema, write-channel type, autonomous-state codomain, protocol class, dependency graph, and separately registered and preregistered claim. Registering that new architecture would not retroactively falsify the declared partial read-only architecture. For audit purposes only, `DEF-DECODER-COMPLETION-CONTRACT` defines the schema of a submitted decoder-completion candidate. It does not assert that such a candidate exists or that any submitted candidate satisfies the contract. The functional stages ``` D_matter -> D_geom -> D_clock ``` and the reading-leg labels ``` D_linear, D_binary, D_quadratic ``` are independent axes. A stage assignment does not identify, merge, or transfer a property between reading legs. A submitted candidate is machine-auditable only when it publishes finite typed manifests with these slots: ``` candidate_id public_pin_id read_convention_id history_equivalence_id region_id coarse_graining_id carrier_manifest[]: carrier_id parent_carrier_id inclusion_or_quotient_map_id equality_id coefficient_object_id record_field_manifest[]: record_id field_id field_type_id role carrier_id domain_id normalization_id equality_id source_item_id write_map_id presence_state absence_basis_item_id emit_rule_id stage_id leg_id stage_manifest[]: stage_id domain_id codomain_id map_id totality_domain_id dependency_item_ids leg_manifest[]: leg_id owned_field_ids domain_id codomain_id map_id bridge_manifest[]: bridge_id source_id target_id domain_id codomain_id map_id dependency_item_ids from_layer to_layer gate_ids quadratic_manifest: coefficient_ring_id effective_carrier_id orbit_to_amplitude_bridge_id gram_id dagger_id transpose_id qcarrier_id q_equality_id q_map_id effect_ids born_pairing_id factorization_map_id physics_manifest: source_id current_id conservation_id propagator_id detector_id measure_manifest: measure_id normalization_id metrology_id scheme_id closure_manifest: write_target_ids feeds_U terminal_output_ids terminality_basis_id obligation_manifest[]: requirement_id owning_item_id value_state basis_item_ids ``` The allowed `stage_id` values are exactly `D_matter`, `D_geom`, and `D_clock`. The allowed `role` values are `READOUT` and `AUXILIARY`. The allowed resolved `leg_id` values are exactly `D_linear`, `D_binary`, and `D_quadratic`. Every output field has exactly one record owner and one stage owner. A `READOUT` field resolves exactly one reading leg; an `AUXILIARY` field uses `NOT_APPLICABLE` with a resolvable basis item. A relation between two legs requires an explicit bridge row and is never inherited from a common stage. Every required identifier-valued slot uses a resolvable public identifier or the literal `UNRESOLVED`. `presence_state` and `value_state` are `RESOLVED`, `UNRESOLVED`, or `NOT_APPLICABLE`; `NOT_APPLICABLE` requires a resolvable basis item. `feeds_U` is `TRUE`, `FALSE`, or `UNRESOLVED`. There is no bare null. Field identifiers are unique. Domains, codomains, equalities, dependencies, layer endpoints, and write targets are explicit. Totality is stated only relative to a named domain. A declared cross-layer map records its public gate identifier; absence of that gate leaves the requirement unresolved. A terminal record `emit_rule_id` is distinct from a write target and does not by itself establish that the output cannot feed `U`. Only fields owned by `D_quadratic` under QUADRATIC-DECODER-DATA are tested for factorization through `Q`. No factorization, status, evidence, or closure is inherited by `D_linear` or `D_binary`, or transferred between `D_matter`, `D_geom`, and `D_clock`. This contract is a schema, not an existence, totality, uniqueness, canonicity, or completeness statement. Syntactic conformance, a resolved identifier, or a submitted candidate is not evidence and cannot change a public status. In particular, the definition does not establish a physical carrier, factorization through `Q`, post-state instrument, canonical geometry, source-current-conservation-propagator-detector chain, physical measure, normalization, metrology, scheme, write port, or completion-wide terminality. Each property remains owned by its registered claim and evidence, and every cross-layer lift remains owned by its public gate. Missing or `UNRESOLVED` fields leave all owning claims unchanged. The contract neither opens nor authorizes a verifier or probe. ### Factor-canonicity audit overlay Within the same definition, a candidate may optionally submit a `factor_canonicity_manifest` for one fixed `stage_id`, one fixed `leg_id`, and one declared scope. This is an overlay on, not a replacement for, the manifests above. Before any classification it freezes the typed datum ```text C_factor = ( source_carriers, output_carriers, declared_domains, source_equalities, output_equalities, candidate_class, candidate_equivalence, factor_maps, gauge_comparisons, scale_comparisons, apparatus_reductions, optional_categorical_coherence, optional_U_congruence, nonconstancy_test, completeness_statement, completeness_method ). ``` A candidate at scale `ell` has the typed form ```text D_ell : Dom_ell -> Y_ell, Dom_ell subset X_ell. ``` Totality is asserted only on the named `Dom_ell`. A factor candidate has the typed shape ```text q : Dom -> QCarrier, im(q) subset QCarrier, q_bar : Dom -> im(q), q_bar(x) = q(x), i_q : im(q) -> QCarrier, the image inclusion, q = i_q o q_bar, F : im(q) -> Y, D : Dom -> Y, D = F o q_bar on Dom. ``` Here `q_bar` is the corestriction of `q` and is surjective by definition. With the equality on `QCarrier` and the equality on `Y` frozen as separate inputs, fiber constancy is exactly the equivalence ```text for all x,x' in Dom, q(x) = q(x') implies D(x) = D(x') if and only if there exists a unique F : im(q) -> Y with D = F o q_bar. ``` The reverse direction uses the factor equation. The forward direction defines `F(q_bar(x)) := D(x)`; fiber constancy gives well-definedness and surjectivity of `q_bar` gives uniqueness. Nonconstancy is a separate requirement and needs exact witnesses `x_0,x_1 in Dom` with `D(x_0) != D(x_1)`, or an equivalent proof that `|im(D)| >= 2`. Syntactic dependence on an input field is not such a certificate. Every admitted comparison is first a typed per-map square ```text r : Dom_s -> Dom_t, D_s : Dom_s -> Y_s, D_t : Dom_t -> Y_t, T_r : Y_s -> Y_t, D_t o r = T_r o D_s. ``` Gauge comparison, scale covariance, apparatus reduction coherence, and optional `U`-congruence are four different obligations. None is inherited from another. - A gauge comparison freezes `g : Dom_s -> Dom_t` and `Gamma_g : Y_s -> Y_t` with `D_t o g = Gamma_g o D_s`. - A scale comparison freezes `R_(b,ell) : Dom_ell -> Dom_(b ell)` and `Z_(b,ell) : Y_ell -> Y_(b ell)` with `D_(b ell) o R_(b,ell) = Z_(b,ell) o D_ell`. - An apparatus reduction freezes `r_X : Dom_P -> Dom_(P')` and `tau_r : Y_P -> Y_(P')` with `D_(P') o r_X = tau_r o D_P`. - Optional `U`-congruence uses a named domain `Dom_U` and freezes `U_U := U|Dom_U : Dom_U -> Dom_U`, `D_U := D|Dom_U : Dom_U -> Y`, `Y_U := D(Dom_U) subset Y`, `D_bar_U : Dom_U -> Y_U`, and the image inclusion `i_U : Y_U -> Y`, with `D_U = i_U o D_bar_U`. The typing of `U_U` includes `U(Dom_U) subset Dom_U`. The exact condition `D_U(x) = D_U(x')` implies `D_U(U_U x) = D_U(U_U x')` for every `x,x' in Dom_U`. Only then is `U_bar(D_bar_U(x)) := D_bar_U(U_U x)` well defined, and only then does `D_bar_U o U_U = U_bar o D_bar_U` hold on `Dom_U`. The unqualified expression `U_bar o D` is not used. This clause supplies no attractor, feedback, purpose, write channel, or terminality result. Strict invariance is the special case only when the source and target domains, decoder components, output carriers, and output transport are all explicitly identified and the output transport is the identity. A forward substitution, an information-losing block map, and an inverse desubstitution are different typed operations. Equality of effects does not imply equality of instruments or post-event operations. Two candidates are equivalent only through an independently frozen typed output isomorphism ```text h : Y_1 -> Y_2 ``` preserving every owned field, equality, normalization, effect, outcome label, orientation datum, and declared transport, and obeying ```text h o D_1 = D_2 ``` on the common declared domain. A bijection invented after seeing the output does not define an admissible equivalence. Before equivalence classes are counted, reflexivity, symmetry, transitivity, and closure of the admitted isomorphisms must be proved on the complete candidate class. For one frozen scope with one common source domain, let `Inv_adm(C_factor)` be an independently defined complete class of admissible readouts that are strictly invariant under the independently frozen invisible-transformation class. Its readouts are typed as ```text D : Dom -> Y_D, D_bar : Dom -> im(D), D_bar(x) = D(x), i_D : im(D) -> Y_D, D = i_D o D_bar, A : Dom -> Y_A. ``` A candidate `D` may be called a maximal invariant only if all of the following are proved: 1. `D` belongs to `Inv_adm(C_factor)` and is strictly invariant under every declared invisible transformation at that scope; 2. every `A` in `Inv_adm(C_factor)` has the same declared source `Dom` as `D`, and there exists a unique typed mediator `A_bar : im(D) -> Y_A` with `A = A_bar o D_bar`; 3. the invisible-transformation class, invariant class, candidate class, and candidate equivalence are each complete at the declared scope; and 4. all dependencies are closed and every input is declared and publicly identified. A non-strict gauge comparison, scale-covariant square, or apparatus-reduction square may be an additional coherence obligation, but it does not discharge strict invariance under an invisible transformation. A universal factorization theorem for a merely covariant family is a maximal-covariant readout theorem, not a maximal-invariant theorem. For an indexed family, every component freezes ```text D_i : Dom_i -> Y_i, D_bar_i : Dom_i -> im(D_i), A_i : Dom_i -> Z_i, A_bar_i : im(D_i) -> Z_i, A_i = A_bar_i o D_bar_i. ``` For every frozen source map `r : Dom_i -> Dom_j`, it also freezes decoder and readout transports with ```text D_j o r = T_D(r) o D_i, A_j o r = T_A(r) o A_i, T_D_bar(r) : im(D_i) -> im(D_j). ``` The component mediators are compatible only when ```text T_A(r) o A_bar_i = A_bar_j o T_D_bar(r) ``` for every admitted `r`. Componentwise factorization without these frozen compatibility equations is not indexed maximality. `Nontrivial maximal invariant` additionally requires the separate exact nonconstancy certificate above. The optional machine-auditable overlay has the slots ```text factor_canonicity_manifest: owner_item_id stage_id leg_id source_carrier_id output_carrier_id domain_id totality_domain_id source_equality_id output_equality_id qcarrier_id q_equality_id q_map_id q_image_id q_corestriction_id q_image_inclusion_id factor_map_id factor_equation_id fiber_constancy_test_id fiber_factor_equivalence_statement_id fiber_factor_equivalence_proof_id nonconstancy_test_id candidate_class_id candidate_membership_test_id candidate_equivalence_id candidate_equivalence_reflexivity_proof_id candidate_equivalence_symmetry_proof_id candidate_equivalence_transitivity_proof_id candidate_isomorphism_closure_proof_id candidate_completeness_statement_id candidate_completeness_proof_id hidden_input_closure_id gauge_square_manifest[]: comparison_id source_domain_id target_domain_id source_decoder_component_id target_decoder_component_id source_output_carrier_id target_output_carrier_id source_map_id output_transport_id membership_test_id commuting_square_id completeness_statement_id completeness_proof_id pre_output_freeze_basis_id strict_invariance_manifest: state = NOT_CLAIMED | DEFINED | UNRESOLVED DEFINED: common_domain_identification_id decoder_component_identification_id output_carrier_identification_id identity_output_transport_id strict_invariance_proof_id NOT_CLAIMED: nonstrict_square_basis_id scale_manifest[]: source_scale_id target_scale_id source_domain_id target_domain_id source_decoder_component_id target_decoder_component_id source_output_carrier_id target_output_carrier_id source_map_id output_transport_id commuting_square_id blocking_origin_id phase_policy_id normalization_id pre_post_convention_id completeness_statement_id completeness_proof_id pre_output_freeze_basis_id strict_invariance_manifest: state = NOT_CLAIMED | DEFINED | UNRESOLVED DEFINED: common_domain_identification_id decoder_component_identification_id output_carrier_identification_id identity_output_transport_id strict_invariance_proof_id NOT_CLAIMED: nonstrict_square_basis_id apparatus_reduction_manifest[]: source_protocol_id target_protocol_id source_domain_id target_domain_id source_decoder_component_id target_decoder_component_id source_output_carrier_id target_output_carrier_id reduction_precondition_id source_transport_id output_transport_id commuting_square_id reduction_equality_id membership_test_id completeness_statement_id completeness_proof_id pre_output_freeze_basis_id U_congruence_manifest: state = DEFINED | NONE | UNRESOLVED DEFINED: U_id U_stable_domain_id U_restriction_id decoder_restriction_id decoder_image_id decoder_corestriction_id decoder_image_inclusion_id decoder_kernel_equality_id congruence_test_id induced_map_id induced_map_equality_id corestricted_factor_dynamics_equation_id NONE: no_U_congruence_claim_basis_id categorical_coherence_manifest: state = NONE | DEFINED | UNRESOLVED DEFINED: object_class_id arrow_class_id identity_arrow_id composition_id arrow_closure_proof_id composition_associativity_proof_id left_unit_law_proof_id right_unit_law_proof_id source_identity_law_id source_composition_law_id output_identity_law_id output_composition_law_id decoder_family_naturality_id group_action_manifest: state = NOT_CLAIMED | DEFINED | UNRESOLVED DEFINED: inverse_assignment_id inverse_closure_proof_id left_inverse_law_proof_id right_inverse_law_proof_id NOT_CLAIMED: no_group_action_claim_basis_id NONE: no_categorical_coherence_claim_basis_id maximality_manifest: state = NOT_CLAIMED | DEFINED | UNRESOLVED DEFINED: common_domain_id decoder_corestriction_id invisible_transformation_class_id strict_invariance_statement_id strict_invariance_proof_id admissible_invariant_class_id invariant_membership_test_id invariant_completeness_statement_id invariant_completeness_proof_id universal_factorization_statement_id universal_factorization_proof_id indexed_family_manifest: state = NONE | DEFINED | UNRESOLVED DEFINED: component_domain_manifest_id component_mediator_manifest_id component_mediator_compatibility_id NONE: single_scope_basis_id nontriviality_manifest: state = NOT_CLAIMED | DEFINED | UNRESOLVED DEFINED: nonconstancy_certificate_id ``` Every identifier-valued field is a resolvable public identifier or `UNRESOLVED`; there is no bare null. `DEFINED` records a typed submitted property and does not mean that the property is proved. A scientific owner and its evidence must decide the corresponding property. `NONE` is the normal categorical-coherence state for the base per-map schema. Categorical `DEFINED` requires every identity, composition, closure, associativity, unit, source, output, and decoder-family field above. A group claim additionally requires every inverse field. A componentwise indexed claim requires its mediator compatibility field; otherwise its state is `UNRESOLVED`. `NOT_CLAIMED` is the normal strict-invariance state for a legitimate non-strict commuting square. It supplies the named non-strict basis and does not turn that square into an invisible transformation. Maximality and nontriviality have independent states. `maximality = NOT_CLAIMED` is required when the submission proves only factorization through a named `q` and no universal invariant property. `nontriviality = NOT_CLAIMED` is required when no exact nonconstancy claim is made. A defined maximality claim may therefore coexist with `nontriviality = NOT_CLAIMED`. The inheritance ```text D_quadratic factorization -/-> D_binary or D_linear factorization D_binary coherence -/-> D_quadratic instrument selection one D_matter field -/-> every MatterData field D_matter closure -/-> D_geom or D_clock closure L5 reduction equivalence -/-> L6 normalization effect equality -/-> instrument equality common numeric value -/-> typed bridge comparison action -/-> adopted gauge ``` is forbidden. A relation between legs still requires an explicit `bridge_manifest[]` row with its complete typing, dependencies, layer endpoints, and gates. This overlay is a schema, not an existence, totality, uniqueness, canonicity, classification, maximality, nontriviality, scale-stability, or completeness result. It supplies no carrier, factor, quotient, gauge, macrodynamics, physical source or event map, instrument, measure, normalization, writeback, terminality, status, evidence, or gate. It neither repairs a fired predecessor nor authorizes a verifier or probe. The terms `nontrivial maximal invariant`, `universal quotient`, `universality class`, and `canonical factor` remain forbidden outputs while any applicable typing, compatibility, completeness, maximality, or nonconstancy item is unresolved. The reading split (READING-SPLIT [D], inline) is therefore a public dictionary at the registered legs, not a completeness theorem: the linear readout is CODEC-TR4, the binary cut drives the census, and the quadratic registration is the Born square. The L1 Thue-Morse sliding-pair density is rho = 1/6 (GYRON-DENSITY [T], registered in section 3). The self similar time quantum (TIME-QUANTUM-TOWER [C], reproduce/foundations-places): on Z/5^k, M_J^(5^k) = i_5 I with i_5^4 = 1 and period exactly 4 x 5^k, computed for k = 1 to 4. No all-k theorem and no generic exponentiation speedup are claimed. TIME-CUT-READING [D] composes only registered public rows. The counter n is cut by theta_n = s_2(n) mod 2 into the selection law i = (z5 + 2 theta_n) mod 5, and RAMIFIED-TM-LIFT realizes this cut as the sign quotient of the four-phase J-channel. The matter channel reads the named isolated pair (theta_(n-1), theta_n) = (0,0), of density 1/6 by GYRON-DENSITY. To audit the bracket, the Thue-Morse recurrence gives (theta_(2m), theta_(2m+1)) = (theta_m, 1 - theta_m), so 00 can occur only as (theta_(2m-1), theta_(2m)). That pair being 00 forces (theta_(m-1), theta_m) = (1,0), whence theta_(2m-2) = theta_(2m+1) = 1. Thus each knot is bracketed 1 00 1 by the cube-free drive. The spatial channel does not sit on noncommutativity of the fired steps, whose commutator subgroup is exactly the fiber translation plane by FIRED-COMMUTATOR-NOGO. It sits on the silent pair a,c, never fired on the sheet, whose commutator reading is carried by CURVATURE-HISTORICAL-TRACE and KERNEL-MACRO-READING while CURVATURE-OPERATOR-CANONICAL remains open. The dimensionless proper-time reading is delta tau hat = 2 pi/5 per tick by METRO-TICK. This dictionary claims no forcing, uniqueness, or completeness, and strengthens none of its component rows. ### QDD Route A dictionary The quadratic leg of `D_matter` gains its exact algebra as public definitions and theorems on the finite balanced piston carrier. The Route A factorization block is L1 exact algebra. The later instrument nonselection theorem is L4 apparatus/support mathematics. Nothing here fills the decoder completion contract, claims an L6 reading, selects a physical instrument, derives the architecture or the effect pair from J, or changes `QUADRATIC-DECODER-DATA`, which remains an open obligation [O]. The physical instrument realization remains the separate obligation `QDD-INSTRUMENT-APPARATUS`. Definitions. ``` DEF-QDD-DOMAIN-K0 K_QDD = {kappa_x = (U^n(0,x))_(n>=0) : x in F_5^6}, equality of complete pointed forward sequences, distinguished head n = 0; the common total domain of the quadratic D_matter leg. DEF-QDD-BALANCED-PISTON ell(0,1,2,3,4) = (0,1,2,-2,-1); for x = (p1,p4,p1p,p4p,q,r), beta_QDD(kappa_x) = (ell(p1),ell(p4),ell(p1p),ell(p4p))^T in V_eff = ell(F_5)^4 subset Q^4; q, r, every later checkpoint, the counter, environment, randomness and dynamic evaluation are forbidden inputs. DEF-QDD-AMPLITUDE-B0 B0 = (1, zeta, zeta^2, zeta^3), zeta = zeta_5; iota_B0(v) = v_0 + v_1 zeta + v_2 zeta^2 + v_3 zeta^3 in K = Q(zeta); Amp_QDD = iota_B0 o beta_QDD. DEF-QDD-COEFFICIENT-Q coefficient ring Q with the trivial involution on the matrix side; the amplitude field K with bar = sigma_4 (zeta -> zeta^4) and Tr = Tr_(K/Q); inv_Q, bar and the Gram adjoint are three distinct typed operations. DEF-QDD-TRACE-PAIRING _tr = (1/5) Tr(x sigma_4(y)); (1/5) Tr(zeta^(a-b)) = delta_(a,b) - 1/5, so the matrix of <.,.>_tr in B0 is I_4 - (1/5) 1 1^T. DEF-QDD-GRAM G = I_4 - (1/5) 1 1^T, the matrix of the trace pairing in B0, on V_eff; G^-1 = I_4 + 1 1^T, G^-1 1 = 5 1; Gram adjoint A^sharp = G^-1 A^T G; a factor-branch helper: the direct write does not name it. DEF-QDD-DAGGER v^dagger = v^T on Q^4. DEF-QDD-TRANSPOSE transpose(A) = A^T on M_4(Q). DEF-QDD-QPAIR Q_QDD(v) = (A_dagger, A_T) = (v v^dagger, v v^T), an ordered pair of two typed slots. DEF-QDD-QCARRIER-EQUALITY QCarrier_QDD = im(Q_QDD | V_eff) subset M_4(Q) x M_4(Q), ordered componentwise rational matrix equality; equal coordinate values do not collapse the two typed slots. DEF-QDD-LOW-LINE lambda_B = 1 + zeta + zeta^2 + zeta^3 = -zeta^4, Tr(lambda_B) = 1, _tr = 4/5; the LOW LINE is Q lambda_B. It is neither the rational line Q.1 nor the trace kernel of K. DEF-QDD-PROJECTOR-LOW E_low = (1/4) 1 1^T; the first member of the frozen ordered effect pair of the EFFECT_SHADOW_MINIMAL owner freeze; ALGEBRAIC_READOUT, not a physical apparatus selection, not a realized outcome, not a post-state instrument, and not claimed to be forced by J. DEF-QDD-PROJECTOR-HIGH E_high = I_4 - E_low; the second member of the frozen ordered pair; the same labels. DEF-QDD-BRANCH-WEIGHT-PAIRING the factor-route Born trace pairing, on the transpose slot A_T = v v^T of QCarrier_QDD: m(A_T) = Tr(A_T G), w_low(A_T) = Tr(E_low A_T G), w_high(A_T) = Tr(E_high A_T G), density A_T G / m(A_T); the owner-frozen Born trace pairing of the EFFECT_SHADOW_MINIMAL freeze, an adopted dictionary input, not derived from J or from the projector identities; a factor-branch helper: the direct write does not name it. DEF-QDD-MATTER-RECORD MatterData_QDD, a pure type schema of five typed fields: support_state in {ZERO_SUPPORT, SUPPORTED}; total_weight in Q_(>=0); branch_weights, an ordered pair (LOW, HIGH) in Q_(>=0)^2, no swap; density_state, the tag ZERO_DENOMINATOR or the tag DENSITY carrying a 4x4 rational matrix; normalized_weight_state, the tag ZERO_DENOMINATOR or the tag NORMALIZED carrying a rational pair. The ZERO branch is fixed as (ZERO_SUPPORT, 0, (0,0), ZERO_DENOMINATOR, ZERO_DENOMINATOR), no division performed. The schema fixes types, tags, the branch order and the ZERO branch only; it names no computation rule, and each branch supplies its own field sources (the direct branch through R_cyc alone). All five fields are L1 exact data; no L6 measure reading of normalized_weight_state is claimed by this block. DEF-QDD-DIRECT-WRITE R_cyc : K -> MatterData_QDD, written from field arithmetic in K, sigma_4, Tr, the trace pairing and the LOW LINE only: R_cyc(0) is the ZERO branch; for w != 0, m_tr(w) = _tr, pi_low(w) = (_tr / _tr) lambda_B, pi_high(w) = w - pi_low(w), w_low = _tr, w_high = _tr, T_w(x) = w _tr, and R_cyc(w) = (SUPPORTED, m_tr, (w_low, w_high), DENSITY(MATRIX_B0(T_w)/m_tr), NORMALIZED((w_low, w_high)/m_tr)), where MATRIX_B0 expresses the operator in the basis B0 by field arithmetic and the trace pairing alone. D_QDD_direct = R_cyc o iota_B0 o beta_QDD : K_QDD -> MatterData_QDD. The independence firewall of the EFFECT_SHADOW_MINIMAL freeze and of the DICTIONARY-DIRECT amendment section 6 holds by construction: neither the map nor its definitional closure names Q_QDD, the Gram matrix, the dagger or transpose slots, the effect pair, the Born pairing or the factor map. DEF-QDD-FACTOR-MAP F_QDD : QCarrier_QDD -> MatterData_QDD by the displayed Gram/projector formulas on the transpose slot. ``` L1 theorems, the L4 instrument nonselection theorem, and the separate apparatus obligation. ``` QDD-ALGEBRAIC-FACTORIZATION [T] D_QDD_direct = F_QDD o Q_QDD o beta_QDD field by field on all 15625 checkpoints; the record is total (25 ZERO_SUPPORT heads, 15600 SUPPORTED), exactly normalized, independent of q and r, dependent on each piston coordinate, constant on each of the 313 Q_QDD-fibres (one of size 25 and 312 of size 50) and injective on QCarrier_QDD; controls: the rational-line reading mismatches on 480 of 625 pistons and omitting G on 540 of 625. The direct write and its definitional closure in the dependency ledger name no factor-side object (the transitive firewall, enforced by the status-separation witness). An identity of the adopted definitions, not an independent readout, not a physical selection, and not a completion, totality or uniqueness claim for D_matter. QDD-PROJECTOR-PAIR-TR4 [T] E_low is the unique G-self-adjoint idempotent with kernel ker Tr_4, since a G-self-adjoint idempotent has image (ker)^perp_G and G^-1 1 = 5 1 gives (ker Tr_4)^perp_G = span(1); {E_low, E_high} is the G-orthogonal resolution of Q^4 along the piston character Tr_4; closed forms m = |v|^2 - s^2/5, w_low = s^2/20, w_high = |v|^2 - s^2/4, s = sum v_i, so w_low and w_high are the squared trace-pairing lengths of the projections onto span(1) and onto ker Tr_4. Linear algebra only; no apparatus, no physical reading, and no uniqueness-from-J: the theorem identifies the pair inside the stated algebraic class and does not force the choice of that class. QDD-QCARRIER-DIAGONAL-BOUNDARY [T] on the frozen V_eff, A_dagger = A_T = v v^T. Both slots remain typed and declared; the current domain does not test their difference; no physical central phase is derived from this equality. The cyclotomic pair (w sigma_4(w), w^2) has 90 distinct Hermitian slots and 313 distinct pairs on the 625 pistons, and 80 Hermitian slots carry more than one record; neither Herm-only nor use of both slots is asserted. QDD-INSTRUMENT-NONSELECTION [T] At L4 apparatus/support scope over V = Q^4, freeze G = I_4 - (1/5) 1 1^T, A^sharp = G^-1 A^T G, E_low = (1/4) 1 1^T and E_high = I_4 - E_low. Every nonzero G-self-adjoint idempotent effect fibre is {W E : W in O(G,Q)}. The ordered raw two-branch fibre is one branchwise O(G,Q) x O(G,Q) orbit, not one diagonal orbit; Gamma_ab = K_a^sharp K_b completely classifies diagonal orbits, with C = K_low^sharp K_high the complete two-branch invariant. Under the frozen pure density-operator post-state definition, physical equivalence inside one nonzero effect fibre is exactly K ~ +/-K. The rational rotations R_t on im(E_high) therefore give an injection Q -> physical post-state instrument classes at fixed effects, fixed branch weights and C = 0. Every rational isometry between subspaces of a positive-definite rational bilinear space extends by rational reflections, so every complete rational two-branch family has a rational orthogonal dilation on the frozen system/pointer type. Existence of an unrestricted rational orthogonal dilation is therefore not an instrument-selection principle. A coupling already controlled by the target projectors is circular as independent-selection evidence. G-self-adjointness and G-positivity select K = E only as a mathematical positive-square-root section, not as a physical law. Finally K^T G K = G E reproduces the frozen occurrence weights globally. No physical selector, L5 realized-event stream, L6 measure, decoder completion or SI statement; SAMPLING NOT PROVIDED, not SAMPLING IMPOSSIBLE. QDD-U-INDUCED-CHANNEL [T] On the registered autonomous carrier Omega = N_0 x F_5^6, write x = (p1,p4,p1p,p4p,q,r), pi(x) = (p1,p4,p1p,p4p), and f(x) = (q,r). The five generator fibers are exactly f(a(x)) = (q,r), f(b(x)) = (-q,-r), f(c(x)) = (1-q,-r), f(d(x)) = (1-q,1-r), f(e(x)) = (2-q,1-r). Since sigma = sum pi(x) + (q+r) + 2 theta_n mod 5, the delay-one fiber is a function only of (f(x_k),theta_k,sum pi(x_k)); every piston-to-fiber influence at delay one passes through the selector. Fiber-to-piston influence is witnessed independently by theta=0, x=000000, y=000020, sigma=0,2, post=0000,2121, and theta=0, x=000002, y=000011, sigma=2, post=2324,2220, difference=0104=(r-r')u_c. This is exact L1 bidirectional algebraic channel structure. It claims no structural factorization for delays d>=2 and no physical measurement, selector, instrument, effect realization, event law, sampling map, normalized measure or decoder completion. QDD-U-INDUCED-FINITE-NONSELECTION [C] Freeze the six nonzero projective F_5-linear fiber functionals, the thirty nonempty proper subsets S of F_5, the 180 records rho_(lambda,S), delays D={1,2,3,4,5}, W with 512 <= n < 2048 over all 15625 seeds, W2 with 2048 <= n < 16384 over the 625 ready-fiber seeds, and the summed W census. The complete exact classification of all 900 record-delay pairs gives INFO=150 on the census and zero realizations of the frozen occurrence law on W, W2 or the census. It further gives FUNCTIONAL=0, ORIENT-POST-COHERENT=0, POST-PURE-STRICT=0, POST-MIXED=0, POST-UNDEFINED-OR-ZERO=900, ZERO-INPUT-MULTIVALUED=900. The frozen-family eligible set is empty, so evaluated/member/outside is 0/0/0; no outside-family conclusion follows. Seed-dependent triples number 271350, orientation-dependent triples number 22500, and the canonical table root is 0baacabc9d94a824c6a9480695c7a37f2762a3a2e773d1161c26816a2dbdee15. This is a multi-layer finite classification: its induced-apparatus conclusion is at L4, over the registered L1 update and the frozen finite L5 windows, for this U, split, record class, delays and seeds only. It supplies no limit, independent physical selector, sampling-impossibility theorem, exclusion of another admissible apparatus class or decoder completion. QDD-INSTRUMENT-APPARATUS [O] after the nonselection theorem only two independent blockers remain: O2, independent physical instrument selection from a public admissible law or coupling class frozen before comparison with E_low and E_high; O1, realized event generation and sampling. An exact rational apparatus reducing to the Lueder pair is exhibited, but its target-controlled coupling is circular as independent-selection evidence. Positive-root uniqueness is mathematics only, equality of effects does not identify post-state instruments, the row remains separate from QUADRATIC-DECODER-DATA and fills no decoder-completion-contract field. SAMPLING NOT PROVIDED; SAMPLING IMPOSSIBLE is not claimed. ``` Disclosure. `|QCarrier_QDD| = 313 = 1 + (5^4 - 1)/2` arises from the sign identification `v ~ -v`, with fibres 25 and 50. CENSUS-313 has the same count and the same 25/50 profile from a different origin, and the two partitions of `F_5^6` share no block. No cross-leg identity is claimed. Over the 312 nonzero classes the normalized pair takes 22 values; the value 1/6 on 12 classes is a numerical witness with no input, threshold, normalization, confirmation or dependency role. Evidence: `reproduce/qdd-route-a`, byte-identical on the public x86_64 and aarch64 jobs, RESULT 15/15 ALL PASS. At L4, `P-QDD-INSTRUMENT-NONSELECTION-1` supplies the theorem-grade written proof and exact two-architecture audit for S1a-S6. It proves an injective rational family of physically distinct post-state instruments at fixed effects, weights and C = 0 and proves rational orthogonal-dilation surjectivity. It does not select a physical family or create an event stream or measure. `P-QDD-INSTRUMENT-U-INDUCED-1` supplies the written L1 channel proof and its complete checkpoint audit, together with the two explicit fiber-to-piston witnesses. The same two-architecture bundle exhausts the frozen 900 record-delay pairs and supplies the finite multi-layer classification above, whose apparatus conclusion is at L4. Neither result closes `QDD-INSTRUMENT-APPARATUS`: O2 independent physical selection and O1 realized event generation or sampling remain STOP. The empty frozen-family eligibility set licenses no conclusion about instruments outside that family. ## 3. The kernel and the census F_5^6, 15625 checkpoint states; the Klein-100 typology; 313 attractors. The finite kernel is the declared checkpoint architecture paired with the algebraic verb. No derivation or uniqueness of this architecture from J or M_J is claimed. ### FIELD-ZERO-NONZERO-MULTIPLICATIVE-CUT [T] Let \(F\) be any field, let \(\varnothing\ne A\subsetneq F\), and put \(d_A=1_A:F\to\{0,1\}\). There is a total Boolean operation \(B:\{0,1\}^2\to\{0,1\}\) satisfying ```text d_A(xy) = B(d_A(x),d_A(y)) for every x,y in F ``` if and only if exactly one of the two oriented cases holds: ```text A = {0}, B = OR; A = F^x, B = AND. ``` In either case \(B\) is unique. Indeed, put \(\epsilon=d_A(0)\). Since both colors occur and \(0y=0\), one has \(B(\epsilon,b)=\epsilon\) for both attained bits. If a nonzero \(u\) also had color \(\epsilon\), then \(d_A(uy)=\epsilon\) for every \(y\); multiplication by \(u\) is bijective, so \(d_A\) would be constant. Thus `d_A(u)=1-epsilon` for every nonzero `u`. All four Boolean input pairs are realized by zero/nonzero factors, so the table is uniquely OR when `epsilon=1` and uniquely AND when `epsilon=0`; the absence of zero divisors proves both converses. The exact evidence is `probes/P-FIELD-ZERO-NONZERO-CUT-1`. The theorem is independent of characteristic and contains no specifically fifth-prime content. Deleting zero changes the problem: on \(F_5^x\), the two quadratic-character orientations are `QR/XNOR` and `NQR/XOR`. This unit-group boundary does not contradict the total-field classification. The result is L1 field and Boolean algebra only. Its shape is a structural cross-reference to the public `ZERO_SUPPORT | SUPPORTED` and `ZERO_DENOMINATOR` branches, not a theorem that any QDD record satisfies the displayed composition law. It selects no field or cut, creates no QDD dependency, and moves no decoder, apparatus, L5 stream, L6 measure, or frontier status. The checkpoint space is Z_5^6 with coordinates (p1, p4, p1p, p4p, q, r) and five involutive generators, all arithmetic mod 5: ``` a swap (p1, p4, p1p, p4p, q, r) -> (p4, p1, p4p, p1p, q, r) b time inversion x -> (-p1p, -p4p, -p1, -p4, -q, -r) c transport piston -> b4(piston) + s_c + r u_c; q -> 1 - q; r -> -r d mirror x -> c_d - x e shifted mirror x -> (c_d + v_e) - x s_c = (2, 1, 2, 1), u_c = (0, 1, 0, -1), c_d = (2, 1, 3, 4, 1, 1), v_e = (0, 0, 0, 0, 1, 0); b4 is the piston part of b. Relations: a^2 = b^2 = c^2 = d^2 = e^2 = id and (bc)^5 = id. Drive: theta_n = s_2(n) mod 2 (Thue-Morse); phase z_6(x) = Tr_6(x), the sum of the six coordinates mod 5; selection law i_n = z_6(psi_n) + 2 theta_n mod 5, with the autonomous update U defined in section 2. ``` For the Thue-Morse drive itself, put `t_n = theta_n = s_2(n) mod 2` and, for every `L >= 1`, ```text n_a(L) = #{0 <= i <= L - 1 : t_i = a}, c_ab(L) = #{0 <= i <= L - 2 : (t_i,t_(i+1)) = (a,b)}, S(L) = n_0(L) - n_1(L), d(L) = 6 c_00(L) - L, ``` with the empty pair census at `L = 1`. The exact substitution seams imply ```text d(2L) = -d(L) + 3S(L) - 6, d(4L) = d(L) - 3S(L). ``` Hence `d(4L) = d(L)` exactly when `L` is even. The least positive counterexample to the unqualified four-step equality is `L = 1`, with `d(1) = -1` and `d(4) = -4`. The signed-affine state `q(L) = (S(L),t_(L-1),t_L)` has exactly the six reachable states ```text (-1,1,0), (0,0,0), (0,0,1), (0,1,0), (0,1,1), (1,0,1). ``` Its complete 96-path four-bit certificate gives the exact cumulative extrema ```text E_1 = 2, E_2 = 4, E_k = 2(floor((k+1)/4) + 2), k >= 3, ``` where `E_k = max_(1 <= L <= 2^k) |d(L)|`. The endpoint obeys `|d(2^k)| = 2` for odd `k` and `4` for even `k`. Consequently ```text d(L) = O(log L), d(L)/L^epsilon -> 0 for every fixed epsilon > 0, c_00(L)/L -> 1/6, c_00(L)/(L - 1) -> 1/6, ``` with the last normalization used only for `L >= 2`. These statements are GYRON-DISCREPANCY-LOG [T] at L1, evidenced by probes/P-GYRON-DISCREPANCY-LOG-3. The finite prefix census is an independent audit, not the proof of any universal clause. The corresponding forward pair-substitution law must retain its two anchoring conventions. In pair order `v = (a,b,c,d) = (v_00,v_01,v_10,v_11)`, put ```text I_L(v) = (0,a+b,c+d,0), I_R(v) = (0,a+c,b+d,0), B(v) = (c,d,a,b), R_L = (I_L+B)/2, R_R = (I_R+B)/2. ``` On the full four-dimensional carrier the maps and their spectra differ: ```text chi_(R_L)(x) = x(x-1)(x-1/2)(x+1/2), chi_(R_R)(x) = x(x-1)(x+1/2)^2. ``` On the stationary subspace `b = c` their restrictions agree. The common forward phase-averaged operator `R` has spectrum `{1,-1/2,0}`. In the normalized stationary affine space `b = c` and `a+b+c+d = 1`, ```text v_* = (1,2,2,1)/6 ``` is the unique fixed point, and `R^n v -> v_*` componentwise for every input in that affine space. At the fixed point, ```text I(v_*) = (0,1/2,1/2,0), B(v_*) = (1/3,1/6,1/6,1/3), v_* = (I(v_*)+B(v_*))/2. ``` Thus the `00` coordinate is `0` in the `I(v_*)` phase, `1/3` in the `B(v_*)` phase, and `1/6` only under the frozen equal phase average. These statements are TM-PAIR-SUBSTITUTION-FIXED-POINT [T] at L1, evidenced by probes/P-GYRON-DISCREPANCY-LOG-3. GYRON-DENSITY [T] records exactly the two compatible L1 density clauses: the two prefix-frequency limits above, and the `00` coordinate `1/6` of the unique normalized equal-phase stationary fixed point. It is not an unqualified finite-prefix invariant: `d(4L) = d(L)` holds iff `L` is even. It is also not a density in either fixed phase. The value is not a six-line cardinal average, Born multiplier, mass density, cosmological parameter, selector weight, physical probability, L5 stream, or L6 measure. The operator `R` is a normalized forward substitution or inflation operator, not coarse-graining, desubstitution, inverse RG, a blocking-origin-independent finite-prefix histogram, a decoder factor or quotient, a physical probability or measure, or an L1-to-L5/L6 bridge. Equality of the `R_L` and `R_R` restrictions holds only on the stationary subspace and does not identify the full maps or their spectra. The census (CENSUS-313 [C], reproduce/census): the full enumeration of all 15625 seeds, warmup 400 ticks and window 300, yields exactly 313 attractors: 312 of size 20 with basins of 50, and one singlet of size 10 with a basin of 25; the attractor support has 6250 states, the basins cover all of Z_5^6, and 313 = 13^2 + 12^2. Every recurrent state lies on the z_6 sheet {1, 4}, so the selection law fires only b, d, and e on the attractor: the recurrent algebra is the mirror algebra (CENSUS-Z5-SHEET [C], reproduce/census). The piston pairing Phi = (b4 d4 on pistons, identity on the fiber) is an involution that permutes the 313 attractors with cycle type 144 transpositions plus 25 fixed points, the fixed points being 24 attractors of size 20 and the singlet, so the classes modulo the pairing number 169 = 13^2 (CENSUS-PAIRING [C], reproduce/census). The hosting formula holds on all 313 (CENSUS-HOSTING [C], reproduce/census): the return group `H_1 = ` has order 10, d o (b e b) is the line translation T_0 = (0, 0, 0, 0, 3, 2), and for every attractor A there is a chosen representative x_A in A such that the exact two-coset formula is `A = H_1 x_A union b H_1 x_A`. The boundary hyperplane: with Tr_4 the piston sum and M the z_6 sheet, the boundary class S_hyp = M cap {Tr_4 = 0} has size 1250 exactly, the charged complement 5000 in two sheets of 2500, the boundary fibers 625 + 625, by full enumeration of all 15625 states (HYPERPLANE-BOUNDARY-CLASS [T], reproduce/hyperplane-codec); the census window realizes S_hyp as the union of the 63 = (p^3 + 1)/2 boundary attractors, 62 of size 20 with the singlet, and the charged sector as the remaining 250 (HYPERPLANE-BOUNDARY-REALIZATION [C], reproduce/hyperplane-codec). The unique M_J readout is the trace character: the step matrix is multiplication by J column by column, its characteristic polynomial is Phi_5(x - 1) with det(2I - M_J) = 5 = p. For x = (x_a, x_b, x_c, x_d) in the power basis over Z, the exact identity is ``` Tr_4(M_J x) = 2 Tr_4(x) - 5 x_c. ``` Therefore Tr_4(M_J x) = 2 Tr_4(x) in F_5, and the scalar multiples of Tr_4 are the only covectors reading any multiplier at all (CODEC-TR4 [T], reproduce/hyperplane-codec). This fixed channel has an exact ramified four-state lift. Put `lambda = 1 - zeta_5` and `J_lambda = [J] mod lambda`. Reduction sends `zeta_5` to 1, hence `J_lambda = 2` in `F_5^*`, with orbit `1, 2, 4, 3` and exact order four. Define ``` Theta_0 = 1, Theta_(2n) = Theta_n, Theta_(2n+1) = J_lambda Theta_n. ``` Binary-length induction gives the unique solution ``` Theta_n = J_lambda^s_2(n). ``` For the sign quotient `q : F_5^* -> F_5^*/{+-1} ~= F_2`, with `q(+-1) = 0` and `q(+-2) = 1`, one has for every `n >= 0` ``` q(Theta_n) = theta_n, Theta_n^2 = (-1)^theta_n. ``` The four-cycle is the digit-1 transition, not the chronological successor. If `c_n = nu_2(n+1)`, then the exact successor law is ``` Theta_(n+1) = Theta_n J_lambda^(1-c_n). ``` Moreover, for every `x_0 in Z^4` with `Tr_4(x_0) = 1 mod 5` and every `k >= 0`, induction on the CODEC-TR4 identity gives ``` [Tr_4(M_J^k x_0)]_5 = J_lambda^k. ``` Thus `k = s_2(n)` realizes `Theta_n` on the one-dimensional fixed `M_J/Tr_4` readout quotient. The channel selects multiplier 2; the sign quotient alone is inversion-blind and also sends `3^s_2(n)` to `theta_n`. Independently, the same integer is the unique universal binary carry weight: ``` x + y = (x XOR y) + 2 (x AND y) for all x,y in N_0, ``` with uniqueness already forced by `x = y = 1`. The frozen formulas also imply, for all `x,y in N_0`, ``` s_2(x) + s_2(y) = s_2(x XOR y) + 2 s_2(x AND y), Theta_x Theta_y = Theta_(x XOR y) omega(x,y), omega(x,y) := (-1)^theta_(x AND y) in {+-1} subset F_5^*. ``` On the XOR group `(N_0, XOR)`, `omega` is the symmetric normalized bicharacter induced by the finite-support bit pairing `sum_i x_i y_i mod 2`. Equivalently, `omega(x,y) = Theta_x Theta_y Theta_(x XOR y)^(-1)`, so it is the normalized 2-coboundary of `Theta` and hence a 2-cocycle. In this factorization, `x AND y` is the exact bit-intersection datum from which the scalar factor set `omega` is read. Together these exact statements form RAMIFIED-TM-LIFT [T] at L1, evidenced by probes/P-RAMIFIED-TM-LIFT-1. The equality of the two integer values is an arithmetic consonance, not a physical carry or phase identification. No constant chronological quarter-turn, full-state four-phase identification, checkpoint factorization, parity on all of `Z_2`, time-arrow reading, or lift to L2-L6 is claimed. The checkpoint-only factorization closes negatively on the frozen full forward carrier ``` C = {U^n(0, psi_0) : n >= 0, psi_0 in F_5^6}. ``` For every initial checkpoint, the five trace laws force `z_6(psi_3) = 1`; the selectors at steps three, four, and five are therefore all one. The same involution `b` acts three times, so `b^2 = id` gives `psi_4 = psi_6`, while the inherited lift has `Theta_4 = 2` and `Theta_6 = 4`. Hence no single-valued `h : F_5^6 -> F_5^*` factors `Theta_n` through the checkpoint on `C` (CARRY-J-CHECKPOINT [T] at L1, probes/P-CARRY-J-CHECKPOINT-1). The theorem decides no restricted carrier, selector offset, physical carry, phase, time, or gravity reading, decoder completeness, parity on all of `Z_2`, or lift to L2-L6. KERNEL-Z6-SYNCHRONIZATION [T] is the exact L1 theorem for the declared autonomous update `U` on `X = F_5^6`. For `z in F_5`, put ``` X_z = {psi in X : z_6(psi)=z}, X_14 = X_1 union X_4, E_n(psi_0)=pr_checkpoint(U^n(0,psi_0)), q_n=4+2 theta_(n-1) mod 5 for n>=1. ``` Here `q_n` is a sheet label, not the checkpoint coordinate `q`. Direct summation of the declared generator coordinates and direct substitution give ``` z_6(a psi)=z_6(psi), z_6(b psi)=-z_6(psi), z_6(c psi)=2-z_6(psi), z_6(d psi)=2-z_6(psi), z_6(e psi)=3-z_6(psi), g_i^2=id. ``` Consequently the selector gives the complete sheet table ``` input sheet z 0 1 2 3 4 t=0 0 4 0 4 4 t=1 2 1 1 3 1. ``` Every displayed arrow is a bijection: on a fixed input sheet the selector chooses one involution, and the source and target sheets both have `5^5=3125` elements. The theorem has four clauses: ``` S1 for every n>=3 and z in F_5, E_n|X_z:X_z->X_(q_n) is a bijection; hence, at each fixed n, E_n:X->X_(q_n) is exactly 5-to-1, with one preimage in every initial sheet; S2 for every n>=1, E_n|X_1 and E_n|X_4 are separate bijections onto X_(q_n); hence, at each fixed n, E_n|X_14 is exactly 2-to-1; S3 for every psi_0 in X, neither (z_6(E_n(psi_0)))_(n>=0) nor (E_n(psi_0))_(n>=0) is eventually periodic; S4 for every psi_0 in X, no finite set Y, self-map H:Y->Y, map pi:Y->X, and y_0 in Y satisfy pi(H^n(y_0))=E_n(psi_0) for all n>=0. ``` Proof. The initial bits are `(theta_0,theta_1,theta_2)=(0,1,1)`. The sheet table sends the complete sheet sets through ``` {0,1,2,3,4} -> {0,4} -> {1,2} -> {1}, ``` and, separately on every initial sheet, the three restrictions compose to a bijection `E_3|X_z:X_z->X_1`. This is S1 at `n=3`. At `n=1`, the restrictions `X_1->X_4` and `X_4->X_4` are bijections, giving the S2 base. For the induction step put `t=theta_(n-1)` and `u=theta_n`. The four possible cases are ``` t u q_n selector i_n q_(n+1) 0 0 4 4 4 0 1 4 1 1 1 0 1 1 4 1 1 1 3 1. ``` In every case the selected generator restricts to a bijection `X_(q_n)->X_(q_(n+1))`. Composition proves S1 for every `n>=3` and S2 for every `n>=1`. The five initial sheets are disjoint and partition `X`; `X_1` and `X_4` are disjoint. The fixed-time multiplicities follow. For S3, suppose `theta_n` has period `p>=1` after `N`, and put `w=s_2(p-1)`. Choose `k` so large that `2^k-p>=N` and `k-w` is even. Since ``` 2^k-p=(2^k-1)-(p-1), s_2(2^k-p)=k-s_2(p-1), ``` we have `theta_(2^k-p)=0`, whereas `theta_(2^k)=1`. The two indices differ by `p`, a contradiction. Thus Thue--Morse is not eventually periodic. By S1, for every seed and `n>=3`, ``` z_6(E_n(psi_0))=q_n=4+2 theta_(n-1) mod 5. ``` The map `t |-> 4+2t mod 5` is injective on `{0,1}`, so the checkpoint trace is not eventually periodic. A periodic checkpoint trajectory would have a periodic image under `z_6`; hence the trajectory is not eventually periodic. This proves S3. For S4, let `|Y|=m`. The existence of `y_0` gives `m>=1`, and two of `H^0(y_0),...,H^m(y_0)` coincide. Determinism of `H` propagates the equality, so its orbit and its image under the fixed map `pi` are eventually periodic, contrary to S3. The exact finite premises, bases, four induction cases, proof schemas, and two independent exact implementations were audited by `probes/P-KERNEL-Z6-SYNCHRONIZATION-1`; `RESULT.md` records `PROOF-SURVIVES` and byte-identical aarch64/x86_64 output. Every multiplicity above is a fixed-time statement. The theorem is L1 only. It makes no unknown-time or unindexed checkpoint-fiber claim, assigns no census meaning to `X_14`, does not make `X` the complete autonomous state, and derives no decoded-log, physical-irreversibility, unique infinite-realization, decoder completion, or L2-L6 statement. `CARRY-J-CHECKPOINT [T]` is lineage, not a logical premise. The same ramified digit recursion has an exact two-branch lift through `C8 -> C4 -> C2`. In `F_25 = F_5[tau]/(tau^2-2)`, put `eta = tau^3`. Then `eta^2 = phi = 3`, `eta^4 = -1`, `eta` has order eight, and its norm to `F_5` is the ramified image `J_lambda = 2`; the roots of `r^2 = phi` are exactly `eta` and `-eta`. On ``, the norm has kernel `{+-1}` and image ``, but no section because both preimages of `J_lambda` have order eight. For either root `r_epsilon`, define the digit recursion ``` Y_0^epsilon = 1, Y_(2n)^epsilon = Y_n^epsilon, Y_(2n+1)^epsilon = r_epsilon Y_n^epsilon. ``` Binary-length induction gives its unique solution and successor law: ``` Y_n^epsilon = r_epsilon^s_2(n), Y_(n+1)^epsilon = Y_n^epsilon r_epsilon^(1-nu_2(n+1)). ``` For both branches and every `n >= 0`, the exact projections are ``` N(Y_n^epsilon) = Theta_n, q(N(Y_n^epsilon)) = theta_n, (Y_n^epsilon)^2 = Theta_n^-1, (Y_n^epsilon)^4 = (-1)^theta_n, Y_n^- = (-1)^theta_n Y_n^+. ``` Thus `Y_n -> Theta_n -> theta_n` is the exact `C8 -> C4 -> C2` tower, and the chronological multiplier is not constant. These statements form SQRT-PHI-DIGIT-LIFT [T] at L1, evidenced by probes/P-SQRT-PHI-DIGIT-1. Neither sign branch is physically selected; no checkpoint identification, physical tick, time arrow, gravity dynamics, coupling, SI scale, or lift to L2-L6 is claimed. The typed clock and gravity bridge remains SQRT-PHI-TIME-GRAVITY [O]. The two-branch lift has an exact branch-invariant bilinear shadow. In the same field `K = F_25 = F_5[tau]/(tau^2-2)`, ``` = F_5^* union tau F_5^*. ``` The even powers are exactly the nonzero base-field axis, and the odd powers are its `tau`-multiple. The eight powers are distinct roots of `x^8-1`, while the four roots of `x^4-1` already lie in `F_5^*`; hence the elements of exact order eight are precisely `{+-tau,+-eta}`. Frobenius is conjugation: ``` Frob(a+b tau) = (a+b tau)^5 = a-b tau. ``` It fixes `F_5`, negates `tau F_5`, preserves norm under sign, and sends `eta` to `-eta`. Therefore ``` Frob(Y_n^+) = Y_n^- for every n >= 0. ``` For either registered branch, form the record `V^epsilon` containing every `Theta_n`, each `Y_n^epsilon` with even `s_2(n)`, and each same-branch product `Y_n^epsilon Y_m^epsilon` when both digit sums are odd. Writing `s=s_2(n)` and `t=s_2(m)` gives ``` s even: Y_n^+ = Y_n^- = phi^(s/2) in F_5, s,t odd: Y_n^+ Y_m^+ = Y_n^- Y_m^- = phi^((s+t)/2) in F_5. ``` Thus `V^+ = V^- =: V` and the whole record is `F_5`-valued. By contrast, the branches differ exactly on odd digit-sum classes, and a mixed-parity product is a nonzero `tau F_5` element whose single branch sign does not cancel. Since both `2` and `phi=3` have order four, the norm channel `Theta_n=2^s` reads exactly `s mod 4`, whereas `V` reads `s mod 8` on its even classes and `s+t mod 8` on its odd-pair classes. The witnesses `n=15,255` have digit sums `4,8`, and the pairs `(1,1),(1,31)` have sums `2,6`. On the even classes carried by `V`, `(Y_n^epsilon)^2=Theta_n^-1`. These statements form C8-BILINEAR-SHADOW [T] at L1, evidenced by the immutable public bundle `probes/P-C8-BILINEAR-SHADOW-2`. Its self-contained all-`n` proof carries the theorem; the finite verifier audits the field, axes, roots, residue classes, branch record, and successor identities. The formal aarch64 execution and required GitHub x86_64 check produced byte-identical output. The theorem selects no branch and asserts no broader physical or gauge equivalence, checkpoint identification, clock, gravity, SI, force, uniqueness, or lift to L2-L6. SQRT-PHI-TIME-GRAVITY [O] remains open. The next carry stratum has an exact pentagonal form in the frozen four-coordinate Hamming frame. On `V = F_2^4`, put ``` w(x) = popcount(x), t(x) = w(x) mod 2, d(x,y) = popcount(x AND y) mod 2, q(x) = binom(w(x),2) mod 2. ``` Then `q` is the second binary weight bit and its polarization is `B(x,y) = t(x)t(y) + d(x,y)`. The form `B` is alternating and nondegenerate, `Arf(q) = 1`, and the five nonzero singular vectors are `P = {1,2,4,8,15}`; distinct members pair to one and their XOR is zero. Consequently ``` O(q) = O^-(4,2) ~= S_5, Sp(4,2) ~= S_6, ``` with `O(q)` the stabilizer of the selected minus-type refinement. The raw intersection form `d` has stabilizer `S_4 x C_2` of order 48 and therefore has no element of order five. The five-fold symmetry belongs to `q`, not to `d` alone. With order five fixed, dimension four is the least binary linear width admitting it, because `ord_5(2) = 4`. The integral bridge is the augmentation root lattice `A4 = {(z_0,...,z_4) in Z^5 : sum z_r = 0}` with the coordinate five-cycle `C`. Under `a_i = e_i-e_0 -> zeta_5^i-1`, `A4` is the ideal `(zeta_5-1)Z[zeta_5]`; `I+C^2` is integrally conjugate to the public `M_J`, and for every `a in (Z/5Z)^*`, ``` char(I+C^a) = Phi_5(X-1), g_a(I+C)g_a^-1 = I+C^a. ``` Thus the unfixed integral-isometry class does not select an exponent, cycle, or orientation. The ramified quotient `A4/(C-I)A4` has order five, and every `I+C^a` acts on it as multiplication by 2; its characteristic shadow is `(X-2)^4 mod 5`. Modulo two, the even-lattice refinement is `q_A(x)=q(x)+B(15,x)`, and the transvection `tau_15(x)=x+B(x,15)15` is the explicit isometry from `(A4/2A4,q_A)` to `(V,q)`. Weyl reduction gives `W(A4) ~= S_5 -> O(q_A)`, and the five nonzero singular classes map to the five powers of `zeta_5`, whose sum is zero. These statements form CARRY-PENTAD [T] at L1, evidenced by `probes/P-CARRY-PENTAD-1`. The theorem is relative to the frozen frame and fixed order-five target. It does not unconditionally select `p=5`, width four, a five-cycle, its orientation or exponent, and it adds no decoder, physical gauge, phase, force, spacetime, entropy, measure, or lift to L2-L6. No coding rate is inferred from this dimension count. The inherited phrase "rate 4/5" is retired from Public Canon v13; any future coding claim must define its alphabet, message space, encoder, decoder, error criterion, and rate from scratch. Six completed public probes now delimit the entropy bridge without closing it. Write ``` F_eps(psi) = g_{z_6(psi) + 2 eps mod 5}(psi). ``` For this program, `K = Q(zeta_5)`, `lambda = 1 - zeta_5`, and `O_(K,lambda)` is the completion of `O_K` at the unique place above 5. `K_TM` is the two-sided Thue-Morse subshift with shift `S_K`, unique substitution probability `m_TM`, and reading `theta(kappa) = kappa_0`. Let `h_lambda` be normalized additive Haar probability on `O_(K,lambda)`, so `h_lambda(O_(K,lambda)) = 1`, and put ```text mu = m_TM x h_lambda, tau_src(kappa,y) = (S_K kappa,Jy). ``` Give the source its product Borel sigma-algebra and `F_5^6` its discrete sigma-algebra. Maps are identified only by equality `mu`-almost everywhere. The typed Route A target of ENTROPY-LAYER-BRIDGE [O] is a measurable total map ```text P_5 : K_TM x O_(K,lambda) -> F_5^6 ``` satisfying ```text P_5(tau_src(kappa,y)) = F_(theta(kappa))(P_5(kappa,y)) ``` `mu`-almost everywhere. This is the complete equality convention; no unregistered uniqueness or canonicity quotient is implied. Let `R` be the public set of 6250 recurrent finite-kernel states and let `W = {512,...,2047}`. For a measurable total `P`, define ```text psi_(P,n)(kappa,y) = P(tau_src^n(kappa,y)), i_(P,n)(kappa,y) = z_6(psi_(P,n)(kappa,y)) + 2 theta(S_K^n kappa) mod 5, nu_(P,W) = (1/1536) sum_(n in W) (psi_(P,n))_* mu. ``` `Law_W(P)` is the exact conjunction ```text nu_(P,W)({psi}) = 1/6250 for psi in R, nu_(P,W)({psi}) = 0 for psi not in R, (1/1536) sum_(n in W) mu(i_(P,n)=j) = (0, 2/3, 0, 1/6, 1/6)_j, (1/1536) sum_(n in W) m_TM(theta(S_K^n kappa)=theta(S_K^(n+1) kappa)=0) = 1/6. ``` Equivalently, the generator coordinates are `(a,b,c,d,e)` in the displayed order. The recurrent component masses are therefore exactly their normalized basin masses: `50/15625` for each size-20 component and `25/15625` for the singlet. `Law_W` is only the frozen finite-window predicate for `512 <= n < 2048`; it asserts no limit and no larger-window law. Let `A_A` be the set of `mu`-almost-everywhere classes of measurable total maps of the displayed type satisfying both exact equivariance and `Law_W`. The archimedean coordinate from the original source never enters this finite cut and is not part of the target type. The literal integer lift of the finite generator presentation does not satisfy `(bc)^5 = 1`: over Z its fifth iterate is ``` (bc)^5(x) = x + (10, 5 - 5r, 10, 5 + 5r, 5, 0). ``` Thus the relation holds in the mod-5 shadow, not in that literal lift (ENTROPY-LIFT-DEFECT [F], probes/P-ENTROPY-BRIDGE-1). On the declared finite carrier, the same probe establishes the window-scoped joint law: component masses equal basin sizes from tick 512, the checkpoint marginal is exactly uniform on the 6250 recurrent states on the frozen window, and the letter and pair masses are exactly `(2/3, 1/6, 1/6)` and `1/6` (ENTROPY-JOINT-CESARO-LAW [C]). The original finite-cylinder machine coverage is narrower than earlier prose stated. Its pure-word system has zero solutions at cursor `c = 0` for every `L = 4..16`, and additionally at `(L,c) = (5,1), (6,1), (6,2)`. The `J`-invariant zero residue embeds each of these exact pure-word obstructions at every lambda-depth. Direct depth tables corroborate the listed small-window cases at their frozen orbit lengths. No other cursor is claimed by that older probe (ENTROPY-CYLINDER-CUT [F], probes/P-ENTROPY-BRIDGE-2). The separate proof-first cursor theorem uses the same exact L5 finite-cylindrical constraint graph, with the global solution count equal to the product of the root-seed counts over all weak components. For every `L = 4..32` and every `c = 0..L-1`, all 522 distinct pure-word systems have global solution count zero. The zero residue is fixed by multiplication by `J` at every finite lambda-depth, and its labelled context graph projects exactly to the pure-word node and edge graph. Thus each of those 522 obstructions holds at every finite lambda-depth by restriction, not by extrapolation from the 27 direct audits at `ell in {4,20,100}`. This is ENTROPY-CYLINDER-NOGO-CURSOR [T], evidenced by probes/P-ENTROPY-CURSOR-CLOSURE-1 with byte-identical aarch64 and x86_64 transcripts. It excludes only the typed finite-cylindrical L5 ansatz in the declared window range. It supplies no non-cylindrical cut, construction of `P_5`, measurable selection, entropy, regularity, canonicity, image-law, measure lift, L6 statement, or physical interpretation. At the tested dyadic scales `k = 0..10`, both renormalized block maps are exactly two-to-one on the recurrent core, one unresolved bit per scale (ENTROPY-BLOCK-HALVING [C]). Two-to-one means coarse graining here is a semigroup and not a group, and the object that ordinarily reads such a semigroup is its fixed points together with the spectrum of its linearization. That object is now computed exactly at the dyadic scales `k = 0..14` for both letters, and it is not a fixed point in the usual sense. Outside two residues the block maps fix no state at all, on the recurrent core or off it. At `k = 0` each letter fixes exactly one recurrent state, the reflection centre `3 (C_D + V_E)` for `eps = 0` and `3 C_D` for `eps = 1`, both in the size-10 component, each with multiplier exactly minus the identity. At every scale with `k = 1 mod 4` the fixed set is exactly the opposite living half `H_(1-eps)`, 3125 states meeting all 313 components, and the multiplier at every one of them is exactly the identity, so no expanding or contracting datum exists there to be read. What the range exhibits instead is a return: the scales carrying a full-half return are exactly those whose block length satisfies `2^k = 2 mod 5`, and the scales with `2^k = 1 mod 5` carry none (ENTROPY-RG-RETURN [C], probes/P-ENTROPY-RG-RETURN-1). The image of every block map on the core is 3125 states across the whole range, which re-audits ENTROPY-BLOCK-HALVING four scales wider than the row itself states. The agreement of that residue class with the order of the ramified digit unit of RAMIFIED-TM-LIFT is an arithmetic consonance and is recorded as one; the computed row rests on the residue arithmetic and the enumerated fixed sets, not on that agreement. No continuum limit, scaling limit, critical exponent, monotone scale function, measure, or all-scale law follows. The two branch images partition the core into halves of 3125 states, and all four restrictions between source and target halves are bijections. Every size-20 attractor splits `(10, 10)`, the singlet splits `(5, 5)`, and there are `312 x 10 + 5 = 3125 = 5^5` living trajectories (ENTROPY-LIVING-SET [C], probes/P-ENTROPY-BRIDGE-3). At the frozen anchors through depth 12 every word-prefix composition has image 3125 with every fiber exactly two; the backward tree is a width-two caterpillar with one death per level, so backward indeterminacy does not accumulate in the declared living construction (ENTROPY-UNIQUE-PAST [C]). Finally, ``` ord(J mod lambda^i) = (4, 20, 20, 20, 20, 20, 100, 100), i = 1..8, Spec(J on O/lambda^5) = {1: 1, 4: 1, 20: 156}, 3125 = 5^5 = |O/lambda^5|. ``` This is ENTROPY-COUNT-MATCH [C]: the depth-five lambda carrier has exactly the living-set cardinality. Cardinality is not a construction of the cut, does not name a new target space, and does not construct an equivariant bijection. The fourth probe resolves a further finite quotient of that living carrier. On every recurrent component and half, the level-`k` vertex groups for `k = 1..10` are cyclic of order five with one constant partition into five-cells: two cells per size-20 component half and one per singlet half, for `625 = 5^4` cells per full living half. Both one-tick branch maps carry cells to cells, and the induced vertex holonomy on the cell quotient is trivial for `k = 1..8` (ENTROPY-PENTAGON-QUOTIENT [C], probes/P-ENTROPY-BRIDGE-4). In the preregistered coherent level-2 gauge, every level-`k` cell map for `k = 0..10` is affine over `F_5`; all 313 components have the same frozen `(a,b)` spectra, with period four for `k = 1..10` (ENTROPY-AFFINE-COCYCLE [C]). This is gauge-specific and does not identify the multipliers gauge-independently with the lambda-digit action. The same probe finds zero component-local cylinder solutions in exactly the 900 frozen cases: the singlet and canonical size-20 component for `L = 4..16`, three cursor positions and eleven stated clocks, plus the singlet at clock four for `L = 17..30` at the same cursors (ENTROPY-COMPONENT-NOGO [C]). This finite enumeration does not quantify over every component, clock, or window. None of these finite results, including ENTROPY-CYLINDER-NOGO-CURSOR, kills the Route A problem. ENTROPY-LAYER-BRIDGE [O] closes positively exactly when `A_A` is proved nonempty by one exhibited exact map. It closes negatively only by a complete theorem `A_A = empty`. Failure of one proposed construction is STOP, not a negative decision. An exact finite-cylindrical exhibition for any registered `(L,c)` with `4 <= L <= 32` at any finite lambda-depth would refute ENTROPY-CYLINDER-NOGO-CURSOR [T] rather than close this row; inside the older ENTROPY-CYLINDER-CUT scope it would also require correction of that [F] row. The mirror probe resolves the finite mirror law on the same living carrier. Each branch letter restricted to its own living half is an involution with cycle type `{1: 1, 2: 1562}` and a unique fixed state in the singlet component. The cross restrictions are mutually inverse in the exact directions `F_1 o F_0 = id` on `H_1` and `F_0 o F_1 = id` on `H_0`. On canonical pentagon cells, each letter fixes the singlet cell and swaps the two cells of every size-20 component-half. In the frozen coherent level-2 gauge, every one-tick cell map is a reflection with multiplier `4 = -1`; the two ordered letter pairs are `((4,0),(4,2))` and `((4,2),(4,0))`, each on exactly 625 source cells (ENTROPY-MIRROR-LAW [C], probes/P-ENTROPY-MIRROR-1). This is a finite, gauge-specific statement. It supplies no all-scale or measurable mirror theorem, gauge-independent normal, equivariant selection family, measure transport, or L2 lift. Macro space is the coupled kernel. Cells couple on the entanglement axis by the two way CSUM transvections, and the wedge w_ij = x0_i x1_j - x0_j x1_i is the inter cell symplectic area. The structure is exact (KERNEL-WEDGE-AFFINITY [T], KERNEL-WEDGE-COUPLING [T], KERNEL-WEDGE-LINEAR-STRATA [T], KERNEL-WEDGE-AFFINE-MIX [T], reproduce/kernel-connectivity): every generator is affine, g(x) = M_g x + v_g with det M_g = 1, a and b linear, c, d, e strictly affine; the two transvections generate SL2(F_5), order 120, the group of the color door bridge, preserving all 15 wedges; the linear sector acts by congruence W -> M_g W M_g^T and preserves the wedge rank strata, with rank 0 exactly on dependent pairs; only the affine translations cross strata, lifting exactly 62480 of the 78125 dependent pairs each for c, d, and e, and none for a or b. The single cell component census over all seventeen recorded generator subsets is exact (KERNEL-CELL-COMPONENTS [C], reproduce/kernel-connectivity), from {ac} at 945 components down to the full verb {abcde} at 1: one cell is connected by the five letters alone. Connected macro space is the affine translational sector breaking inter cell symplectic parallelism, organized by the SL2(F_5) coupling, with the wedge automorphism a completing transitivity (KERNEL-MACRO-READING [D]); this refines the commutator reading of space. Connectivity of `{a, c, d, e}` on every coupled power `(F_5^6)^k`, `k >= 2`, is exact (KERNEL-CONNECT-ALL-K [T], probes/P-KERNEL-CONNECT-ALL-K-1). Let `Gamma = ` and let `U` be the smallest `Gamma`-invariant subspace of `F_5^6` containing the translations `{v_c, v_d, v_e}`. Exact closure gives `dim U = 6`. For every affine letter `h in {c, d, e}`, the frozen commutator identity is `D_h R_(i-1)^4 D_h R_(i-1) = t_(delta_i tensor v_h)`; conjugation gives `D_g t_(delta_i tensor u) D_g^(-1) = t_(delta_i tensor M_g u)`. Thus the extracted cell factors close to all of `U`, and the two-way ring transvections transport them to every cell, so the translation subgroup is all of `(F_5^6)^k` for every `k >= 2`. The lower bound is sharp: at `k = 1` the same four letters have nine components. This is an L1 state theorem for the declared coupled carrier; it supplies no continuum, measure, or physical lift. TM-SHEET-SYNCHRONIZING-GRAPH [T] is the exact L1 theorem for the sheet pair induced by the declared generators. Direct substitution in the sheet laws above gives the two letter maps on `F_5`, ``` T_0 = (0,4,0,4,4), T_1 = (2,1,1,3,1), ``` a word `w = w_1 ... w_k` acting in temporal order, `T_w = T_(w_k) o ... o T_(w_1)`, with image `R(w) = T_w(F_5)`, and `theta_n = s_2(n) mod 2` driving the Thue-Morse language. The theorem has eight clauses: ``` G1 for every finite binary word w, |R(w)| = 1 iff w contains 011 or 110; the minimal synchronizing words are exactly 011 and 110, and R(11) = {1,3}; G2 the transformation automaton with identity start has exactly 9 states and 8 nonempty-word maps; its support quotient has 7 states and exactly two 2-to-1 identifications, on support {0,4} the maps (0,4,0,4,4) and (0,4,4,4,4), on support {1,2} the maps (2,1,2,1,1) and (2,1,1,1,1); the quotient is well defined on every support class; G3 the Thue-Morse factors of length at most 16 are exactly the factors of mu^4(ab) for ab in {00,01,10,11}, mu(0)=01, mu(1)=10; the factor counts for lengths 1..16 are 2 4 6 10 12 16 20 22 24 28 32 36 40 42 44 46; every factor of length 9 synchronizes; G4 the unique nonsynchronizing factor of length 8 is the palindrome w* = 10100101, the unique 11-free factor of length 8; no 11-free factor of length 9 to 16 exists; for lengths 3 to 16 a factor fails to synchronize iff it contains no 11; the unique two-sided neighborhood is 1 w* 1; G5 T_w*(0) = 2 and T_w*(z) = 1 for z != 0, so the preimage partition is {0} | F_5^x, the zero orientation partition of FIELD-ZERO-NONZERO-MULTIPLICATIVE-CUT, as a partition coincidence only; G6 eps(T_w*(z)) = 1 iff z = 0 in the quadratic-class bit eps(1) = eps(4) = 0, eps(2) = eps(3) = 1; R(01) = R(w*) = {1,2} while T_01 = (2,1,2,1,1) differs from T_w*, the pre-final maps being (0,4,0,4,4) and (0,4,4,4,4) of equal support {0,4}; G7 over the two-sided Thue-Morse subshift the skew product (theta, z) -> (S theta, T_(theta_0) z) has the invariant graph z = 4 + 2 theta_(-1); every length-9 factor composes to the constant map with value 4 + 2 u_last, so the graph is the unique invariant graph, unconditionally; the canonical start 011 reaches it in exactly 3 letters, and 9 is sharp, witnessed by |R(w*)| = 2; G8 (4 + 2 theta) mod 5 = (-(-1)^theta) mod 5, so on the graph z_(n+1) = -Theta_n^2 in the RAMIFIED-TM-LIFT phase; the quadratic-class bit is removed by a 2-to-1 merge at every synchronization edge and the clock bit survives only as the sign inside {1,4}. ``` Proof shape. Either reset word sends each of the seven reachable subset states to a singleton, and a word avoiding both walks the five nonsingleton states of the subset graph and never leaves them; the equivalence was additionally checked over all 32766 binary words of length at most 14. Every Thue-Morse factor of length at most 16 lies in a concatenation of two adjacent mu^4 blocks, and mu^4 of each occurring pair is itself a factor, so the block language is exact. For the graph, any invariant assignment h satisfies h(theta) = T_(theta_(-1)) o ... o T_(theta_(-9)) applied to an arbitrary value, and every such nine-letter composite is the constant 4 + 2 theta_(-1). The eight clauses were verified by `probes/P-TM-SHEET-SYNCHRONIZING-GRAPH-1`: sixteen exact gates, exit zero, and byte-identical aarch64 and x86_64 output against the recorded expected stream. The theorem is L1 only. It assigns no census meaning to the value pair {1,4}, reconstructs no counter, claims no completeness, uniqueness, or totality for any decoder reading, makes no physical-irreversibility, probability, or measure statement, and recovers no four-phase lift value. KERNEL-Z6-SYNCHRONIZATION [T] is lineage: the sheet laws and the table are re-derived above, not imported. The partition coincidence of G5 carries no derivation in either direction. ## 4. The two places The total-ramification census among full quartic cyclotomic fields is exact (QUARTIC-CYCLOTOMIC-TOTAL-RAMIFICATION-CENSUS [T], `probes/P-QUARTIC-CYCLOTOMIC-TOTAL-RAMIFICATION-CENSUS-1`). For `K_n = Q(zeta_n)` one has ```text phi(n) = 4 iff n in {5,8,10,12}, Q(zeta_10) = Q(zeta_5). ``` After quotienting by equality of fields, the class is exactly `{K_5,K_8,K_12}`, with ```text disc(K_5) = 5^3, disc(K_8) = 2^8, disc(K_12) = 2^4 3^2. ``` Using `O_n = Z[zeta_n]`, the exact ramified-prime profiles are ```text 5 O_5 = p_(5,5)^4, p_(5,5)=(1-zeta_5), (e,f,g)=(4,1,1), 2 O_8 = p_(8,2)^4, p_(8,2)=(1-zeta_8), (e,f,g)=(4,1,1), 2 O_12 = P_(12,2)^2, (e,f,g)=(2,2,1), 3 O_12 = P_(12,3)^2, (e,f,g)=(2,2,1). ``` The displayed discriminants exclude every other ramified rational prime. Consequently the total-ramification locus in the frozen class is exactly `{(K_5,5),(K_8,2)}`. A unique prime above a rational prime is not enough: the two `K_12` controls have `g=1` but `e=f=2` and are not total. At the total primes the residue fields and their unit groups are ```text O_5/p_(5,5) = F_5, F_5^x = C_4, O_8/p_(8,2) = F_2, F_2^x = C_1. ``` The inherited axiom element reduces to `J mod p_(5,5)=2`, which has exact order four and generates `F_5^x`. The non-total controls are `O_12/P_(12,2)=F_4` with unit group `C_3` and `O_12/P_(12,3)=F_9` with unit group `C_8`. This is L1 exact arithmetic only. It does not select degree four, prove that full cyclotomic fields exhaust any broader CM or admissible class, classify ramification in all number fields, identify a unique physical field or place, strengthen TWO-PLACE-PHYSICS [D], or define a decoder, Born measure, physical bit, clock, force, observable, or lift to L2-L6. A second exact theorem answers a different minimization question (ABELIAN-CM-UNIQUE-EVEN-BIT-DISCRIMINANT-MINIMUM [T], `probes/P-ABELIAN-CM-UNIQUE-EVEN-BIT-DISCRIMINANT-MINIMUM-1`). For a finite Galois CM extension `K/Q`, write `G_K=Gal(K/Q)` and let `c_K` be its canonical CM complex conjugation. Let `A` be the class, up to `Q`-isomorphism, in which ```text G_K is abelian, |Hom(G_K,C_2)| = 2, including the trivial character, chi(c_K) = +1 for every chi:G_K -> C_2. ``` Then ```text absDisc(K) >= 125 for every K in A, absDisc(K) = 125 iff K is Q-isomorphic to Q(zeta_5). ``` The standard inputs are the central nonidentity CM involution, the finite abelian square-kernel identity, the totally imaginary Minkowski bound, Kronecker--Weber and the abelian Dirichlet character-field correspondence, the conductor-discriminant theorem, primitive quadratic characters and fundamental discriminants, and the cyclotomic discriminant formula. Here is the proof. For a finite abelian group, the intersection of the kernels of all maps to `C_2` is `G_K^2`. The nonidentity involution `c_K` is killed by every quadratic character, hence `c_K=tau^2` for some `tau` of exact order four. Thus `4` divides `[K:Q]`. For a totally imaginary degree-`n` field, Minkowski gives ```text absDisc(K) >= (pi/4)^n n^(2n)/(n!)^2. ``` Using `pi>3`, put `M(n)=(3/4)^n n^(2n)/(n!)^2`. The exact checkpoints are ```text M(8) = 21233664/1225 > 125, M(n+1)/M(n) = (3/4)(1+1/n)^(2n) >= 3. ``` Therefore every smaller or tied competitor has degree four. An abelian group of order four is `C_4` or `C_2 x C_2`; the latter has four maps to `C_2`, including the trivial map. The unique-bit condition forces `G_K=C_4`. By Kronecker--Weber and the abelian character-field correspondence, `K` is cut out by a primitive Dirichlet character `psi` of exact order four. It is odd. Let `epsilon` be the primitive character associated to the pointwise square `psi^2`; it is the unique nontrivial quadratic character and is even. The conductor-discriminant theorem for abelian extensions gives the field-discriminant identity ```text absDisc(K) = f(psi)^2 f(epsilon). ``` There is no primitive order-four character of conductor below five and no nontrivial even primitive quadratic character of conductor below five, so `f(psi)>=5` and `f(epsilon)>=5`. The pure 2-primary branch has the stronger floor ```text f(psi) >= 16, f(epsilon)=8, absDisc(K) >= 16^2 8 = 2048, ``` and therefore no pure 2-primary competitor lies below 125. Hence `absDisc(K)>=5^2 5=125`. Equality forces both conductors to be five. An order-four character modulo five is faithful on `(Z/5Z)^x=C_4`, so its field is the full `Q(zeta_5)`. Conversely `Q(zeta_5)` belongs to `A` and has discriminant `5^3=125`. The two L1 results now supply two independent answers to "why five": ```text ramification answer: among full quartic cyclotomic fields, the complete total-ramification locus is {(K_5,5),(K_8,2)}; minimum answer: in the abelian Galois CM unique-even-bit class A, K_5 is the unique absolute-discriminant minimizer. ``` The answers use different frozen classes and are not a physical selection chain. Total ramification is not a premise of the minimum theorem; the class `A` and discriminant minimization are not claimed to be forced by `J`, the decoder, or Nature. In particular `K_8` has Galois group `C_2 x C_2` and does not satisfy the unique-bit premise. Neither theorem promotes TWO-PLACE-PHYSICS [D], derives a write/read assignment, or supplies a decoder, Born measure, physical bit, clock, force, observable, or lift to L2-L6. ### The alternating trace-form pencil Put `K = Q(j)`, `O_K = Z[j]`, `K+ = Q(sqrt5)`, and `O_K+ = Z[phi]`. Define ```text lambda_1 = j - j^-1, lambda_2 = j^2 - j^-2, L = {lam in O_K : conj(lam) = -lam}. ``` In the basis `1,j,j^2,j^3`, `conj(q,r,s,t) = (q-r,-r,t-r,s-r)`. Hence `conj(lam) = -lam` is equivalent to `r = 2q` and `s+t = 2q`. Writing `v=s-q` gives ```text (q,2q,q+v,q-v) = q lambda_1 + v lambda_2. ``` Direct reduction gives `lambda_1 phi = lambda_1 + lambda_2`. Since `phi^-1 = phi-1` and `Z[phi] = Z + Z phi^-1`, this proves ```text L = Z lambda_1 + Z lambda_2 = lambda_1 Z[phi], lambda_2 = lambda_1 phi^-1. ``` For `lam` in `L`, set ```text Omega_lam(h,k) = Tr(lam h conj(k))/5. ``` Trace invariance under conjugation gives `Omega_lam(k,h) = -Omega_lam(h,k)` and `Omega_lam(h,h) = 0`. The different and codifferent are ```text D_K = (5/(1-j)), D_K^-1 = ((1-j)/5). ``` The identities ```text N(lambda_1) = N(1-j) = 5, lambda_1 = j(1-j)(1+j+j^2) ``` show `(lambda_1) = (1-j)`. The trace-dual criterion therefore makes every `Omega_lam` integer-valued and makes it unimodular exactly when `(lam) = (lambda_1)`. More explicitly, if `lam = lambda_1 eta` with `eta` in `Z[phi]`, multiplication by `eta` in the first trace argument gives ```text det(Omega_lam) = N_(K/Q)(eta) det(Omega_1) = N_(K+/Q)(eta)^2 det(Omega_1). ``` The direct Gram matrix below has Pfaffian one, so `det(Omega_1)=1`. Hence `det(Omega_lam)=N_(K+/Q)(eta)^2`, and the criterion follows. For `Omega_(a,b) = a Omega_1 + b Omega_2`, exact trace reduction gives ```text Omega_1 = [[ 0, 1, 0, 0], [-1, 0, 1, 0], [ 0,-1, 0, 1], [ 0, 0,-1, 0]], Omega_2 = [[ 0, 0, 1,-1], [ 0, 0, 0, 1], [-1, 0, 0, 0], [ 1,-1, 0, 0]]. ``` The entries are linear in `a,b`, so the Pfaffian is a binary quadratic form. Its exact values at `(1,0)`, `(0,1)`, and `(1,1)` are `1`, `-1`, and `-1`. These three values determine all three coefficients and prove ```text Pf(Omega_(a,b)) = a^2-a b-b^2 = N_(K+/Q)((a-b)+b phi). ``` Since `a lambda_1+b lambda_2 = lambda_1((a-b)+b phi)`, the unimodular pencil members are exactly the parameters of norm `+1` or `-1`. To classify them, take a unit of `Z[phi]`, change its sign, and multiply by a power of `phi` so that its positive real value `delta` satisfies `1<=delta x . ``` The unimodular locus is therefore the Pell unit orbit. Induction from `phi^2=phi+1` gives `phi^n=F_n phi+F_(n-1)` for `n>=1`, and hence ```text Pf(Omega_(F_(n+1),F_n)) = (-1)^n. ``` For every `u` in `O_K^x`, commutativity gives ```text Omega_lam(u h,u k) = Omega_(lam u conj(u))(h,k). ``` Thus the unit action on the pencil factors through `N_(K/K+)(u) = u conj(u)`. The unit ranks of `K` and `K+` are both one, while the relative norm sends each real-subfield unit to its square. Its kernel therefore has rank zero and consists of roots of unity. The roots in `K` are exactly ```text mu_10 = {+j^k,-j^k : 0<=k<5}, ``` and all ten have relative norm one, so this is exactly the kernel. Since `J conj(J) = 2-phi = phi^-2`, its pullback action on the parameter lattice has the coordinate columns ```text lambda_1 phi^-2 = lambda_1-lambda_2, lambda_2 phi^-2 = -lambda_1+2 lambda_2. ``` Thus, in the ordered basis `lambda_1,lambda_2`, its matrix is ```text A_J = [[1,-1],[-1,2]]. ``` It has determinant `1`, trace `3`, characteristic polynomial `t^2-3t+1`, and eigenvalues `phi^2` and `phi^-2`. Parameter multiplication by `phi` has matrix ```text B_phi = [[1,1],[1,0]], det(B_phi) = -1, ``` and `A_J = B_phi^-2`. Finally, for every four by four alternating matrix `W`, ```text Pf(M^T W M) = det(M) Pf(W). ``` If `M` is integral with determinant `1`, `W` is a nonzero pencil member, and `M^T W M = mu W`, cancellation gives `mu^2 = 1`. Hence `mu` is `+1` or `-1`. The parameter map is injective because a zero trace form would make `Tr(lam h)=0` for every `h` in `O_K`, and nondegeneracy of the trace pairing would force `lam=0`. A scalar multiplier from unit multiplication must therefore equal `u conj(u)`. This relative norm is positive at both real embeddings, so unit multiplication realizes only `mu=1`, exactly for the ten roots of unity. Multiplication by `J` instead moves the pencil by `A_J`; it does not scale a fixed member by `phi^-2`. Conjugation is not multiplication by a unit and satisfies ```text C^T Omega_lam C = -Omega_lam. ``` These statements form CM-ALTERNATING-PENCIL [T] at L1, evidenced by `probes/P-CM-ALTERNATING-PENCIL-1`. Separately, the exact arithmetic uses two places, disjoint over Q: Q(zeta_5) cap Q(zeta_8) = Q. Their physical assignment is the TWO-PLACE-PHYSICS dictionary [D]: ``` v_5 WRITES: home Q(zeta_5), real floor Q(sqrt5), state F_5^6; geometry, gravity, the native pentit substrate, the native magic C_5 (order 5, prime 5). v_2 READS the quadratic mode: home Q(zeta_8), real floor Q(sqrt2), with i and the foreign qubit magic m_8 = zeta_8 (order 8, prime 2); Clifford, Born, the outside read. They meet only over Q. The dictionary reads the layer boundary as the field boundary; it does not prove that this reading is unique. ``` The exact field facts are DEGREES-BY-PRIME [T] (reproduce/foundations-places): sqrt5 lives at zeta_5 by (2 phi - 1)^2 = 5, while sqrt2 and i live at zeta_8 by (m_8 + m_8^-1)^2 = 2 and m_8^2 = i, and neither sqrt2 nor i lies in Q(sqrt5), the unique quadratic subfield of Q(zeta_5). The linear, quadratic, cubic, and foreign-magic assignments are readings of these facts under TWO-PLACE-PHYSICS [D]. The Z2 symmetries split arithmetically (Z2-PLACES-SPLIT [T], reproduce/foundations-places): there is one involution at zeta_5, the single order 2 element of its cyclic Galois group, and a complete Klein four at zeta_8, where every nontrivial element is an involution; 5 = (2 + i)(2 - i) in Z[i], with conjugation swapping the factors. TWO-PLACE-PHYSICS [D] separately reads CP from the zeta_5/Gaussian conjugation and T from the Thue-Morse reversal, hence CPT as their composition; this is a dictionary, not a theorem identifying the physical CPT operator. The same dictionary assigns forces to sqrt5, spin to sqrt2, and charge conjugation to i. i is bilocated (I-BILOCATED [D], reproduce/foundations-places): the order 4 element of F_5* at v_5 and zeta_8^2 at v_2, identified only over Q, never merged. TWO-PLACE-PHYSICS also declares the Born square as the descent reading for cross-place transfers; its values are carried by the Born rows, and no completeness claim is made. The silver ring facts (SILVER-RING-FACTS [C], reproduce/foundations-places): inside F_25 = F_5(sqrt 2), where the step collapses to the doubling J = 2, tau = sqrt(J) has tau^4 = -1 with ord(tau) = 8, and F_25* is cyclic of order 24; norm, orders, cyclicity, and census are finite computations. The silver sibling (SILVER-SIBLING [D], reproduce/foundations-places) is the dictionary reading resting on those facts: m_8 = zeta_8 = sqrt(i) at prime 2 with the silver unit 1 + sqrt2 of norm -1 mirrors tau = sqrt(J) at prime 5; sqrt(i) is the square root of the axiom read at the foreign place. The central phase of the quadratic lift is exact (CENTRAL-LIFT-PHASE [T], `probes/P-CENTRAL-LIFT-PHASE-1`). In `K = Q(zeta_5)` with `O_K = Z[zeta_5]`, principal `zeta_5 = exp(2 pi i/5)`, and ```text phi = -(zeta_5^2 + zeta_5^3), J = phi^-1 zeta_5, zeta_10 = -zeta_5^3, s = zeta_10 / sqrt(phi), g_J = diag(s,s^-1), A_J = diag(J,1) = s g_J, ``` define, for `A in GL_2(C)`, ```text H_A(X) = A X A^dagger / |det A|, X in Herm_2(C), S_A(Y) = A Y A^T / |det A|, Y in Sym_2(C). ``` The principal square root obeys ```text s^2 = J, s^5 = -phi^(-5/2), g_J^5 = -diag(phi^(-5/2),phi^(5/2)), g_J^10 = diag(phi^-5,phi^5). ``` Thus the fifth power retains the central minus sign on the spinor but induces the displayed pure boost on `Herm_2(C)`, where `-I` acts trivially. This is a projective action statement, not a physical-clock or tick identification. For every nonzero complex scalar `c`, the normalized actions separate as ```text H_(cA) = H_A, S_(cA) = (c^2/|c|^2) S_A. ``` Writing `X = [[u,w],[conjugate(w),v]]`, the square-root-free J action and its fifth power are ```text H_(A_J): (u,v,w) -> (phi^-1 u, phi v, zeta_5 w), H_(A_J)^5: (u,v,w) -> (phi^-5 u, phi^5 v, w). ``` With `B = diag(J,J^-1)`, one has `A_J^2 = J B`, hence `H_(A_J^2) = H_B`, while `S_(A_J^2) = zeta_5^2 S_B`. The Hermitian slot forgets the scalar and the symmetric slot retains its central phase. Finally the unit-scalar phase map ```text O_K^x -> C^x, c -> c/conjugate(c) ``` has image exactly `mu_5`. Kronecker's unit-circle lemma places every image in the roots of unity of `K`, reduction modulo `1-zeta_5` removes the negative `mu_10` coset, and `c = zeta_5^a` attains all of `mu_5`. The inherited identity `1-J = -zeta_5^2` lies in `mu_10 \ mu_5`, so it is not the normalized symmetric-square phase of an `O_K` unit scalar. These are L4 quadratic-support theorems only. They define no positive, Born, or causal cone, boundary, split-unit projector, rigidity or common carrier, integral tick, physical time, bit, U(1), electromagnetic channel, decoder `Q` or `QCarrier`, `MatterData`, L5 stream, L6 measure, or cross-layer lift. In particular QUADRATIC-DECODER-DATA [O] remains STOP and unchanged. The golden and silver numbers are the positive roots of the two simplest metallic laws `x^2 = t x + 1`, at `t = 1` and `t = 2` (METAL-TRACE-CASCADE [T], `probes/P-METAL-TRACE-1`). Both are units of norm `-1`. The discriminant `t^2+4` gives the fundamental discriminants 5 and 8, so the ramified prime of each law divides its own discriminant. Their powers obey their native integer recurrences: ``` phi^n = F_n phi + F_(n-1), delta^n = P_n delta + P_(n-1), ``` where `F` is Fibonacci and `P_(n+1)=2P_n+P_(n-1)` is Pell. At the silver place, `1+m_8^2 = 1+i = sqrt2 m_8 = sqrt(2i)`, so `(1+i)^2=2i`, while its absolute field norm in `Q(zeta_8)` is 4; it is not a unit. At the golden place `N(J)=1`. The claim is scoped to `t in {1,2}`: trace does not select a number field for general `t` because `t=4` also has squarefree discriminant kernel 5. No run/record, time-arrow, force, or other physical reading is included in this theorem. ## 5. The force is the curvature The finite Weyl commutator is exact on all five basis states (FORCE-WEYL-HOLONOMY [T], reproduce/force-born-dictionary): ``` Z X Z^-1 X^-1 = j I; j has exact order 5 and arg j = 2 pi / 5. ``` FORCE-AS-CURVATURE [D] reads this holonomy as a force curvature and, through AXIOM-PROJECTION-DICTIONARY [D], reads the two J projections as the two abelian force channels. This physical assignment is not part of FORCE-WEYL-HOLONOMY [T] and no uniqueness of the force dictionary is claimed. The gravity dictionary currently places three labeled ingredients side by side: the cell action 864 pi at k = d(d + 1) = 12, the quadrupole-power comparison against the Hulse-Taylor binary pulsar (99.83 +/- 0.16 percent [measured comparison], source SRC-PSR-B1913), and the declared kernel coefficient A_GD = 1/(8 pi) with lambda = 216 pi. Their conjunction is a GRAVITY-BRIDGE-LAW reading [D], not an additional theorem and not a public value of G. Its exact displayed identity is: ``` G_nat = d^3 = 27 = 864 pi x 2 / (4 pi x 16). ``` The Coulomb layer separates the finite computation from the continuum reading. On the finite decoder graph C4, the Moore-Penrose Green kernel is ``` 16 G = circ(5, -1, -3, -1), L G = I - (1/4) 1 1^T, ``` with zero row sum (COULOMB-GREEN-COMPUTATION [C], reproduce/force-born-dictionary). The continuum Green law 1/(4 pi r) is the same-propagator dictionary: gravity reads the modulus and Coulomb the argument (COULOMB-PROJECTION [D]); it is a scaling reading, not a value asserted on finite C4. The sign structure is the polar dictionary (FORCE-POLAR-SIGN [D]): mass is a modulus, one sign, universal attraction; charge is an argument, two signs. The exact chain rows and the classical dictionary are separate. The chain layer is at T and MAXWELL-CLOSED is its D reading (reproduce/maxwell): the Bianchi identity holds identically in the 32 edge symbols and gauge invariance is an identity (MAXWELL-BIANCHI); Gauss is the boundary equation on the closed spatial torus, with a constructive dipole (MAXWELL-GAUSS-CHAIN); the inhomogeneous pair closes with conservation an identity in the 96 face symbols (MAXWELL-AMPERE-CHAIN). The obstruction pair is counted in p (MAXWELL-OBSTRUCTION-P): Gauss is solvable iff the total charge vanishes mod 5; the current pair iff the current is conserved and all four winding numbers vanish mod 5. The abelian face dictionary (ABELIAN-FACE-DICTIONARY [D], reproduce/force-born-dictionary) reads the electric half as P = J/2 on the 12 electric faces from the bare tick. At public v1 scope, the magnetic axiom pair is an explicit input to the dictionary; no uniqueness or selection theorem is claimed. Given that input, stay or twist once reads ``` Psi(k) = 1 + zeta^k, mu(k) = |Psi(k)|^2 / 10 = w(k)/10 = (2/5, (3+sqrt5)/20, (3-sqrt5)/20, (3-sqrt5)/20, (3+sqrt5)/20) ``` entrywise exact in Q(sqrt5) (BORN-FACE-WEIGHTS [T], reproduce/born-faces); the slot 4 pair lands on the sigma_2 Galois image exactly; the tilt magnitude is exactly sqrt5/5, with sign set by the Thue-Morse slot orientation; the Legendre symbol (2|5) = -1 is realized on this face. Coupling seeds: the declared decomposition has exact Gram weights 1/p on the trace and conformal directions and 1 on the spatial base, with 3/4 = d/(d + 1) (MEASURE-SPATIAL-ONLY [T]). STRONG-SEED [D] reads these weights as coupling roots on the spatial gauge sector: alpha* = 1/p and the strong root is 1 x 3/4 x 1 = 3/4; the seed ratio EM to strong is 15 : 4. Dark energy w = -14/15. Running and scheme are frontier rows (ALPHA-S-RUNNING, SCHEME-DICTIONARY). ## 6. Alpha and the observable register ``` alpha = 5 S / ((8 pi)^2 sqrt(s)), sqrt(s) = (3 - phi)^(1/4), S = (1 + X/5)^-5, X = 1 / (32 pi^2 phi^4) alpha^-1 = 137.035999190; CODATA 2022: 137.035999177(21) source SRC-CODATA-2022-ALPHA [measured comparison] ``` (ALPHA-FORM at D; the enclosure witness ALPHA-VALUE-DIGITS at C; reproduce/alpha-value. The digit string and the CODATA window are labeled witnesses of the committed form; no value is claimed beyond it.) The exact lemma (ALPHA-SEED [T], reproduce/alpha-exact-lemma): the dimensionless trace target is exact, alpha* = 1/p, via the cyclotomic Galois-trace Gram G = p I - 1 1^T with normalized spectrum {1/p (once), 1 (p - 2 times)}, the all ones trace direction the eigenvector; exact for p in {3, 5, 7, 11, 13}. Scope: alpha* = 1/p is the dimensionless seed; the physical alpha^-1 lives at the bridge level. Prefactor unification (ALPHA-PREFACTOR-UNIFICATION) [T]: with B_{2,chi5} = 4/5 exactly, the Gauss sum tau = 2 phi - 1 = sqrt5 exactly in Z[zeta_5], and L(2, chi5) = 4 pi^2 / (25 sqrt5), the witness formula and the formula above are one formula; no independent L value enters the alpha sector; the prefactor reads (1 + J Jbar)^(1/4): the gravity modulus inside the electromagnetic number. Weinberg (WEINBERG-FORM at D on the exact layer WEINBERG-TREE and HYPERCHARGE-LAW at T; reproduce/weinberg): sin^2 theta_W = 3/13 + 1/(32 pi^2 phi^4), the committed form with its comparison fenced; the tree value 3/13 = deg_f / (V + 1) with deg_f = Tr(J) = 3 and V = 12 = d(d + 1). All seven standard model hypercharges follow from Y_F = (B - L) + 2 T_3^R; B_quark = 1/3 = 1/dim ker(Tr), the trace kernel of dimension 3. alpha is built from L and X, both pi even, hence delta free (section 7). ## 7. The mass ladder and the parity law With the single SI anchor m_e (the committed forms at MASS-LADDER-FORMS, D; the theorem layer at T; reproduce/mass-ladder): ``` C_mu_tau = 89/5 = 18 - 1/p (MU-TAU-COEFFICIENT) [T] mu_mu = 2688/13 - (89/5) alpha^2 (-0.023 sigma) [measured comparison] source SRC-CODATA-2018-MUON mu_tau = 3477 + 240 (89/5) alpha^2 (-0.011 sigma) [measured comparison] source SRC-CODATA-2018-TAU exchange identity: delta mu_tau + 240 delta mu_mu = 0 [T, exact in Q] (MU-EXCHANGE-IDENTITY) mu_p = 6 pi^5 (1 + alpha^2 / 3) the single bare pi^5, an odd carrier mu_n = mu_p + deg_v / chi - Delta_EM; a numerical comparison is not retained until deg_v / chi and Delta_EM are public; Delta_EM stays open (NEUTRON-DELTA-EM) the electron at the Dirac step: det = 1 + m_D^2 = 5 = p ``` PROTON-RESIDUAL-IS-QCD [O] is only a typed future residual between the registered formal proton expression and a future QCD dynamics output. It contains no comparison with a measured proton moment. Once the QCD carrier and action, output carrier, residual codomain, normalization, equality, and inference schema are frozen, let `Der_QCD` be the class of all total exact inference maps satisfying those data and the registered residual equation. The row closes positively exactly when `Der_QCD` is proved nonempty by an exhibited derivation. It closes negatively only by an exact theorem `Der_QCD = empty`, equivalently by proving that every admissible derivation violates a named frozen law. Failure of one proposed derivation is STOP, not negative. The row is STOP while any part of the schema is incomplete. Empirical comparison would require a separately registered source, comparison window, and inference rule. The parity law (PARITY-LAW) [T] lives in a formal observable register, not under ordinary complex conjugation. Let ```text R = A[pi, pi^-1], iota_pi(pi) = -pi, iota_pi(a) = a for every a in A, ``` where `A` contains the pi-free coefficients and the other named formal generators. Multiplicative extension makes `iota_pi` an involutive algebra automorphism, and a monomial `c pi^k u`, with `u` in `A`, has eigenvalue `(-1)^k`. This is a formal grading operation; it is not complex conjugation on the complex numbers, which fixes the real number pi. At this public scope, thirteen named entries are even and delta free: alpha^-1, sin^2 theta_W, w = -14/15, L, Omega_b = pi^2/200, the slip X, mu_mu, mu_tau, G, PMNS, the dark matter ratio, the zeta_K residue, and the neutrino register 341/10. Three odd delta carriers sit at degrees pi^1, pi^3, and pi^5: the capacity 2 pi/phi^2, the Kahler capacity 64 pi^3 phi^2, and the proton 6 pi^5. The neutron is the unique mixed composite among these named forms. No larger parity census is claimed. The bridge defect delta = 6 phi^2 - 5 pi = (9 + 3 sqrt5) - 5 pi is nonzero by Lindemann-Weierstrass (about +2.4 x 10^-4, a labeled gap witness) (BRIDGE-DEFECT). The exact bridges are xi phi^2 = 5 = p, script-Q phi^2 = 2 pi, and script-Q / xi = 2 pi/5 = arg J. The PMNS sector waits on the mass mechanism frontier. ## 8. The measure and Born The Born quartet unpacks into four registered results, reproduce/born-quartet: BORN-HALF-ANGLE [T], BORN-RESIDUAL-SPLIT [T], SPIN-BISECTOR [T], and BORN-ORDER-STAIRCASE [T]. The Born unit group of the read place residual algebra A_8 = Z[zeta_8]/5 is cyclic of order 24, and the unit square root of a phase is the Born normalized bisector: (1 + u)^2 = u N_B(1 + u) on every Born unit; the bisector normalizes to a unit square root exactly on the 11 non antipodal units of the norm gated half, the squares of the unit group (at u = -1 the bisector vanishes, so normalization is undefined there); zeta_8 itself is ungated, so its halving forces the quadratic step (BORN-HALF-ANGLE). The Born faithful residuals are A_4 = Z[i]/5 and A_8 = Z[zeta_8]/5, split rings whose two factors are swapped by conjugation, the Born involution (BORN-RESIDUAL-SPLIT). The quarter turn spinor is R = (1 - B)/sqrt2 with det R = 1, carried exactly as det(1 - B) = 2 and (1 - B)^2 = -2 B over Z and as an order 8 element of SL_2(F_25), the finite shadow of the double cover (SPIN-BISECTOR). Each Born halving is one quadratic step: the staircase of root orders 4 -> 8 -> 16 is first realized at F_5 -> F_25 -> F_625, with minimality against every smaller degree (16 divides none of 4, 24, 124), gated by the square condition on the bisector norm (BORN-ORDER-STAIRCASE). No positivity is claimed in the finite algebra. Probability and measurement language belongs to the MEASURE-BORN-VERB dictionary [D], not to any of the four theorem rows. The measure dictionary reads the Born square of the verb (MEASURE-BORN-VERB [D], reproduce/force-born-dictionary): w(k) = |1 + zeta^k|^2, with its exact identities carried by BORN-FACE-WEIGHTS [T] (reproduce/born-faces). The identikit: eight preregistered strikes died first class and carved the survivor clause by clause: integer amplitudes, a quadratic Born reading, Galois breaking, irrational in position; counts may enter only as amplitudes. The kernel to cell dictionary (KERNEL-CELL-DICTIONARY [D]): time is the clock tick; space is the three F_5 directions of the trace kernel, isotropic under the Galois Gram; the fiber is the Z_5 edge variable with F = dA; deposits are additive, valued in fifths, flux quantum zeta_5; amplitudes are unimodular; the measure is the Born square. The substrate knit (SUBSTRATE-KNIT [T], reproduce/force-born-dictionary) is the exact matrix statement: C_+ = I + S and C_- = C_+^T obey C_+ C_+^T = circ(2, 1, 0, 0, 1) with spectrum w; the position and Fourier bases have squared overlap 1/5 on all 25 pairs; the Plancherel masses are 2 and 10, in the ratio p = 5. Reading C_+ and C_- as two abelian faces is part of KERNEL-CELL-DICTIONARY [D]. ## 9. The photon and the electron The quantum is an integer path count, one bit per tick, on an RP Hilbert space. The photon window has exact coordinates (PHOTON-WINDOW-COORDINATES [T], reproduce/photon-electron): the point w = (4, phi^2, phi^-2, phi^-2, phi^2) and its Kramers-Wannier dual w_hat = 5 (2, 1, 0, 0, 1) with w_check = w exactly, w not self dual, and the tilt w(1) - w(2) = sqrt5. The quantum costs one bit per tick on every center, w_hat(1)/w_hat(0) = 1/2, and every class |k| >= 2 is exactly closed for p >= 5 (PHOTON-UNIVERSAL-BIT [T]). Monopole charge exists only in lumps of five (MONOPOLE-FIFTHS [T]); the elementary monopole costs between 17 and 21 occupied faces (MONOPOLE-COST [C]). The frozen photon-window proof route used the two-integer gap 32 < 2401 < 131072: straight runs cost 5 bits per segment, the 1 x K ladder bound is exactly 9K + 8, and the greedy incidence identity holds (KAPPA-BOUNDS [T]); the nine-shape library realizes the exact bound table with minimum 31/8 bits per segment and the integer margin 2^31 > 7^8 (KAPPA-SHAPES [C]). Its remaining universal occupancy lemma now closes negatively (PHOTON-KAPPA-LEMMA [F], probes/P-PHOTON-KAPPA-LEMMA-1): on the frozen finite-support ternary L4 carrier, the admitted connected current j_* has L = 3240 and an exhibited ternary filling n_* with partial n_* = 5j_*, |supp(n_*)| = 7993, and 2^7993 <= 7^3240. Thus F_occ(j_*) <= 7993; equality and optimality are not claimed. The exact exclusion lemma rules out every coprime positive a,b with 2^(4a) > 2401^b from satisfying b F_occ(j) >= a L(j) for every j in ParentWorldline. Because this occupancy bound was a necessary conjunct, PHOTON-WINDOW-PROOF [F] is falsified as the frozen compound route. The independent electric-face roughening question remains undecided. No Froehlich-Spencer import, massless Coulomb phase, continuum limit, photon propagator, or physical-photon conclusion follows from this result. The center split: conjugation swaps the residue channels exactly for p = 3 mod 4 and preserves them for p = 1 mod 4, with Gauss sums g^2 = -3, +5, -7 (CENTER-SPLIT-RECIPROCITY [T]); the same verb freezes p = 2 and leaves p = 3 with full dual support (CENTER-SPLIT-CLOSURE [T]); with the declared self duality import (the 4D Z_N window opens only for N >= 5), the first prime passing both doors is p = 5 (CENTER-SPLIT-SELECTION [D]). The photon belongs to the null light branch, E+ E- = 0 identically. The electron: g = 2 = (2 pi/5) / (pi/5) is the vertex flux over the spinor half angle, the reading ELECTRON-G-TREE [D] on exact pillars: sixteen identities with the pivot 1 - J = e^(-i pi/5) and the polylog ladder halving 2 pi/5 -> pi/5 -> pi/10 (ELECTRON-G-RATIO [T], reproduce/photon-electron); one drive arg j, spinor pi/5 per step through the diagonal bridge, ratio 2 at every step, closure pair (5, 10), R^5 = -I (ELECTRON-G-DOUBLE-COVER [T]). The scalar nominated by the still-open first-order question is already exact arithmetic. Write u = J Jbar and q = script-Q. J-MODULUS-CHORD [T] gives u phi^2 = 1, while BRIDGE-DEFECT [T] gives q phi^2 = 2 pi. Hence q - 2 pi u = u(q phi^2 - 2 pi) - q(u phi^2 - 1) = 0, and therefore J Jbar / script-Q = 1/(2 pi) (QUANT-SCHWINGER-TARGET [T]). This theorem fixes only the target scalar. It does not identify that scalar with [alpha^1]((g_e(alpha)-2)/2), construct a substrate coupling, fix an action normalization or regularization, or close QUANT-SUBSTRATE [O]. The electron sign is measured into existence (ELECTRON-SIGN [D]): the charge sign is the orientation, ghost {a, c} against live {b, d}, of the single gyron anchored engagement of a defect worldline, pair neutrality is realized through the annihilation sink, and the sign Z_2 equals the initial datum [z_6(B_0) = 4]. Seven exhaustive public laws carry it (ELECTRON-SIGN-LAWS [T], reproduce/photon-electron): the count parity as initial datum; the forced event slot with the phase locked forever after; the annihilation census decided by tick 3; the eps ledger summing to zero; one Z_2 across the five readouts; the closed 20 state pair cycles with equal stable supports; and the transient shift and Galois images. The Dirac ladder closes [DIRAC-LADDER at D, the theorem layer at T; reproduce/dirac-ladder]: ``` G1 u = 2 phi^n, v = 2 phi^-n, u v = 4 the light cone [T] (LADDER-LIGHTCONE) G2 N(phi) = -1 the spinor floor [T] (SPINOR-FLOOR) G3 the ladder root is the cat map carrier: F^2 = C over Z; mod 5 one Jordan block; towers of order 20 = 4p and 10 = 2p, -I at half [D, T] (FIB-ROOT-CARRIER at D on the exact ties FIB-ROOT-TIES at T) G4 D_J(m_D) = S (I + i m_D X), zero free parameters: streaming is the counter, the coin modulus is the rest rung, the phase i is the plenum ellipticity; massless is the one sided shift; det = 1 + m_D^2 = 5 for the electron; the mass shell is exact; the rest coin has infinite order [D, T] (DIRAC-STEP at D on the theorem layer DIRAC-STEP-THEOREMS at T) G5 the ladder Z_2 (conductor 5) and the spin Z_2 (conductor 8) are distinct place attached Galois involutions (LADDER-SPIN-PLACES) [T] ``` The checkerboard between the rungs is a Gaussian tower [D at the reading, within DIRAC-LADDER; T at the tower, CHECKERBOARD-GAUSS-TOWER]: one binomial pair per case; the diagonal lives in Z with pair weight -4, the cross in iZ; the totals are (1 + 2i)^n with recursion x^2 - 2x + 5 and c^2 + d^2 = 5^n; zone edges -5I for the electron and -I for the photon. No eta identity is inferred from this tower; the inherited naming clause is not part of Public Canon v47. The fermionizer Phi_f(s) = 1 - 2^(1-s) (FERMIONIZER) [T]: the two that makes matter out of light. One beat is one boost times one alternator tick (LADDER-ALTERNATOR-BASIS) [T]; the alternator is breath at one scale and Thue-Morse at every scale (TM-BREATH-TOWER) [T]. ### MOBIUS-TM-PRIME2-BRIDGE [T] This subsection is L1 arithmetic only. To avoid a type collision with the global field element `tau = sqrt(J)`, write ```text tau_TM(n) = (-1)^s_2(n), n >= 0, c_TM = mu * tau_TM ``` for the function denoted `tau` in the public probe and its Dirichlet convolution with the Moebius function. Arithmetic functions below have domain the positive integers; `1` is the constant-one arithmetic function and ```text (f*g)(n) = sum_(d|n) f(d)g(n/d), D_p f(n) = f(pn). ``` First, for every prime `p`, every complex-valued arithmetic function `f`, and `g = mu*f`, ```text D_p f = f ``` on all positive integers if and only if `g(n)=0` whenever `p|n`. For the forward direction write `n=p^a r`, with `a>=1` and `p` not dividing `r`. Only divisors with zero or one factor `p` survive under `mu`, so ```text g(p^a r) = sum_(e|r) mu(e) [f(p^a r/e)-f(p^(a-1)r/e)] = 0. ``` Conversely, Moebius inversion gives `f=1*g`. If `g` vanishes on multiples of `p`, then ```text f(pn) = sum_(d|pn, p not dividing d) g(d) = sum_(d|n, p not dividing d) g(d) = f(n). ``` The binary recurrences ```text tau_TM(2n) = tau_TM(n), tau_TM(2n+1) = -tau_TM(n) ``` therefore imply, by the preceding equivalence at `p=2`, ```text c_TM(2n) = 0, n >= 1. ``` Since `tau_TM=1*c_TM`, ```text tau_TM(n) = sum_(d|n) c_TM(d) = sum_(d|n, d odd) c_TM(d). ``` Thus `c_TM` is the unique arithmetic function supported on the odd positive integers whose divisor sum is `tau_TM`. The identity `mu(2x)=-mu(x)` used in the paired proof is restricted to odd `x`; its unrestricted form is false already at `x=2`. For odd squarefree `n=product_(p in P)p`, with `I` the identity operator, the surviving divisor cube gives ```text c_TM(n) = product_(p|n)(D_p-I)tau_TM(1) = sum_(S subseteq P) (-1)^(|P|-|S|) tau_TM(product_(p in S)p). ``` This is the top Boolean mixed difference of binary digit parity over the prime-divisor cube. It is not a multiplicativity statement: exactly `c_TM(3)=2`, `c_TM(5)=2`, and `c_TM(15)=-2`. For every odd prime `p` and `k>=1`, ```text c_TM(p) = tau_TM(p)+1 in {0,2}, c_TM(p^k) = tau_TM(p^k)-tau_TM(p^(k-1)) in {-2,0,2}. ``` For every positive integer `n`, the odd-step recurrence also gives the exact divisor recursion ```text sum_(d|2n+1)c_TM(d) = -sum_(d|n)c_TM(d). ``` For `|x|<1`, put `F_TM(x)=sum_(n>=0)tau_TM(n)x^n`. Splitting even and odd indices gives `F_TM(x)=(1-x)F_TM(x^2)`, and iteration with absolute convergence gives ```text F_TM(x) = product_(j>=0)(1-x^(2^j)). ``` The bound of `|c_TM(m)|` by the number of positive divisors of `m` makes the Lambert rearrangement absolutely convergent for `|x|<1`. The odd-supported reconstruction then gives the bridge ```text product_(j>=0)(1-x^(2^j))-1 = sum_(d>=1, d odd)c_TM(d)x^d/(1-x^d), |x|<1. ``` Finally, for `Re(s)>1`, define ```text T_TM(s) = sum_(n>=1) tau_TM(n)n^(-s), C_TM(s) = sum_(n>=1) c_TM(n)n^(-s), T_TM,odd(s) = sum_(n odd) tau_TM(n)n^(-s), zeta_odd(s) = (1-2^(-s))zeta(s). ``` The bound `|c_TM(n)|<=sum_(d|n)|mu(d)|<=d(n)` makes the series and their rearrangements absolutely convergent in this half-plane. Moebius inversion and the prime-2 recurrence give ```text T_TM(s) = zeta(s)C_TM(s), T_TM(s) = T_TM,odd(s)/(1-2^(-s)), zeta_odd(s)C_TM(s) = T_TM,odd(s), C_TM(s) = T_TM,odd(s)/zeta_odd(s). ``` The last division is valid because the absolutely convergent Euler product over odd primes makes `zeta_odd(s)` nonzero for `Re(s)>1`. These six clauses form MOBIUS-TM-PRIME2-BRIDGE [T], evidenced by `probes/P-MOBIUS-TM-PRIME2-1`. The written proofs supply the universal quantifiers; the two-architecture verifier is an audit of finite instances and coefficient identities. Its general parity and size checks are guards, not additional theorem clauses. At the same argument, the cancelled factor `1-2^(-s)` is not the neighboring FERMIONIZER factor `Phi_f(s)=1-2^(1-s)`. The trivial shifted identity `1-2^(-s)=Phi_f(s+1)` is unused and creates no dependency on or physical identification with FERMIONIZER, LADDER-ALTERNATOR-BASIS, or TM-BREATH-TOWER, and supplies no matter or light reading. No RH or zero-location result, Nyman--Beurling or Baez--Duarte completeness, analytic continuation outside the displayed domains, `J` coupling, `p=5` selection, decoder, Born, observer, force, spacetime, SI, physical vacuum or other physical interpretation, pointwise or averaged/Cesaro Moebius--Thue--Morse orthogonality, Sarnak-type correlation, asymptotic cancellation such as `sum_(n<=N)mu(n)tau_TM(n)=o(N)`, multiplicativity of `c_TM`, or lift to L2--L6 is asserted. ### TM-MULTIPLICATION-CARRY-DEFECT [T] This is a second, logically standalone L1 theorem. It reuses the locally typed arithmetic functions `tau_TM(n)=(-1)^s_2(n)` and `c_TM=mu*tau_TM`, but does not import MOBIUS-TM-PRIME2-BRIDGE as a premise. To distinguish ordinary integer-multiplication carries from every other Canon use of `kappa`, write `kappa_(2,mul)` for the quantity called `kappa_2` in the public probe. For positive integers ```text a = sum_i a_i 2^i, b = sum_j b_j 2^j, a_i,b_j in {0,1}, ``` ordinary schoolbook multiplication has raw column multiplicities ```text r_k = sum_(i+j=k)a_i b_j, ab = sum_k r_k 2^k, sum_k r_k = s_2(a)s_2(b). ``` Set `q_(-1)=0`, put `r_k=0` after the last raw column, and iterate until the carry vanishes: ```text u_k = r_k+q_(k-1), z_k = u_k mod 2, q_k = (u_k-z_k)/2. ``` The `z_k` are the unique binary digits of `ab`. Summing `r_k+q_(k-1)=z_k+2q_k` through the final carry gives ```text kappa_(2,mul)(a,b) := sum_k q_k = s_2(a)s_2(b)-s_2(ab) >= 0. ``` Equivalently, each unit normalization move `2*2^k -> 2^(k+1)` lowers the coefficient mass by one. Every complete forward normalization has the same unique binary endpoint, so `kappa_(2,mul)` is its normalization-order- independent number of unit moves. It is not an algorithm-dependent count of grouped carry events. Now put ```text P = s_2(a) mod 2, Q = s_2(b) mod 2, K = kappa_(2,mul)(a,b) mod 2, R = s_2(ab) mod 2. ``` Reduction of the carry-mass identity modulo two and the Boolean identity `(P AND Q) XOR P XOR Q = P OR Q` give ```text R = (P AND Q) XOR K, tau_TM(ab)tau_TM(a)tau_TM(b) = (-1)^((P OR Q) XOR K). ``` Thus `tau_TM` is multiplicative at the pair `(a,b)` exactly when `K=P OR Q`. For distinct odd primes `p,q`, direct expansion of `c_TM=mu*tau_TM` gives ```text c_TM(pq) = tau_TM(pq)-tau_TM(p)-tau_TM(q)-1. ``` With `P=s_2(p) mod 2`, `Q=s_2(q) mod 2`, and `K=kappa_(2,mul)(p,q) mod 2`, the preceding sign law yields the complete abstract table ```text P Q K | c_TM(pq) 0 0 0 | -2 0 0 1 | -4 0 1 0 | 0 0 1 1 | -2 1 0 0 | 0 1 0 1 | -2 1 1 0 | 0 1 1 1 | 2 ``` Therefore ```text c_TM(pq) in {-4,-2,0,2}, c_TM(pq)=0 iff K=0 and (P OR Q)=1. ``` The table classifies the Boolean states; it does not assert that every state is realized by an odd-prime pair. A zero shadow is not zero carry: ```text kappa_(2,mul)(3,11)=4, c_TM(33)=0; kappa_(2,mul)(7,17)=0, c_TM(119)=0; kappa_(2,mul)(3,113)=3, c_TM(339)=-4. ``` For every prime `p`, only the divisors `1,p` survive in the convolution at `p^2`, hence ```text c_TM(p^2)=tau_TM(p^2)-tau_TM(p). ``` The parity law at `a=b=p` then proves ```text c_TM(p^2)=0 iff kappa_(2,mul)(p,p) is even. ``` Finally let `n=product_(i=1)^m p_i`, `m>=1`, be odd and squarefree. For every subset `S` of `{1,...,m}`, define ```text P_i = s_2(p_i) mod 2, n_S = product_(i in S)p_i, A_S = product_(i in S)s_2(p_i), Delta_mul(S) = A_S-s_2(n_S), K_S = Delta_mul(S) mod 2, ``` with `n_empty=A_empty=1`, `Delta_mul(empty)=0`, and the empty AND equal to one. The raw multiproduct has coefficient mass `A_S`; the same unit-move argument proves `Delta_mul(S)>=0` and ```text s_2(n_S) mod 2 = (AND_(i in S)P_i) XOR K_S. ``` Expanding every squarefree divisor directly in `c_TM=mu*tau_TM`, with `S` the support of `n/d` and hence `mu(d)=(-1)^(m-|S|)`, therefore gives the complete carry-parity field on the divisor Boolean cube: ```text c_TM(n) = sum_(S subseteq {1,...,m}) (-1)^(m-|S|) (-1)^((AND_(i in S)P_i) XOR K_S). ``` These five clauses form TM-MULTIPLICATION-CARRY-DEFECT [T], evidenced by `probes/P-TM-MULTIPLICATION-CARRY-DEFECT-1`. The written proof carries the universal quantifiers; the two-architecture verifier audits exact finite instances and the complete three-bit table. Here `kappa_(2,mul)` is distinct from decoder-seed and photon-worldline uses of `kappa`, from the finite-vector-space `e_2` carry layer of CARRY-PENTAD, and from the chronological `nu_2(n+1)` carry cocycle of RAMIFIED-TM-LIFT. No multiplicativity of `c_TM`, RH or zero-location result, Nyman--Beurling or Baez--Duarte completeness, analytic continuation, `J` coupling, `p=5` selection, FERMIONIZER dependency, decoder, measure, Born, observer, force, spacetime, SI, physical vacuum, matter, light, entanglement, curvature, physical interaction or other physical interpretation, or lift to L2--L6 is asserted. In particular, `c_TM=0` means neither zero carries nor trivial multiplication. ### The Hankel divisor block of c_TM This block of subsections is L1 arithmetic only, continuing the locally typed symbols `tau_TM(n)=(-1)^s_2(n)` and `c_TM=mu*tau_TM`. Fix an odd squarefree positive integer with prime set `P` and `k=|P|`, and write, for subsets `S,T` of `P`, ```text n_S = product_(p in S) p, K_P(S,T) = c_TM(n_S n_T), Kxor_P(S,T) = c_TM(n_(S XOR T)), R_P = K_P - Kxor_P, W(S,T) = 2^(|T|-|S|) if S subseteq T, else 0. ``` `P` is called extremal when `tau_TM(n_Z)=(-1)^(|Z|+1)` for every subset `Z`, the equality locus of the divisor-count bound `|c_TM(n_P)|<=2^k`. Inertia triples are written with named fields `NEG ZERO POS`. ### TM-HANKEL-DIVISOR-BRIDGE [T] For the Hankel kernel `H(m,n)=c_TM(mn)/(mn)` on positive integers, the compression `H_P` to the divisor cube of `n_P` satisfies, with `D_P=diag(n_S)` positive, ```text (D_P H_P D_P)(S,T) = c_TM(n_S n_T) = K_P(S,T), ``` so by Sylvester's law of inertia `inertia(H_P)=inertia(K_P)`: the integer matrix `K_P` carries the entire inertia content of the divisor block. Since `n_S n_T = n_(S XOR T) n_(S AND T)^2` exactly, the block splits as ```text K_P = Kxor_P + R_P, ``` where `Kxor_P` is an XOR circulant diagonalized by the `2^k` Walsh characters and the defect `R_P` is supported exactly on the pairs with `S AND T` nonempty. This forms TM-HANKEL-DIVISOR-BRIDGE [T], evidenced by `probes/P-TM-HANKEL-K3-TRANSFER-1`; the written proof is the two displayed identities, and the probe audits them on 246 prime sets, extremal and not. ### TM-HANKEL-SQUAREFUL-RANK-NOGO [T] `c_TM(1)=-1` is odd, and `c_TM(N)` is even for every `N>=2`, because `c_TM(N)` is a sum of `2^omega(N)` unit terms. This parity is table-general: it uses no property of `tau_TM` beyond taking values in `{-1,+1}`. Hence the empty row and column of `R_P` vanish, the nonempty block of `R_P` is congruent to the identity modulo 2 with odd determinant, and ```text rank R_P = 2^k - 1, ker R_P = span(e_empty) ``` exactly. The same parity applies at every intersection order: for every `U subseteq P` the layer transform ```text d_U(z) = sum_(V subseteq U) (-1)^(|U|-|V|) c_TM(n_z n_V^2) ``` on subsets `z` of `P` minus `U` has `d_U(empty)` odd and `d_U(z)` even for nonempty `z`, its XOR-structured layer matrix is congruent to the identity modulo 2 with odd determinant and full rank `2^(k-|U|)`, and the exact layer inversion ```text K_P(S,T) = sum_(V subseteq S AND T) d_V(S XOR T) ``` holds entrywise. Consequently no low-rank compression of the squareful defect exists, globally or at any single intersection order: the defect is two-adically unimodular on its carrier. This forms TM-HANKEL-SQUAREFUL-RANK-NOGO [T], evidenced by `probes/P-TM-HANKEL-K3-TRANSFER-1`. ### TM-HANKEL-EXTREMAL-WITT-SKELETON [T] `W` is unimodular, upper triangular, and congruent to the identity modulo 2. On every extremal `P`, the extremal cube values `c_TM(n_T)=-(-2)^(|T|)` force the exact tensor decomposition `Kxor_P = -(m^(tensor k))` for `m=[[1,-2],[-2,1]]`, and the local identity `u^T m u = diag(1,-3)` for `u=[[1,2],[0,1]]` tensorizes to the integral congruence ```text W^T Kxor_P W = diag_(S subseteq P) ((-1)^(|S|+1) 3^(|S|)). ``` The empty row and column of `W^T R_P W` vanish, so the empty direction splits off at the constant `-1` for the entire pencil `Kxor_P + s R_P`. The balance `NEG = POS = 2^(k-1)` of the skeleton is read off the diagonal. This forms TM-HANKEL-EXTREMAL-WITT-SKELETON [T], evidenced by `probes/P-TM-HANKEL-K3-TRANSFER-1`; the probe verifies the congruence on 1109 extremal sets with `p < q <= 1000`, four triples, and chains at `k = 4` and `k = 5`. ### TM-HANKEL-K2-TRANSFER [T] For an extremal pair `{p,q}` put `A=tau_TM(p^2)`, `B=tau_TM(q^2)`, `D=tau_TM(p^2 q)`, `E=tau_TM(p q^2)`, `F=tau_TM(p^2 q^2)`, all units. In the `W` basis the pencil `Kxor_P + s R_P` equals exactly ```text row empty: ( -1, 0, 0, 0 ) row {p}: ( 0, 3+sA, 0, s(A+D) ) row {q}: ( 0, 0, 3+sB, s(B+E) ) row {p,q}: ( 0, s(A+D), s(B+E), -9+s(3D+3E+F) ) ``` with `|3D+3E+F| <= 7`. For `0 <= s <= 1` the two middle pivots are at least 2, and Schur elimination leaves ```text h(s) = -9 + s(3D+3E+F) - s^2 (A+D)^2/(3+sA) - s^2 (B+E)^2/(3+sB) <= -9 + 7s <= -2 < 0, ``` so the pencil determinant never vanishes on `[0,1]` and the inertia is constantly `NEG 2 ZERO 0 POS 2`, uniformly over all 32 squareful sign patterns. The balanced transfer is therefore universal at `k = 2`. This forms TM-HANKEL-K2-TRANSFER [T], evidenced by `probes/P-TM-HANKEL-K3-TRANSFER-1`. ### TM-HANKEL-K3-UNIVERSAL-TRANSFER [F] The universal proposal that the balanced transfer persists at `k = 3`, that is, that every extremal triple has `K_P` inertia `NEG 4 ZERO 0 POS 4`, is false. The extremal triple ```text 147965 = 5 . 101 . 293 ``` has `K_P` inertia `NEG 5 ZERO 0 POS 3`, determinant `-3840`, pencil constant term `3^12`, and exactly one root of `det(Kxor_P + s R_P)` in the open interval `(0,1)`: an interior crossing, not an endpoint degeneracy. Among all 157 extremal triples with `n <= 200000` this is the unique nonbalanced case; the witnesses `1942781 = 83 . 89 . 263` and `11743733 = 149 . 269 . 293` behave identically, and among the 99 triples with `p < q < r <= 300` exactly three fail. Every witness holds by two independent exact integer paths on two architectures. The false universal statement is registered as TM-HANKEL-K3-UNIVERSAL-TRANSFER [F], evidenced by `probes/P-TM-HANKEL-K3-TRANSFER-1`; the firing does not reopen TM-HANKEL-K2-TRANSFER and does not weaken the finite `k = 3` classification below. ### TM-HANKEL-K3-TWO-SCALAR-CLASSIFICATION [C] Abstract setting: ternary sign tables on `{0,1,2}^3` with the extremal binary face fixed leave 19 free squareful signs, and `K` is linear in them. Exhaustive facts over all `2^19` tables, byte-identical on two architectures: writing `G_6` for the weight-at-most-2 block in the `W` basis, the sign of `det G_6` obeys the rigidity trichotomy ```text det G_6 < 0 iff G_6 inertia NEG 3 ZERO 0 POS 3, det G_6 = 0 iff G_6 inertia NEG 2 ZERO 1 POS 3, det G_6 > 0 iff G_6 inertia NEG 2 ZERO 0 POS 4, ``` with census `32398 / 110 / 260` on the 15-bit substrate, exact 16x lift `518368 / 1760 / 4160` to the `2^19` strata, and a unique all-minors-zero pair-Schur configuration, of inertia `NEG 2 ZERO 0 POS 1`. The full-block classes refine by the sign of `det K`: ```text det K > 0 iff K inertia NEG 4 ZERO 0 POS 4 (522462 tables), det K = 0 iff K inertia NEG 4 ZERO 1 POS 3 (51 tables), det K < 0 iff K inertia NEG 5 ZERO 0 POS 3 (1775 tables), ``` and the two-scalar law holds over the whole domain: ```text FAIL iff det G_6 < 0 and det K <= 0. ``` The three real witnesses satisfy `det G_6 < 0` and `det K < 0` through the abstract linear form, with `det K` equal to the direct integer values `-3840`, `-768`, `-9856`. This is TM-HANKEL-K3-TWO-SCALAR-CLASSIFICATION [C], a computation at the stated finite, fully enumerated scope, evidenced by `probes/P-TM-HANKEL-K3-TRANSFER-1`. The finite local case lists behind the trichotomy, the four-value bound table with its 32 diagonal and 16 coupling cases, are printable and theorem-eligible, but this row stays at computation grade until they are separately written and reviewed. ### TM-HANKEL-K3-QUADRATIC-INVARIANT-SUFFICIENCY [C] The 19 free cells carry the `S_3` representation `6 trivial + sign + 6 standard`. The six linear orbit sums do not decide the transfer class: they produce 3584 buckets of which exactly 58 are mixed. The canonical quadratic layer of 28 invariants, namely the six orbit sums, the 21 Gram pairings of the six standard-component vectors, and the squared circulation, is sufficient: it produces 88352 buckets with zero mixed, so the transfer decision factors through the quadratic invariant map. Burnside's count gives exactly ```text (2^19 + 3 . 2^12 + 2 . 2^7)/6 = 89472 ``` orbits, by the formula and by direct enumeration, so the decision factors through a proper quotient of the orbit space, merging 1120 orbit distinctions and no class pair. Sufficiency means factoring through the invariant map; no claim is made that the deciding function is itself a polynomial of degree two. This is TM-HANKEL-K3-QUADRATIC-INVARIANT-SUFFICIENCY [C], evidenced by `probes/P-TM-HANKEL-K3-TRANSFER-1`. The seven rows above assert no RH or zero-location result, no Weil positivity or explicit-formula connection, no Nyman--Beurling or Baez--Duarte statement, nothing about the operator `H` beyond the displayed finite compressions, no `J` coupling, no `p=5` selection, no decoder, measure, Born, observer, force, spacetime, physical, or SI reading, and no L2--L6 lift. The successor question at `k = 4` is not a registered claim of this Canon. ## 10. Relativity as counting The group generated by the icosahedral rotations and the J boost is dense in SO+(3,1); the boost rapidity is ln phi. The exact boost reading split (BOOST-READING-SPLIT [T]; reproduce/observer-boost) is ``` C_n = phi^n + phi^-n, S_n = phi^n - phi^-n, (C_n, S_n) = (L_n, sqrt(5) F_n) for n even, (C_n, S_n) = (sqrt(5) F_n, L_n) for n odd. ``` Indeed psi = -phi^-1, while the Binet identities are L_n = phi^n + psi^n and sqrt(5) F_n = phi^n - psi^n; substitution proves the split for every nonnegative integer n. Consequently beta_n = S_n/C_n obeys the exact addition law beta_(m+n) = (beta_m + beta_n)/(1 + beta_m beta_n), and C_n^2 - S_n^2 = 4 is the parity-resolved form of L_n^2 - 5 F_n^2 = 4 (-1)^n. The count ladder (BOOST-COUNT-LADDER [D]) reads n as the substrate's integer rapidity count and beta_n as the decoder velocity; the dictionary rests on the exact split and the exponent law B_J^m B_J^n = B_J^(m+n). The observer alternator (OBSERVER-ALTERNATOR [D]) reads mu_4 as 1 + 3, whereas multiplication by lambda = -1 has the two orbits {1, -1} and {i, -i}, giving the substrate partition 2 + 2. In the commuting diagonal model B_J = diag(phi, phi^-1), A = diag(1, -1), the name boost axis (BOOST-AXIS [D]) is the reading of the two exact A eigendirections. No claim that any of these three existing dictionary choices is forced is made. The fixed `beta_1` walk has a theorem-grade exact drift. Put ``` A_1 = 1/sqrt(5) [[1, 2], [2, -1]], Sigma = diag(-1, 1), D = A_1 Sigma A_1 = 1/5 [[3, -4], [-4, -3]]. ``` Then `D^2 = I`, `tr(D) = 0`, and `spec(D) = {-1,+1}`. For ``` S(z) = diag(z,z^-1), W(z) = S(z) A_1, t = tr(W) = (z-z^-1)/sqrt(5), h = t/2, r^2 = (18+z^2+z^-2)/20, lambda_+ = h+r, lambda_- = h-r, p_+ = W-lambda_- I, p_- = W-lambda_+ I, ``` the following identities hold division-free in `Q(sqrt(5))[z,z^-1,r]/(r^2-(18+z^2+z^-2)/20)`: ``` W^2 = tW+I, W^-1 = W-tI, lambda_+ lambda_- = -1, r^2-h^2 = 1, p_+^2 = 2r p_+, p_-^2 = -2r p_-, p_+ p_- = p_- p_+ = 0, p_+-p_- = 2rI. ``` On `z=exp(ik)`, both bands are unitary. With `rho=lambda_-/lambda_+`, ``` |1-rho|^2 = 4r^2, r^2 - 4/5 = cos^2(k)/5, ``` so the squared step gap never closes and its minimum is `16/5`. The exact diagonal drift is obtained from ``` G = p_+ D p_+ + p_- D p_- = -(z+z^-1)/sqrt(5) (2W-tI), V_inf = G/(4r^2), V_inf^2 = [beta_1^2 cos^2(k) / (1-beta_1^2 sin^2(k))] I. ``` It obeys `V_inf(1)=-beta_1 A_1`, and its spectrum fills exactly `[-beta_1,beta_1]`. If ``` P_+ = p_+/(2r), P_- = -p_-/(2r), V_bar_N = (1/N) sum_(j=0)^(N-1) W^-j D W^j, ``` the division-free conjugation kernel gives, for every `N>=1`, ``` ||V_bar_N-V_inf|| <= 1/(Nr) <= sqrt(5)/(2N). ``` This is DRIFT-IS-THE-READ [T] at L5. The final reading statement is an implication under two declared premises: ``` P1 the read is translation covariant and acts fiberwise in k; P2 the read window is long, N >> 1. ``` P1 and P2 are not derived. The theorem proves the exact operator identities and all-`N` bound and says what a read satisfying those premises returns. It does not prove decoherence, an environment, collapse, a Born coupling, measurement, L6, SI, coin adoption, or unique physics. The corresponding integer coin classification is also exact. For positive odd `n`, ``` beta_n = L_n/(sqrt(5) F_n), 1-beta_n^2 = 4/(5F_n^2). ``` An integer-normalized positive-orientation alternator coin has ``` A(a,b) = 1/sqrt(5) [[a,b],[b,-a]], a,b in Z_(>0), a^2+b^2 = 5. ``` The positive solutions are `(a,b)=(1,2),(2,1)`. Equivalently, integrality on an odd rung implies `F_n | L_n`; then `L_n^2-5F_n^2=-4` gives `F_n^2 | 4`, and monotonicity of the positive Fibonacci numbers leaves exactly `n=1,3`. Thus the complete admissible pair is ``` {beta_1,beta_3}, beta_1=1/sqrt(5), beta_3=2/sqrt(5). ``` In rapidity coordinate `x=eta/log(phi)`, distinguish the closed ranges used for completeness from the open bands used for cost: ``` n in Z, w>0, C_(n,w) = [n-w,n+w], I_(n,w) = (n-w,n+w). ``` The closed family covers the line exactly when `w>=1/2`. For integer `w>=1`, the open-band multiplicity is `2w` away from a rung and `2w-1` on a rung. Hence ``` coin w generic multiplicity rung multiplicity beta_1 1 2 1 beta_3 3 6 5 ``` At `w=1/2`, the closed intervals still cover, but the open bands have multiplicity one only away from the seams and multiplicity zero at `Z+1/2`; the half-rung is therefore only a generic single tiling. Its velocity `tanh(log(phi)/2)=sqrt(5)-2=phi^-3` is not integer-admissible. Under the registered boost addition law, the closed rapidity cover maps to a complete composed velocity cover of `(-1,1)`. Repeating the exact spectral comparison at `beta_3` gives ``` coin r_min^2 uniform constant min |1-rho|^2 beta_1 4/5 sqrt(5)/2 16/5 beta_3 1/5 sqrt(5) 4/5. ``` Therefore, on the complete frozen pair, ``` S1 minimum generic open-band multiplicity -> beta_1 uniquely, S2 minimum worst-case uniform constant -> beta_1 uniquely, S3 maximum coherent half-width -> beta_3 uniquely. ``` This is COIN-SELECTION-CONDITIONAL [T]. S1 and S2 are distinct definitions that agree on this pair; no general equivalence or experimental independence is asserted. S3 is the exact counter-ranking. The theorem adopts no selector. Its use as a velocity-read comparison is bounded by BOOST-COUNT-LADDER [D]; that dictionary is not a proof premise of the T row. The Canon adopts S1/S2 under one name: ``` MINIMAL-READ usporne cteni; choose the complete integer-admissible coin minimizing both generic covering multiplicity and worst-case uniform-read constant MAXIMAL-REACH nejdelsi dosah; choose the admissible coin with the largest coherent range ``` COIN-MINIMAL-READ [D] selects `beta_1` by MINIMAL-READ. MAXIMAL-REACH is the named, unadopted counter-selector and selects `beta_3`. The agreement of the two minimum-cost criteria is evidence on the frozen pair, not a derivation of the premise and not a general equivalence. The row fires if a complete exact decoder derivation uniquely forces `beta_3`, or if an exact counterexample overturns the admissible-pair or ranking theorem; merely naming another preference that chooses `beta_3` does not fire it. MINIMAL-READ-DERIVATION [O] owns the unresolved L5-to-L1 selection boundary. It asks whether the complete registered decoder architecture, without adopting MINIMAL-READ, uniquely forces `w=1`, `beta_1`, and its minimum-read property. GATE-L5-L1-MINIMAL-READ closes positively only when the complete typed carrier, cover-to-output map, admissible protocol class, accumulator/equality rule, redundancy theorem, and dependency graph uniquely force that result. It closes negatively only when the complete admissible class is proved nonempty and either contains fully compliant `beta_1` and `beta_3` realizations or uniquely forces `beta_3`. It is STOP while any listed type, map, class, rule, graph, or layer endpoint is incomplete. Failure of the no-feedback route or any one favored route is only STOP unless it classifies the complete admissible class. READ-REDUNDANCY-PRIME-SUPPORT [T] settles what absence of feedback does and does not buy. A typed projection funnel carries the sixfold cover over Z with zero arithmetic nodes, and over Q every multiplicity carries, so acyclicity, no-feedback, integrality and totality together obstruct nothing; dropping totality voids every bound as well, since the diagonal common-value map carries every multiplicity. For anonymous total accumulators over the localization Z_S the bound is exact: multiplicity m carries if and only if every prime divisor of m lies in S. Necessity is the weight-1 stratum equation m c_(1) = 1 in Z_S, since (1) is the only weight-1 partition and every monomial symmetric basis element has diagonal degree equal to its weight; sufficiency is (x_1 + .. + x_m)/m. At the registered places {2} and {2,5} the sixfold read therefore costs the prime 3, which the constant ring does not carry, and the twofold read costs only the prime 2. The criterion is a prime-support selector, not a smaller-is-better principle: on the pair {6,10} over {2,5} it selects the larger multiplicity, and on {2,10} it is nonunique. The row informs MINIMAL-READ-DERIVATION [O] and closes nothing. The two decoder definitions that row lists as missing, the cover-to-output map and the accumulator equality rule, are reduced to two named bits with known prices, anonymity and totality; the row itself stays O and STOP by its own decision text, since one clause pair is not the complete registered decoder class, and COIN-MINIMAL-READ [D] remains the adopted dictionary. No lift is performed and GATE-L5-L1-MINIMAL-READ is untouched. The boost ladder above reads the units of `F = Q(sqrt5)`. The three rows below carry the arithmetic structure the whole field supplies under the same reading, at L1 and with no physical lift. ### ARITHMETIC-RAPIDITY-DECOMPOSITION [T] For `x = a + b sqrt5` in `F*` write `conj` for the nontrivial automorphism, `N(x) = x conj(x)` for the norm with sign, and `rho(x) = x/conj(x)` for the multiplicative rapidity avatar; `rho` and `N` are multiplicative. Setting `t = a` and `s = b sqrt5`, ```text t^2 - s^2 = a^2 - 5 b^2 = N(x) with sign, ``` so `N` is the Minkowski invariant of the pair, `N > 0` timelike and `N < 0` spacelike. The F-rational null locus is empty, `a^2 = 5 b^2` forcing `a = b = 0` by irrationality of `sqrt5`, while the real completion keeps the full cone `t = +- s`; neither statement may be quoted without the other. The rest locus `eta = 0` is exactly `ab = 0`. The rational norm-one points are dense in `SO+(1,1)`, not discrete, by the parametrization `t -> ((1+5t^2)/(1-5t^2), 2t/(1-5t^2))`; the discrete rapidity lattice is supplied by the units alone, `O_F*/{+-1} = ` with `eta(phi^n) = n log phi`. Since `phi^n = (L_n + F_n sqrt5)/2` with `L_n^2 - 5 F_n^2 = 4(-1)^n`, Lucas is always the time reading and `sqrt5` Fibonacci always the space reading in these coordinates, and the alternator `N(phi) = -1` exchanges the timelike and spacelike unit sheets. The positive-inverse coordinates of BOOST-READING-SPLIT above and the signed Galois coordinates here are bridged by `sigma-(phi^n) = (-1)^n phi^-n`: the parity swap there and the fixed Lucas time reading here are one identity read in two conventions, with no conflict. This forms ARITHMETIC-RAPIDITY-DECOMPOSITION [T], with written proof in `probes/P-ARITH-RAPIDITY-1/PREREG.md` and the pinned two-architecture verifier as its finite audit, evidenced by `probes/P-ARITH-RAPIDITY-1`. ### SPLIT-PRIME-RAPIDITY-CLASS [T] For a rational prime `p` split in `F`, a prime ideal above it is generated, by class number one, by some `pi` with `N(pi) = +- p`. The class `r = [eta(pi)]` in `R/(log phi)Z` does not depend on the generator, any other being `+- phi^n pi`; conjugation negates it. The rational prime therefore carries canonically only the unordered pair ```text R(p) = { r, -r } in ( R/(log phi)Z ) / {+-1}, ``` and nothing finer without an extra choice. The class is decided exactly and without logarithms: `[eta(x)] = [eta(y)]` iff `rho(x)/rho(y) = +- phi^(2n)`, and membership of a norm-one integral `w` in `+- phi^(2Z)` is decided by `|Tr(w)| = L_(2m)`, strictly increasing in `|m|`, followed by exact comparison. Anchors: every inert `p` has `rho(p) = 1` and class zero exactly; the ramified generator `sqrt5` has `rho(sqrt5) = -1` and `eta` exactly `0`, not merely `0` modulo the lattice. This forms SPLIT-PRIME-RAPIDITY-CLASS [T], with written proof in `probes/P-ARITH-RAPIDITY-1/PREREG.md` and the pinned verifier as its finite audit, evidenced by `probes/P-ARITH-RAPIDITY-1`. ### SPLIT-PRIME-RAPIDITY-CONSTRUCTION-AGREEMENT [C] For all 146 split `p < 2000`, two structurally independent constructions return a generator of norm `+p` in `Z[phi]`: a Diophantine Pell sweep on `a^2 - 5 b^2 = +- 4 p`, and a Euclidean gcd of `p` and `sqrt5 - r` at the canonical root `0 < r < p/2`, with every division step asserted norm-decreasing rather than assumed. The canonical unordered classes agree in every case, `R1(p) = R2(p)`. The oriented split is data and gates nothing: 70 pairs agree oriented and 76 only after conjugation, as expected since the Pell sweep fixes no orientation; only the unordered `R(p)` is canonical. Byte-identical stdout on two architectures. This forms SPLIT-PRIME-RAPIDITY-CONSTRUCTION-AGREEMENT [C], evidenced by `probes/P-ARITH-RAPIDITY-1`. ### SPLIT-PRIME-RAPIDITY-INDEPENDENCE [T] Let `p_1, ..., p_k` be pairwise distinct rational primes, all split in `F`, let `P_i = (pi_i)` be a prime ideal above `p_i` with either orientation, and let `r_i = [eta(pi_i)]`. If integers `m_i` satisfy ```text m_1 r_1 + ... + m_k r_k = 0 in R/(log phi)Z, ``` then every `m_i = 0`. Equivalently, real lifts `t_1, ..., t_k` together with `log phi` are linearly independent over `Q`. The proof needs only what is already registered above. Put `x = prod pi_i^(m_i)`; since `rho` is multiplicative and `eta` additive, `[eta(x)] = 0`, so by SPLIT-PRIME-RAPIDITY-CLASS `rho(x) = +- phi^(2n)`. Passing to fractional ideals and using that `phi` is a unit, ```text prod_i P_i^(m_i) conj(P_i)^(-m_i) = (1), ``` and the `2k` ideals `P_i`, `conj(P_i)` are pairwise distinct because each `p_i` splits and residue characteristics separate distinct indices. The group of fractional ideals is free abelian on the primes, so every exponent vanishes. Replacing `P_i` by `conj(P_i)` negates `t_i`, so no orientation is chosen anywhere. Carried consequences: no split class is torsion, the map `p -> m R(p)` is injective for every `m >= 1` and never the zero class, the integer-multiple orbit equidistributes in the `k`-torus by Weyl's criterion, and after one choice of orientation the classes generate a free abelian subgroup of infinite rank. The equidistribution statement concerns multiples of a fixed finite set of classes and says nothing about the distribution of primes as `p` grows. This forms SPLIT-PRIME-RAPIDITY-INDEPENDENCE [T], with written proof in `probes/P-SPLIT-PRIME-INDEPENDENCE-1/PREREG.md` and the pinned verifier as its finite audit, evidenced by `probes/P-SPLIT-PRIME-INDEPENDENCE-1`. ### REDUCED-SPLIT-GENERATOR-HEIGHT [T] Among the generators `+- phi^n pi` of a prime ideal above a split `p`, exactly one up to sign and conjugation has ```text eta in ( -(log phi)/2, (log phi)/2 ), ``` the endpoints being unattainable because `eta = +-(log phi)/2` means `2 r = 0`, which the preceding row forbids. For that reduced generator `|sigma+| |sigma-| = p` and `|eta| < (log phi)/2` give `min |sigma| > sqrt(p) phi^(-1/2)`, and the smallest split rational prime is `11`, so both embeddings exceed one; both comparisons are exact in `Z[phi]`. The generator is an algebraic integer of degree two with monic minimal polynomial, so its absolute logarithmic height is ```text h(pi) = (1/2) log p, ``` and hence `h(pi/conj(pi)) <= log p`. The arithmetic content is the pair of exact embedding comparisons; the height value then follows from the definition of the height and the norm identity. The reduction is not cosmetic: without it the height is unbounded at fixed class, because `pi -> phi^a pi` multiplies the avatar by `(-1)^a phi^(2a)` and leaves the class unchanged. This forms REDUCED-SPLIT-GENERATOR-HEIGHT [T], with written proof in `probes/P-SPLIT-PRIME-INDEPENDENCE-1/PREREG.md` and the pinned verifier as its finite audit, evidenced by `probes/P-SPLIT-PRIME-INDEPENDENCE-1`. ### SPLIT-PRIME-RAPIDITY-QUANTITATIVE-SEPARATION [T] Put `L = log phi`. For every split rational prime `p`, fix an oriented prime ideal `P_p` above `p` and a generator satisfying ```text (pi_p) = P_p, N(pi_p) = +p, eta_p = (1/2) log(pi_p/conj(pi_p)). ``` The common sign is chosen so that both real embeddings of `pi_p` are positive. This positive-norm convention applies uniformly to every generator in this row. Given any finite list `(p_j,epsilon_j,a_j)`, with `epsilon_j` in `{+1,-1}` and `a_j` integral, first merge repetitions and conjugated orientations over each rational prime: ```text c_p = sum_(j:p_j=p) epsilon_j a_j. ``` Delete the zero coefficients. For every resulting nonzero finite vector `c`, set ```text P(c) = prod_p p^|c_p|, S = sum_p c_p eta_p. ``` Then ```text dist(S,L Z) >= asinh(1/(2 sqrt(P(c)))). ``` More precisely, let `n` be the unique nearest integer to `S/L` and put `delta = S-nL`. The parity refinement is ```text n even: |delta| >= asinh(sqrt(5)/(2 sqrt(P(c)))), n odd: |delta| >= asinh(1/(2 sqrt(P(c)))). ``` In the odd branch, `P(c) = -1 (mod 5)` improves the numerator `1` to `2`. To prove the result, form the totally positive element ```text x = prod_(c_p>0) pi_p^c_p prod_(c_p<0) conj(pi_p)^(-c_p), y = phi^(-n) x, D_c = y-(-1)^n conj(y). ``` Its embeddings give ```text D_c = 2 sqrt(P(c)) sinh(delta), sinh^2(dist(S,L Z)) = D_c^2/(4P(c)). ``` Writing `y=a+b phi` and `T=Tr(y)=2a+b`, one has `D_c=b sqrt(5)` when `n` is even and `D_c=T` when `n` is odd. The determinant cannot vanish: if it did, `x/conj(x)=phi^(2n)`, but direct fractional-ideal valuation gives ```text v_(P_p)((x/conj(x))) = c_p, ``` whereas the unit on the right has valuation zero at every `P_p`. Thus every `c_p` would vanish. Consequently the exact determinant lattices are ```text n even: |D_c| in sqrt(5) Z_(>0), n odd: |D_c| in Z_(>0). ``` The same argument applied to `2c` excludes a nearest-lattice tie. The norm identities ```text T^2-5b^2 = +4P(c) for even n, T^2-5b^2 = -4P(c) for odd n ``` give the displayed bounds and the odd congruence refinement. Positive-norm even-unit gauge changes on a fixed oriented ideal preserve the parity branch, `D_c^2/P(c)`, and the metric; simultaneous orientation reversal and coefficient reversal also preserve it. No global orientation of the rational split primes is selected. For two distinct public unordered split-prime classes, the minimizing signed channel is unique up to simultaneous conjugation. Equality of the sum and difference gaps would make twice one class vanish, contradicting SPLIT-PRIME-RAPIDITY-INDEPENDENCE [T]; this still selects no global orientation. The universal numerator `1` is sharp, but no asymptotic least-gap law is asserted. The positive-norm lifts ```text pi_11=3+phi, pi_41=6+phi, phi^-1 pi_11 pi_41=-9+19phi, pi_421=19+4phi, pi_431=19+5phi, phi^-1 pi_421 pi_431=-190+381phi ``` give norm-minus-`451` and norm-minus-`421*431` half-band translates of trace one. For the positive-norm `421/431` sum channel, with ```text d_L = dist(eta_421+eta_431,L Z), d_2L = dist(eta_421+eta_431,2L Z), ``` the corresponding numerical controls are only ```text d_L = 0.0011737895036417... d_2L = 0.4800380355559618... ``` No mixed-sign decimal comparison enters this row. The theorem is falsified by one permitted nonzero merged vector with `D_c=0`, by a failure of the exact metric or bound, by dependence on a permitted even-unit gauge change, or by failure of the parity-resolved absolute-value condition ```text n even: |D_c| in sqrt(5) Z_(>0), n odd: |D_c| in Z_(>0). ``` The sign of `D_c` is unrestricted: a valid determinant can be negative, as in the ordinary control `-182 sqrt(5)`, and negativity alone is not a falsifier. This forms SPLIT-PRIME-RAPIDITY-QUANTITATIVE-SEPARATION [T], with written proof in `probes/P-SPLIT-RAPIDITY-QUANTITATIVE-SEPARATION-1/PREREG.md` and the pinned two-architecture verifier as its finite exact audit, evidenced by `probes/P-SPLIT-RAPIDITY-QUANTITATIVE-SEPARATION-1`. ### SPLIT-RAPIDITY-FEJER-GRAM-BOUND [T] Let `A` be a nonempty finite set of distinct oriented split prime-power addresses ```text a=(p,m,epsilon), m>=1, epsilon in {+1,-1}, p^m<=X, beta_a=epsilon m eta_p in R/(L Z). ``` If `|A|>=2`, define ```text delta_A = min_(a!=b) dist(beta_a-beta_b,L Z). ``` The effective vector of an address difference is nonzero and has product budget at most `X^2`, so the preceding theorem gives ```text delta_A >= asinh(1/(2X)). ``` For every integer `K>=0`, define the height-one normalized Fejer kernel and its Gram matrix by ```text Phi_K(u) = 1/(K+1) sum_(|h|<=K) (1-|h|/(K+1)) exp(2 pi i h u) = 1/(K+1)^2 (sin(pi(K+1)u)/sin(pi u))^2, (A_K)_(ab) = Phi_K((beta_a-beta_b)/L). ``` At integral `u` the quotient has its continuous value `1`. Then ```text ||A_K-I||_(2->2) <= pi^2 L^2/(12(K+1)^2 delta_A^2) <= pi^2 L^2/(12(K+1)^2 asinh^2(1/(2X))). ``` For a singleton, `A_K=[1]` and the norm is zero; no spacing minimum is defined or needed. Writing `||u||=dist(u,Z)`, ```text 0 <= Phi_K(u) <= min(1,1/(4(K+1)^2 ||u||^2)). ``` The half-open circular shell `j delta_A <= dist(beta_a-beta_b,L Z) < (j+1) delta_A` contains at most one address on each side of `beta_a`. Every off-diagonal row sum is therefore at most ```text 2 sum_(j>=1) L^2/(4(K+1)^2 j^2 delta_A^2) = pi^2 L^2/(12(K+1)^2 delta_A^2), ``` and Schur's test proves the result. Uniformly over families of such finite sets, the second bound has the precise asymptotic form ```text (pi^2 L^2/3+e_X) (X/(K+1))^2, |e_X| <= C X^-2 for all X>=X_0, ``` for constants `C,X_0` independent of the finite family. Thus `K(X)/X -> infinity` is sufficient. The asymptotic coefficient is not a finite replacement for the exact `asinh` denominator. The characters `chi_h(a)=exp(2 pi i h beta_a/L)` are abstract parts of the displayed definition. They are not identified with a Hecke character, an `xi_h` or `xi_(2h)` family, or an `End` mode. For any later integral phase multiplier `nu`, the exact budget is ```text P(nu c) = P(c)^|nu|. ``` It is unchanged only for `|nu|=1`; `nu=0` is the excluded diagonal and `|nu|>=2` requires a new spacing derivation. A nonintegral phase lies outside the carrier. The `491/1429` controls keep the two normalizations separate. Their ordinary signed channels have determinant rungs `|b|=22` and `|b|=182`, hence even-branch determinants `+-22 sqrt(5)` and `+-182 sqrt(5)`; these audit only the base sign channels, and one valid value is `-182 sqrt(5)`. The actual doubled-phase falsifier uses the effective vector `(2,2)`. With ```text x=(20+7phi)(34+13phi), P=491*1429, phi^-3 x^2=-313768+627565phi, Tr(phi^-3 x^2)=29, sinh^2(d_2)=841/(4P^2) < 1/(4*1429^2), ``` the numerator `841` falsifies reuse of the unchanged `X` spacing budget after phase doubling. It is not a counterexample to the preceding separation theorem, whose correct doubled budget is `P^2`, and it does not by itself assert a concrete operator-norm violation. This row is falsified by a declared finite set with spacing below `asinh(1/(2X))`, by failure of the two-points-per-half-open-shell count, or by a finite `A_K` violating either displayed norm bound; for a singleton it is falsified exactly by `A_K != [1]`. Importing an external phase normalization without its correct product budget is outside the row, not a repair of its threshold. This forms SPLIT-RAPIDITY-FEJER-GRAM-BOUND [T], with written proof in `probes/P-SPLIT-RAPIDITY-QUANTITATIVE-SEPARATION-1/PREREG.md` and the same pinned two-architecture verifier as its finite exact audit, evidenced by `probes/P-SPLIT-RAPIDITY-QUANTITATIVE-SEPARATION-1`. Both rows are L1 arithmetic and finite split-address analysis only. They add no unrestricted least-gap or prime-distribution result, Pell parametrization, inert or ramified diagonalization, gamma or polar term, completed-zeta compression, Weil positivity, RH or zero-simplicity statement, decoder, measure, physical or SI reading, or L2--L6 lift. ### SUZUKI-LOCAL-CAPACITY-NOGO [T] One classical no-go complex on the screw function of the Riemann zeta function, in Suzuki's normalization. Freeze ```text P(t) = sum_(p^k <= e^t) (log p) p^(-k/2) (t - k log p) t >= 0 S(z) = sum_(k >= 0) z^k / (4k+1)^2 0 < z < 1 A(t) = 8 (cosh(t/2) - 1) - alpha t + C - 4 e^(-t/2) S(e^(-2t)) t > 0 alpha = (log pi - psi(1/4)) / 2, C = psi'(1/4) / 4 Psi = A - P, K_F(s,t) = F(s) + F(t) - F(|s-t|). ``` The positivity criterion for Psi and the screw kernel K_A - K_P are Suzuki's; the curvature closed form and the plastic transition below are Mittermeier's, reproduced here by an independent implementation with no novelty claimed. Six statements, written proofs pinned in the public probe, finite gates certified by outward interval arithmetic at scale `2^-192`: ```text N1 for every locally finite event family {(tau_q, omega_q)} the direct-sum curve Y_t = (+)_q omega_q 1_[tau_q,t] has ||Y_t||^2 = F(t), = F(min(s,t)), orthogonal increments and ||Y_t - Y_u||^2 = F(t) - F(u); prime powers give P R2 A''(t) = e^(t/2) + e^(-t/2) - e^(-t/2)/(1 - e^(-2t)); the sign changes once, at log rho with rho^3 = rho + 1, rho in (13/10, 4/3) N3 no c0, c1 and locally finite Borel measure mu >= 0 give A = c0 + c1 t + integral (t-a)_+ dmu on (0, inf): a certified three-point convexity violation at (1/20, 1/4, 1/2) empties the nonnegative ramp class N4 A(1/4) > A(1/2), so dA is not a nonnegative measure and no filtration model exists on (0, log 2); on [log 2, 4/5] the single exact prime ramp exceeds the capacity increment, so increment domination dP <= dA and every nonnegative per-place budget die at the first event q = 2 N5 4A(3) - A(6) < 0 < A(6) and 4P(3) - P(6) < 0 < P(6), with e^3 in (20, 23) and e^6 in (401, 409): both screw kernels are separately indefinite; only the difference can be a screw geometry N8 T Z_t = Y_t with ||Z_t||^2 = A, ||Y_t||^2 = P and ||T|| <= 1 - delta forces Psi >= delta (2 - delta) A against Psi = o(e^(t/2)) from the prime number theorem, so ||T|| = 1 ``` Together: the completion capacity is not a positive superposition of prime-type ramp atoms, admits no filtration or per-place domination reading, and is not itself a screw geometry; every Gram realization dominating the prime curve has operator norm exactly one and is nonlocal in `t`. The prime-event counting frame is motivation only. No statement about the Riemann hypothesis or its zeros, no least-gap or prime-distribution result, no decoder, measure, physical or SI reading, no J-coupling, and no L2--L6 lift. ### SUZUKI-PRIME-FREE-WINDOW [C] `A(t) > 0` for every `t` in `[1/128, 45/64]`, where `Psi = A` on `(0, log 2)`: a 100-leaf adaptive outward-interval cover with zero undecided leaves, an independent code path reproducing one corner of Mittermeier's certified event-segment strip; the interval `(0, 1/128)` carries no gate. ### SUZUKI-EVENT-COUNT [C] `N(10^6) = 78734` prime-power events `p^k <= 10^6`, by direct enumeration and by `sum_k pi(floor(10^(6/k)))`, exact integer equality of two independent counting paths. At L1, `P-SUZUKI-LOCAL-CAPACITY-NOGO-1` supplies the written proofs, the byte-identical two-architecture audit, twelve certified gates and the two finite computations above, with frozen attribution to Suzuki and to Mittermeier in its preregistration. It makes no statement about the Riemann hypothesis. ## 11. The pentit ring and the magic boundary In F_25 = F_5[tau]/(tau^2 - 2), the ramified images are J = 2 and phi = 3; tau^4 = -1, tau^6 = phi, and therefore sqrt(phi) = tau^3 (PENTIT-ROOT-FACTS [T], reproduce/pentit-p5-closure). Calling tau the square root of the axiom, sqrt(J) = tau, is the gate-line dictionary (PENTIT-ROOT-READING [D]); it does not identify the argument with the clock. The exact magic-prime gate is the complete norm ladder N(tau^k) = 3^k = 2^-k in F_5*, the order-four source i_5 = 2 reaching -1 after two steps, and the reciprocity sqrt(J) sqrt(phi) = tau^4 = -1 with phi = J^-1 (MAGIC-PRIME-GATE [T], reproduce/pentit-p5-closure). Define V_+ as the sign quotient F_5*/{+-1}; its two classes are {1, 4} and {2, 3}, so V_+ is cyclic of order two (QUBIT-FROM-F5 [T], reproduce/pentit-p5-closure). The native magic is the cubic C_5 (order 5, prime 5), while the foreign read uses m_8 of order 8 at prime 2. For n in {5, 8}, define ``` E_n(a,b) = Re(zeta_n^(a-b)), S_n = abs(E_n(a0,b0) + E_n(a0,b1) + E_n(a1,b0) - E_n(a1,b1)), M_n = max S_n over (Z/nZ)^4. ``` Complete exact enumeration, independently certified after quotienting by the common phase shift, gives ``` M_5 = 1/2 + sqrt(5), M_8 = 2 sqrt(2), 2 < M_5 < M_8. ``` This is BELL-MAGIC-BOUNDARY [T] at the stated finite-functional scope, evidenced by probes/P-BELL-MAGIC-BOUNDARY-1. It is not an unrestricted Bell cap, a theorem about local-variable models, a continuous quantum optimum, or a Tsirelson claim. The legacy modulus bound involving phi is a different observable. ### Product composition of the quadratic pair Let \(K\) be a field of characteristic not two with involution `c`, let `bar(V)` be the `c`-twist of `V`, and identify `Sym^2(V)` with the `+1` eigenspace of factor interchange. Put ```text H(V) = V tensor bar(V), S(V) = Sym^2(V), Q(V) = H(V) direct-sum S(V), H(v) = v tensor bar(v), S(v) = v tensor v, Q(v) = (H(v),S(v)). ``` Canonical factor reorderings define the typed matched maps ```text boxtimes_H: H(V) tensor H(W) -> H(V tensor W), boxtimes_S: S(V) tensor S(W) -> S(V tensor W). ``` The exact evidence is `probes/P-QPAIR-SYM2-TENSOR-DEFECT-1`. **QPAIR-PRODUCT-COMPOSITION [T].** The componentwise law ```text mu((A,B),(C,D)) = (A boxtimes_H C, B boxtimes_S D) ``` is natural, associative, symmetric, and unital, with unit `Q(1)=(1,1)`, and satisfies ```text Q(v tensor w) = mu(Q(v),Q(w)). ``` This is an exact product-vector law and extends bilinearly on the matched carrier sectors. It does not assert surjectivity onto the symmetric squares of entangled vectors. **QPAIR-CROSS-SECTOR-NONDESCENT [T].** The reciprocal change of factorization `(v,w) -> (lambda v,lambda^-1 w)` fixes the composite tensor. It leaves `H tensor H` and `S tensor S` invariant, while the cross sectors have weights ```text H tensor S: c(lambda)/lambda, S tensor H: lambda/c(lambda). ``` For \(K=Q(\zeta_5)\), `c(zeta_5)=zeta_5^-1`, and \(\lambda=\zeta_5\), the weights are \(\zeta_5^3\) and \(\zeta_5^2\), neither one. Thus neither nonzero cross sector assignment arising from nonzero `v,w` descends to a function of the composite pure tensor. Retaining the matched sectors is a typed descent through factorization gauge, not an arbitrary deletion of invariant data. **QPAIR-SYM2-TENSOR-DEFECT [T].** If `char(K) != 2` and \(\dim V=\dim W=2\), then ```text Sym^2(V tensor W) = (Sym^2(V) tensor Sym^2(W)) direct-sum (Lambda^2(V) tensor Lambda^2(W)), 10 = 9 + 1. ``` The linear span of product squares is exactly the first, nine-dimensional summand. Freeze the reorder ```text R((v tensor w) tensor (v' tensor w')) = (v tensor v') tensor (w tensor w'). ``` On the reordered space let `alpha` swap the two `V` factors, let `beta` swap the two `W` factors, and put ```text P_++ = (1+alpha)(1+beta)/4, P_-- = (1-alpha)(1-beta)/4. ``` For ```text x = a e0 f0 + b e0 f1 + c e1 f0 + d e1 f1, kappa = (e0 wedge e1) tensor (f0 wedge f1), u wedge v = u tensor v - v tensor u, ``` the missing projection is ```text P_-- R(x tensor x) = ((ad-bc)/2) kappa. ``` The line transforms by `det(g)det(h)`, is fixed by `SL_2 times SL_2`, and vanishes exactly on the product cone. The missing line is a determinant/concurrence direction in the quadratic symmetric target. It is not a Bell state and not the ordinary two-qubit singlet in \(V\otimes W\). After a complex norm is supplied, normalized Bell states are inputs maximizing the ordinary concurrence `2|ad-bc|`. The `9+1` result is confined to the symmetric slot and does not assert that a full Hermitian-plus-symmetric informational or entanglement defect is one-dimensional. These are L1 product-composition statements. They create no `BELL-CAUSAL-ACCOUNTING` row, no dependency on `QUADRATIC-DECODER-DATA`, no bridge to rational \(V_{\rm eff}\), and no decoder, observable, Born rule, instrument, L5 stream, or L6 measure. The `zeta_5` instance is only an exact factor-gauge witness and carries no fifth-prime physical or selection content. `QDD-QCARRIER-DIAGONAL-BOUNDARY` and `QUADRATIC-DECODER-DATA` retain their registered scopes and statuses. The Fibonacci category with central charge c = 14/5 is mathematical background; its physical reading fired: the phibit is abelian Z_5, not the tau anyon (PHIBIT-NOT-TAU [F], reproduce/hyperplane-codec). The boundary proof is finite: the phibit fusion ring is the group ring of Z_5, five simples, all invertible with dimension 1; the Fibonacci ring obeys tau tau = 1 + tau, its dimension satisfies d^2 = d + 1 with no rational root, d = phi, and tau has no fusion inverse; five invertible simples cannot land on a ring with one. The dead physical reading is archived, not deleted. ## 12. The color door The theorem layer in this section is finite group, representation, and invariant theory. COLOR-LADDER-DICTIONARY [D] reads its D5 to 2I to E8 ladder as the nonabelian color door and reads color su(3) on the traceless endomorphisms of the three dimensional trace kernel. The dictionary rests on the exact rungs below; it proves neither a unique color assignment nor QCD running, confinement, or the measure lift from the core to the full SL_3(F_5) carrier. Rungs 1 and 2 (COLOR-RETURN-D5 and COLOR-TORSOR-HOLONOMY [T], COLOR-SPLIT-12 [D], reproduce/color-ladder): the return group is D5 of order 10 with integer Plancherel mass M(E) = 10 |E|. The 312 size-20 attractors have a free D5 half; the singlet half is the five reflection axes. The recurrent spatial holonomy is a Klein four group, orientation is tick parity, and the pairing involution Phi lies in SL_3(F_5). Its eigenspaces give the dictionary split 3 = 1 + 2. Rung 3 is the binding negative turn. COLOR-DYNAMICAL-COLOR [F] (reproduce/color-ladder): within the registered dynamical candidate families the generated census symmetry group has order 20 and its special-linear spatial image is only {I, Phi} = Z_2; the non abelian dynamical-color falsifier fired. The surviving kinematical statement is exact. COLOR-KIN-NORMALIZER [T]: the filtered product-affine normalizer has 40 x 480 = 19200 elements and every one permutes the 313 attractor supports. COLOR-KINEMATICAL-GL2 [D]: its special-linear image is read as GL_2(F_5), order 480, on the antisymmetric plane, embedded by g -> (det g)^-1 direct-sum g along 3 = 1 + 2. Rung 4 (COLOR-CORE-2I [T], reproduce/color-ladder): the special-linear core has order 120, is perfect, has center {I, -I}, and has class sizes (1, 1, 12, 12, 12, 12, 20, 20, 30), hence is 2I = SL_2(F_5). The block trace by element order is {1:2, 2:3, 3:4, 4:0, 5:2, 6:1, 10:3}; the pentagonal values {2, 3} are the ramified shadow of the golden pair. Rung 5 (COLOR-GOLDEN-TABLE [T], reproduce/color-ladder): the full 9 x 9 character table is exact over Q(sqrt5), orthogonal by rows and columns, and Galois stable. The core has one involution, so D5 does not embed; the loop pair lifts through Dic_5 and reflections acquire order 4. The spin-lifted pair reads the icosahedral edge module, the pentagonal spin weights are the Born squares of the golden amplitudes, and the 5-regular block traces are the ramified Brauer shadow of the golden spin row. The corresponding marked-pair uniqueness proposal has now been decided. Under the pinned cover SL_2(F_5) -> PSL_2(F_5), the full A1 to A7 admissibility predicate, and simultaneous conjugacy by SL_2(F_5), exact enumeration gives 240 admissible triples in four inequivalent classes of size 60. Central retwist pairs classes 1 with 2 and 3 with 4, and no such pair is conjugate. Thus the dicyclic witness exists but is not forced: SPIN-LIFT-FORCED [F], evidenced by probes/P-SPIN-LIFT-FORCED-1. A coarser quotient identifying retwists or base relabelings is a different question. Rung 6 (COLOR-MCKAY-E8 [T], reproduce/color-ladder): tensoring by the spin row gives affine E8 with marks equal to representation dimensions. Its closed-walk moments equal the Catalan numbers through degree 10; finiteness first appears at degree 12 by one, 133 = 132 + 1. The verb weights are {4 phi^-2, 4 phi^2}, with product 16. Rung 7 (COLOR-MOMENT-FINGERPRINT [T], reproduce/color-ladder): the moment series is N/D, where N and D are the matching polynomials of finite and affine E8, and ``` 120 m_n = 2^(n+1) + 40 + 24 L_n ``` for even n. Also 1 - J = -zeta_5^2 is a primitive tenth root of unity, joining the D5 order to the J shadow without a new parameter. Rung 8 (COLOR-SPECTRAL-INVARIANTS [T], reproduce/color-ladder): the partial-fraction weights are exactly the class masses |C|/120. The Molien series is ``` (1 + t^30) / ((1 - t^12)(1 - t^20)), ``` so the invariant degrees are (12, 20, 30) = (vertices, faces, edges) with one relation at 60. The Platonic excess is 1/2 + 1/3 + 1/5 - 1 = 1/30 = 1/h(E8). Rung 9 (COLOR-DICKSON-RAMIFICATION [T], reproduce/color-ladder): at the ramified place the invariant forms are ``` E_6 = u^5 v - u v^5, D_20 = (u^5 - u v^4)^4 + v^20. ``` They are algebraically independent and the modular invariant ring is free on degrees (6, 20). In general (V, F, E) = (2(p+1), p(p-1), p(p+1)) and V - E + F = 2 identically. Rung 10 (COLOR-KLEIN-REDUCTION [T], reproduce/color-ladder): T^2 + H^3 = 1728 f^5 holds exactly in Z[u,v], with 1728 = 12^3. At the ramified place invariant reduction has shape (alpha E_6^2, beta D_20, gamma E_6^5); freeness forces beta = 0 and gamma^2 = 3 alpha, with good alpha in {2, 3}. The modular invariant tower is the Artin-Schreier tower t -> t^5 - t. Rung 11 (COLOR-INTEGRAL-LIFT [T], reproduce/color-ladder): the two matrices ``` S = ((0, -1), (1, 0)), T = ((zeta, 1), (0, zeta^4)) ``` close to exactly 120 matrices over Z[zeta_5], and reduction modulo (1 - zeta) is a bijection onto SL_2(F_5). The engine identity is (1 - zeta)^4 = 5(zeta^2 - zeta - zeta^3). The Hessian reduces to the Dickson form; in the Klein gauge the explicit orbit realizes gamma^2 = 3 alpha. ### The integral quadratic pair and its relative closure The following definitions fix the carrier and admissible category of this subsection. **DEF-QPAIR-SPIN-CARRIER.** Put ```text K = Q(zeta_5), O_K = Z[zeta_5], c(zeta_5) = zeta_5^-1, V_spin = O_K^2. ``` The coordinates of `V_spin` are independent. Scalar multiplication `u(z1,z2)=(u z1,u z2)` is defined on the carrier for `u in O_K`, and is a carrier automorphism for `u in O_K^x`. Statements below using `u in K^x` are made after scalar extension to `K^2`. This is neither a diagonal two-place image of one field element nor the rational carrier \(V_{\rm eff}\) of `DEF-QDD-QPAIR`. The marked binary-icosahedral action uses exactly `COLOR-INTEGRAL-LIFT`: ```text S0 = ((0,-1),(1,0)), T0 = ((zeta_5,1),(0,zeta_5^-1)), G = . ``` **DEF-QPAIR-HERM-SLOT.** On `V_spin` define `H(v)=v c(v)^T`. Under the principal embedding its linearly readable coordinate space \({\cal H}\) has real dimension four. **DEF-QPAIR-SYM-SLOT.** Define `S(v)=vv^T` and `Q(v)=(H(v),S(v))`. The realification \({\cal S}\) of the three complex symmetric coordinates has real dimension six. **DEF-QPAIR-MIXED-C4.** The rational-linear state action `Phi(z1,z2)=(z2,c(z1))` satisfies `Phi^2=c` componentwise and `Phi^4=1`. **DEF-QPAIR-ADMISSIBLE-LINEAR-CLASS.** Regard the coordinates of `H` and `S` as formal homogeneous real quadratic polynomials on \((C^2)_R\), subsequently restricted to \(O_K^2\). The class \({\cal A}_{\rm rel}\) consists of real linear subspaces `E` such that \({\cal H}\subseteq E\) remains a linearly readable typed slot and `E` is stable under pullback by every marked `g in G` and by `Phi`. Quotients by pointwise coincidence, nonlinear image actions, arbitrary set reconstruction, decoder factorization, normalization, and a presumed group product between the two pullback families are excluded. The exact evidence for the seven rows below is `probes/P-QPAIR-C4-2I-MINIMALITY-1`. **QPAIR-HERM-INTEGER-NONDESCENT [T].** For `v=(1,1)` and `v'=zeta_5 v`, ```text H(v)=H(v'), H(Phi v)_12=1, H(Phi v')_12=zeta_5^2. ``` Consequently no total set map on `im(H)` descends `Phi` on \(O_K^2\). For nonzero vectors over `K` the complete fiber is ```text H^-1(H(v)) = K^1 v, K^1 = {u in K^x : u c(u)=1}. ``` Define `cont(z1,z2)=z1 O_K+z2 O_K`. Since \(O_K^x=\mu_{10}\times\langle\varphi\rangle\), with \(\varphi=-(\zeta_5^2+\zeta_5^3)\), for every nonzero `v in O_K^2` the vectors `w` satisfying both `H(w)=H(v)` and `cont(w)=cont(v)` are exactly \(\mu_{10}v\). The qualification is essential: for `a=2+zeta_5`, the integral vectors `(c(a),0)` and `(a,0)` have equal Hermitian slots but distinct content and are not related by \(\mu_{10}\). **QPAIR-TRANSPOSE-FIBER-REDUNDANCY [T].** For `v,w in K^2`, `S(v)=S(w)` iff `w=+-v`. Because `H(-v)=H(v)`, there is a unique set map `F:im(S)->im(H)` with `F(S(v))=H(v)`, and `im(Q)->im(S)` is bijective. Thus the pair carries exactly the set-theoretic information of `S`. This supplies no polynomial, rational, linear, typed-natural, or admissible factorization. **QPAIR-TYPED-MIXED-C4-CLOSURE [T].** For ```text H=((h11,h12),(h21,h22)), S=((s11,s12),(s12,s22)), h21=c(h12), ``` the exact typed action is ```text H'=((h22,s12),(c(s12),h11)), S'=((s22,h21),(h21,c(s11))). ``` Define `T_Q(H,S)=(H',S')`. It is rational-linear, satisfies `T_Q Q(v)=Q(Phi v)` and `T_Q^4=1`, and gives the cycle `h12 -> s12 -> c(h12) -> c(s12) -> h12`. The least `Phi`-stable space containing \({\cal H}\) has dimension six, so `Phi` alone does not force the full pair. **QPAIR-SYM2-2I-IRREDUCIBLE [T].** Relative to the marked public character row `2a`, ```text Sym^2(2a)=3a ``` is absolutely irreducible and factors through \(A_5\); its Galois companion is `Sym^2(2b)=3b`. For the pullback convention `(g.q)(v)=q(g^-1 v)` on binary quadratics, the seed `s(x,y)=xy` obeys `T0.s=s-zeta_5 y^2` and its subsequent `S0` orbit supplies `x^2`. Hence the orbit spans `K{x^2,xy,y^2}`. **QPAIR-MINIMAL-2I-CLOSURE-OF-HERM-UNDER-MIXED-C4 [T].** Write `z1=a+ib` and `z2=c+id`. The four Hermitian real coordinates are ```text a^2+b^2, c^2+d^2, ac+bd, bc-ad, ``` and the six symmetric real coordinates are ```text a^2-b^2, 2ab, ac-bd, ad+bc, c^2-d^2, 2cd. ``` These ten coordinates recover every real quadratic monomial on \(C^2\): ```text dim_R H = 4, dim_R S = 6, H intersection S = 0, absolute coordinate determinant = 64. ``` Since `Phi^* H12=z1 z2`, every \(E\in{\cal A}_{\rm rel}\) contains both real coordinates of the seed. The marked `2I` orbit forces the full realification of `Sym^2(V^*)`; complex coefficients act on real and imaginary parts by real two-by-two matrices. Therefore \({\cal H}\oplus{\cal S}\subseteq E\). The pair space is itself stable under both pullback families, so it is the least member of \({\cal A}_{\rm rel}\), of real dimension ten. This is relative minimality of a fixed readable `H` slot, not absolute, slot-count, or information-theoretic minimality. **QPAIR-2I-ONLY-PAIR-FORCING [F].** The universal proposition that `2I`-equivariance alone forces a pair is false: the single symmetric slot already obeys `S(gv)=gS(v)g^T`. With ```text epsilon=((0,1),(-1,0)), Theta(Y)=Y epsilon^-1, ``` one also has `Theta(gYg^T)=g Theta(Y) g^-1`. This is an adjoint presentation of the one-slot counterexample. It does not identify the standard trace-zero Hermitian slice with an invariant module in the displayed nonunitary marked basis, and no invariant graph or quotient involving that slice is claimed. **QPAIR-MIXED-C4-NORMALIZES-2I [F].** The proposition that `Phi` normalizes the marked `2I` is false: ```text Phi T0 Phi^-1(z1,z2) = (zeta_5^-1 z1, zeta_5^-1 z2 + c(z1)). ``` The conjugate contains `c(z1)`, is not `K`-linear, and is not in `G`. Simultaneous stability under the two pullback families supplies neither `2I semidirect-product C4` nor `2I times C4`, and it is not the distinct registered `COLOR-CM-2I-SEMILINEAR-PAIR`. All statements in this block are L1 carrier algebra. They create no bridge to rational \(V_{\rm eff}\), `DEF-QDD-QPAIR`, MatterData, a Born pairing, decoder write map, physical `U(1)`, instrument, L5 stream, or L6 measure, and they move no QDD or color-selection parent. The marked CM closure of Rung 11 (COLOR-CM-2I-SEMILINEAR-PAIR [T], probes/P-CM-2I-QCARRIER-1) is relative to exactly the displayed marked integral representative. Put ```text K = Q(zeta_5), tau(zeta_5) = zeta_5^2, sigma = tau^2, F = K^sigma = Q(sqrt(5)), phi = (1 + sqrt(5))/2, G = , rho(S) = S, rho(T) = T, rho^a(g) = a(rho(g)), V = K^2 direct-sum K^2, Pi(g) = diag(rho(g),rho^tau(g)), q = zeta_5 - zeta_5^4, C0 = ((1,q),(-q,1)). ``` Field automorphisms act entrywise on matrices and vectors, `M^dagger = sigma(M)^T`, `N_K/F(x) = x sigma(x)`, and `H0 = sum_(g in G) rho(g)^dagger rho(g)`. An admitted tau-semilinear structure is an invertible `nu_B(v) = B tau(v)` satisfying `nu_B Pi(g) = Pi(g) nu_B`. Its frozen pair-coordinate equivalence is ```text B' = A B tau(A)^-1, A = diag(r I2,s I2), r,s in K^x. ``` For `gamma in Gal(K/Q)`, marked twist-isomorphism means that there is one `P_gamma in GL2(K)` satisfying ```text P_gamma gamma(rho(S)) P_gamma^-1 = rho(S), P_gamma gamma(rho(T)) P_gamma^-1 = rho(T). ``` An invariant sigma-Hermitian form on the single branch means a matrix `H` with `H^dagger = H` and `rho(g)^dagger H rho(g) = H` for every `g in G`. The four conclusions are: 1. The marked twist-isomorphism stabilizer of `rho` is exactly `{1,sigma}`. The elements `tau` and `tau^3` exchange the golden trace values `phi^-1` and `-phi`. This stabilizer equality neither identifies unmarked conjugacy classes nor supplies an outer automorphism, an `F`-form, or a descent datum; the norm obstruction below forbids normalizing the sigma-intertwiner to an involution. 2. The ordered pair `Pi` has Q-valued character and a Galois-stable K-isomorphism class through explicit intertwiners. Its branches are absolutely irreducible and inequivalent with scalar endomorphism algebras, and the exact block spaces are ```text Hom_G(rho^tau,rho) = 0, Hom_G(rho^sigma,rho) = K C0, End_G(rho^tau) = K I2, Hom_G(rho^sigma,rho^tau) = 0. ``` These facts supply neither a Q-form nor a coherent C4 descent datum. 3. Every admitted structure has the exhaustive antidiagonal form ```text B = ((0, a C0), (d I2, 0)), a,d in K^x. ``` The intertwiner cocycle is `mu(a) = N_K/F(a) (-phi^2) I2`, so its invariant class under the displayed equivalence is `[-1]` in `F^x/N_K/F(K^x)`. Total positivity of nonzero CM norms excludes order four universally. For `B0 = ((0,C0),(phi I2,0))`, the map `nu = B0 tau` obeys `nu^4 = -I4` and `nu^8 = I4`, so eight is the smallest attainable finite order. This does not say that every admitted `nu` has finite order. 4. On the single `rho` branch the invariant sigma-Hermitian forms are exactly the F-line `F H0`, which contains the totally positive definite form `H0`. For the chosen form `H_pair = diag(H0,tau(H0))`, the explicit order-eight map satisfies ```text B0^dagger H_pair B0 = phi^2 tau(H_pair), ``` where determinants and positivity force the multiplier `phi^2`. No claim is made that the full space of invariant forms on the pair `V` is one dimensional. These are L4 representation-theoretic statements only. They assert neither uniqueness nor selection of the displayed marked lift; SPIN-LIFT-FORCED [F] remains unchanged. They do not define decoder `Q` or `QCarrier`, a decoder Gram, an orbit-to-amplitude map, `MatterData`, a physical U(1), or an L5-L6 measure lift. QUADRATIC-DECODER-DATA [O] and COLOR-MEASURE-SELECTION [O] remain STOP and unchanged. COLOR-MEASURE-TRANSPORT [T] (reproduce/color-ladder): the golden dual measure is transported from D5 onto the 2I core through the exact character, verb, and spectral class measures of rungs 5 to 8. COLOR-MEASURE-SELECTION [O] is the remaining L4-to-L6 lift from that core. It is STOP until the target carrier, orbit equivalence, observable sigma-algebra, group action, total normalization, equality, and selection constraints are public. The existing finite core theorems do not determine those data. Positive closure requires one canonical normalized lift compatible with the frozen action and observables; negative closure requires an exhaustive exact classification yielding no lift or more than one inequivalent lift under every registered constraint. ## 13. Gravity and cosmology The bridge law alpha B g = 1; G proportional to alpha^20 with the geometric factor (32/33)^2 = (2^5/(2^5 + 1))^2; m_e the single SI anchor (GRAVITY-BRIDGE-LAW at D: the equation layer, with g = 2^5 phi^2 sqrt(3 - phi) carried exactly through its square; the SI value of G stays on the frontier; reproduce/gravity-chain). The Kahler capacity V_geo = (4 pi)^3 phi^2 = 64 pi^3 phi^2 (KAHLER-CAPACITY) [T]: from the single Kahler metric h(z) = (J Jbar)^-1 (1 + |z|^2)^-1, the order 2 jet gives Fubini-Study (4 pi)^3 and the order 0 jet gives h(0) = phi^2, since J Jbar = phi^-2 exactly. The FRW rank 1 canonical form (FRW-CANONICAL-FORM) [T]: H^2 = 72 pi rho_phys with lambda = 216 pi from the rank 1 lapse action; the fiber multiplier k_f = 1 is forced by the master closure against G_nat = 27 = d^3, with the cell volume 864 pi carrying exactly one fiber 2 pi. The public theorem stops at the displayed lapse, Hamiltonian, and fiber identities; it makes no claim here about a unique source projector, the amplitude ansatz rho = rho_0 ell^2, or E_total = 0. The ell-G wall is a mechanism wall: ell_P / lambda_e = (32/33) alpha^10 / sqrt(g), with the exponent identity G_T = alpha^(20 + sigma) (GRAVITY-BRIDGE-LAW). At the homogeneous L5 rational-coefficient scope, let `chi = log a` and write `rho_q` proportional to `a^(-q)`. With `d = 3`, the registered FRW continuity identity gives ``` q = 3(1 + w_q), {(q,w_q) in Q^2: q = 3(1 + w_q)} = {(q, q/3 - 1): q in Q}. ``` In particular, `(q,w_q) = (0,-1)` and `(1/5,-14/15)` are distinct exact solutions. Therefore the displayed continuity identity alone does not select a unique density character (DE-TRACE-DENSITY-UNDERDETERMINATION [T], probes/P-DE-TRACE-DENSITY-1). The numeral `1/5` is only a rational witness; it is not identified with `gamma_tr`. No trace-Gram dictionary, physical trace-to-density transport, source projector, inhomogeneous extension, SI normalization, or empirical fit is supplied, and DE-CONFORMAL-WEIGHT [O] remains open. Cosmology (reproduce/cosmology-register): the exact deformation J -> J e^(i eps) freezes J Jbar at linear order (TT-LINEAR-ZERO [T]). COSMOLOGY-READING-DICTIONARY [D] reads that identity as zero linear tensor response about the declared isotropic background and hence as r_T = 0 at that dictionary layer; it also reads the gyron density as the mass-ladder prefactor and the registered forms below as cosmological observables. It is not a uniqueness or full perturbation theorem. The tilt n_s - 1 = -p alpha = -5 alpha is at H (NS-TILT, falsifier live at CMB-S4); the bilinear TT decoder permits induced tensor power at quadratic field order (TT-QUADRATIC-INDUCED) [D]; a numerical r_T(k) waits on the vector state normalization (TT-VECTOR-STATE-NORMALIZATION). Dark energy w = -14/15; Omega_b = pi^2 / 200; Omega_DM / Omega_b = 18 p^3 ln^2(phi) / pi^4, and the dark matter ratio 5 : 1 follows from Thue-Morse pair statistics (COSMOLOGY-REGISTER at D, the committed forms with fenced comparisons). For DE-CONFORMAL-WEIGHT [O], its `w = -14/15` entry is only a committed target and consistency witness. Neither COSMOLOGY-REGISTER [D] nor COSMOLOGY-READING-DICTIONARY [D] selects `Delta_DE`, supplies a trace/source map, or authorizes `Delta_DE := gamma_tr`; using either row as a selection premise is circular. COSMOLOGY-READING-DICTIONARY [D] reads the L1 stationary equal-phase sliding-pair value 1/6 from GYRON-DENSITY as the mass-ladder prefactor. Its proton and cosmology assignments belong only to that dictionary. GYRON-DENSITY itself is not a mass density, cosmological parameter, Born multiplier, physical probability, or L5-to-L6 measure statement. Independently, freeze the six golden projective lines with representatives ``` v1 = (0,1,phi), v2 = (0,1,-phi), v3 = (1,phi,0), v4 = (1,-phi,0), v5 = (phi,0,1), v6 = (phi,0,-1). ``` Put `K = Q(phi)`, `r = phi + 2`, `Pi = vi vi^T/r`, and give the six lines equal cardinal weight. On `Sym2(K^3)` define ``` M(A) = (1/6) sum_i Tr(Pi A) Pi, P1(A) = Tr(A) I3/3, P5(A) = A - P1(A). ``` The rank-one projectors sum to `2 I3` and their centered projectors form a rank-five regular simplex. The direct sum defining `M` is Galois-stable and descends to a rational endomorphism of `Sym2(Q^3)`. The complete exact commutant of the rational `so(3)` action is spanned by `P1` and `P5`, and ``` M = (1/3) P1 + (2/15) P5. ``` Thus its scalar-to-per-channel coefficient ratio is `5:2`, its scalar-to-total-traceless trace-mass ratio is `1:2`, and `M != (1/6) I6`. For the negative control put ``` c1 = (1,1,1), c2 = (1,1,-1), c3 = (1,-1,1), c4 = (-1,1,1), Qi = ci ci^T/3, M_cube(A) = (1/4) sum_i Tr(Qi A) Qi. ``` Then `(1/4) sum_i Qi = I3/3`, but `M_cube` is not in the rational `so(3)` commutant. Thus second-order isotropy alone does not imply the Sym2 centrality (GOLDEN-SIX-LINE-SYM2-FRAME [T] at L1, probes/P-TM-SYM2-FRAME-1). The six lines and their equal weights are frozen inputs, not a selection from J, U, a checkpoint, or a Thue-Morse orbit. The coefficient `1/6` in `M` is only the cardinal average over six lines; it is not GYRON-DENSITY, a clock density, or a Born multiplier. This theorem supplies no Thue-Morse measure, Born halving, physical probability, L5 stream, or L6 measure lift. At the frozen L5 selector surface, TM-SYM2-PROJECTIVE-FOURFOLD [T], evidenced by probes/P-TM-SYM2-MEASURE-1, records 48 exact selectors in four free projective-gauge orbits of size 12. The v16 N2 threshold therefore fired NONCANONICAL. The same classification proves that every selector has the uniform mathematical pushforward nu_s(v_i) = 1/6 and the common operator M_s = (1/3)P1 + (2/15)P5 with ratio 5:2. These selector-independent outputs do not select a representative and are not a physical probability or Born theorem. TM-SYM2-SEMILINEAR-TWOFOLD [T], evidenced by probes/P-TM-SYM2-SEMILINEAR-GAUGE-1, identifies the exact order-24 comparison image Gamma_sl = ker(chi_Q chi_F), its twelve exponent-one elements of additive character (1,1), and its two free selector orbits of size 24. The remaining bit is epsilon_read = chi_Q chi_F. TM-SYM2-REVERSAL-CLOSURE [T], evidenced by probes/P-TM-SYM2-REVERSAL-CLOSURE-1, proves that drive-word reversal R acts by precomposition and toggles epsilon_read: the translations of N, R, and NR are (0,1), (1,0), and (1,1); the N and R transports are nonrealizable over Q(sqrt5) at both frozen exponents, while NR is realizable at exponent one; and the candidate mixed action generated by Gamma_sl postcomposition and R precomposition has one orbit of size 48. ### TM-SYM2-SPECTRAL-COHERENCE [T] On the frozen length-three Thue-Morse window carrier and complete 48-selector class, define the signed transfer operator by `(L_s)[x,w] = v_{s(w)} . v_{s(x)}` on the twelve frozen transfer edges and zero elsewhere. Every one of the 1128 unordered selector pairs admits an exact similarity witness ```text L_t = D P^-1 L_s P D, P in {id, N}, D diagonal with entries in {+1,-1}. ``` Thus the 48 operators form one signed-permutation similarity class. Their common characteristic polynomial and spectrum are ```text det(x I - L_s) = x^2 (x^2 - phi^2) (x^2 - 3 phi^2), spec(L_s) = {0, 0, +phi, -phi, +sqrt(3) phi, -sqrt(3) phi}. ``` The frozen battery is orbit-constant and does not separate the two `epsilon_read = chi_Q chi_F` classes: the edge-sign product is `+1`, the two transfer-triangle sign sum and the all-triple sign sum are both zero, and the characteristic polynomial is common to all 48 selectors. The exact auxiliary census gives 60 exponent-one semilinear realizations among all 720 line permutations, 12 in the 48-element centralizer, and 60 linear realizations; the line two-graph splits its 20 triples as 10 and 10, and Galois conjugation flips every pairwise dot sign. These are L5 statements only. They provide no L5-to-L6 bridge, physical measure, Born reading, preferred selector, enlarged gauge, or uniqueness outside the frozen class (probes/P-TM-SYM2-SPECTRAL-COHERENCE-1). Only the forbidden full torsor W merges the two residual classes as a postcomposition gauge. Reversal merges them only through drive-word precomposition, so R does not realize the missing postcomposition coset. The old N2 result is preserved, not repaired. Gamma_sl and N, R, and NR remain comparison actions and are not adopted as gauge. Reading orientation epsilon_read is retained as typed L5 data, not quotiented. TM-SYM2-MEASURE [F] records the exact frozen S_TM boundary. Its projective selector gate fired NEGATIVE/N2 because four gauge orbits survive, so the route emits no canonical selector stream and never reaches its Born branch. This falsifies that frozen compound route, not the frame theorem, the stationary law, GYRON-DENSITY, the common mathematical pushforward, or every future TM-to-measure definition. ### TM-SYM2-PHYSICAL-MEASURE [D] The surviving TM-SYM2 measure route is closed only as a physical dictionary, not as a selector theorem. Two L5 objects are frozen as definitions for this bridge. First, ```text C_sel = Sel_class/G, |C_sel| = 4, Q_word = W3/, |Q_word| = 3, omega(a,b,c) = c-a. ``` `C_sel` is the four-class selector-gauge record and `Q_word` is the three-shell word quotient. They are different types and no map `C_sel -> Q_word` is introduced. The complete L5 source retains the whole `C_sel` record, `epsilon_read = chi_Q chi_F`, the current word in `W3`, and `omega`. On `W3`, `omega` is anti-invariant under both complement `N` and reversal `R`, and the rational joint `(-1,-1)` function sector is one-dimensional, spanned by `omega`. The `N`, `R`, and `NR` actions remain comparison actions and are not adopted as gauge. This is `DEF-TM-SYM2-ORIENTATION-SOURCE`. Second, put `j = zeta_5`, use the exact five-point Fourier convention ```text F(a)_k = sum_(r in Z/5Z) a_r j^(rk), v_t = delta_t + delta_(t+1), t in Z/5Z. ``` The separately frozen monomial verb-lift class obeys ```text F(v_t)_k = j^(tk) (1+j^k). ``` For `k != 0`, `1+j^k = sigma_(3k)(J)` with `sigma_a(j)=j^a`; the `k=0` slot is separately `2` and is not called a Galois conjugate. Exact inverse Fourier transform returns the two-term coefficient vectors. This is `DEF-TM-SYM2-MONOMIAL-VERB-LIFT`. The public `ABELIAN-FACE-DICTIONARY [D]` fixes the corresponding face moduli, but it does not select a phase lift. The distinction is necessary: the frozen negative control conjugates only the `k=1` spectral slot, preserves all five pointwise spectral moduli, and has full inverse-Fourier support with unequal coefficient Born weights. Therefore the modulus data alone do not force the halving and no uniqueness among all amplitude lifts is claimed. Within the frozen monomial class, every `v_t` has two equal nonzero coefficient amplitudes. `MEASURE-BORN-VERB [D]` constrains the physical read to the Born square of this typed verb. Normalized coordinate square on its support is independent of `t` and of sheet order and, only after that equality is proved, is ```text Born_t = (1/2,1/2). ``` For the other factor, import only the public L5 stationary law on `W3` from `TM-SYM2-PROJECTIVE-FOURFOLD [T]`. Its already known six-line pushforward and `M_TM` are not construction inputs. Form the three `N`-orbit marginals of that window law and then apply the derived two-sheet Born conditional. This gives a total normalized word measure. Only after construction and normalization its six word weights, and hence the six line weights under every selector chart, are read as ```text mu_B(w) = 1/6, w in W3. ``` Because the word measure is constant, every one of the 48 frozen selector charts has the same normalized pushforward. No selector representative is chosen and no postcomposition gauge is enlarged. The complete L5 source still contains `epsilon_read`; the scalar L6 measure is orientation-blind only as a proved output of the total map. The fired N2 conclusion of `TM-SYM2-MEASURE [F]` remains terminal and is not repaired. This closes `GATE-L5-L6-TM-SYM2-BORN-MEASURE` as a `DICTIONARY_LIFT` at status D. The exact finite algebra and the two-architecture audit are evidenced by `probes/P-TM-SYM2-BORN-HALVING-1`. The status does not rise above D because the physical Born-of-the-verb assignment is the registered dictionary reading and the monomial lift is frozen here as the typed bridge input, not selected from all same-modulus lifts. `GYRON-DENSITY` is not a dependency or confirmation of this bridge. No `M_TM` confirmation, `D_matter`, decoder-completion, SI, or all-lift uniqueness claim follows. The conformal mode prefactor K_chi5 = 1/(864 pi) is derived at the homogeneous L5 scope from the single layer 5 action, with c_hom = 12 K_chi5 = 1/(72 pi) (CONFORMAL-PREFACTOR) [D]; the inhomogeneous scalar action and the SI clause stay open (FRW-INHOM, METRO-EDGE-SCALE). N is the cosmic clock, carried at its committed labels within COSMOLOGY-REGISTER. ## 14. The gravitational wave program The TT decoder is the complex squaring map with kernel {+-1}, the spin double cover; volume neutrality is det = 1 - |h|^2 exactly; one propagation law c = 1 - s^2: breathing +1, photon 0, TT square -3 (TT-SQUARING-DECODER) [D; reproduce/coupling-metrology]. Writing `v = v_1 + i v_2` and `h = h_+ + i h_x`, the same square gives `h_+ = v_1^2 - v_2^2` and `h_x = 2 v_1 v_2`. For the exact input-frame rotation `v_1' = a v_1 - b v_2`, `v_2' = b v_1 + a v_2`, with `a^2 + b^2 = 1`, the pair transforms by ``` ((a^2-b^2, -2ab), (2ab, a^2-b^2)), ``` the doubled-angle spin-2 law; conjugation fixes `h_+` and negates `h_x`. The two coordinates are read as plus and cross (POL-READ [D]). This readout introduces no independent propagation law, source map, detector convention, action or state normalization, helicity selection, or numerical tensor ratio. The finite moment boundary (TT-VECTOR-MOMENT-UNDERDETERMINATION) [T; probes/P-TT-VECTOR-MOMENT-UNDERDETERMINATION-1] is entirely L1. Let `X = Z/5`, `z = zeta_5`, `w_x = v_x^2`, and `S_w(k) = sum_r E[w_r conjugate(w_0)] z^(-kr)`. Define ```text A v_x = z^(t_x), with the t_x iid uniform on Z/5, B_m v_x = z^(t_0+mx) epsilon_x, m in Z/5, with t_0 uniform and the epsilon_x iid uniform on {+1,-1}. ``` The six laws `A,B_0,...,B_4` are translation invariant, have deterministic pointwise modulus one, and agree on the mean, `C = delta`, `P = 0`, and every polynomial functional of total degree at most three. Their squared-readout spectra are ```text A: S_w(k) = 1, B_m: S_w(k) = 5 delta_(k,2m mod 5), Bmix: S_w(k) = (5/4)(1-delta_(k,0)), ``` where `Bmix` is the uniform mixture over nonzero `m`. For every distinct pair of the six extremal laws the minimal separating degree is exactly four, and the separator set is exactly the twenty monomials `v_x^2 conjugate(v_y)^2`, `x != y`. Their values are zero under `A` and `z^(2m(x-y))` under `B_m`, so they read the full `Z/5` family index. At degree five the separators from `A` are exactly the ten fifth powers `v_x^5` and `conjugate(v_x)^5`, with values one under `A` and zero under every `B_m`; no two `B` laws separate at any odd degree. For `u in (Z/5)^x`, the site action `rho_u` sends `m` to `u^-1 m`, the pointwise coefficient action `gamma_u` sends `m` to `u m`, and the two actions commute. Their diagonal fixes every extremal law and `Bmix`. Consequently `B_(3k_0 mod 5)` is diagonally invariant and has its squared-readout peak exactly at `k_0`. Finally, any zero-mean state with `P_xx = 0` and `|v_x|^2 = a^2` almost surely at `a > 0` has `E[|v_x|^4] = a^4`, whereas Isserlis closure requires `2a^4`; hence `K_xx = -a^4` and no Gaussian or Wick closure exists at fixed modulus. All arithmetic is exact in `Q(zeta_5)`. This theorem neither chooses a state nor supplies a normalization, physical tensor identification, source, propagation law, helicity, detector response, observable or numerical `r_T(k)`, and it makes no L2--L6 claim. The inherited WKB3 ringdown grade has no public source or reproducing test and is therefore not a Canon claim. Stage A, the Schwarzschild TT endpoint: the Regge-Wheeler coefficients (1, 0, -3) are forced at scope, V_2 = f (L/r^2 - 6M/r^3) (SCHWARZSCHILD-TT-ENDPOINT) [T at the displayed family scope]; no wider uniqueness theorem is claimed. Stage B uses the explicit dictionary inputs mu = 1 and Z_L2 = 1/2 (TT-QUADRATIC-GERM [D]); neither the action germ, a Gaussian-state boundary, nor a Stage B pullback is derived by that bookkeeping identity. No numerical mu corridor is retained without a public shadow-to-mu inference rule. The emission map and the quasinormal mu decision after such a rule remain open (TT-SOURCE, QNM-LEAVER-MU). ## 15. Couplings, instruments, and metrology The density against the Gram form is rho_psi = psi psi^dagger G / (psi^dagger G psi), with trace exactly one, and the Born value is the branch G norm (COUPLINGS-DETERMINE [T at this finite identity scope], reproduce/coupling-metrology). No instrument-uniqueness theorem or gyron-carrier no-go is asserted by this identity. The covariant canonical form: the dressing coefficient is the DeWitt norm, 12 = d(d + 1) at lambda = -1, and the level 1 to level 2 normalization inheritance is closed, the chain of twelves exact (DEWITT-TWELVES) [T at scope]. No general metrological-admissibility theorem is asserted. Public Canon v22 replaces the former compound row by a typed reduction parent, one higher-rank finite-state child, and an exhaustive residual row, all at status O and all STOP. The tick clause alone is closed dimensionless: delta tau hat = 1/5 cycle = 2 pi/5 per tick (METRO-TICK) [T at scope]; the remainder is the canonical selector on the commutator phi ladder and the SI clause (METRO-EDGE-SCALE). The dressing insertion bookkeeping carries the exact witness 72 alpha^4 (about 0.204 ppm, labeled) with the form decision gated on the integer crossing count (DRESS-CROSSCOUNT). No end-to-end Lorentz closure is asserted; the former compound A2/A3/K6 row is retired until its operators can be registered as separate gates. `QUADRATIC-DECODER-DATA` [O] asks for a publicly typed action on data; no unregistered closure of state-update, Gram, dagger, or data-effect clauses is asserted. The rational finite-state boundary itself is exact (METRO-FINITE-STATE-RATIONALITY [T] at L5). Let `q >= 2`, let `S` be the finite state set of an accessible `q`-DFAO with transition function `delta`, and let `w in Q^S` be its output vector. Define the integer transition-count matrix ``` B[s,t] = #{d in {0,...,q-1} : delta(s,d) = t}, P = B/q. ``` Every row of `B` sums to `q`, so `P` is a nonnegative rational row-stochastic matrix and `||P^m||_infinity = 1` for every `m >= 0`. Consequently eigenvalue 1 of `P`, equivalently eigenvalue `q` of `B`, is semisimple: a nontrivial Jordan chain there would make `||P^m||_infinity` grow at least linearly. Write the rational characteristic polynomial as ``` chi_B(x) = (x - q)^r h(x), h(q) != 0. ``` Bezout gives `u,v in Q[x]` with ``` u(x)(x - q)^r + v(x)h(x) = 1. ``` Thus `E = v(B)h(B)` is the rational `q`-primary spectral projector. The preceding semisimplicity gives ``` E^2 = E, BE = qE, E 1 = 1. ``` If the normalized uniform-word averages converge entrywise, ``` P^m w = q^(-m) B^m w -> L 1, ``` then `E P^m w = Ew` for every `m`, while passage to the limit gives `Ew = E(L 1) = L 1`. Since `E` and `w` are rational and `S` is nonempty, `L` is rational. This is conditional rationality of an already existing common limit only. It supplies no convergence criterion, discrepancy bound, selector theorem, Folner equivalence, cross-layer lift, or physical unit. A matrix whose row sums are not `q` is outside the declared `q`-DFAO input schema; it is not a counterexample to the theorem. The v22 metrology split is definition-only and obeys four binding rules. First, let `U_RF` be the L5 class of tuples ```text P = (q,a,r,S,A0,{delta_(i,u)},enc_q,w), q >= 2, a >= 1, r >= 1, S finite, empty != A0 subseteq S, delta_(i,u): S -> S, i in {1,...,a}, u in {0,...,q-1}, w: S -> Q^r. ``` The tuple has ordered input and output bases and a fixed digit and padding convention recorded by `enc_q`. No commutation, normalization, or L6 lift is built into `U_RF`. Let `S_reach(P)` be the closure of `A0` under every single-digit map `delta_(i,u)`, equivalently its `Sigma*` orbit. Iterated tuples of coordinate digit words give the same orbit, but a single tuple need not. For `v = (v_1,...,v_a)`, coordinate 1 acts first in `delta_v = D_a(v_a) o ... o D_1(v_1)`, and the pointwise L5 stream is `Stream_P(s,n) = w(state_P(s,n))`. A reduction arrow `R: P -> P'` is typed by an exact predicate `Pre_R(P)`, transports of allowed starts and input indices, and an exact rational output transport `tau_R` taking `w` to `w'`. An arrow is admitted only when `Pre_R` holds and the transported L5 streams intertwine pointwise for every allowed start and input. The declared allowed arrows are: 1. state relabeling by a bijection `phi: S -> S'` transporting `A0`, every digit map, and `w' = w o phi^(-1)`; 2. restriction to `S_reach(P)`; 3. the multi-action Nerode quotient `s ~_V t iff w(delta_v(s)) = w(delta_v(t))` for every tuple `v` of coordinate digit words. It is automatically a congruence for coordinate 1; the exact finite precondition is congruence for every digit map in coordinates `i >= 2`. Under that precondition the quotient maps and output `w'([s]) = w(s)` are well defined; 4. coordinate permutation transporting the ordered input basis, digit maps, input indices, and boxes while leaving the ordered output basis fixed. Write `P approx_red P'` exactly when `P` and `P'` are joined by a finite zig-zag of admitted allowed arrows. This is the generated reduction equivalence. Flattening the `N^a` geometry, erasing named coordinate digit-word actions, arbitrary factor weights, output-dependent regrouping, and replacing boxes by an unrelated ordering are forbidden. Common `q^k` blocking remains undecided unless all length-`k` words, padding, every exponent-residue vector, the pointwise stream, the decision, and the terminal value are transported exactly. METRO-REDUCTION-ARROWS [C], evidenced by the immutable `probes/P-METRO-REDUCTION-ARROWS-4` bundle, registers the exact preconditions and transports of these four arrows and their `tau_R = identity` pointwise L5-stream equality, closing obligations A and C only. Its frozen four-state witnesses distinguish the one-shot tuple image from the `Sigma*` closure, exhibit the necessary higher-coordinate congruence proviso, and separate a transported-basis permutation from the basis-fixed lookalike. Exact exhaustion found zero congruence counterexamples among exactly 1,024 two-state and 4,251,528 three-state protocols in the frozen `q = 2, a = 2, r = 1` binary-output family. The local and GitHub records are both x86_64, so this computation-only result is C, not T. METRO-REDUCTION-CALCULUS [O] remains the parent for obligations B, D, and E: complete exact witnesses for the forbidden catalogue; common `q^k` blocking with all length-`k` words, padding, exponent-residue vectors, the pointwise stream, scientific decision, and terminal value transported; and completeness of `approx_red` for the registered class. It remains STOP. Neither row owns normalization, L6, or another cross-layer gate. Second, METRO-ADMISSIBILITY-DIM [O] is the child on an `N^a`-indexed commuting digit-word system, not an additive action. Its input has ```text q >= 2, a >= 2, r >= 1, S finite, empty != A0 subseteq S, delta_(i,u): S -> S, i in {1,...,a}, u in {0,...,q-1}, W5: S -> Q_(>=0)^r, ``` with an ordered input basis, a separately ordered output basis, a fixed digit and padding convention, exact leading-zero behavior, and commuting actions between distinct coordinates. The L5 endpoint is the complete raw stream ```text L5Stream_s(n) = W5(state(s,n)) in Q_(>=0)^r, n in N^a. ``` The L6 endpoint is the tagged exact space ```text Y_r = ZERO | PROBABILITY(Delta_r(Q)), Normalize(y) = ZERO if sum(y) = 0, PROBABILITY(y/sum(y)) otherwise, ``` with tagged equality and the maximum metric inside the probability tag. For every allowed reachable start and translated box `R(t,N) = product_i {t_i,...,t_i+N_i-1}`, form the normalized average `A_R(s)`. The scientific decision `Adm_direct(P)` is INADMISSIBLE or the unique `ADMISSIBLE(L)` for which one exact algorithm supplies an effective modulus uniform in every translation, allowed start, and box with `min_i N_i` large enough. `Cert_joint(P,c,d)` is an independently defined proof relation, not a second admissibility law. Its schema must cover the relevant invariant submodule, simultaneous primary and peripheral data, individual digit maps, the allowed-start observable quotient, Jordan and terminal sectors, q-adic boundary and residue data, and an effective translated-box modulus. Soundness, completeness, and decision coherence require ```text Adm_direct(P) = d iff exists c: Cert_joint(P,c,d), ``` with one terminal `L` across all valid admissible certificates. The child alone owns GATE-L5-L6-METRO-NORMALIZATION and is STOP until the stream semantics, complete certificate schema, exact checker, reductions, all-parameter proof or fully bounded finite surface, translated-box theorem, and effective modulus are public. The fixed-length two-state factorwise counterexample is an inline control against replacing joint convergence by factorwise power convergence; it is not a child theorem or architecture report. Third, let `C_dim` be exactly the subclass of `U_RF` typed by METRO-ADMISSIBILITY-DIM [O]. Residual METRO-ADMISSIBILITY [O] owns exactly: ```text R1 = U_RF \ C_dim: every rank-one finite rational protocol, every higher-rank finite rational protocol with noncommuting coordinate actions, and every finite rational source or readout not of the child's typed form; R2 = non-finite-state streams; R3 = unbounded-memory adaptive protocols; R4 = stochastic protocols without an exact reduction into U_RF; R5 = irrational carriers or readouts; R6 = cross-layer normalization outside C_dim; R7 = physical units; R8 = every protocol carrying two or more of R1 through R7. ``` This is the exhaustive ownership cover of the registered METRO universe. The row is STOP until R1 through R8 each has a typed child and decision condition, and it closes only when all eight children close. Fourth, METRO-FINITE-STATE-RATIONALITY [T] is a base dependency only. It proves conditional rationality of an already existing one-dimensional limit inside part of R1. It proves neither convergence, R1 closure, the reduction calculus, nor the higher-rank child. ## 16. p = 5 and the wall The retained public root selector is exact: for a positive prime p, (p - 2)/(p + 1) = 1/2 if and only if p = 5 (P5-ROOT-SELECTION [T], reproduce/pentit-p5-closure), since clearing the positive denominator gives p - 5 = 0. No further coincidences are claimed as independent support for selecting p = 5. The two logarithmic axes: pi on the argument (c odd), ln phi on the modulus (c even, transcendental by Baker); linearly independent over the algebraic numbers (LOG-AXES-INDEPENDENCE [T]). No algebraic independence claim is made. The exact second rung of the archimedean wall is the following real-part identity (WALL-LI2-RUNG [T]). Let `sigma_a(zeta_5) = zeta_5^a` for `a in {1,2,3,4}`. For the principal branches, import the classical identities ``` Li_2(z) + Li_2(1 - z) = pi^2/6 - Log(z) Log(1 - z), Re Li_2(e^(i theta)) = pi^2/6 - pi theta/2 + theta^2/4, 0 <= theta <= 2 pi. ``` At `z = sigma_1(J) = phi^-1 e^(2 pi i/5)`, one has `1 - z = e^(-pi i/5)`, ``` Log(z) = -ln phi + 2 pi i/5, Log(1 - z) = -pi i/5, Re(Log(z) Log(1 - z)) = 2 pi^2/25, Re Li_2(1 - z) = 23 pi^2/300. ``` Therefore ``` Re Li_2(sigma_1(J)) = pi^2(1/6 - 2/25 - 23/300) = pi^2/100. ``` At `z = sigma_2(J) = phi e^(-pi i/5)`, one has `1 - z = e^(3 pi i/5)`, ``` Log(z) = ln phi - pi i/5, Log(1 - z) = 3 pi i/5, Re(Log(z) Log(1 - z)) = 3 pi^2/25, Re Li_2(1 - z) = -13 pi^2/300. ``` Therefore ``` Re Li_2(sigma_2(J)) = pi^2(1/6 - 3/25 + 13/300) = 9 pi^2/100. ``` Complex conjugation gives the same real parts for `a = 4` and `a = 3`, respectively. Hence ``` Re Li_2(sigma_a(J)) = pi^2/100 for a in {1,4}, Re Li_2(sigma_a(J)) = 9 pi^2/100 for a in {2,3}, sum_(a=1)^4 Re Li_2(sigma_a(J)) = pi^2/5, expanding / contracting = 9, (pi^2/5) / zeta(2) = 6/5, pi^2/5 - zeta(2) = pi^2/30. ``` This is a Galois-orbit real-part sum, not a field trace of the transcendental `Li_2` values. It asserts nothing about their imaginary parts, a substrate coupling, the Schwinger coefficient, an action, normalization, regularization, physical observables, or uniqueness. The closed form behind this rung holds on every root circle (WALL-CIRCLE-LEMMA [T]). For every integer `N >= 3` and `1 <= a <= N - 1`, put ``` z = 1 + zeta_N^a, alpha = Arg(1 - z) = pi(2a - N)/N, psi = -alpha = pi(N - 2a)/N. ``` Since `1 - z = -zeta_N^a`, one has `|1 - z| = 1`; moreover `-pi < alpha < pi`, so the principal logarithm is unambiguous and ``` Li_1(z) = -Log(1 - z) = -i alpha = i psi. ``` There is one direct midpoint case. If `2a = N`, then `z = 0`, and the defining values `Li_1(0) = Li_2(0) = 0` give the claimed formula without using `Log(z)` or `Arg(z)`. Assume now `2a != N`. Use the same Euler reflection and unit-circle boundary identities displayed above. Represent `1 - z = exp(i theta)` by ``` theta = alpha if alpha > 0, theta = alpha + 2 pi if alpha < 0, ``` so `0 < theta < 2 pi`. The strict principal range for `alpha`, not the later algebraic reduction, selects this representative. From ``` z = 1 - exp(i theta) = -2 i sin(theta/2) exp(i theta/2) ``` one gets `Arg(z) = theta/2 - pi/2`. Taking real parts in Euler reflection therefore gives ``` Re Li_2(z) = alpha Arg(z) + pi theta/2 - theta^2/4 = alpha^2/4 = psi^2/4 = pi^2 (N - 2a)^2/(2N)^2 ``` in both sign cases. Together with the direct midpoint case this proves the formula without excluding any `a`. Exact summation of the rational coefficients, ``` sum_(a=1)^(N-1) (N - 2a)^2 = N(N - 1)(N - 2)/3, ``` yields the full nontrivial-root sum ``` sum_(a=1)^(N-1) Re Li_2(1 + zeta_N^a) = pi^2 (N - 1)(N - 2)/(12N). ``` For `N = 5`, reducing the exponent in `sigma_a(J) = 1 + zeta_5^(2a)` reproduces exactly the registered values `pi^2/100` for `a in {1,4}` and `9 pi^2/100` for `a in {2,3}`, with Galois-orbit real-part sum `pi^2/5` and channel ratio 9. For composite `N` the displayed full nontrivial-root sum is not called a field trace. The theorem is about real parts only and adds no imaginary-part, substrate, action, normalization, regularization, physical-observable, or uniqueness claim. The pentagon root-filter normalization is exact (PENTAGON-NORMALIZATION [T], probes/P-PENTAGON-WEIL-1). Let `j = zeta_5 = (J - 1)^3` and, for `Re(s) > 1`, put ``` c(n) = sum_(a=1)^4 j^(a n) = 5[5 divides n] - 1, P_0(s) = sum_(n>=1) c(n)n^(-s), f_5(s) = 5^(1-s) - 1. ``` Absolute expansion of the four polylogarithms gives `P_0(s) = f_5(s) zeta(s)`. With the classical meromorphic continuation of `zeta` and its standard completion explicitly imported, define ``` Z_J(s) = MerCont_(Re(s)>1)(P_0(s)/f_5(s)) = zeta(s), xi_J(s) = (1/2)s(s-1)pi^(-s/2)Gamma(s/2)Z_J(s) = xi(s). ``` The zeros `s_k = 1 - 2 pi i k/log 5`, `k in Z`, of `f_5` are root-filter artifacts removed by the normalization; the pole and trivial zeros are handled by the imported standard zeta completion. In natural series order, `P_0(1) = -log 5`. On `Re(s) > 1`, ``` coeff(-P_0'/P_0)(n) = Lambda(n) - (log 5)n[n = 5^m], ``` so the artificial `5^m log 5` tower is subtracted, not read as prime data. The unnormalized standard completion is `Xi_raw(s) = f_5(s)xi(s)` and does not have a constant unit-modulus root number: exactly `f_5(2)/f_5(-1) = -1/30`. This is a normalization identity. Meromorphic continuation, the zeta functional equation and divisor, the standard `xi` completion, the Euler product, and the treatment of the pole and trivial zeros are named classical imports, not results of the verifier. No Weil test space, positive form, operator realization, or statement about RH follows. The standard finite-dimensional toral Haar-Koopman carrier cannot realize the full Li norm ladder (J-LI-TORAL-HAAR-NOGO [T], probes/P-J-LI-TORAL-HAAR-1). For every `d >= 1` and every `A in GL_d(Z)` with no root-of-unity eigenvalue, let `U_A f = f o T_A` on `L^2(T^d, Haar)`. There is no vector `v` such that ``` || sum_(k=0)^(n-1) U_A^k v ||^2 = lambda_n ``` for every `n >= 1`, where `lambda_n` is the standard Li sequence. The proof is by contradiction. An exact realization makes every `lambda_n` nonnegative, so Li's criterion forces RH inside the argument. On that forced branch, for each positive ordinate `gamma` of a nontrivial zero, with multiplicity `m_gamma`, put ``` theta_gamma = 2 arctan(1/(2 gamma)), sigma_xi = sum_(gamma>0) m_gamma/(gamma^2 + 1/4) (delta_(exp(i theta_gamma)) + delta_(exp(-i theta_gamma))). ``` The standard Li zero-sum formula, second differences, and Fourier uniqueness then force only the symmetrized vector spectral measure `mu_v + iota_* mu_v`, with `iota(z) = conjugate(z)`, to equal the purely atomic `sigma_xi`. The toral character law `U_A e_m = e_(A^T m)` has no finite nonzero orbit under the carrier hypothesis: a repeat would make an eigenvalue of `A` a root of unity. Every nonconstant orbit is therefore a bilateral-shift sector, whose vector measures have no atoms. This contradicts the atoms of `sigma_xi` away from `1`. The contradiction is unconditional for the assumed realization. RH is a forced intermediate consequence, not an assumption or conclusion. The standard Li coefficients and zero-sum normalization, Li's criterion in both directions, convergence of the fixed-`n` zero sum under RH, Fourier uniqueness, the unitary spectral theorem including its atom/eigenspace correspondence, and the toral character/bilateral-shift decomposition are named classical imports. The proof is all-`n`; finite fits are not excluded. For the named TWIST-J specialization, `d = 4` and `A = M_J`; its exact characteristic data and eigenvalue moduli `phi, phi, phi^-1, phi^-1` place it in the root-of-unity-free class. Gauss's lemma and the standard cyclotomic minimal-polynomial facts used by the probe's alternative exact exclusion audit are named imports. The Riemann-von Mangoldt law and Stieltjes partial summation belong only to the probe's conditional accumulation audit; that asymptotic is not promoted here. This does not exclude matrices with root-of-unity eigenvalues, non-Haar, non-Koopman, infinite-dimensional, boundary, scattering, or enlarged carriers. No replacement realization, moment or cocycle bridge, Weil-positive form, decoder, physical lift, or RH result is asserted. The unsymmetrized `mu_v` is not claimed unique. The conditional spectral target is a proof step, not a separately registered claim. This carrier exclusion does not alter or falsify the public algebraic and finite-periodic statements about `M_J`. These three verdicts delimit three declared carrier classes. They do not complete the plenum's single-unitary map, because the compact-boundary cocycle-vector route remains open. - E8 functional-calculus carrier. Here `det_1` is the ordinary trace-class Fredholm determinant `product_j (1 - t^2 a_j)`, with no exponential regularization. If `det_1(I - t^2 A) = Xi(t)/Xi(0)` for positive trace-class `A = f(Delta_E8)`, equality of zero multisets forces the nonzero spectrum `gamma^-2` with zeta-zero multiplicities. The weakened explicit Riemann-von Mangoldt bound gives `N(200) >= 68` and multiplicity at most 18 for each ordinate up to 200, hence at least four distinct small-multiplicity eigenvalues. But `Theta_E8 = E4` makes every nonzero E8 shell have size `240 sigma_3(n) >= 240`; only the one-dimensional zero shell can supply a smaller multiplicity. Therefore no operator in the frozen functional-calculus class realizes the determinant (J-LI-E8-SHELL-MULTIPLICITY-NOGO [T], probes/P-MCKAY-THETA-CARRIER-1). MCKAY-THETA-FUNCTIONAL-CALCULUS-CARRIER [F] records exactly that fired route. Non-shell-constant operators and the use of the theta identity as a positivity source are not excluded. - lambda-adic boundary Koopman, Hilbert-Schmidt subroute. For any unit `u` and any `n >= 1` with `u^n != 1`, multiplication by `u` leaves every additive character above the finite level `v_lambda(u^n - 1)` non-fixed. Thus `I - U_u^n` has infinite Hilbert-Schmidt norm. For `J = 1 + zeta_5^2`, the complex embedding `zeta_5 = exp(2 pi i/5)` gives `|J| = phi^-1 != 1`; hence `J` is not a root of unity and the divergence holds for every `n >= 1`. The spectrum is pure point on roots of unity with unbounded multiplicity, and its frozen order tower is `4, 20, 20, 20, 20, 20, 100, 100`, independently re-derived in the Eisenstein model `x^4 - 5x^3 + 10x^2 - 10x + 5` of discriminant 125 (J-LI-LAMBDA-HAAR-HS-NOGO [T], probes/P-R2-LAMBDA-HAAR-1). LAMBDA-BOUNDARY-HS-KOOPMAN [F] closes only the Hilbert-Schmidt-perturbation and S2 forms. It does not close the cocycle-vector form LAMBDA-COCYCLE-ANGLES [H]. - lambda-adic scaling shift. The value group of `K_lambda^*` is `Z`. The unitary `U f(x) = 5^(-1/2) f(lambda x)` maps the valuation shell `S_n` to `S_(n-1)` and is a bilateral shift of infinite multiplicity and homogeneous Lebesgue type. Therefore every vector spectral measure is absolutely continuous, incompatible with the purely atomic measure forced by an exact all-n Li cocycle realization; `I - U^n` is non-Hilbert-Schmidt for every `n >= 1`, and every declared discrete-time tensor composite `U tensor V` remains absolutely continuous by convolution. For the boundary vector the ladder increments increase strictly toward `phi^2` without attaining it. After the one free rescale `|c|^2 = lambda_1`, the second rung `lambda_1 (2 + 2r)` is disjoint from `lambda_2 = 4 lambda_1 - M_1` by exact intervals (J-LI-LAMBDA-SHIFT-NOGO [T], probes/P-R2-SCALING-SHIFT-1). LAMBDA-DISCRETE-SCALING-SINGLE-UNITARY-CARRIER [F] records this exact discrete-time class and its declared unitary tensor composites. It does not cover distribution or trace formalisms, moment-functional constructions, or genuinely global adelic objects. The first completed-zeta K-side rung is pinned at finite enclosure scope (O-R2-K-JUNCTION-PIN [C], probes/P-R2-K-JUNCTION-PIN-1). For `K = Q(zeta_5)`, define ``` lambda_1^K = 1 + (3/2) log 5 - 2 log(2 pi) - gamma + sum_(chi != chi_0 mod 5) L'/L(1,chi), lambda_1^Q = 1 + gamma/2 - (1/2) log(4 pi). ``` Exact rational interval arithmetic gives the computed enclosures ``` lambda_1^K in [0.304595618542798635262524662701, 0.304595618542798635262524662702], sum_(chi != chi_0 mod 5) L'/L(1,chi) in [1.143408547611871901089216, 1.143408547611871901089217], lambda_1^Q in [0.023095708966121033814310, 0.023095708966121033814311]. ``` The identity `(3/2) log 5 = 6 log s_J + 3 log phi` is preserved interval by interval. The literal modulus-five principal-character convention is a disjoint negative control at `[0.706955096651323728912714, 0.706955096651323728912715]`. This computed pin constructs no R2 carrier, proves no higher Li rung, and makes no RH claim. The E8 multiplicity obstruction, the compact-boundary Hilbert-Schmidt obstruction, and the scaling-shift spectral-type obstruction are now closed only at those declared scopes. In particular, the compact lambda-adic boundary cocycle-vector route remains LAMBDA-COCYCLE-ANGLES [H]. The program also continues in the moment-functional (Weil positivity) frame and in genuinely global constructions. RH remains O; none of these results proves, assumes, or falsifies RH. LAMBDA-COCYCLE-BRANCH-COLLAPSE [T]. On the critical line the Cayley factor is not merely of modulus one, it is the Cayley-angle unit itself: ```text 1 - 1/rho = e^(i alpha_gamma), alpha_gamma = 2 arctan(1/(2 gamma)), 1/rho^2 = -(1 - 1/rho)/(1/4 + gamma^2). ``` Apply the single identity `X^(n+1) + X^(n-1) - 2 X^n = X^(n-1)(X-1)^2` to the Bombieri-Lagarias representation. Inside the declared class, write `t_n = lambda_(n+1) + lambda_(n-1) - 2 lambda_n`, `M = 2 lambda_1`, and `n_A = 4 . 5^A`; then ```text M - t_n = sum_(gamma>0) 4 sin^2(n alpha_gamma / 2)/(1/4 + gamma^2), M = 2 lambda_1 = sum_(gamma>0) 2/(1/4 + gamma^2). ``` Every summand is nonnegative and bounded, so `0 <= M - t_n <= 2 M` is a necessary consequence of membership at every `n`. These are pointwise necessary bounds only. Through this residual-bound test, a finite family of exact Li values or rigorous interval enclosures can contradict membership only by proving one of the displayed inequalities false. Satisfying any finite family of them has no converse implication: it neither constructs a cocycle vector nor proves that the finite profile is realizable. No general finite-profile nonfalsifiability or realization theorem is claimed. If `M - t_(n_A)` fails to tend to zero, then for some `delta > 0` it stays at least `delta` on infinitely many `A`. Convergence of the majorant gives a finite ordinate window whose tail is below `delta/2`, and a finite pigeonhole argument gives one fixed ordinate in that window for which `dist(n_A alpha_gamma/(2 pi),Z)` stays bounded away from zero for infinitely many `A`. LAMBDA-COCYCLE-GRID-EQUIVALENCE [T]. The grid in that row is not an assumption about zeta; it is the point spectrum of the operator. Multiplication by `J = 1 + zeta_5^2` is a Haar-preserving automorphism of `O_lambda`, so `U_J` permutes the character basis by `y -> J y`. The trivial character `y = 0` is fixed. It is the unique length-one orbit because `J - 1 = zeta_5^2` is a unit. Every nontrivial character has an exact positive level `k`, and its orbit has size `ord_(lambda^k)(J)`. The residue of `J` modulo `lambda` has order four, so every such order is divisible by four; since `|(O/lambda^k)^x| = 4 . 5^(k-1)`, Lagrange's theorem makes every nontrivial orbit length `4 . 5^a`. Moreover, ```text ord_(lambda^(4m))(J) = 4 . 5^m, m >= 1. ``` Thus every `4 . 5^a` occurs: `n_0 = 4` occurs at level one, and `n_A = 4 . 5^A` occurs at level `4A` for `A >= 1`. The complete cycle-length set is `{1} union {4 . 5^a : a >= 0}`. A cycle of length `d` contributes exactly the `d`-th roots of unity. The length-one cycle contributes only eigenvalue `1`, whose angle zero already lies in the grid, so `U_J` has pure point spectrum with eigenvalue angle set exactly `2 pi (1/4) Z[1/5]`, every angle attained. The tail/grid step is an all-real statement. For `x in R/Z`, ```text dist(4 . 5^A x, Z) -> 0 if and only if x lies in (1/4) Z[1/5] modulo Z. ``` Put `y = 4x`. Choose nearest integers `m_A` and signed errors `e_A = 5^A y - m_A`, so `|e_A| -> 0`. The quantity `m_(A+1) - 5m_A = 5e_A - e_(A+1)` is an integer tending to zero, hence it vanishes for every sufficiently large `A`. Therefore `e_(A+1) = 5e_A` eventually. Since `e_A -> 0`, this forces `e_A = 0` eventually, so `y` lies in `Z[1/5]`. The converse is immediate because a sufficiently large power of five annihilates the denominator. Combined with the finite-window localization above, this proves that failure of the tail limit `t_(n_A) -> M` is exactly the angle-grid obstruction. Conversely, suppose the Riemann hypothesis holds and every Cayley angle lies on that grid. Define a symmetric atomic measure `mu` by assigning mass `(1/2)/(1/4 + gamma^2)` to each of `+alpha_gamma` and `-alpha_gamma`, with multiplicity, and combine masses when atoms coincide. Then `mu(R/(2 pi Z)) = lambda_1`, and every atom is an eigenvalue angle. Choose one unit eigenvector for each distinct atom and weight it by the square root of the combined mass; their orthogonal sum is a vector `v` of squared norm `lambda_1`. Put `f_n = ||sum_(k=0)^(n-1) U_J^k v||^2`. Then `f_0 = lambda_0 = 0` and `f_1 = lambda_1`, while the atomic construction gives `f_(n+1) + f_(n-1) - 2f_n = t_n` for every `n >= 1`. Thus the two sequences agree identically. Membership in the cocycle class is therefore equivalent to RH together with the grid condition, equivalently to every nontrivial zero having the form `rho = 1/(1 - xi)` with `xi != 1` and `xi^(4 . 5^a) = 1` for some integer `a >= 0`. The wall is one archimedean wall. Its exact second rung is WALL-LI2-RUNG [T], the real-part identity above. The quantum substrate gates remain separate: the Larmor clause and the Schwinger physical realization carry the exact arithmetic target J Jbar / script-Q = 1/(2 pi) (QUANT-SCHWINGER-TARGET [T]), while deriving that scalar as the first-order electron coefficient from a typed substrate coupling remains open inside QUANT-SUBSTRATE [O]. WALL-LI2-RUNG does not supply that coupling. The non abelian measure lift (COLOR-MEASURE-SELECTION) and the shared 2 pi U(1) circle remain at their registered scopes. ## 17. Engineering witnesses Unregistered engineering readouts are excluded from the normative Canon. Their non-canonical disposition record is `notes/ENGINEERING.md`; nothing in that note is evidence for a public claim. The standing bar for a computation-only theorem is byte-identical stdout on two architectures through the public activation gate. ## 18. The frontier The live obligations and hypotheses of the program. Each identifier is a registry row with status O or H; the registry carries for every row a concrete falsifier or a decision condition: what closes it positively and what closes it negatively. The COIN-MINIMAL-READ [D] paragraph below is status-separation context for its live O owner, not a frontier row. - LAMBDA-COCYCLE-ANGLES [H]. The compact lambda-adic boundary route remains open only in cocycle-vector form: an exact realization would require a vector `v in L^2(O_lambda,Haar)` with `||sum_(k=0)^(n-1) U_J^k v||^2 = lambda_n` for every `n >= 1`. Such a vector exists exactly when the Riemann hypothesis holds and every Cayley angle `2 arctan(1/(2 gamma))` lies in `2 pi (1/4) Z[1/5]`, equivalently when every nontrivial zero is `rho = 1/(1 - xi)` with `xi != 1` and `xi^(4 . 5^a) = 1` for some integer `a >= 0`. The hypothesis is fired by a disproof of the Riemann hypothesis or by one exact arithmetic exclusion of a nontrivial zero from that form. With `t_n = lambda_(n+1) + lambda_(n-1) - 2 lambda_n`, `M = 2 lambda_1`, and `n_A = 4 . 5^A`, failure of the tail limit `t_(n_A) -> M` localizes to one such off-grid ordinate. The pointwise inequalities `0 <= M - t_n <= 2 M` are necessary only: an exact finite violation contradicts membership, while satisfaction at finitely many indices does not decide the row. No general finite-profile nonfalsifiability or realization theorem is claimed. The tail localization and the all-real tail/grid equivalence are supplied respectively by LAMBDA-COCYCLE-BRANCH-COLLAPSE [T] and LAMBDA-COCYCLE-GRID-EQUIVALENCE [T] in section 16. COIN-MINIMAL-READ [D] is the adopted L1 dictionary premise that selects `beta_1` from the complete integer-admissible pair by the two distinct minimum-cost criteria proved in COIN-SELECTION-CONDITIONAL [T]. MAXIMAL-REACH remains the named unadopted alternative and selects `beta_3`. The dictionary does not say that the decoder architecture forces this choice. MINIMAL-READ-DERIVATION [O] is the corresponding open L5-to-L1 decision. It requires a complete typed decoder class, cover-to-output map, accumulator/equality rule, redundancy theorem, dependency graph, and layer typing. A complete derivation uniquely forcing `w=1` and `beta_1` closes it positively; a complete nonempty class supporting both coins, or uniquely forcing `beta_3`, closes it negatively. Missing structure or failure of only one proposed route is STOP. The v36 ledger change is signed term by term: ```text claims: 220 + 1 Kappa falsification row = 221, T: 118, D: 41, C: 24, H: 3, all unchanged, O: 24 - 1 falsified compound route = 23, F: 10 + 1 Kappa row + 1 compound route = 12, live H/O: 27 - 1 = 26. ``` Public Canon v36 declares PHOTON-KAPPA-LEMMA [F] at L4 and changes PHOTON-WINDOW-PROOF from O to F. The public two-architecture `P-PHOTON-KAPPA-LEMMA-1` certificate exhibits one admitted connected current with `L=3240` and one ternary filling of support 7993 satisfying `2^7993<=7^3240`. Only `F_occ<=7993` is used. The exact integer exclusion lemma therefore rules out every admissible rational Kappa coefficient from the frozen universal occupancy proposition. The `F` label classifies that failed positive proposition; the exclusion lemma itself is true. The parent was a conjunction. Falsification of its necessary Kappa conjunct falsifies the frozen compound route regardless of the separate roughening question. Issue #201 and electric-face roughening remain open, undecided, and unregistered; no successor is created automatically. KAPPA-SHAPES [C] and MONOPOLE-COST [C] remain unchanged. No filling equality or optimality, pump family, asymptotic family, broader carrier, Froehlich-Spencer import, massless or Coulomb phase, continuum limit, propagator, physical photon, cross-layer lift, or physical measure is claimed. The explanatory README of the older `reproduce/photon-electron` bundle is updated to point to the later public Kappa probe while preserving that reproduction's own exact scope and unchanged verifier. Its 14 existing evidence consumers are re-pinned to the recomputed bundle hash with explicit evidence-change history events. The public Kappa probe supplies the new and parent evidence rows. Mandatory architecture CI replays every public probe and minimal reproduction because this fold changes `canon/`. The v35 ledger change is signed term by term: ```text claims: 220 unchanged, T: 118, C: 24, O: 24, F: 10, all unchanged, D: 40 + 1 dictionary retype = 41, H: 4 - 1 retyped dictionary = 3, live H/O: 28 - 1 = 27. ``` Public Canon v35 reclassifies COIN-MINIMAL-READ from H to D at byte-identical scope and byte-identical falsifier. The registered sentence already adopts `beta_1` by the MINIMAL-READ dictionary on the complete proved pair and does not assert that the decoder architecture forces that choice. The public evidence bundle remains `probes/P-BOOST-COHERENCE-1`; no probe file, probe verifier, or evidence row changes, and no new formal scientific execution or evidence is introduced. The status-separation audit moves with the fold, and mandatory architecture CI replay creates no evidence record. MINIMAL-READ-DERIVATION [O] remains the separate live L5-to-L1 decision and stays O, ROOT, STOP, and owner of GATE-L5-L1-MINIMAL-READ. MAXIMAL-REACH remains named and unadopted. No derivation, closure, theorem, selector, dependency edge, action-layer lift, decoder completion, or physical claim is added. The v34 ledger change is signed term by term: ```text claims: 219 + 1 theorem = 220, T: 117 + 1 abelian CM discriminant-minimum theorem = 118, D: 40, C: 24, H: 4, O: 24, F: 10, all unchanged, live H/O: 28 unchanged. ``` Public Canon v34 registers ABELIAN-CM-UNIQUE-EVEN-BIT-DISCRIMINANT-MINIMUM [T] at L1 on the immutable `probes/P-ABELIAN-CM-UNIQUE-EVEN-BIT-DISCRIMINANT-MINIMUM-1` bundle. The theorem status is earned by the complete owner-accepted proof; the accepted proof and byte-identical x86_64 and aarch64 transcripts audit its seven exact gates, with the full execution-integrity record retained in the evidence bundle. The fold adds only the unique absolute-discriminant minimum in the frozen abelian Galois CM unique-even-bit class and the concise Core synthesis of the two independent "why five" answers. It adds no claim that the class, total ramification, or discriminant minimization is physically selected, no physical-field uniqueness, TWO-PLACE-PHYSICS promotion, decoder, measure, or L2-L6 lift, and no live H or O row moves. The v33 ledger change is signed term by term: ```text claims: 218 + 1 theorem = 219, T: 116 + 1 quartic cyclotomic total-ramification theorem = 117, D: 40, C: 24, H: 4, O: 24, F: 10, all unchanged, live H/O: 28 unchanged. ``` Public Canon v33 registers QUARTIC-CYCLOTOMIC-TOTAL-RAMIFICATION-CENSUS [T] at L1 on the immutable `probes/P-QUARTIC-CYCLOTOMIC-TOTAL-RAMIFICATION-CENSUS-1` bundle. The theorem status is earned by the complete independently accepted proof; the byte-identical x86_64 and aarch64 transcript and the independently frozen blind breaker audit its exact certificates. The fold adds only the complete `phi(n)=4` index and field quotient, discriminants, ramification profiles, total-prime locus, residue-unit groups, and inherited `J` compatibility. It adds no degree-four or broader CM selection, physical-field uniqueness, TWO-PLACE-PHYSICS promotion, decoder, measure, or L2-L6 lift, and no live H or O row moves. The v32 ledger change is signed term by term: ```text claims: 217 + 1 theorem = 218, T: 115 + 1 central-lift phase theorem = 116, D: 40, C: 24, H: 4, O: 24, F: 10, all unchanged, live H/O: 28 unchanged. ``` Public Canon v32 registers CENTRAL-LIFT-PHASE [T] at L4 on the immutable `probes/P-CENTRAL-LIFT-PHASE-1` bundle. The theorem status is earned by the independently accepted proof in its preregistration; the byte-identical x86_64 and aarch64 transcript audits the exact E1-E3 certificates. The fold adds only the projective fifth, normalized Herm/Sym scalar separation, square-root-free J action, central `zeta_5^2` Sym factor, exact unit-phase image `mu_5`, and the `mu_10 \ mu_5` obstruction. It adds no Herm2 cone, rigidity, common carrier, decoder data, U(1), measure, or physical tick, and no live H or O row moves. The v31 ledger change is signed term by term: ```text claims: 216 + 1 theorem = 217, T: 114 + 1 marked-CM semilinear-pair theorem = 115, D: 40, C: 24, H: 4, O: 24, F: 10, all unchanged, live H/O: 28 unchanged. ``` Public Canon v31 registers COLOR-CM-2I-SEMILINEAR-PAIR [T] at L4 on the immutable `probes/P-CM-2I-QCARRIER-1` bundle. The theorem status is earned by the independently accepted proof in its preregistration; the byte-identical x86_64 and aarch64 verifier transcript audits its exact certificates. The result is relative to the displayed marked representative and creates no marked-lift selector, decoder `QCarrier`, U(1) dictionary, or measure lift. No live H or O row moves, and both adjacent STOP obligations remain unchanged. The v30 ledger change is signed term by term: ```text claims: 216 unchanged, statuses: 0 T-LOCK, 114 T, 40 D, 24 C, 4 H, 24 O, 10 F, all unchanged, live H/O: 28 unchanged, evidence: 216 unchanged. ``` Public Canon v30 registers no claim and retires none. It republishes the v29 ledger unchanged so that a tagged tree passes its own activation readback and the release can carry assets. The v29 gap is recorded in `notes/canon/V29-RELEASE-ERRATUM.md` and is not repaired by this version. The v29 ledger change is signed term by term: ```text claims: 215 + 1 computation = 216, C: 23 + 1 renormalization-return computation = 24, T: 114, D: 40, H: 4, O: 24, F: 10, all unchanged, live H/O: 28 unchanged. ``` Public Canon v29 registers ENTROPY-RG-RETURN [C] at L5 on the immutable `probes/P-ENTROPY-RG-RETURN-1` bundle, whose local aarch64 leg and required GitHub x86_64 leg produced the same 13-gate PASS transcript byte for byte. The row is closed at birth inside the declared range and no live H or O row moves. No continuum limit, scaling limit, critical exponent, monotone scale function, measure or all-scale law is registered with it. The v28 ledger change is signed term by term: ```text claims: 214 + 1 theorem = 215, T: 113 + 1 read-redundancy theorem = 114, D: 40, C: 23, H: 4, O: 24, F: 10, all unchanged, live H/O: 28 unchanged. ``` Public Canon v28 registers READ-REDUNDANCY-PRIME-SUPPORT [T] at L5 on the immutable `probes/P-READ-REDUNDANCY-1` bundle, whose local aarch64 leg and required GitHub x86_64 leg produced the same 16-gate PASS transcript byte for byte. The universal classification rests on the short symbolic proof that the verifier audits; the three exhaustion legs corroborate it inside their frozen boxes and establish nothing beyond them. Public Canon v28 also amends the scope and falsifier of TM-SYM2-PHYSICAL-MEASURE, which was O at that release. The row previously required a future bridge to preserve mu_i = 1/6 and M_TM = (1/3)P1 + (2/15)P5 and to derive the typed factorization 1/6 = (1/2)(1/3); those are the values such a bridge would be trying to establish, so requiring them in advance stated the answer before the question. The amendment moves them to the result side and adds the explicit CIRCULAR rejection, following CANON13-SCOPE-DE-CONFORMAL-WEIGHT. The fired N2 boundary, the complete 48-selector class in four free orbits of 12, the retained reading orientation, the prohibition on enlarging gauge, and the three exact classifications are all preserved without change, and the row stays O and STOP. No status or count moves except the history. The v27 ledger change is signed term by term: ```text claims: 210 + 2 theorems + 1 hypothesis + 1 obligation = 214, T: 111 + 2 exact boost/coin theorems = 113, H: 3 + 1 adopted coin premise = 4, O: 23 + 1 derivation owner = 24, live H/O: 26 + 2 = 28. ``` Public Canon v27 registers DRIFT-IS-THE-READ [T] at L5 and COIN-SELECTION-CONDITIONAL [T] at the exact L1/L5 comparison boundary. The first row packages the fixed `beta_1` reflection, division-free spectral skeleton, nonclosing gap, exact drift and range, and all-`N` uniform bound. Its read interpretation is conditional on P1 and P2; neither premise is derived or physically anchored. The second row proves the complete pair `{beta_1,beta_3}`, endpoint-qualified cover and multiplicity statements, exact uniform constants and squared gaps, and the S1/S2/S3 rankings. The same immutable `probes/P-BOOST-COHERENCE-1` bundle is the sole evidence for both theorem rows and for the current D/O split. Its one formal aarch64 execution and required GitHub x86_64 reproduction produced the same 8-group, 477-check PASS transcript byte for byte. No probe file is amended and no `P-BOOST-COHERENCE-1` probe verifier or formal execution is rerun by this fold. COIN-MINIMAL-READ [D] adopts `beta_1` by MINIMAL-READ and records MAXIMAL-REACH as the unadopted selector of `beta_3`. MINIMAL-READ-DERIVATION [O] separately owns the OPEN_SELECTION gate and remains STOP on incomplete decoder, class, map, accumulator, redundancy, graph, and layer data. No premise-to-theorem promotion, P1/P2 anchoring, decoherence, Born, L6, SI, measurement, collapse, environment, or unique physics is added. The absent incubation identifiers `O-COIN-CANONICAL` and `O-DECOHERENCE-CLAUSE` receive no public retirement or history event. TM-SYM2-PHYSICAL-MEASURE and every other v26 row remain unchanged. The v26 ledger change is signed term by term: ```text claims: 208 + 1 L1 theorem + 1 L5 computation = 210, T: 110 + 1 C8 bilinear-shadow theorem = 111, C: 22 + 1 reduction-arrow computation = 23, O: 23 unchanged, live H/O: 26 unchanged. ``` Public Canon v26 registers C8-BILINEAR-SHADOW [T] at its exact L1 boundary. The proof identifies the two nonzero axes of ``, the exact order-eight elements and Frobenius conjugation, proves the all-`n` branch swap and the branch-invariant record, and shows its strict mod-8 refinement of the norm channel. The existing public probe supplies byte-identical aarch64 and GitHub x86_64 audit output. No probe file is amended, and no branch selector, physical gauge equivalence, checkpoint, clock, gravity, SI, force, uniqueness, or L2-L6 statement is introduced. Public Canon v26 also registers METRO-REDUCTION-ARROWS [C] at its exact L5 boundary. The computation freezes the preconditions and transports of the four admitted arrows and their pointwise `tau_R = identity` L5-stream invariance, closing obligations A and C only. Its public prospective-pinned bundle passed all 17 gates and the exact two-state and three-state exhaustion. The local and GitHub records are both x86_64 and byte-identical, so this earns reproduction but not a two-architecture theorem gate and remains C. METRO-REDUCTION-CALCULUS remains O and STOP on obligation B, common `q^k` blocking with decision and terminal-value transport in obligation D, and `approx_red` completeness in obligation E. No L6, cross-layer, physical, or SI claim is added. The v25 ledger change is signed term by term: ```text claims: 208 unchanged, T: 109 + 1 KERNEL-Z6 theorem = 110, O: 24 - 1 closed owner = 23, live H/O: 27 - 1 = 26. ``` Public Canon v25 promotes KERNEL-Z6-SYNCHRONIZATION from O to T at its unchanged L1 boundary. The self-contained proof establishes the fixed-time 5-to-1 and 2-to-1 sheet laws, non-eventual-periodicity of every checkpoint trace and trajectory, and the impossibility of any finite autonomous realization. The existing public probe audits the exact finite premises, complete induction cases, and two implementations; its aarch64 and x86_64 outputs are byte-identical. No new run, physical interpretation, decoder completion, or L2-L6 lift is introduced. The v24 ledger change is signed term by term: ```text claims: 206 + 2 L1 theorems = 208, T: 107 + 2 = 109, live H/O: 27 unchanged, decoder factor-canonicity definition: +0 claims, +0 dependencies, +0 gates. ``` The Gyron-owned transaction registers GYRON-DISCREPANCY-LOG [T] and TM-PAIR-SUBSTITUTION-FIXED-POINT [T], and corrects GYRON-DENSITY [T] without changing its status or the separate dictionary consumers of its value. The public proof bundle passed its sole native aarch64 execution and the first GitHub x86_64 replay byte identically. The corrected density row is re-pinned to that public probe. The decoder-owned transaction extends only `DEF-DECODER-COMPLETION-CONTRACT` with the optional factor-canonicity audit overlay. It creates no decoder-universality row, scientific claim, evidence, dependency, gate, probe, result, or status change. No Gyron-to-decoder dependency is added. Both transactions retain their firewalls. The forward pair-substitution operator is not coarse-graining or a decoder factor, and the decoder schema does not inherit existence, canonicity, maximality, nontriviality, physical meaning, or cross-layer closure from a commuting square. The v23 ledger change is signed term by term: ```text live H/O: 26 + 1 KERNEL-Z6 owner = 27, claims: 205 + 1 = 206, O: 23 + 1 = 24. ``` Public Canon v23 registered KERNEL-Z6-SYNCHRONIZATION as one L1 owner-definition row at status O. It froze the four candidate synchronization clauses and their exclusions but supplies no proof, computation, probe, gate, run, physical reading, or L2-L6 lift. No existing claim changes status; no candidate clause is promoted. The v22 ledger change is signed term by term: ```text live H/O: 25 - 1 retired KC3 - 1 consumed METRO row + 3 typed METRO rows = 26, claims: 203 - 1 KC3 - 1 old METRO + 3 split rows + 1 entropy cursor theorem = 205, T: 106 + 1 = 107, H: 4 - 1 = 3, O: 21 - 1 + 3 = 23. ``` The metrology and STOP-surface subfold changes no scientific status. The only new scientific status in v22 is ENTROPY-CYLINDER-NOGO-CURSOR [T], earned by its pinned two-architecture public probe. ``` MEASUREMENT AND METROLOGY METRO-REDUCTION-CALCULUS typed protocol objects, reduction preconditions, w transport, and generated equivalence; STOP METRO-ADMISSIBILITY-DIM commuting digit-word N^a child with exact raw L5 stream, tagged L6 normalization, direct translated-box decision, and joint certificate; owns the L5-to-L6 gate and remains STOP METRO-ADMISSIBILITY exhaustive R1 through R8 residual: U_RF minus C_dim, including rank-one, noncommuting, and other out-of-child finite rational protocols; non-finite-state, unbounded adaptive, non-reducible stochastic, irrational, out-of-child cross-layer, physical-unit, and mixed-class protocols; STOP METRO-EDGE-SCALE the canonical selector on the commutator phi ladder; the SI clause (the second and the meter over the single m_e bridge) DRESS-CROSSCOUNT the integer crossing count per observable; witness 72 alpha^4, labeled QUADRATIC-DECODER-DATA the typed quadratic/Born D_matter action and its exact factorization through Q; carrier, bridge, Gram, dagger, transpose, QCarrier equality, effects, MatterData schema, write map, domain, and complete dependencies remain open; linear, binary, reconstruction, and post-state instrument claims are outside this row QDD-INSTRUMENT-NONSELECTION the L4 rational fibre, diagonal-orbit and dilation classification; fixed effects, weights and C = 0 leave infinitely many physical post-state classes, and dilation existence does not select an instrument QDD-INSTRUMENT-APPARATUS only O2 independent physical instrument selection and O1 realized event generation / sampling remain; SAMPLING NOT PROVIDED, not impossible; fills no completion-contract field THE WALL QUANT-SUBSTRATE the Larmor gate and the Schwinger physical-realization gate; the target scalar is exact arithmetic and its production as the first-order coefficient remains open COLOR-MEASURE-SELECTION the typed canonical normalized L4-to-L6 lift; target carrier, equivalence, observables, action, normalization, and constraints are STOP GRAVITATIONAL WAVE TT-SOURCE the emission map QNM-LEAVER-MU the quasinormal mu decision TT-VECTOR-STATE-NORMALIZATION the only gate yielding a numerical r_T(k) COSMOLOGY FRW-INHOM the inhomogeneous sector, the named classical horizon NS-TILT n_s - 1 = -5 alpha; falsifier live, CMB-S4 DE-CONFORMAL-WEIGHT the typed homogeneous dictionary, if any, selecting the dark-energy density character; FRW continuity alone is nonunique and no trace-Gram dictionary is authorized COLOR ALPHA-S-RUNNING the running above the 3/4 seed SCHEME-DICTIONARY exact seeds to measured couplings; source-seed domain, named scheme, scale and threshold, total map, source manifests, and dependency graph are STOP GENERATIONS-L3 the generation structure at the L3 frontier PLENUM AND KERNEL SQRT-PHI-TIME-GRAVITY the typed clock and gravity bridge remains after the exact L1 digit lift CURVATURE-OPERATOR-CANONICAL whether the public architecture determines one equivalence class of spatial-curvature operator after its carrier, measure, projection, and commutator type are frozen ENTROPY BRIDGE ENTROPY-LAYER-BRIDGE Route A asks whether A_A contains one measurable total P_5: K_TM x O_(K,lambda) -> F_5^6 with exact equivariance and Law_W for 512 <= n < 2048; STOP ENTROPY-CYLINDER-NOGO-CURSOR exact finite-cylinder no-go at every cursor for L = 4..32, with structural zero-residue transport to every lambda-depth; T boundary MATTER NEUTRON-DELTA-EM the interior compression channel PROTON-RESIDUAL-IS-QCD typed exact QCD residual only; carrier, normalization, equality, and inference are STOP OBSERVER MINIMAL-READ-DERIVATION decide whether a complete typed decoder forces the beta_1 minimum read; L5-to-L1 gate is STOP EMPIRICAL HORIZON DESI DR3 (w = -14/15); MOLLER (sin^2 theta_W); future shadow measurements after a public inference rule; CMB-S4 (the tilt) ``` ## 19. Verification and the registry Every registered claim lives in canon/REGISTRY.tsv with its status, scope, canon section, evidence, and falsifier. Claim identifiers are status neutral and never change; statuses move in the registry. Evidence is inline (a self contained derivation in this document), a minimal reproduction under reproduce/, or a named external manifest. Run any reproduction from the repository root: ``` python3 reproduce/kernel/verify.py python3 reproduce/alpha-exact-lemma/verify.py python3 reproduce/born-faces/verify.py python3 reproduce/census/verify.py ``` Each must exit 0 with byte identical stdout against its EXPECTED.txt. During synthesis, a status-labelled statement without a registry identifier is unfinished draft material, not a public claim. Before the synthesis pull request opens, it must be registered with evidence, rewritten as a definition or remark, or removed. The cutover audit records historical provenance only; it supplies neither evidence nor status to a public claim. Simplicity is the ultimate perfection. Truth is not rude. Truth is just true.