SOURCE: https://physics.magflowmeters.com/gates/dossiers/sg1.html
======================================================================

SG-1 — geometry / shape selection — dossier & ledger 

 ← Gates scoreboard · Jump to closure ledger 

 Gate dossier — SG-1 — geometry / shape selection

 Question: Why this exact 13-dimensional shape and no other? 
 Status (fixed): REDUCED-TO-AXIOM · ANCHORED +1 
 Full 13-dimensional treatment: the frozen arena M4 x K6=SU(3)/T2 x S2 x S1_Y/Z2, all three layers, full precision. Self-contained — every value derived or stated inline. 

 Required endpoint (2026-07-06)

 Status: CLOSED / REDUCED-TO-AXIOM .

 Nothing left. Anchored on: 

 Shape: the frozen 13D carrier — 4D spacetime × K6 = SU(3)/T² (color) × S2 (weak) × the folded hypercharge circle S1Y/ℤ₂

 Granularity: the simplicity/economy measure (finite records ⇒ a description-length bit-cost per tuned quantity) — this is the load-bearing root here; it decides the shape comparison

 Scale: the high-energy boundary package (grand-unification scale ~1016 GeV, compactification radius, threshold) — not the deciding root here

 Observables: None consumed as numeric input — the gate is spectrum-neutral (the Standard-Model content cancels on both sides of the economy comparison). It references, but does not derive, the observed Standard-Model content E that the shape must carry, and the ~13–14 measured reals used in the honest economy count (not 4).

 Dissolution: No hidden derivation is claimed. The residual bottoms on the named value-free axiom/common-currency rule rather than an unbounded obligation.

 This is the current gate-level endpoint. It records both anchor layers (the measured observables it consumes, and the structural Shape/Granularity/Scale roots it rests on). The detailed dossier below is retained in full.

 Executive summary & honest status

 Headline (the sentence a skimmer should retain). The complete thirteen-dimensional arena on which the entire the framework construction is built — \(\mathfrak{B}_{\rm active} = \mathcal{M}_4 \times K_6 \times S^2 \times S^1_Y/\mathbb{Z}_2\) , with \(K_6 = SU(3)/T^2\) the full \(A_2\) flag manifold and \(D = 4+6+2+1 = 13\) — is not an arbitrary or hand-tuned shape assembled to fit a known answer: it is frozen byte-for-byte, machine-reproduced by an independent target-blind re-run, and three of its four gauge-carrying pieces are forced within a fully declared grammar once the observed Standard Model spectrum \(E\) is taken as given. The weak sector must live on \(S^2\) because no torus or abelian factor of any dimension carries a non-abelian \(SU(2)\) isometry; hypercharge must live on the orbifold interval \(S^1_Y/\mathbb{Z}_2\) because a closed odd circle would retain mirror fermions already excluded by the LEP \(Z\) -width, and only the \(\mathbb{Z}_2\) quotient projects them out; and color must live on \(K_6=SU(3)/T^2\) because, among the family of \(SU(3)\) cosets, the maximal torus \(T^2\) is the unique isotropy subgroup that is purely abelian and therefore injects no spurious extra non-abelian gauge factor — a claim demonstrated not by assertion but by building the rival \(\mathbb{CP}^2=SU(3)/U(2)\) branch end-to-end and watching it break at the next gate by over-producing gauge content. What is not forced, and is instead one clearly named, still-open axiom, is the aggregation rule that lets a "one extra dimension" cost and a "one extra measured anchor" cost be compared on a single numerical scale at all. That single missing rung — not a hidden flaw in the geometry, not a fudge in the fit — is the entire distance between "selected" and "derived," and it is why the gate's grade is fixed at REDUCED-TO-AXIOM, ANCHORED +1, and will not be represented as more, or as less, than that anywhere in this dossier.

 1. The precise claim

 SG-1 is the gate that commits the object every downstream gate (SG-2 through SG-10, and every wall, gap, and observable check built on top of them) is evaluated against. It certifies exactly three things, and only these three:

 Specificity. \(\mathfrak{B}_{\rm active}\) is a fully specified, layer-complete, three-layer object — pinned simultaneously in the \(\times\) Stage layer (the metric manifold and its bundles: \(\mathcal{M}_4 \times K_6 \times S^2 \times S^1_Y/\mathbb{Z}_2\) , carrying all 13 metric dimensions), the \(\oplus\) Rulebook layer (the finite admissibility data: \(F^+_{\rm finite} \oplus C_{\rm admiss}\) , the flavor chamber and the anti-fitting firewall, 0 metric dimensions but fully loaded with structure — the modulus \(\tau=\omega\) , the generation basis, the sector projectors, the constraint set C1–C14), and the \(\otimes\) Actors layer (the bundles and operators that actually act on the geometry: \(E_{\rm matter} \oplus E_{\rm gauge} \oplus E_{\rm Higgs} \oplus E_{\rm proton}\) , 0 metric dimensions, non-trivial content). This object is committed before any downstream gate is evaluated — the freeze happens first, the physics checks happen second, and that ordering is enforced mechanically by the freeze-before-compare barrier that is itself part of \(C_{\rm admiss}\) .

 No-layer-smuggling. No gate anywhere in this program may close using content that was not declared in its proper layer. This is the \(\times/\oplus/\otimes\) discipline, and it is not a stylistic preference — it rests on a proof (the B2 proper-subset null-space result): every proper subset of the three layers, taken alone, closes zero of the gates that are supposed to be closed inside the declared category. A residual or a "closure" seen under a truncated version of the object — Stage only, or Stage+Rulebook without Actors — is by definition an artifact of the truncation, not a fact about the physics, and this dossier uses the complete three-layer object throughout for exactly that reason.

 Reproducibility. Every primitive and derived object in the branch is content-addressed (SHA-256), and an independent, target-blind re-run regenerates every hash from scratch, finding byte-identical agreement with the frozen record across all 33 geometry-description rows, deterministically on repeated execution. This machine witness is finished and is the one component of SG-1 that is genuinely and fully derived in the strict sense — but it should be read for exactly what it is: a hash match is an audit anchor , a guarantee that the object a critic attacks today is the same object the gates were actually run against. It is not itself a piece of physics and it closes no physical question by itself; it lowers no assumption floor.

 The technical witness behind SG-1 is therefore best summarized as a freeze-and-reproduce certificate, not a derivation of the geometry from first principles. It pins the object down so it cannot be quietly retuned later to rescue a downstream result. That pinning is real and is exactly what SG-1 delivers — it should not be mistaken for an explanation of why nature chose this shape , which is the harder question this dossier also engages, honestly, below.

 2. The object under test, all three layers, full precision

 The committed branch is
$$
\mathfrak{B} {\rm active}
=\underbrace{\big[\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2\big]} {\times\ \text{Stage (metric, 13 dims)}}
\ \oplus\ \underbrace{\big[F^+ {\rm finite}\oplus C {\rm admiss}\big]} {\oplus\ \text{Rulebook (0 dims)}}
\ \otimes\ \underbrace{\big[E {\rm matter}\oplus E_{\rm gauge}\oplus E_{\rm Higgs}\oplus E_{\rm proton}\big]}_{\otimes\ \text{Actors (0 dims)}}\,,
$$
with \(K_6=SU(3)/T^2\) the full \(A_2\) -type flag manifold and \(S^1_Y/\mathbb{Z}_2\) the active orbifold interval obtained from the parent circle by the reflection \(\theta\mapsto-\theta\) . Only the \(\times\) -Stage layer carries metric dimension: \(D=4+6+2+1=13\) exactly. The \(\oplus\) -Rulebook and \(\otimes\) -Actors layers are non-metric — zero-dimensional — but load-bearing; a reading of the object that keeps only the \(\times\) -Stage factors is an incomplete object, and any residual computed under that truncation is an artifact of the truncation, not a property of the geometry.

 The gauge-routing ledger this object encodes is fixed and unambiguous. \(\mathcal{M}_4=\mathbb{R}^{3,1}\) is observed spacetime, an observational primitive, not derived. \(K_6=SU(3)/T^2\) (real dimension 6) sources color \(SU(3)_c\) through its left-isometry algebra \(\mathfrak{su}(3)\) . \(S^2\) (real dimension 2) sources weak \(SU(2)_L\) through its isometry \(\mathfrak{su}(2)\) — explicitly not through any \(SU(2)\) subgroup sitting inside \(SU(3)\) . \(S^1_Y/\mathbb{Z}_2\) (dimension 1, folded to an interval) sources hypercharge \(U(1)_Y\) together with the chirality filter that removes mirror fermions. The zero-dimensional \(F^+\) chamber carries the flavor/Yukawa finite-operator structure.

 At the level of exact invariants a reviewer can check independently: \(K_6\) 's Killing-form-normal-metric curvature at the symmetric chamber center \(\vec u=(1,1,1)\) gives \(\mathrm{Ric}_i=5/12\) , \(\mathrm{Scal}=5/2\) , \(|\mathrm{Ric}|^2=25/24\) , \(|\mathrm{Riem}|^2=23/12\) (ratio \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75\) , and \(|\mathrm{Ric}|^2/\mathrm{Scal}^2=1/6\) — the geometric origin of the corpus's " \(\kappa=1/6\) "), with \(\mathrm{Scal}/\mathrm{Ric}_i=6=\dim K_6\) in both curvature normalizations used across the corpus (the dimensionful \(R_6\) -normalization, where \(\mathrm{Ric}_i=1/(2R_6^2)\) and \(\mathrm{Scal}=3/R_6^2\) ; and the dimensionless Killing-form normalization used for all exact rationals). The cubic invariants are \(K_1 = R_{ab}{}^{cd}R_{cd}{}^{ef}R_{ef}{}^{ab} = -113/72\) and \(K_2 = R_{abcd}R_{aecf}R_{ebfd} = -5/72\) ; and \(|\nabla\mathrm{Riem}|^2=1/4\neq0\) certifies that \(K_6\) is homogeneous but not locally symmetric (zero second-Bianchi violations). \(K_6\) admits exactly four invariant Einstein metrics — the normal metric \((1,1,1)\) and the three Kähler–Einstein permutations of \((1,1,2)\) — reproducing the classical \(SU(3)/T^2\) result independently; off-center the space is non-Einstein, which is precisely the squashing degree of freedom the selector eliminates. Euler characteristics are exact topological integers: \(\chi(K_6)=6\) (equal to \(|S_3|\) , the number of Weyl chambers, as expected for a full flag manifold), \(\chi(S^2)=2\) , \(\chi(S^1_Y/\mathbb{Z}_2)=1\) . These rational values are frozen negative controls in the corpus: \(|\mathrm{Riem}|^2\) is \(23/12\) and must never be reported as \(31/147\) , nor as \(60\) (the latter belongs to the unrelated manifold \(S^6\) ).

 Two further exact structural facts anchor the object: the electroweak/color center identification is the finest faithful quotient — Smith normal form of the charge-character matrix returns invariant factors \([1,6,6]\) , giving \(G_{\rm SM}=(SU(3)_c\times SU(2)_L\times U(1)_Y)/\mathbb{Z}_6\) — and the hypercharge sum rule \(\sum_f Y_f^2 = 10/3\) per generation is exact. Volumes are likewise exact or sixteen-significant-figure: \(V_{K_6,0}=(2\pi)^3/\sqrt3 = 143.2118575035129\) , \(\mathrm{Vol}(S^1_Y/\mathbb{Z}_2)=\pi R_0 = 1/(2M_U) = 5.0\times10^{-17}\ \mathrm{GeV}^{-1}\) exactly, and \(\mathrm{Vol}(X_{\rm active}) = 3.704417261398702\times10^{-148}\ \mathrm{GeV}^{-9}\) , feeding the Planck-normalization identity \(M_{\rm Pl}^2 = M_*^{11}\,\mathrm{Vol}(X_{\rm active})\) with \(M_* = 7.467050992135091\times10^{16}\) GeV — a derived higher-dimensional scale, fixed by the geometry plus the measured \(M_{\rm Pl}=1.2209\times10^{19}\) GeV, not an independent input.

 The branch's freeze is not asserted but witnessed: the branch content hash, manifest meta-hash (33 rows), and orbifold freeze were independently recomputed in a target-blind re-run — the verification script's own comparator was bypassed and all 33 geometry-description hashes were regenerated from scratch — with exit code 0 and byte-equal agreement, deterministically across reruns. This reproducibility leg (labeled R6 in the residual ledger below) is the one genuinely and fully derived component of the gate: a content-addressed self-witness. It lowers no assumption floor and it closes no physics; it proves the object is the object, and says nothing about whether that object is uniquely forced.

 3. The explicit non-claims

 Because this is the gate most exposed to overclaiming — "why this universe and not another" is the single most tempting sentence in the whole program to overstate — the honest boundary of SG-1 is stated here without hedging in either direction.

 Not minimal-in-any-absolute-sense. SG-1 does not claim 13D is the minimal or absolutely forced geometry. The forcedness result is selector-minimal and category-relative only , evaluated against a named, declared family of competitor branches under one declared grammar — not against every architecture logically conceivable. The uncomputable version of that stronger question (absolute Kolmogorov-style minimality, \(K(T)\) ) is a unicorn : a universal negative over an unbounded search space, formally undecidable, and this dossier dissolves it as a limit on all knowledge rather than banking it as solved or leaving it as an unlabeled gap.

 Not derived from first principles. The geometry is declared and then frozen; the freeze certificate proves fidelity to that declaration, not necessity of the declaration. The word "derived" is reserved in this dossier only for the pieces that earn it: the reproducer's hash match (mechanical); the weak carrier \(S^2\) and hypercharge carrier \(S^1_Y/\mathbb{Z}_2\) (derived within the declared grammar , given \(E\) ); and the color carrier \(K_6\) (derived on a named shelf of alternatives). Selection is not derivation; category-relative is not absolute; given- \(E\) is not a derivation of \(E\) .

 Not "4 inputs → 22 outputs." That historical headline is known-overstated and explicitly retired here. The honest accounting is roughly 4 hard anchors \(\{M_{\rm Pl},\alpha_i(M_Z),y_t,|V_{us}|\}\) plus roughly 9–10 further injected reals (sector normalizations \(N_d=2.4\times10^{-2}\) , \(N_e=1.02\times10^{-2}\) , \(N_\nu=1\) ; the CKM/Berry \(\delta\) -phase triple; the Higgs holonomy angle \(\theta_H^\star\) ), for a total charged cost of about 13–14 reals — a real, still-favorable economy (roughly \(1.8\times\) , not the previously circulated \(\sim4\times\) ), but this dossier uses only the corrected number throughout.

 Not the unique conceivable carrier of color. \(K_6=SU(3)/T^2\) is the unique clean, purely abelian-isotropy carrier under the named representation-theoretic shelf, not a proof that no other conceivable sub-six-dimensional homogeneous space could do the job; full-shelf completeness (item R4/N.4 below) is explicitly open.

 Not a discriminating success. Recovering the Standard Model gauge group is a filter every serious rival framework passes — Kaluza–Klein, string/M/F-theory, noncommutative geometry, and lattice constructions all recover \(SU(3)\times SU(2)\times U(1)\) from some internal structure. This dossier reports that as a tie, not a differentiator.

 Does not derive \(E\) or \(\mathcal{M}_4\) . The chiral Standard Model content \(E\) and observed spacetime \(\mathcal{M}_4\) are both consumed as inputs to SG-1's argument, not produced by it. The family index \(\chi(K_6,E)=-3\) (three generations) is a given-E bookkeeping record inherited from a separate gate (SG-3); SG-1 only records it.

 The dead rival stays dead. \(\mathbb{CP}^2=SU(3)/U(2)\) was built out end-to-end and shown to break at the next gate by over-producing gauge content; nothing here reopens it.

 No falsifier, not a frontier. SG-1 carries no falsifier of its own and sits among roughly thirty already-RESOLVED gates; its status is never blended with the handful of separate gates elsewhere in the program that do carry live, numerically sharp tensions (an up-quark-mass pull, an irreducibility question, and similar) — those remain distinct objects with distinct grades.

 4. The honest current grade

 The fixed grade for SG-1 is REDUCED-TO-AXIOM / ANCHORED +1 . This grade is given, not argued for here, and this dossier does not move it in either direction. Concretely, the terminal reflects a specific, load-bearing structure: three of the four gauge-carrying pieces of the geometry are forced within the declared grammar , labeled DERIVED-WITHIN-GRAMMAR — the weak-sector carrier \(S^2\) by a hand-checkable general theorem about isometry groups of abelian/torus spaces; the hypercharge carrier \(S^1_Y/\mathbb{Z}_2\) by an index-theoretic argument using the Atiyah–Patodi–Singer index on the interval \([0,\pi]\) , returning \(n_L=+3\) , \(n_R=0\) and confirming no mirror fermions survive; and the color carrier \(K_6=SU(3)/T^2\) by the representation-theoretic fact that among cosets \(SU(3)/R\) , the maximal torus is the unique isotropy subgroup with abelian centralizer, injecting no spurious non-abelian gauge factor, with the rival \(\mathbb{CP}^2\) choice explicitly built and explicitly broken as the negative control.

 What keeps the gate from crossing into a fully forced, from-nothing derivation is a single remaining seam: the claim that the whole 13D branch is the economy winner — cheapest in bits — over its competitor family rests on comparing two structurally different kinds of cost (a discrete dimension-count and a continuous measured-anchor count) on one common numerical scale, and that comparison requires an aggregation rule — named in this program BR-Arch, the Archimedean common-currency axiom — that is demonstrably not forced by any of the three complete geometric roots (Shape, Scale, Granularity) run at full precision and full three-layer discipline. A second, logically consistent total order on the identical space of candidate branches exists — dimension-first lexicographic ordering — under which a lean four-dimensional effective theory beats the 13D branch outright regardless of anchor cost, and nothing in the geometry itself excludes that ordering. Because the space of orderings is provably exhausted by exactly two classes (Archimedean and non-Archimedean, via Hahn's embedding theorem: Archimedean if and only if rank-1, i.e. embeddable in a single copy of \(\mathbb{R}\) ), adopting the Archimedean branch is a genuine, nameable, non-vacuous new axiom — paid for at cost +1 on the floor — rather than a theorem recovered for free from the frozen geometry. That is precisely what REDUCED-TO-AXIOM / ANCHORED +1 means here: the chain of reasoning bottoms out cleanly on one declared, closed-candidate-class axiom, not on an unbounded regress and not on a silent assumption.

 It is equally important to say what this grade is not . It is not OPEN — the axiom is named, its necessity is proven (via the explicit lex countermodel), its candidate class is closed (via Hahn), and its target-blindness is independently confirmed (the Causal-Order screen treats the aggregation rule as a cross-theory ranking convention rather than a signal object, so it cannot have been reverse-engineered from a desired answer). It is not CERTIFIED-IRREDUCIBLE — that terminal is reserved for results proven to have no possible lever under any completion, and here a lever plainly exists in principle (prove or refute the Archimedean property directly) even though it has not yet been pulled. And it is not a DERIVED/RESOLVED +0 result — the +1 is real, it is charged honestly, and no wording in this dossier presents the geometry as forced-from-nothing. ANCHORED +1 is its own legitimate, durable terminal — not a discount applied to DERIVED, and not a euphemism for OPEN.

 5. What this dossier establishes, and what it does not

 This dossier establishes, with full inline derivations and exact numerical values, that the specific 13-dimensional arena \(\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2\) together with its complete rulebook and actor content is: (i) precisely and reproducibly frozen, machine-verified byte-for-byte by an independent target-blind re-run across all 33 geometry-description hashes; (ii) internally consistent at the level of exact curvature invariants, topological Euler characteristics, and representation-theoretic classification, cross-checked against known mathematical results for the \(SU(3)/T^2\) flag manifold (the four-Einstein-metric count, the \(\chi(K_6)=6\) Weyl-chamber count); (iii) the leanest-bit-cost branch, by a corrected margin of roughly \(1.8\times\) at transparent cost (≈13–14 charged reals versus a ≈25-real target ledger for the standard effective-field-theory alternative), over a first-pass survey of ten rival dimensional/architectural branches (ten lose, one fails structurally, zero are refuted — a first-pass survey, explicitly not yet a certified exhaustive classification); and (iv) forced, within its declared grammar and given the observed spectrum \(E\) , in three of its four gauge-routing choices, with the fourth piece of "forcedness" — the cross-kind cost-aggregation rule needed to declare an outright winner rather than merely an incommensurable comparison — resting on one explicitly named axiom, BR-Arch, that this dossier does not claim to have proven.

 This dossier does not establish that this geometry is the unique geometry compatible with the observed Standard Model. It does not establish that the four measured anchors \(\{M_{\rm Pl},\alpha_i(M_Z),y_t,|V_{us}|\}\) must take the specific numerical values they do — that is disclosed as a stated scope boundary, not a solved physics problem. It does not establish that the completeness search over all conceivable sub-six-dimensional homogeneous carriers of \(SU(3)\) has been exhausted (item R4/N.4 stays open, as does the actor-layer no-competitor question, R8). It does not establish or attempt to establish the Archimedean common-currency axiom as a theorem — that remains the single decisive open seam of this gate, named R5, reported here as an honest, bounded, in-principle-falsifiable target rather than smoothed over, with its equally honest twin outcome (a "REFUTED-ECONOMY" world in which a clean 4D effective theory genuinely wins) stated as a live possibility this dossier does not suppress. And the entire argument is deliberately spectrum-neutral throughout: the observed matter content \(E\) appears identically on both sides of the economy comparison and cancels, so no Standard Model number anywhere in this gate has been tuned to hit a known answer.

 6. Endpoint preview

 Every forcedness chain in this gate bottoms out cleanly — the three gauge-carrying factors on the observed spectrum \(E\) (given, not derived, and canceling out of the economy comparison so no number is smuggled in), and the single live seam (R5) on one finite, named, closed-candidate-class Archimedean axiom (BR-Arch) charged honestly at floor +1 against the Granularity root — so that SG-1 closes as REDUCED-TO-AXIOM, ANCHORED +1 , with the absolute-uniqueness question correctly dissolved as an uncomputable universal-negative limit on all knowledge rather than mislabeled as an unfinished derivation.

 The community gap & state of the art

 1. The open problem, stated precisely

 The Standard Model of particle physics, in its minimal renormalizable form, carries roughly 19–25
free real parameters that are simply written down and fit to data: three gauge couplings, the Higgs
quartic and mass parameter, nine charged-fermion Yukawa couplings (equivalently masses), four CKM
mixing parameters (three angles, one phase), and — once neutrino mass is included — a further set of
neutrino parameters (masses, PMNS mixing angles, phases), together with the QCD vacuum angle
 \(\bar\theta\) and the strong-CP bound on it. None of this parameter count is explained by the Standard
Model itself; it is simply the input the theory needs to match experiment. Separately, and just as
unexplained, are the qualitative facts: why spacetime is \(3+1\) -dimensional, why the gauge group is
exactly \(SU(3)_c\times SU(2)_L\times U(1)_Y\) (quotiented, as it happens, by a \(\mathbb{Z}_6\) that is
itself unexplained within the bare SM), and why there are exactly three chiral generations of quarks
and leptons rather than one, two, or seventeen.

 SG-1 addresses the geometric wing of this problem in its sharpest form: given that one is willing
to posit additional structure beyond \(3+1\) dimensions and beyond the bare SM gauge group — i.e.,
given that one adopts some version of the Kaluza–Klein idea that internal geometry sources gauge
symmetry — why this particular internal geometry and not another? This is the shape-selection
problem: among the very large space of compact internal manifolds, bundles, and admissibility rules
that a dimensional-reduction framework could in principle adopt, what singles out one specific
object, and on what basis can that selection be defended as anything more than an after-the-fact
accommodation of the observed low-energy physics?

 This is not a narrow or parochial question. It is a standing, unresolved feature of every unification
program that has ever proposed extra structure to explain the gauge group and matter content: the
program successfully recovers \(SU(3)\times SU(2)\times U(1)\) and chiral fermions from some choice
of internal data, but it does not explain why that internal datum, rather than one of the many
alternatives compatible with the same general framework, is the one nature uses.

 2. History of the problem across frameworks

 Kaluza–Klein and coset-space dimensional reduction (CSDR). The idea that a compact internal
manifold's isometry group becomes a 4D gauge symmetry after dimensional reduction is the founding
Kaluza–Klein observation, generalized from the original \(U(1)/S^1\) case to non-abelian gauge groups
and systematized as coset-space dimensional reduction: one picks a coset \(G/H\) with \(G\) the isometry
group and \(H\) the isotropy (little) group at a point, the 4D gauge fields surviving reduction are
those valued in the centralizer of \(H\) inside a chosen bulk gauge group (the CSDR centralizer rule),
and the surviving 4D fermion content is fixed by how \(H\) embeds in the bulk gauge and Lorentz
structure. The chronic, unsolved problem in this tradition is that many cosets \(G/H\) can be made to
reproduce the Standard Model gauge group under suitable embedding choices, and the choice of which
coset, which embedding, and which bulk gauge group to start from is itself unconstrained by any
first-principles criterion internal to CSDR — it is supplied by the model-builder, not derived. This
gate's use of coset language — \(K_6=SU(3)/T^2\) , the flag manifold of \(A_2\) type, against the
competing coset \(CP^2=SU(3)/U(2)\) — sits squarely inside this coset-space-dimensional-reduction
tradition, and the "which coset" degeneracy it must confront is the same degeneracy CSDR has never
resolved from first principles alone.

 String theory, M-theory, F-theory: the landscape. In string-type compactifications the analogous
freedom is enormously larger. Compactifying on a Calabi–Yau-type internal manifold, or on other
higher-dimensional internal geometries proposed within the M-theory and F-theory programs, the 4D
gauge group, chiral matter content, and Yukawa couplings all depend on the choice of compactification
manifold's topology, the choice of gauge bundle or brane configuration threading it, and the choice
of background flux. This freedom is the origin of what the string-theory community itself calls the
"landscape": a very large space of topologically and flux-distinguished vacua, no member of which is
singled out by any known dynamical selection principle internal to the theory. The shape-selection
problem for string/M/F-theory compactifications is thus the most widely recognized instance, in the
whole beyond-Standard-Model literature, of exactly the problem SG-1 addresses for its own much
smaller thirteen-dimensional construction: a mechanism that is capable of producing the right kind of
answer, with no internal principle that forces the specific choice actually used.

 Noncommutative geometry (NCG) — the spectral-triple approach. The Connes–Chamseddine spectral
approach derives the Standard Model — including its gauge group, its fermion representation content,
and constraints on its Higgs sector — from a choice of finite noncommutative geometry (a finite
spectral triple) tensored onto ordinary 4D spacetime, subject to a package of axioms (reality, a
first-order condition, a grading) that constrain but do not by themselves uniquely fix the finite
algebra. This is, among the frameworks surveyed here, the one that comes closest to a genuine forcing
argument, because its axiom package is comparatively restrictive. But the open problem is structurally
parallel to the coset and landscape cases: the axioms are chosen with the Standard Model's answer
already in view, and a full classification proving that the specific finite algebra used is the
unique (or minimal-cost) finite spectral triple satisfying those axioms and reproducing the observed
chiral content is not established. This is exactly the open comparison this gate's own residual
ledger names under actor-layer minimality (the demand to rule out a lower-cost finite NCG-type
operator as a competing realization) — an explicitly open item here, not a comparison this dossier
claims to have won.

 Lattice and other non-geometric approaches. Approaches that take the SM gauge group and \(D=4\) as
given inputs and study the resulting dynamics non-perturbatively do not attempt to answer the shape-
selection question at all; they are relevant to this survey only as confirmation that "reproduce the
observed SM gauge group" is a bar that can be met by frameworks that make no geometric-selection
claim whatsoever, underscoring that meeting the bar is not itself a discriminating achievement.

 3. Why recovering \(SU(3)\times SU(2)\times U(1)\) is a tie, not a discriminator

 A point this gate insists on stating plainly, rather than letting stand as an implicit advantage, is
that recovering the observed Standard Model gauge group does not by itself favor one internal-
geometry framework over another. Coset-space dimensional reduction can recover it, for numerous
choices of coset. String/M/F-theory compactifications can recover it, at many points in the
landscape. Noncommutative geometry recovers it, given its axiom package. It is a filter every mature
internal-geometry framework can be tuned to pass — a necessary condition for viability, not a
distinguishing virtue relative to the other frameworks that also pass it. A dossier that reported
"this geometry reproduces the SM gauge group" as if that were a discriminating success would be
overstating the case: it is a tie across the field. What differs from framework to framework is not
 whether the SM gauge group can be recovered, but how expensive, how forced, and how uniquely the
recovery is achieved — and it is that cost-and-forcedness question, not the recovery itself, that
constitutes the actual open problem this gate targets.

 4. The state of the art: what "best existing bound" means here

 No framework in this comparison class has a theorem proving its internal geometry is the unique or
 forced choice among all logically conceivable alternatives, so there is no existing "bound" in the
ordinary sense of a rigorous inequality separating viable from excluded shapes. The strongest results
available anywhere in this literature, including in this program's own corpus, are one of three
weaker things.

 (a) A necessity argument for a sub-piece of the geometry, holding fixed a declared grammar of
admissible constructions. This is the pattern behind the weak-sector result used in this gate:
 given that one insists the weak gauge factor arise from an isometry algebra and not be inserted by
hand, no torus or other purely abelian-isometry space can supply it, because no manifold whose full
isometry algebra is abelian has a non-abelian \(\mathfrak{su}(2)\) subalgebra among its Killing vectors
— hence the weak sector needs a genuinely non-abelian isometry carrier such as \(S^2\) . This kind of
"no-go for the cheaper alternative" result is real, hand-checkable, and closes off whole classes of
competing shapes, but it is conditional on the declared grammar (isometry-sourced gauge fields) and
says nothing about whether that grammar itself is forced.

 (b) A representation-theoretic uniqueness result on a named, explicitly restricted shelf of
candidates. Within the coset toolkit specifically, restricting to homogeneous spaces \(SU(3)/H\) and
asking which isotropy subgroups \(H\) are purely abelian — so that the CSDR centralizer construction
injects no spurious extra non-abelian gauge factor beyond the desired color \(SU(3)_c\) — the maximal
torus \(T^2\) is the unique such choice: \(T^2\) is abelian and self-centralizing in \(SU(3)\) 
( \(C_{SU(3)}(T^2)=T^2\) ), whereas the alternative isotropy \(H=U(2)\) (giving the coset
 \(CP^2=SU(3)/U(2)\) ) is not abelian — \(U(2)=(SU(2)\times U(1))/\mathbb{Z}_2\) contains a non-abelian
 \(SU(2)\) factor that, by the same centralizer rule, becomes gauge-active in its own right and so
over-produces gauge symmetry beyond the color group one set out to obtain. This is a real, checkable
representation-theory fact, and it is the strongest structural result available in this program's own
toolbox: it is stronger than a merely heuristic preference because it was tested end-to-end — the
 \(CP^2\) branch was built out explicitly and confirmed to break, over-producing \(SU(2)+U(1)\) gauge
content at the gate that checks gauge-symmetry survival, exactly as the abelian-isotropy argument
predicts. But this uniqueness is explicitly on a named shelf : it is uniqueness among cosets
 \(SU(3)/H\) with \(H\) ranging over subgroups of \(SU(3)\) , not uniqueness among all six-real-dimensional
(or lower-dimensional) manifolds whatsoever that could in principle carry a color-sized isometry.
Whether some other sub-six-dimensional carrier, outside the \(SU(3)/H\) coset family entirely, could
also deliver a clean abelian-isotropy color sector without the centralizer over-production pathology
is a full-shelf completeness question that has not been answered — it is recorded honestly in this
program's own ledger as an open item (full-shelf completeness, labeled N.4) rather than asserted away.

 (c) An economy or minimum-description-length (MDL) comparison against a finite, explicitly
enumerated ladder of rival constructions , rather than a completeness proof against the space of
 all conceivable constructions. This is the character of the strongest available "bound" for the
overall shape, not just its color factor: a spectrum-neutral bit-cost comparison in which the observed
SM data \(E\) is charged identically on every side of the ledger — and therefore cancels, so no fit to
 \(E\) is smuggled into the comparison — and only the cost of specifying the geometry itself (number of
continuous tuned parameters versus number of discrete structural choices) is compared. Under such a
ledger, the 13-dimensional branch beats every member of an explicit, finite comparison ladder: ten
rival constructions lose, one fails structurally, and none refutes the 13D branch outright, for a
transparent-cost margin of roughly \(1.8\times\) . This is a genuine result, but it is explicitly not a
proof that no cheaper construction exists anywhere in the unenumerated space of possible geometries;
and, more fundamentally, the very act of comparing costs "across kinds" — a geometric-dimension label
against a continuous-parameter bit — presupposes that the two kinds of cost can be placed on a common
numerical scale at all. That presupposition is not free; it is exactly this gate's own named open
axiom, discussed in the sections that follow, and no result in the wider literature discharges it
either, because none of the frameworks surveyed here poses its shape-selection argument as an
explicit bit-cost ledger in the first place — the more common practice, across CSDR, string, and NCG
alike, is to assert that a construction is "natural," "minimal," or "elegant" without specifying the
cost metric under which that judgment is being made.

 Put together: the state of the art, across the field surveyed here, is that no result anywhere proves
a compact internal geometry sourcing the Standard Model is the unique or forced choice among all
logically possible alternatives. The best available results are (a) conditional no-go theorems that
exclude cheaper alternatives within a declared grammar , (b) uniqueness results on a named,
explicitly restricted shelf of candidates, and (c) economy comparisons against a finite, enumerated 
ladder of named rivals. Every one of these three kinds of result leaves an explicit, acknowledged
residual: the grammar itself is not shown to be forced; the shelf is not shown to be complete; the
ladder is not shown to be exhaustive. This is not a defect unique to this gate's approach — it is the
honest ceiling of the shape-selection question wherever it has been posed in this literature.

 5. Why every prior attempt falls short — itemized

 It is worth being explicit about why, mechanistically, each class of prior attempt falls short of a
forced or unique answer, because the shortfall in each case is different and understanding the
difference is what motivates the specific axiomatic move this gate ultimately makes.

 Coset-space dimensional reduction fails to be forced because the choice of coset, embedding, and
 bulk gauge group is external input, not output. The CSDR machinery computes the consequences of a
 choice of \(G/H\) and embedding; nothing internal to that machinery tells you which \(G/H\) to start
 from. This is the most direct version of the shape-selection gap, and it is the one this gate's own
 abelian-isotropy-uniqueness result partially — but only partially — answers, because that result
 still operates inside a CSDR-style grammar (isometry-sourced gauge fields, coset reduction, the
 centralizer rule) that is itself an input choice, not something derived from something more
 primitive.

 String/M/F-theory fails to be forced because the landscape is large and no accepted dynamical
 selection principle picks a single vacuum from it. This is a categorically harder version of the
 shape-selection problem than CSDR's, because CSDR at least deals with a small, hand-enumerable
 family of low-dimensional cosets, while the landscape's scale defeats even an attempt at exhaustive
 comparison.

 NCG fails to be forced because its defining axioms are chosen with the answer in view. The
 reality condition, the first-order condition, and the specific grading used are reasonable
 mathematical choices, but none is derived from a principle more primitive than "these axioms happen
 to make the Standard Model's algebra come out close to uniquely." The axioms function as a declared
 grammar, exactly analogous to CSDR's declared reduction rules, not as first-principles constraints
 derived from something more basic than the target.

 Economy/MDL-style arguments — including this gate's own — fail to be forced because "cost" is not
 self-evidently a single, cross-kind-commensurable quantity. Even after granting that discrete
 structural choices should be cheap and continuous tuned parameters should be expensive — a
 reasonable and widely shared intuition — nothing forces the exchange rate between "one more
 dimension" and "one more fitted real number" to be finite in both directions. A rival bookkeeping
 convention exists and is a perfectly consistent total order on the identical space of candidate
 constructions: rank constructions first by number of spacetime dimensions, and only break ties 
 by counting anchor/parameter cost. Under that dimension-first lexicographic convention, any
 four-dimensional effective field theory beats a thirteen-dimensional construction outright,
 regardless of how many continuous parameters the 4D theory must fit, simply because a smaller
 dimension count lexicographically dominates before any parameter count is consulted. Nothing in the
 ordinary toolkit of geometric or field-theoretic reasoning excludes this lexicographic convention as
 illegitimate — it is a logically coherent, non-Archimedean ordering. By Hahn's embedding theorem (an
 order is Archimedean if and only if it embeds in a single copy of the real line, i.e. is rank-1;
 every non-Archimedean order is instead built from rank two or higher, of which the dimension-first
 lexicographic order is the canonical rank-two example) this alternative sits in a well-understood,
 exhaustive alternative class to the additive, single-currency bit-cost ledger this gate's economy
 argument requires. No result anywhere in the CSDR, string-landscape, or NCG literature addresses
 this ordering ambiguity, because none of those literatures poses shape-selection as an explicit
 bit-cost comparison in the first place; the ambiguity is native to the economy-argument strategy
 itself, wherever it is deployed.

 6. What this leaves as the honest open problem, and what this gate does and does not claim to close

 Surveying the state of the art shows that the field-wide "why this shape" question decomposes into
at least two logically separate sub-questions that prior work has not kept fully apart: (i) within a
declared grammar of admissible constructions, is the observed shape forced or merely preferred? and
(ii) is any one grammar itself the uniquely correct one to declare, as opposed to a competing,
equally coherent bookkeeping convention? None of the frameworks surveyed above — CSDR, string/M/F-
theory, or NCG — has resolved question (i) completely for its own preferred construction (each leaves
an explicit shelf-completeness or landscape-completeness gap), and none has attempted question (ii) at
all, because none poses its selection argument in a form where an alternative grammar (an alternative
cost bookkeeping, an alternative axiom package) is made explicit enough to be compared against.

 Assessed honestly against that backdrop, this gate's contribution is threefold. First, it freezes one
specific, fully three-layer-specified geometric object — content-addressed and independently,
target-blindly re-verified — so that it cannot be silently re-tuned to rescue results computed
downstream of it, a discipline that goes beyond the informal prose-and-equations description typical
of how a "chosen" compactification is usually presented in the wider literature. Second, it supplies,
on a named and explicitly bounded shelf (homogeneous cosets \(SU(3)/H\) ), a genuine representation-
theoretic necessity argument for the color factor that is reinforced by an explicit negative control
rather than left as an assertion: the \(CP^2\) branch is not merely disfavored on aesthetic grounds, it
is built end-to-end and shown to break, over-producing gauge content, exactly as the centralizer rule
predicts. Third, it makes question (ii) explicit, by isolating the single load-bearing axiom — an
Archimedean common-currency assumption about how structural and numerical cost aggregate — that the
economy argument needs and cannot derive from anything more primitive, naming it, bounding it, and
charging it openly as a paid axiom (+1 on the anchoring floor) rather than letting it pass silently as
an unexamined convention, which is what happens by default whenever a framework in this literature
informally calls its own construction "minimal," "natural," or "elegant" without specifying the cost
metric under which that judgment is being made.

 What this gate does not claim, and what no result surveyed above supports anyone else claiming
either, is that the observed thirteen-dimensional shape is the unique geometry compatible with the
observed low-energy world in some grammar-independent, axiom-free sense. That stronger claim remains
open across the entire field, not merely here — and the honest terminal this gate reaches,
REDUCED-TO-AXIOM / ANCHORED +1, reflects exactly that boundary: three of the four gauge-carrying
pieces of the geometry are forced within the declared grammar and given the observed spectrum \(E\) 
(the weak-sector no-go, the hypercharge mirror-fermion exclusion, and the color abelian-isotropy
uniqueness on its named shelf, the last backed by an explicit built-and-broken rival), while the
single remaining step — treating the resulting economy comparison as an outright win rather than an
incommensurable comparison — rests on one named, closed-candidate-class axiom (the Archimedean
common-currency rule) that this dossier does not claim to have proven, and states as the field's
shared, still-open residual rather than smoothing it into a stronger claim than the evidence supports.

 7. A note on this program's own prior overstatement, corrected here

 Part of the honest state-of-the-art record is that earlier internal presentations of this same gate
overstated its own result, and the corrected numbers are the ones used throughout this dossier. A
previously circulated headline describing the construction as "4 inputs \(\to\) 22 outputs," implying a
roughly four-anchor cost, is retired: the audited cost is closer to four measured anchors
( \(M_{\rm Pl}\) , the gauge couplings \(\alpha_i\) at \(M_Z\) , the top Yukawa \(y_t\) , and the CKM element
 \(|V_{us}|\) ) plus roughly nine to ten further injected real parameters (the sector normalizations
 \(N_d\approx2.4\times10^{-2}\) , \(N_e\approx1.02\times10^{-2}\) , and \(N_\nu=1\) ; the threshold triple
 \((\delta_1,\delta_2,\delta_3)=(+4.8424,-3.1112,-1.7313)\) ; and the Wilson-line/Hosotani modulus
 \(\theta_H^\star\) ), for a total honestly charged cost of roughly 13–14 real numbers. Correspondingly, a
previously circulated "roughly \(4\times\) margin, beating ten rivals outright" headline is also
retired: the audited, transparent-cost margin is roughly \(1.8\times\) (approximately \(13\) – \(14\,b\) bits
for the 13D branch against approximately \(25\,b\) bits for the standard effective-field-theory target,
where \(b=\log_2(1/\Delta_0)\) is the per-tuned-real bit cost at the declared cell resolution
 \(\Delta_0\) ), and the underlying survey is explicitly a first-pass classification — ten rival branches
lose, one fails structurally, zero are refuted outright — not yet a certified exhaustive
classification of the full competitor family. Both corrections are stated here as part of the
community-facing record precisely because a prior inflated number, left uncorrected, would itself
become a piece of unreliable state-of-the-art literature for the next reviewer to inherit.

 The frozen 13D arena at full precision

 SG-1 is the gate that commits the object every other gate in this ledger is scored against. Before any claim about forcedness, economy, or selection can be evaluated, the arena itself has to be written down completely, at all three layers, with every constant pinned to exact rational or 16-significant-figure precision. That pinning is the content of this section. Nothing here is an input tuned to a downstream result — the numbers below are frozen geometric facts about a specific coset space, a specific sphere, and a specific orbifold, computed once from the Killing form and the RG/threshold data, and re-used identically everywhere downstream. The freeze is committed as a content-addressed SHA-256 branch digest, a manifest meta-digest spanning all 33 geometry-description rows, and a separate orbifold-freeze digest; a target-blind independent re-run regenerates all 33 geometry-description digests without the original script's comparator and finds them byte-equal, deterministically, across reruns. That reproducibility is a freeze-and-reproduce certificate, not a derivation — it guarantees the object a reviewer attacks is the object the gates actually ran on, and it is the one leg of SG-1 that is genuinely, mechanically DERIVED (R6, CLOSED).

 1. The object, in full: three layers, one active branch

 The active branch is the layered object

 \[
\mathfrak{B}_{\rm active}
=\underbrace{\big[\,\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2\,\big]}_{\times\ \text{STAGE (metric, 13 dims)}}
\ \oplus\ \underbrace{\big[\,\mathcal{F}^+_{\rm finite}\oplus \mathcal{C}_{\rm admiss}\,\big]}_{\oplus\ \text{RULEBOOK (0 dims)}}
\ \otimes\ \underbrace{\big[\,\mathcal{E}_{\rm matter}\oplus \mathcal{E}_{\rm gauge}\oplus \mathcal{E}_{\rm Higgs}\oplus \mathcal{E}_{\rm proton}\,\big]}_{\otimes\ \text{ACTORS (0 dims)}}
\]

 with \(K_6 = SU(3)/T^2\) the full \(A_2\) flag manifold and \(S^1_Y/\mathbb{Z}_2\) the active orbifold interval. Compressed mnemonic used elsewhere in the corpus, \(\mathcal{M}_{\rm GUT} = \mathcal{M}_4\times K_6\times S^2\times S^1_Y\times F^+\) with \(K_{\rm gauge}\equiv K_6\times S^2\times S^1_Y\) , is a correct shorthand but visually hides the \(\oplus/\otimes\) layers — they are never dropped when the object is used for real.

 Only the \(\times\) -layer carries metric dimension:

 \[
D = \dim\mathcal{M}_4 + \dim K_6 + \dim S^2 + \dim S^1_Y = 4 + 6 + 2 + 1 = 13 \quad [\text{EXACT}].
\]

 The \(\oplus\) (rulebook) and \(\otimes\) (actors) layers are non-metric — zero-dimensional — but load-bearing: a \(\times\) -only reading of this object is an incomplete object, and SG-1's no-layer-smuggling certification means no downstream gate may be closed by content that lives in a layer other than the one it is declared in. \(F^+\) in particular is a finite/operator chamber, not a propagating metric factor : its Cartan-torus modulus \(\tau\) is chamber data (a discrete choice), not a Kaluza–Klein tower, and it therefore contributes 0 to \(D\) even though it carries real physical content (the flavor structure, §5 below).

 Gauge-routing ledger — which factor sources which force (the physical content of the \(\times\) -layer): 

 Factor 
 Real dim 
 Metric 
 Primitive/derived 
 Physical role 
 Gauge group sourced 

 \(\mathcal{M}_4=\mathbb{R}^{3,1}\) 
 4 
 Minkowski 
 primitive (observational) 
 observed spacetime 
 — 

 \(K_6=SU(3)/T^2\) 
 6 
 Weyl-rigid invariant (normal at center) 
 primitive 
 color source; spin- \(\mathbb{C}\) family index 
 \(SU(3)_c\) via left-isometry \(\mathfrak{su}(3)\) 

 \(S^2\) 
 2 
 round 
 primitive 
 weak source; doublet routing 
 \(SU(2)_L\) via isometry \(\mathfrak{su}(2)\) 

 \(S^1_Y\) 
 1 
 flat 
 primitive 
 parent hypercharge circle 
 \(U(1)_Y\) via isometry 

 \(S^1_Y/\mathbb{Z}_2\) 
 interval 
 induced quotient 
 derived ( \(\theta\mapsto-\theta\) ) 
 chirality / no-mirror filter 
 \(U(1)_Y\) + chirality 

 Binding physical fact, stated once and held throughout the dossier: gauge forces here are literally isometries of the internal metric factors — nothing more mysterious than that. \(SU(2)_L\) comes from \(S^2\) and from nowhere else; in particular it is not any \(SU(2)\subset SU(3)\) sitting inside \(K_6\) . \(K_6\) carries \(SU(3)_c\) and only \(SU(3)_c\) . This separation of carriers is what makes the shape's economy argument (§6 below, and the full ladder in the forcedness section) meaningful: each force has its own dedicated geometric home, and mixing them (e.g. routing weak isospin through a subgroup of \(K_6\) , as \(CP^2\) 's isotropy does) is a distinguishable, falsifiable structural choice — not a relabeling.

 \(\mathcal{M}_4=\mathbb{R}^{3,1}\) is a declared observational primitive (labeled AXIOM SG1-α downstream): it is never counted as a forced rung or cited as a derivation. Observed 3+1 spacetime is an input to the branch, not an output of it, and SG-1 does not pretend otherwise.

 2. Convention note that governs every curvature number below

 Two internally-consistent metric normalizations coexist on this arena and both are used downstream; a curvature number is only meaningful once the normalization is stated, and this dossier tags every one.

 (A) Frozen physical ( \(R_6\) ) normalization. The internal radius is the derived compactification radius \(R_6\) (chamber-center value \(R_6=R_0\) , §3 below). Curvature carries physical units of GeV \(^2\) : \(\mathrm{Ric}_i = 1/(2R_6^2)\) , \(\mathrm{Scal}=3/R_6^2\) . This is the normalization used for every dimensionful downstream quantity — Planck normalization, KK spectra, threshold radii.

 (B) Killing-form normal metric. \(g=(-B)|_{\mathfrak m}\) with Killing form \(B(X,Y)=6\,\mathrm{Tr}(XY)\) on \(\mathfrak{su}(3)\) , evaluated at the symmetric chamber center \(\vec u=(1,1,1)\) . Curvature is dimensionless: \(\mathrm{Ric}_i=5/12\) , \(\mathrm{Scal}=5/2\) . This is the normalization in which the exact-rational invariants below ( \(|\mathrm{Riem}|^2\) , the weight-6 cubic products, the heat-kernel \(a\) -coefficients) are computed and stored.

 The bridge (scale-invariant, identical in both). Ratios of curvature invariants do not depend on normalization. The load-bearing one is
$$
\frac{\mathrm{Scal}}{\mathrm{Ric}_i} = 6 = \dim K_6 \qquad\text{in BOTH normalizations:}\quad \frac{3/R_6^{2}}{(1/2)/R_6^{2}}=6 \ \ \text{(A)}, \qquad \frac{5/2}{5/12}=6\ \ \text{(B)}.
$$
Likewise \(|\mathrm{Ric}|^2/\mathrm{Scal}^2 = 1/6\) and \(|\mathrm{Riem}|^2/\mathrm{Scal}^2 = 23/75\) in both. This bridge is what lets a physical (dimensionful) statement and a purely combinatorial (dimensionless, rep-theoretic) statement about the same coset agree exactly — it is not a coincidence but an identity of ratios under rescaling \(g\mapsto\lambda g\) .

 3. Radii and scale: how the arena is set dimensionfully

 The internal geometry is not dimensionless in an absolute sense — it is pinned to a physical length scale via the unification/threshold RG chain and the two measured anchors \(M_Z\) and \(M_{\rm Pl}\) . This is where "derived, not fitted" is verified concretely.

 \[
M_U = 1.0\times10^{16}\ \text{GeV}\quad[\text{DERIVED}] — \text{closure target } \alpha_i^{-1}(M_U)=\alpha_j^{-1}(M_U),\ \text{residual } 9.6\times10^{-11}\ \text{(numerical-pipeline floor)}.
\]

 \[
R_0 \equiv (2\pi M_U)^{-1} = 1.591549430918954\times10^{-17}\ \mathrm{GeV}^{-1}\quad[\text{DERIVED}].
\]

 At the symmetric chamber center \(\vec u=(1,1,1)\) (defined below), \(R_6=R_2=R_0\) :

 \[
R_6\equiv R_{K_6} = 1.591549430918954\times10^{-17}\ \mathrm{GeV}^{-1},\qquad R_2\equiv R_{S^2} = 1.591549430918954\times10^{-17}\ \mathrm{GeV}^{-1}.
\]

 The hypercharge circle radius carries an extra factor of \(\tfrac12\) from the active \(\mathbb{Z}_2\) orbifold halving:

 \[
R_Y \equiv R_{S^1_Y} = 7.957747154594768\times10^{-18}\ \mathrm{GeV}^{-1} = \tfrac12 R_0.
\]

 There is also a Cartan-torus radius living inside the \(F^+\) chamber (not a propagating dimension, but part of the pinned object): \(R_{T^2_{\rm Cartan}} = R_0\sqrt{2/\sqrt3} = R_0\sqrt2\,3^{-1/4} = 1.710231163476377\times10^{-17}\ \mathrm{GeV}^{-1}\) , evaluated at the modular fixed point \(\tau=\omega\) (§5 below).

 Squashing chamber: the \(K_6\) metric is not rigid to a single scale — it admits an anisotropic squashing \(\vec u=(u_1,u_2,u_3)\in[1/2,3/2]^3\) , Weyl-rigid, with chamber-center witness \(u_1=u_2=u_3=1\) . Off-center points fail Weyl-rigid admissibility and are eliminated by the selector; the center is the value every \(K_6\) -dependent gate in this dossier actually uses. This is the squashing input — the one continuous internal-geometry degree of freedom this arena carries, and it is pinned to its symmetric (Einstein) value, not left floating.

 The two measured comparison scales that anchor the whole ladder are \(M_Z = 91.1876\) GeV [MEASURED, PDG, \(\pm0.0021\) ] and \(M_{\rm Pl}=1.2209\times10^{19}\) GeV [MEASURED, ordinary — not the reduced \(\bar M_{\rm Pl}=M_{\rm Pl}/\sqrt{8\pi}\) ]. Both are inputs, honestly declared as such; nothing about the shape-selection argument in SG-1 hides them as "derived."

 4. \(K_6=SU(3)/T^2\) curvature at full precision — the color carrier

 \(K_6\) is the full \(A_2\) flag manifold, \(SU(3)/T^2\) , where \(T^2\) is the maximal torus of \(SU(3)\) . Its role is to source \(SU(3)_c\) by left-isometry and to carry the spin- \(\mathbb{C}\) structure whose index gives the family count. Every number in this subsection is [Killing-norm] at the symmetric center unless marked [ \(R_6\) -norm], and every one is exact — a rational number, not a numerical fit.

 Root system ( \(A_2=\mathfrak{su}(3)\) ). Cartan basis \((h_1,h_2,h_3)\) with \(h_1+h_2+h_3=0\) ; simple roots \(\alpha_1=(1,-1,0)\) , \(\alpha_2=(0,1,-1)\) , and \(\alpha_1+\alpha_2=(1,0,-1)\) . The three positive roots are \(\{\alpha_1,\alpha_2,\alpha_1+\alpha_2\}\) , half-sum \(\rho=\tfrac12\sum_{\alpha>0}\alpha=(1,0,-1)\) with \(\|\rho\|^2=2\) (Killing normalization). The Weyl group is \(S_3\) , order 6 — and note this integer resurfaces exactly as the Euler characteristic below, which is not a coincidence: a full flag manifold's Euler characteristic always equals the order of its Weyl group.

 Tangent decomposition. \(T(K_6)=\mathfrak m_1\oplus\mathfrak m_2\oplus\mathfrak m_3\) with \(\dim_{\mathbb R}\mathfrak m_i=2\) for each of the three root planes ( \(\alpha_3\equiv\alpha_1+\alpha_2\) ). The \((-B)\) -orthonormal basis is \(\{X_{ij}=E_{ij}-E_{ji},\,Y_{ij}=i(E_{ij}+E_{ji})\}/\sqrt{12}\) over the three pairs \((01),(12),(02)\) , with Killing form \(B=6\,\mathrm{Tr}\) .

 Invariant metric and general-chamber Ricci (Wang–Ziller/Nomizu). With independent scale parameters \(x_1,x_2,x_3\) on \(\mathfrak m_1,\mathfrak m_2,\mathfrak m_3\) :

 \[
\mathrm{Ric}_1=\frac{x_1^2-x_2^2+6x_2x_3-x_3^2}{12\,x_1x_2x_3},\quad
\mathrm{Ric}_2=\frac{-x_1^2+6x_1x_3+x_2^2-x_3^2}{12\,x_1x_2x_3},\quad
\mathrm{Ric}_3=\frac{-x_1^2+6x_1x_2-x_2^2+x_3^2}{12\,x_1x_2x_3},
$$
$$
\mathrm{Scal}=\frac{x_1x_2+x_1x_3+x_2x_3-\tfrac16(x_1^2+x_2^2+x_3^2)}{x_1x_2x_3}.
\]

 Solving for Einstein points ( \(\mathrm{Ric}_1=\mathrm{Ric}_2=\mathrm{Ric}_3\) ) on this three-parameter family recovers exactly 4 invariant Einstein metrics on \(SU(3)/T^2\) : the fully symmetric normal metric \((1,1,1)\) , plus the Kähler–Einstein metric \((1,1,2)\) and its three permutations \((1,2,1)\) , \((2,1,1)\) — four points total, a classic result in homogeneous-space geometry, reproduced here independently as a validation of the computational engine, not asserted by fiat. Off the Einstein locus the space is non-Einstein — this is exactly the squashing chamber of §3, and the selector's chamber-center choice \(\vec u=(1,1,1)\) lands precisely on the normal Einstein point.

 Curvature at the symmetric center \(\vec u=(1,1,1)\) , both normalizations: 

 Quantity 
 [ \(R_6\) -norm] 
 [Killing-norm, exact] 

 \(\mathrm{Ric}_1=\mathrm{Ric}_2=\mathrm{Ric}_3\) 
 \(1/(2R_6^2)=1.973920880217872\times10^{33}\ \mathrm{GeV}^2\) 
 \(5/12\) 

 \(\mathrm{Scal}(K_6)\) 
 \(3/R_6^2=1.184352528130723\times10^{34}\ \mathrm{GeV}^2\) 
 \(5/2\) 

 \(\mathrm{Scal}/\mathrm{Ric}_i\) 
 \(6\ (=\dim K_6)\) 
 \(6\ (=\dim K_6)\) 

 Metric-scale-invariant curvature ratios (identical in both normalizations — the load-bearing numbers used across the dossier): 

 \[
\mathrm{Scal}^2 = \frac{25}{4} = 6.25,\qquad
\|\mathrm{Ric}\|^2 = \frac{25}{24} = 1.041666666666667,
$$
$$
\|\mathrm{Riem}\|^2 = \frac{23}{12} = 1.916666666666667,\qquad
\frac{\|\mathrm{Riem}\|^2}{\mathrm{Scal}^2} = \frac{23}{75} = 0.3066666666666667,\qquad
\frac{\|\mathrm{Ric}\|^2}{\mathrm{Scal}^2} = \frac16 = 0.1666666666666667.
\]

 The last ratio is the number this corpus calls " \(\kappa=1/6\) " wherever it appears downstream — it is exactly \(\|\mathrm{Ric}\|^2/\mathrm{Scal}^2\) on \(K_6\) at the Einstein center, nothing more exotic.

 Anti-drift negative controls (frozen, never to be silently changed): \(\|\mathrm{Riem}\|^2/\mathrm{Scal}^2=23/75\) is confirmed and is never \(31/147\) ; \(\|\mathrm{Riem}\|^2\) is never \(60\) — that value belongs to the round unit \(S^6\) , a topologically and metrically distinct 6-manifold, and its appearance in any computation flags a wrong-manifold error. These controls exist precisely so that a later gate cannot silently substitute a friendlier curvature value.

 Cubic / derivative / weight-6 invariants (Killing-norm, Einstein center, all exact rationals): 

 \[
K_1 = R_{ab}{}^{cd}R_{cd}{}^{ef}R_{ef}{}^{ab} = 8\,\mathrm{tr}(R_{\rm op}^3) = -\frac{113}{72},\qquad
K_2 = R_{abcd}R_{aecf}R_{ebfd} = -\frac{5}{72},
$$
$$
\|\nabla\mathrm{Riem}\|^2 = \frac14\ \text{(via Nomizu; 2nd-Bianchi check passes with 0 violations)},
$$
$$
\mathrm{Scal}^3=\frac{125}{8},\quad \mathrm{Scal}\cdot\|\mathrm{Ric}\|^2=\frac{125}{48},\quad \mathrm{Scal}\cdot\|\mathrm{Riem}\|^2=\frac{115}{24},\quad \mathrm{Ric}^3=\frac{125}{288},\quad \mathrm{Ric}\cdot\|\mathrm{Riem}\|^2=\frac{115}{144}.
\]

 The nonvanishing of \(\|\nabla\mathrm{Riem}\|^2=1/4\) is a physically consequential fact, not a curiosity: it certifies that \(K_6\) is homogeneous but not locally symmetric — the curvature tensor is covariantly non-constant. This is precisely why the graviton heat-kernel coefficient \(a_6\) (§7 below) carries an off-diagonal Gelfand–Tsetlin ladder term that a locally symmetric space would not have, and why that leg remains an honestly bounded OWED computation rather than a closed value.

 Topology. \(\chi(K_6)=6\) [EXACT/topological] — equal to \(|S_3|\) , the order of the Weyl group, exactly as expected for a full flag manifold (the number of Weyl chambers). \(\chi(S^2)=2\) , \(\chi(S^1_Y/\mathbb{Z}_2)=1\) [EXACT/topological]. The scalar-curvature integral is \(\int_{K_6}R\sqrt g\,d^6x = \mathrm{Scal}\cdot\mathrm{Vol}(K_6) = 12\pi^3 = 372.0753201635977\) under the Killing-form-absorbing normalization, or \((2\pi)^3\sqrt3 = 429.6356725105388\) under the pure \(\sqrt g\,d^6x\) normalization at \(R_6=1\) ; both forms are recorded so a reviewer working from either convention finds the matching number.

 5. Volumes and the Planck normalization

 The exact volume formulas, evaluated at the frozen radii:

 \[
\mathrm{Vol}(K_6)(\vec u) = V_{K_6,0}\,R_6^6\sqrt{u_1u_2u_3},\qquad V_{K_6,0}=\frac{(2\pi)^3}{\sqrt3} = 143.2118575035129,
$$
$$
\mathrm{Vol}(S^2)=4\pi R_2^2,\qquad \mathrm{Vol}(S^1_Y)_{\rm parent}=2\pi R_Y,\qquad \mathrm{Vol}(S^1_Y/\mathbb{Z}_2)_{\rm active}=\pi R_Y.
\]

 Evaluated at \(\vec u=(1,1,1)\) , \(R_6=R_2=R_0\) :

 \[
\mathrm{Vol}(K_6) = V_{K_6,0}R_0^6 = 2.327554010848277\times10^{-99}\ \mathrm{GeV}^{-6},
$$
$$
\mathrm{Vol}(S^2) = 4\pi R_0^2 = 3.183098861837907\times10^{-33}\ \mathrm{GeV}^{-2},
$$
$$
\mathrm{Vol}(S^1_Y)_{\rm parent} = 2\pi R_0 = 1.000000000000000\times10^{-16}\ \mathrm{GeV}^{-1}\ \big(\text{exact} = 1/M_U\big),
$$
$$
\mathrm{Vol}(S^1_Y/\mathbb{Z}_2)_{\rm active} = \pi R_0 = 5.000000000000000\times10^{-17}\ \mathrm{GeV}^{-1}\ \big(\text{exact} = 1/(2M_U)\big).
\]

 The exactness of the \(S^1_Y\) volumes is not accidental bookkeeping: the \(2\pi\) in the volume formula cancels the \(2\pi\) built into \(R_0=1/(2\pi M_U)\) , leaving the clean \(1/M_U\) (parent circle) and \(1/(2M_U)\) (active, post-orbifold interval) — a direct algebraic consequence of how the radius was defined from the unification scale.

 Multiplying the three internal factors gives the total internal volume:

 \[
\mathrm{Vol}(X_{\rm parent}) = \mathrm{Vol}(K_6)\,\mathrm{Vol}(S^2)\,\mathrm{Vol}(S^1_Y)_{\rm parent} = 7.408834522797404\times10^{-148}\ \mathrm{GeV}^{-9},
$$
$$
\mathrm{Vol}(X_{\rm active}) = \mathrm{Vol}(K_6)\,\mathrm{Vol}(S^2)\,\mathrm{Vol}(S^1_Y/\mathbb{Z}_2)_{\rm active} = 3.704417261398702\times10^{-148}\ \mathrm{GeV}^{-9}.
\]

 This 9-dimensional internal volume \(\mathrm{Vol}(X_{\rm active})\) is the object that fixes the higher-dimensional Planck scale \(M_*\) once the ordinary (measured) \(M_{\rm Pl}\) is supplied:

 \[
M_{\rm Pl}^2 = M_*^{\,D-2}\,\mathrm{Vol}(X_{\rm active}),\qquad D=13,
$$
$$
M_*^{11} = \frac{M_{\rm Pl}^2}{\mathrm{Vol}(X_{\rm active})} = 4.023836152402511\times10^{185}\ \mathrm{GeV}^{11},\qquad M_* = 7.467050992135091\times10^{16}\ \mathrm{GeV}.
\]

 \(M_*\) is therefore fixed by the geometry plus the one measured anchor \(M_{\rm Pl}\) — it is not an independent input, and it is not the same number as \(M_U\) (the two agree only to within an order of magnitude, as expected for two physically distinct thresholds — one a unification scale from gauge RG closure, the other a Kaluza–Klein/string-scale-like quantity from dimensional reduction of gravity). Under the reduced-Planck convention, the left-hand side is rescaled by \(1/(8\pi)\) ; the geometry on the right-hand side is completely unchanged, so this is purely a bookkeeping convention, not a second derivation.

 6. Discrete/topological structure: three generations, \(\mathbb{Z}_6\) charge quantization, chirality

 Family count. The spin- \(\mathbb{C}\) index on \(K_6\) twisted by the matter bundle \(E\) is \(\chi(K_6,E)=-3\) , i.e. three left-handed families. This number is given-E — it is a downstream result (properly SG-3's, recorded here only for arena-completeness) computed from the spin- \(\mathbb{C}\) structure whose Chern class is fixed to reproduce it; SG-1 does not derive the observed spectrum \(E\) , it only certifies that the branch carries the machinery ( \(S^{\rm spin^c}_{K_6}\) , the twist) capable of producing a family index at all.

 Charge quantization. The full gauge group realized on this arena is not simply the product \(SU(3)_c\times SU(2)_L\times U(1)_Y\) but its finest faithful quotient:

 \[
G_{\rm SM} = \big(SU(3)_c\times SU(2)_L\times U(1)_Y\big)/\mathbb{Z}_6,
\]

 with generator \(z=(\omega_3,-1,\zeta_6)\) identifying the center \(\mathbb{Z}_3\subset SU(3)_c\) , \(\mathbb{Z}_2\subset SU(2)_L\) , and a sixth root of unity in \(U(1)_Y\) . This is not asserted — it is certified by direct computation: the Smith normal form of the charge-character matrix has invariant factors \([1,6,6]\) , so the annihilator of the jointly-trivially-acting subgroup is \(\mathbb{Z}_6\) , the finest (largest) faithful quotient admissible — no coarser and no finer identification is consistent with the observed charge spectrum. \(\sum_f Y_f^2 = 10/3\) per generation [EXACT], and the hypercharge lattice is \(Y\in\tfrac16\mathbb{Z}\) , with \(Q=T_3+Y\) .

 Chirality / no-mirror mechanism. This is where the orbifold factor \(S^1_Y/\mathbb{Z}_2\) earns its physical keep. A closed odd-dimensional circle factor by itself would retain both handedness sectors, producing mirror fermions that are excluded by the LEP measurement of the \(Z\) width; the \(\mathbb{Z}_2\) orbifold identification \(\theta\mapsto-\theta\) projects the mirrors out. The chirality projector on the boundary is

 \[
P_\chi = \tfrac12\big(1+\gamma_5\Gamma_8\big),
\]

 with \(\Gamma_8\) the chirality operator on the internal 8-dimensional spinor bundle \(S(K_6)\otimes S(S^2)\otimes S(S^1_Y)\) . The Atiyah–Patodi–Singer index computed on the active interval \([0,\pi]\) returns \(n_L=+3\) , \(n_R=0\) — three left-handed families surviving, zero right-handed mirrors. At the level of the orbifold's two fixed points \(\theta=0,\pi\) : fields with parity \((+,+)\) — \(Q_L\) , \(L_L\) — have zero modes; fields with parity \((-,-)\) — \(u_R,d_R,e_R,\nu\) — have zero modes via the sector projectors \(\Pi_u,\Pi_d,\Pi_e,\Pi_\nu\) (part of the \(\oplus\) -layer rulebook, §8 below); every mirror-parity combination is forbidden, i.e. no mirror zero mode survives. The resulting Standard-Model hypercharge assignments are \(Y(Q_L)=+1/6\) , \(Y(u_R)=+2/3\) , \(Y(d_R)=-1/3\) , \(Y(L_L)=-1/2\) , \(Y(e_R)=-1\) , \(Y(H)=+1/2\) — the ordinary Standard Model hypercharges, reproduced from the orbifold parity structure rather than assumed.

 The orbifold defect itself is computed with the equivariant (Donnelly) heat-kernel trace, not an ordinary boundary condition: the reflection \(g\) -trace at each of the two isolated fixed points is \(1/|1-(-1)|=1/2\) , summing to a total reflection trace of \(1\) across both fixed points, giving orbifold traces \(K^{\pm}=\tfrac12 K_{\rm circle}\pm\tfrac12\) for even/odd parity respectively, with a per-fixed-point \(a_0\) defect of \(+1/4\) (parity \(+\) ) or \(-1/4\) (parity \(-\) ).

 7. Weak carrier \(S^2\) : representation content

 \(S^2\) carries the round metric \(ds^2_{S^2}=R_2^2(d\theta^2+\sin^2\theta\,d\phi^2)\) , Euler characteristic \(\chi(S^2)=2\) . Its isometry group \(SU(2)\) is the sole source of \(SU(2)_L\) in this arena — the dossier's binding statement that weak isospin is not hiding inside any \(SU(2)\subset SU(3)\) is made concrete here: it is the isometry of a different metric factor entirely. Dirac/Laplace eigenvalues on \(S^2\) are \(\ell(\ell+1)/R_2^2\) for \(\ell\ge |N|/2\) with degeneracy \(2\ell+1\) , where \(N\) is the monopole (spin- \(\mathbb{C}\) twist) charge:

 Monopole sector \(N\) 
 \(SU(2)_L\) representation routed 
 Physical role 

 \(0\) 
 \(\mathbf 1\) singlet 
 weak-singlet routing 

 \(1\) 
 \(\mathbf 2\) doublet 
 \(Q_L\) , \(L_L\) 

 \(2\) 
 \(\mathbf 3\) triplet 
 \(W^\pm, W^0\) adjoint 

 \(\ge3\) 
 \((N{+}1)\) -plet 
 higher KK thresholds 

 This is how the arena assigns the observed \(SU(2)_L\) doublet/singlet structure to specific monopole sectors on \(S^2\) , rather than inserting the representation content by hand at the 4D level.

 8. Color carrier: \(K_6\) representation theory and KK spectrum

 The quadratic Casimir on \(K_6=SU(3)/T^2\) for Dynkin labels \((p,q)\) (Killing normalization) and dimension are

 \[
C_2(p,q) = \frac{p^2+q^2+pq+3p+3q}{3},\qquad \dim(p,q)=\frac{(p+1)(q+1)(p+q+2)}{2}.
\]

 The lowest representations, all exact:

 \((p,q)\) 
 \(\dim\) 
 \(C_2\) 
 zero-weight mult \(m_0\) 
 Physical role 

 \((0,0)\) 
 1 
 \(0\) 
 1 
 trivial/scalars 

 \((1,0)\) 
 \(\mathbf3\) 
 \(4/3\) 
 0 
 quark color triplet, KK matter 

 \((0,1)\) 
 \(\bar{\mathbf3}\) 
 \(4/3\) 
 0 
 anti-quark triplet 

 \((1,1)\) 
 \(\mathbf8\) 
 \(3\) 
 2 
 \(SU(3)\) adjoint (gluons); lowest nonzero scalar harmonic, 16 modes at \(C_2=3\) 

 \((2,0)\) 
 \(\mathbf6\) 
 \(10/3\) 
 0 
 symmetric 2-index 

 \((2,1)\) 
 \(\mathbf{15}\) 
 \(16/3\) 
 0 
 mixed symmetry 

 \((3,0)\) 
 \(\mathbf{10}\) 
 \(6\) 
 1 
 totally symmetric 3-index 

 \((2,2)\) 
 \(\mathbf{27}\) 
 \(8\) 
 3 
 — 

 \((3,3)\) 
 \(\mathbf{64}\) 
 \(15\) 
 4 
 — 

 Peter–Weyl decomposes \(L^2(K_6,E_\mu)=\bigoplus_{(p,q)}V_{(p,q)}\otimes\mathrm{Hom}_{T^2}(V_{(p,q)},E_\mu)\) ; the scalar-sector KK multiplicity at each level equals the zero-weight multiplicity \(m_0(p,q)\) . KK masses are

 \[
m^2_{(p,q),\rm vec} = \frac{C_2(p,q)+\Delta_{\rm vec}}{R_6^2},\qquad
m^2_{(p,q),\rm Dirac} = \frac{C_2(p,q)+\|\rho\|^2+\Delta_{\rm spin^c}}{R_6^2},\quad \|\rho\|^2=2,
\]

 with \(\Delta_{\rm spin^c}\) the twist fixed so the chiral zero-mode count reproduces the family index \(-3\) .

 One item is honestly flagged OPEN at the representation-theory level rather than smoothed over: the off-diagonal (hopping) connection matrix elements mixing the five Weyl-inequivalent \(T^2\) weight classes on \(\mathrm{Sym}^2_0\) (the graviton bundle) are, in principle, exact SU(3) Gelfand–Tsetlin ladder matrix elements — a standard closed-form lowering-operator formula — but they have not yet been enumerated in the atlas. This is a bounded, named computation-debt (the " \(a_6\) graviton wall," §9 below), not an in-principle obstruction.

 9. Heat-kernel data: the \(\oplus/\otimes\) -layer operator content of \(K_6\) 

 Convention: \(K(t)\sim(4\pi t)^{-d/2}\sum_k a_{2k}t^k\) , densities per unit volume, with the exact product rule \(a_{2k}(M_1\times M_2)=\sum_{i+j=k}a_{2i}(M_1)a_{2j}(M_2)\) . All values below are [Killing-norm] at the Einstein center.

 Scalar heat-kernel ratios: 

 Space 
 \(a_2/a_0\) 
 \(a_4/a_0\) 
 \(a_6/a_0\) 

 \(K_6\) scalar 
 \(5/12\) 
 \(11/120\) 
 OWED (Gilkey constants; underlying curvature invariants certified) 

 \(S^2\) scalar ( \(r=1\) ) 
 \(1/3\) 
 \(1/15\) 
 \(4/315\) 

 \(S^6\) round unit (calibration control) 
 \(5\) 
 \(12\) 
 \(1139/63\) 

 The \(S^6\) row is a calibration control: it confirms the general \(a_4\) formula returns exactly \(12\) on the round unit 6-sphere, a passed check that \(K_6\) is being correctly distinguished from \(S^6\) (reinforcing the \(\|\mathrm{Riem}\|^2\ne60\) negative control of §4). \(K_6\) 's vector (tangent) bundle heat-kernel traces are \(\mathrm{tr}\,a_2=0\) , \(\mathrm{tr}\,a_4=-47/360\) with \(E=\mathrm{Ric}\) , \(\Omega=\mathrm{Riem}\) , \(\mathrm{tr}(\Omega_{ab}\Omega^{ab})=-|\mathrm{Riem}|^2\) .

 Bundle endomorphisms ( \(\Delta_{\rm bundle}=\nabla^*\nabla+E_{\rm bundle}\) , Einstein center \(\mathrm{Ric}=\tfrac5{12}g\) ) — this is the \(\otimes\) -Actors layer of the arena made explicit, operator by operator:

 Bundle 
 \(E\) (Weitzenböck endomorphism) 
 Spectrum/traces 

 scalar 
 \(E=0\) 
 — 

 vector/1-form (Hodge) 
 \(E=\mathrm{Ric}=\tfrac5{12}\,\mathrm{Id}\) 
 eigenvalue \(5/12\) , mult 6; \(\mathrm{tr}\,E=5/2\) , \(\mathrm{tr}\,E^2=25/24\) 

 graviton \(\mathrm{Sym}^2\) (dim 21) 
 \((E_Lh)_{ab}=\mathrm{Ric}_{ac}h^c{}_b+\mathrm{Ric}_{bc}h^c{}_a-2R_{acbd}h^{cd}\) 
 Lichnerowicz spectrum \(1/6\,(\times6)\) , \(5/12\,(\times6)\) , \(7/6\,(\times6)\) , \(17/12\,(\times2)\) , \(5/3\,(\times1,\text{pure-trace})\) 

 graviton TT \(\mathrm{Sym}^2_0\) (dim 20) 
 same, transverse-traceless 
 \(1/6\,(\times6)\) , \(5/12\,(\times6)\) , \(7/6\,(\times6)\) , \(17/12\,(\times2)\) ; \(\mathrm{tr}\,E_L=40/3\) , \(\mathrm{tr}\,E_L^2=241/18\) 

 The \(a_6\) two-route status — the arena's honest bounded hole. Route A (Gilkey/Lichnerowicz on \(\mathrm{Sym}^2_0\) ) needs the certified \(E_L\) spectrum plus \(\Omega=\mathrm{Riem}\) plus the Gelfand–Tsetlin off-diagonal hopping term flagged OWED in §8; it is blocked at that stratum. Route B (ghost + vector reconstruction) has its scalar backbone banked at \(a_6/a_2^3=7936/39375\) across three or more independent engines, but its graviton leg is likewise OWED. The two routes have not yet reached agreement — the graviton \(a_6\) coefficient is a documented computation-debt at a named stratum, not an in-principle gap, and is never quoted as a value in this dossier.

 10. \(F^+\) finite/operator chamber — the \(\oplus\) -layer, full precision

 \(F^+\) is non-metric (0 real dimensions) but carries the entire flavor structure of the theory. Its data tuple is \(\{\tau=\omega,\ \mathcal G_{\rm gen},\ \Pi_u,\Pi_d,\Pi_e,\Pi_\nu,\ O_u,O_d,O_e,O_\nu,\ \phi_i,\ N_i,\ \mathcal N_i,\ \mathrm{RG}\}\) .

 The Cartan-torus modulus sits at the order-3 modular fixed point:

 \[
\tau=\omega=e^{2\pi i/3} = -\tfrac12+i\tfrac{\sqrt3}{2} = -0.5000000000000000+0.8660254037844386\,i,
\]

 and the generation basis \(\mathcal G_{\rm gen}=\mathrm{span}\{g_1,g_2,g_3\}\) over \(\mathbb C\) has \(\dim_{\mathbb C}=3\) , matched to the family index \(-3\) of §6. Four orthogonal sector projectors \(\Pi_u,\Pi_d,\Pi_e,\Pi_\nu:\mathcal G_{\rm gen}\to\mathcal G_{\rm gen}\) satisfy \(\Pi_i\Pi_j=\delta_{ij}\Pi_i\) , each of rank 3 — these are the very projectors invoked in §6 to route the right-handed zero modes. The CKM holonomy phase is \(\delta_{\rm CKM}=-2\pi/3=-2.094395102393195\) rad \(=-120.0^\circ\) (Wolfenstein-aligned to \(+60.0^\circ\) ); the lepton Berry phase is \(+2\pi/3=+120.0^\circ\) , giving a leptonic CP-violating output \(\delta_{CP}^\ell\approx 260.2\pm10^\circ\) from the second-cycle Berry phase.

 Chamber Boltzmann factors , all exact functions of \(\tau=\omega\) :

 \[
\kappa = e^{-\pi\sqrt3} = 0.004333420509983131,\qquad K_{tb}^{\rm crit}=e^{-\pi\sqrt3/16}=0.7117081304239685,
$$
$$
\eta_{BK} = \frac{1}{32\pi\,e^{\sqrt3/(24\pi)}} = 0.009721281516312024,\qquad \frac{1}{\eta_{BK}}=32\pi\,e^{\sqrt3/(24\pi)}=102.8670961047707.
\]

 Action ladders and sector-level normalizations (the mechanism that turns a discrete ladder into the observed mass hierarchy): up-sector ladder \(a_u=(2,1,0)\) with \(N_u=1\) (fixes the up-anchor via \(y_t\) ); down-sector \(a_d=(4/3,2/3,0)\) with \(N_d=2.4\times10^{-2}\) (fixes \(m_b\) at \(M_Z\) ); charged-lepton \(a_e=(2,4/3,0)\) with \(N_e=1.02\times10^{-2}\) (fixes \(m_\tau\) at \(M_Z\) ); neutrino \(a_\nu=(1,1/2,0)\) , \(N_\nu\) structural. The diagonal chamber operators \((O_i)^{aa}=N_i\kappa^{a_i^{(a)}}\) evaluate, at \(\tau=\omega\) , to:

 \[
O_u = \mathrm{diag}(1.877853331634246\times10^{-5},\ 4.333420509983131\times10^{-3},\ 1),
$$
$$
O_d = \mathrm{diag}(1.695582872666127\times10^{-5},\ 6.379184034340682\times10^{-4},\ 2.4\times10^{-2}),
$$
$$
O_e = \mathrm{diag}(1.915410398266931\times10^{-7},\ 7.206227208831040\times10^{-6},\ 1.02\times10^{-2}),
$$
$$
O_\nu = \mathrm{diag}(4.333420509983131\times10^{-3},\ 6.582872101129666\times10^{-2},\ 1).
\]

 The Yukawa map is \((Y_i)^{ab}=N_i\langle g_a|O_i|g_b\rangle\) , \(i\in\{u,d,e,\nu\}\) , with the chamber angle \(\theta_F\) (a DFT-on- \(\mathbb Z_3\) rotation) fixed by the \(|V_{us}|\) anchor. The binding rule that keeps this predictive rather than fitted: normalizations are sector-level only — a single \(N_i\) per sector, never a separate \(N_{i,a}\) per family member — so the within-sector hierarchy pattern \(\kappa^{a^{(a)}}\) is a genuine prediction of the ladder exponents, not a per-family fit; all phases are read from the order-3 holonomy at \(\tau=\omega\) and the second-cycle Berry phase, none tuned post-comparison.

 The Higgs sits in this arena as a Wilson-line (Hosotani) mode with integer winding \(n_H=1\) around a cycle of radius \(R_\gamma\sim R_0\) (center value \(1.591549430918954\times10^{-17}\ \mathrm{GeV}^{-1}\) ); \(n_H=0\) would give no VEV, so \(n_H=1\) is the minimal nonzero admissible winding. The Hosotani potential \(V_{\rm Hos}(\theta_H)=-\frac{3}{64\pi^6R_\gamma^4}\sum_{n=1}^\infty\frac{1}{n^5}[N_b-N_f]\cos(n\theta_H)\) converges absolutely (an \(n^{-5}\) tail), guaranteeing a finite Higgs mass; the post-RG outputs are \(v_{\rm pred}=246.02\pm3.5\) GeV, \(m_h=123.82\pm1.8\) GeV, \(\lambda_H = m_h^2/(2v^2)=0.12722\pm0.00181\) .

 The \(\oplus\) -layer rulebook also carries the admissibility firewall \(\mathcal C_{\rm admiss}\) : selector v3 (Search/Compare/Judge/Reconcile/Decide), constraint set C1–C14, the freeze-before-compare barrier (comparison data loaded only after the freeze — the discipline that keeps this whole gate target-blind), anomaly-cancellation trace identities, the no-mirror parity table of §6, the Wilson-line winding rule, and the FCNC/mediator no-go theorem \(\Pi_qM\Pi_\ell=0\) for any sector-respecting operator \(M\) (the proton-safety projector identity, combined with BRST decoupling and KK-number conservation).

 11. \(\otimes\) -Actors layer: the standard bundle/operator index

 Each physical object in this arena is pinned at all three layers simultaneously — × Stage (manifold + bundle base), ⊕ Rulebook (scheme/convention/boundary/projector/grading), ⊗ Actors (connection \(\nabla\) , endomorphism \(E\) , operator domain, readout). The total Hilbert space factorizes as

 \[
\mathcal H_{\rm total} = \mathcal H_{\mathcal M_4}\otimes\mathcal H_{K_6}\otimes\mathcal H_{S^2}\otimes\mathcal H_{S^1_Y/\mathbb Z_2}\otimes\mathcal H_{F^+}\otimes V_{\rm gauge}\otimes V_{\rm spin}\otimes V_{\rm flavor},
$$
$$
\mathcal E_{\rm matter} = S_{3,1}\otimes S^{\rm spin^c}_{K_6}\otimes L_Y\otimes V_{SU(3)}\otimes V_{SU(2)}\otimes V_{F^+}.
\]

 The full index of objects this gate touches, at all three layers:

 Bundle/operator 
 × Stage (base) 
 ⊕ Rulebook 
 ⊗ Actors 

 Scalar Laplacian \(\Delta_0\) 
 \(K_6\) (and each × factor) 
 Killing-norm normal metric, Einstein center; \(\overline{\rm MS}\) 
 \(\nabla=\) Levi-Civita (Nomizu), \(E=0\) ; domain \(C^\infty(K_6)\) ; readout spectrum \(C_2(p,q)/R_6^2\) 

 Vector/Hodge Laplacian 
 \(T^*K_6\) 
 1-form grading, same metric 
 \(\nabla=\) LC, \(E=\mathrm{Ric}=\tfrac5{12}\mathrm{Id}\) (mult 6); Weitzenböck 

 Graviton \(\mathrm{Sym}^2_0\) 
 \(\mathrm{Sym}^2_0T^*K_6\) (dim 20) 
 TT gauge, Lichnerowicz grading 
 \(E_L\) spectrum \(\{1/6,5/12,7/6,17/12\}\) ; GT-hopping off-diagonal OWED 

 Dirac \(\slashed D_{K_6}\) (spin- \(\mathbb C\) ) 
 \(S^{\rm spin^c}_{K_6}\) 
 spin- \(\mathbb C\) structure, Chern class fixed to family index \(-3\) 
 \(\nabla=\) spin- \(\mathbb C\) connection; \(m^2=(C_2+\|\rho\|^2+\Delta_{\rm spin^c})/R_6^2\) 

 Dirac/Laplace \(S^2\) 
 \(S^{\rm spin^c}_{S^2}\) , monopole sector \(N\) 
 monopole grading \(N\in\{0,1,2,\dots\}\) 
 eigenvalues \(\ell(\ell+1)/R_2^2\) , $\ell\ge 

 Hypercharge line bundle 
 \(L_Y\) on \(S^1_Y/\mathbb Z_2\) 
 \(\mathbb Z_2\) orbifold parity, \(Y\in\tfrac16\mathbb Z\) , \(\mathbb Z_6\) center 
 KK momentum \(p_\theta=(n+\alpha)/R_Y\) , twist \(\alpha\in\{0,Y\}\) 

 Gauge \(\mathcal E_{\rm gauge}\) 
 \(T^*\mathcal M_4\otimes\mathrm{ad}(P)\) on \(\mathcal M_4\times K_{\rm gauge}\) 
 BRST/FP gauge-fixing, Gribov domain 
 \(A,F,\rho_{\rm rep}\) , KK tower; \(Q_{\rm BRST}\) off-shell → \(\mathcal H_{\rm phys}\) cohomology 

 Higgs \(\mathcal E_{\rm Higgs}\) 
 \(L_\gamma\otimes V_{SU(2),\rm doub}\) on cycle \(\gamma\) 
 Wilson-line winding \(n_H=1\) , Hosotani grading 
 holonomy \(\theta_H\) ; readout \(=V_{\rm Hos}\) minimum 

 Proton \(\mathcal E_{\rm proton}\) 
 \(\Pi_qE_{\rm matter}\otimes\Pi_\ell E_{\rm matter}\) 
 sector-orthogonal partition 
 four-fermion domain; identity \(\Pi_qM\Pi_\ell=0\) 

 12. Thresholds and the gauge-coupling routing (how the arena connects to \(M_Z\) )

 The one-loop Standard Model beta coefficients (GUT-normalized \(\alpha_1=\tfrac53\alpha_Y\) ), fixed entirely by SM content (3 chiral generations + 1 Higgs doublet + SM gauge sector), are

 \[
b_1^{\rm SM}=\frac{41}{10}=4.1,\qquad b_2^{\rm SM}=-\frac{19}{6}=-3.166666666666667,\qquad b_3^{\rm SM}=-7.
\]

 The Kaluza–Klein threshold packets, summed over every compact factor of the arena via the heat-kernel ledger, give the threshold vector

 \[
(\delta_1,\delta_2,\delta_3) = (+4.8424,\ -3.1112,\ -1.7313)\ \pm\ 1.6\times10^{-3},
\]

 assembled from named packets: \(K_6\) matter contributes \((0,0,+0.79)\) ; \(S^2\) matter contributes \((0,+0.92,0)\) ; the \(K_6\) weak/color gauge+ghost net contributes \((0,-4.02,-2.49)\) ; the \(S^1_Y/\mathbb Z_2\) hypercharge gauge packet contributes \((-0.84,0,0)\) ; the hypercharge zero-mode matter packet ( \(\sum Y^2=10/3\) per generation \(\times3\) ) contributes \((+3.214,0,0)\) ; the Higgs Wilson line ( \(n_H=1\) ) contributes \((+1.047,-0.211,0)\) ; and the orbifold boundary at \(\theta\in\{0,\pi\}\) contributes \((+1.4214,+0.1998,-0.0313)\) . Feeding this threshold vector into two-loop \(\overline{\rm MS}\) running from \(M_Z=91.1876\) GeV closes the three inverse couplings at \(M_U\) to a residual of \(9.6\times10^{-11}\) (the numerical-pipeline floor), comfortably inside the propagated PDG uncertainty band of order \(10^{-3}\) .

 Gauge couplings themselves are routed as \(g_A^{-2}=M_*^{D-2}\int_{X_{\rm int}}\sqrt g\,|\xi_A(y)|^2\,d^{D-4}y\) , with \(SU(3)_c\leftarrow K_6\) , \(SU(2)_L\leftarrow S^2\) , \(U(1)_Y\leftarrow S^1_Y/\mathbb Z_2\) , and \(1/g_{\rm em}^2=1/g_1^2+1/g_2^2\) at \(M_Z\) . It must be stated plainly, because it bears directly on what SG-1 does and does not certify: the three \(\alpha_i^{-1}(M_Z)\) values are declared anchors , not first-principles predictions — these routing integrals are internal consistency links between the higher-dimensional and 4D couplings, not a derivation of the coupling values themselves. What SG-1's arena does certify at the gauge level is the algebra — that the surviving 4D gauge symmetry, after all KK towers and orbifold projections are accounted for, is exactly \(\mathfrak{su}(3)_c\oplus\mathfrak{su}(2)_L\oplus\mathfrak{u}(1)_Y\) with charge quantization \(\mathbb Z_6\) , nothing more and nothing less.

 13. Dynkin indices and the four anchors that close the ledger

 Exact group-theory constants used throughout the threshold and Yukawa calculations: \(T_{\rm adj}(SU(3))=3\) , \(T_{\rm adj}(SU(2))=2\) , \(T(\mathbf3)=T(\mathbf2)=1/2\) , \(\sum_fY_f^2=10/3\) per generation.

 The arena's only genuinely free numerical inputs are the four headline anchors:

 \[
\{\,M_{\rm Pl},\ \alpha_i(M_Z),\ y_t,\ |V_{us}|\,\} \quad [\text{MEASURED}],
\]

 from which every radius, volume, curvature invariant, Casimir, and chamber operator quoted above is derived or exact-topological — none is independently free. This dossier does not claim these four alone close the entire observable ledger without further cost: an honest accounting (developed in the forcedness section of this gate) also charges roughly 9–10 additional injected reals (the sector normalizations \(N_d,N_e,N_\nu\) , the threshold triple \(\delta_i\) , the Higgs modulus \(\theta_H^\star\) ), for a total honest branch cost of order 13–14 measured reals. That accounting belongs to the economy argument proper; here the point is narrower and purely geometric: every curvature number, every volume, every Casimir, and every heat-kernel coefficient in this section is a computed consequence of the frozen shape and the four anchors — not a separately tunable dial.

 14. What this arena physically carries — summary of the pinning

 Collecting the full picture: the ×-Stage is a 13-dimensional metric product, \(\mathcal M_4\times K_6\times S^2\times S^1_Y/\mathbb Z_2\) , whose three internal factors are dedicated, non-overlapping isometry sources for the three observed gauge groups — \(K_6=SU(3)/T^2\) for color, \(S^2\) for weak isospin, the orbifolded circle for hypercharge plus the chirality filter that removes mirror fermions. The ⊕-Rulebook is the zero-dimensional but load-bearing finite/operator chamber \(F^+\) (the flavor structure: generation basis, sector projectors, the Boltzmann-factor mass ladder, the CKM/lepton phases, all evaluated at the single modular fixed point \(\tau=\omega\) ) plus the admissibility firewall \(\mathcal C_{\rm admiss}\) that keeps the whole construction target-blind. The ⊗-Actors layer is the concrete set of connections, endomorphisms, and operator domains — the scalar/vector/graviton Laplacians on \(K_6\) , the spin- \(\mathbb C\) Dirac operators, the gauge and Higgs bundles, the proton-safety projector — each pinned with its own grading and readout. Every curvature invariant is an exact rational at the Killing-form Einstein center; every radius and volume is a 16-significant-figure derived quantity tied to the unification scale \(M_U\) and the measured \(M_{\rm Pl}\) ; the topology ( \(\chi(K_6)=6=|S_3|\) , the \(\mathbb Z_6\) charge quotient with Smith normal form \([1,6,6]\) , the \(n_L=3,n_R=0\) chirality index) is exact and certified. The only genuinely open items at the geometric level — never smoothed over, never quoted as values — are the graviton \(a_6\) heat-kernel coefficient (blocked at the Gelfand–Tsetlin hopping stratum) and the scalar \(a_6/a_0\) Gilkey constant; both are named, bounded computation-debts, not in-principle obstructions, and neither is used anywhere in this dossier's central claims. This is the complete, frozen, three-layer object that SG-1 commits — the fixed point every downstream gate, and the shape-selection/economy argument that gives SG-1 its REDUCED-TO-AXIOM / ANCHORED +1 terminal, is evaluated against.

 Construction I - the deep-root anchoring

 This construction runs the gate's single live forcedness seam — the claim that the frozen 13-dimensional branch is the outright bit-cost winner over its competitor family — through all three complete roots (Shape, Scale, Granularity), each applied at full three-layer discipline and full precision, and then through the four Layer-2 admissibility screens (Invariance, Record-Interface, Causal-Order, Nonseparability). The exercise is diagnostic, not decorative: a residual or an unresolved comparison seen under a truncated object — say, the \(\times\) -Stage dimension count alone, with the \(\oplus\) -Rulebook and \(\otimes\) -Actors layers silently dropped — would be an artifact of the truncation, not a fact about the geometry. Running the complete object at every root removes that possibility and lets the seam be located precisely rather than approximately.

 The branch under test throughout is the frozen three-layer object

 \[
\mathfrak{B}_{\rm active}
=\underbrace{\big[\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2\big]}_{\times\ \text{Stage, 13 dims}}
\ \oplus\ \underbrace{\big[F^+_{\rm finite}\oplus C_{\rm admiss}\big]}_{\oplus\ \text{Rulebook, 0 dims}}
\ \otimes\ \underbrace{\big[E_{\rm matter}\oplus E_{\rm gauge}\oplus E_{\rm Higgs}\oplus E_{\rm proton}\big]}_{\otimes\ \text{Actors, 0 dims}}\,,
\]

 with \(K_6=SU(3)/T^2\) the full \(A_2\) flag manifold and \(S^1_Y/\mathbb{Z}_2\) the active orbifold interval; \(D=4+6+2+1=13\) exactly, carried by the \(\times\) -Stage alone, with the \(\oplus\) and \(\otimes\) layers non-metric but load-bearing. The result, stated in advance so each step below can be checked against it: none of the three roots is truncated, none forces the missing comparison, and the seam survives all three passes intact as one clean, closed-candidate-class axiom — never as an unbounded regress and never as a silently smuggled assumption.

 I.1 Why a root-by-root pass is the right diagnostic here

 The forcedness argument for SG-1 is a description-length economy comparison : the frozen 13D branch is claimed to be cheaper, in bits, than its rivals — a family that includes other dimension counts, other coset choices, and the "clean 4D effective field theory with everything put in by hand" alternative. The comparison is built to be spectrum-neutral: the observed Standard Model chiral content \(E\) sits on both sides of the ledger, so it contributes identically to the 13D branch's cost and to each rival's cost and cancels out of the difference . What remains, per side, decomposes into two structurally different kinds of charge:

 a dimension/structural cost — how many discrete choices (a coset, a \(\mathbb{Z}_6\) identification, an integer spin- \(\mathbb{C}\) index) the branch's Shape must specify, charged at \(O(1)\) bits per discrete choice under the codebook below; and

 an anchor/tuning cost — how many continuous, measured real numbers the branch must inject to reach the observed spectrum, each charged at \(b=\log_2(1/\Delta_0)\) bits, where \(\Delta_0>0\) is the finite measurement resolution (cell width) of a generator.

 The content of "13D wins" is the claim that a small number of expensive anchor-bits ( \(n_{13}\approx13\) –14 real numbers: the four anchors \(\{M_{\rm Pl},\alpha_i(M_Z),y_t,|V_{us}|\}\) plus roughly nine to ten injected reals — \(N_d=2.4\times10^{-2}\) , \(N_e=1.02\times10^{-2}\) , \(N_\nu=1\) , the threshold \(\delta\) -triple, \(\theta_H^\star\) ) beats a family of rivals that trade a handful of extra structural bits for a larger anchor bill (a first-pass ladder survey records 10 LOSE / 1 FAIL(structural) / 0 REFUTED , honest margin \(\approx1.8\times\) at transparent cost — \(\approx13\) – \(14\cdot b\) against the target ledger \(T\approx25\cdot b\) reals: 3 gauge couplings, 9 charged-fermion masses, 4 CKM parameters, 6 neutrino parameters, 2 electroweak parameters, 1 strong-CP bound — not the retired \(\approx4\times\) headline). But "beats," stated that way, presupposes that dimension-bits and anchor-bits can be added on one scale at all. That presupposition is exactly what each root is interrogated for below: does Shape supply it, does Scale supply it, does Granularity supply it? The answer, root by root, is no — each root does real, non-trivial work on the gate, but none of that work is the missing aggregation rule, and the honest boundary of what each root buys is stated precisely so nothing is smuggled across it.

 I.2 Shape, run completely (all three layers): delivers identity, not a scoring rule

 What "running Shape completely" means here. Shape is evaluated at all three layers simultaneously, not at the \(\times\) -Stage metric-dimension count alone — that would be exactly the truncation this program is built to avoid. The complete Shape object is:

 \(\times\) Stage — the manifold-plus-bundle content: which factors ( \(\mathcal{M}_4=\mathbb{R}^{3,1}\) , \(K_6=SU(3)/T^2\) , \(S^2\) , \(S^1_Y/\mathbb{Z}_2\) ), which radii ( \(R_0\equiv(2\pi M_U)^{-1}=1.591549430918954\times10^{-17}\,\mathrm{GeV}^{-1}\) at the chamber center, \(R_Y=7.957747154594768\times10^{-18}\,\mathrm{GeV}^{-1}\) post- \(\mathbb{Z}_2\) -halving), which connection background ( \(K_6\) 's Weyl-rigid squashing moduli \(\vec u\in[1/2,3/2]^3\) , chamber center \(\vec u=(1,1,1)\) ).

 \(\oplus\) Rulebook — the scheme, convention, boundary condition, projector, and grading data: the \(\mathbb{Z}_2\) orbifold parity \(\theta\mapsto-\theta\) on \(S^1_Y\) ; the admissibility constraint set \(C_{\rm admiss}\) (selector v3, constraints C1–C14, the freeze-before-compare barrier, the FCNC/mediator no-go \(\Pi_qM\Pi_\ell=0\) ) that eliminates off-chamber, non-Weyl-rigid squashings and illegal deformations; the finite chamber \(F^+_{\rm finite}\) 's Cartan-torus modulus fixed at the order-three point \(\tau=\omega=e^{2\pi i/3}=-\tfrac12+i\tfrac{\sqrt3}{2}\) ; the two-loop \(\overline{\rm MS}\) RG scheme used to define the threshold vector.

 \(\otimes\) Actors — the connection \(\nabla\) (Levi-Civita/Nomizu on \(K_6\) , spin- \(\mathbb{C}\) connections on the matter bundles), the bundle endomorphisms \(E_{\rm matter}\) , \(E_{\rm gauge}\) , \(E_{\rm Higgs}\) , \(E_{\rm proton}\) , the chirality projector \(P_\chi=\tfrac12(1+\gamma_5\Gamma_8)\) on the 8-dimensional internal spinor bundle \(S(K_6)\otimes S(S^2)\otimes S(S^1_Y)\) , and the readout maps (the Atiyah–Patodi–Singer index on \([0,\pi]\) returning \(n_L=+3\) , \(n_R=0\) ; the Yukawa map \((Y_i)^{ab}=N_i\langle g_a|O_i|g_b\rangle\) ).

 Only the \(\times\) -Stage layer carries metric dimension: \(D=4+6+2+1=13\) exactly. The \(\oplus\) and \(\otimes\) layers are non-metric — zero-dimensional in the dimension count — but load-bearing in exactly the sense that matters for this root: they are part of what makes a candidate branch admissible at all. A candidate that specified the right \(\times\) -Stage manifold but the wrong \(\oplus\) -Rulebook parity (say, no \(\mathbb{Z}_2\) quotient, so mirror fermions survive and the LEP \(Z\) -width constraint is violated) or the wrong \(\otimes\) -Actors isotropy would simply not be in the admissible set. This is not hypothetical: the rival internal space \(CP^2=SU(3)/U(2)\) was built out to the same three-layer completeness and shown to fail structurally rather than merely to score worse — its isotropy group \(U(2)=(SU(2)\times U(1))/\mathbb{Z}_2\) is non-abelian and sits inside color \(SU(3)\) , so by the coset-space-dimensional-reduction centralizer rule it is gauge-active and over-produces \(SU(2)+U(1)\) at the gauge-recovery gate. That is a hard structural exclusion, not an economy comparison — a genuine data point for what complete Shape can deliver, distinct from the softer economy question this construction is about.

 What complete Shape delivers. Running Shape at all three layers answers the question "which records are admissible at all" — it is a domain restriction . It is the root that certifies, for instance, that \(K_6=SU(3)/T^2\) has exactly four invariant Einstein metrics (the normal metric at \(\vec u=(1,1,1)\) plus the three Kähler–Einstein permutations of \((1,1,2)\) ), that \(\chi(K_6)=6\) exactly (equal to \(|S_3|=6\) , the order of the Weyl group, as required for a full flag manifold), that \(\chi(S^2)=2\) and \(\chi(S^1_Y/\mathbb{Z}_2)=1\) exactly, and that the curvature invariants at the Killing-form-normal center are the exact rationals \(\mathrm{Ric}_i=5/12\) , \(\mathrm{Scal}=5/2\) , \(\mathrm{Scal}/\mathrm{Ric}_i=6=\dim K_6\) , \(|\mathrm{Ric}|^2=25/24\) , \(|\mathrm{Riem}|^2=23/12\) (never \(31/147\) , never \(60\) — the frozen negative controls, \(60\) being the round unit \(S^6\) 's value, a different manifold entirely), \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75\) , \(|\mathrm{Ric}|^2/\mathrm{Scal}^2=1/6\) , the cubic curvature invariants \(K_1=R_{ab}{}^{cd}R_{cd}{}^{ef}R_{ef}{}^{ab}=-113/72\) and \(K_2=R_{abcd}R_{aecf}R_{ebfd}=-5/72\) , and \(|\nabla\mathrm{Riem}|^2=1/4\neq0\) (so \(K_6\) is homogeneous but not locally symmetric, with zero second-Bianchi violations). It is also the root that certifies the global center structure: the Smith normal form of the charge-character matrix on \(\mathbb{Z}_3\times\mathbb{Z}_2\times\mathbb{Z}_6\) has invariant factors \([1,6,6]\) , so \(G_{\rm SM}=(SU(3)_c\times SU(2)_L\times U(1)_Y)/\mathbb{Z}_6\) is the finest faithful quotient — no coarser or finer identification is admissible — and that \(\sum_f Y_f^2=10/3\) per generation. All of this is real, certified, three-layer-complete work, and it is precisely what lets the economy comparison in Section I.4 even be stated meaningfully: without a well-posed admissible-record set, "count the bits of branch \(A\) versus branch \(B\) " would not be well-defined.

 What complete Shape cannot deliver. Shape, run at all three layers, is a statement about which combinations of \((\times,\oplus,\otimes)\) data are legal records — it says nothing about how to compare the magnitude of one legal record's cost against another's. In particular, it cannot express a cross-layer, cross-kind magnitude inequality of the form "one unit of dimension-count is worth more or less than one unit of anchor-precision." Concretely: Shape tells us that \(\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2\) is an admissible 13-dimensional record and that a hypothetical bare \(\mathbb{R}^{3,1}\) four-dimensional effective field theory carrying the same Standard Model content by hand, with \(T\approx25\) free real parameters, is also an admissible record, in the trivial sense that both specify a legal three-layer object. Shape has no machinery — none of its three layers is built to carry a numeric weighting function over records — that would let it declare one of these two admissible records "cheaper" than the other. Adopting an additive-over-lexicographic scoring rule at the Shape level, so that Shape itself decided the winner, would be target-anchoring : quietly building the desired answer (13D wins) into the very definition of admissibility, rather than deriving it. The complete-Shape verdict is therefore recorded honestly as:

 Shape verdict: PASS / no purchase on the Archimedean question. Shape is necessary — without it there is no well-posed comparison — but it is not sufficient, and no re-reading of any of its three layers, however carefully expanded, changes that. This is a genuine finding, not a hedge: it tells the gate precisely where not to look for the missing rung.

 I.3 Scale, run completely (the full \(M_{\rm Pl}\) -anchored discipline): exposes the two competing cost axes

 What "running Scale completely" means here. Scale is the root that fixes the physical normalization of every dimensionful quantity in the branch against the one truly external anchor, the ordinary (not reduced) Planck mass \(M_{\rm Pl}=1.220900000000000\times10^{19}\) GeV (the reduced convention \(\bar M_{\rm Pl}=M_{\rm Pl}/\sqrt{8\pi}\approx2.4357\times10^{18}\) GeV is the alternative, unused here), together with the full boundary/UV package the branch commits to. Concretely, running Scale completely means carrying through the entire derived chain:

 \[
M_{\rm Pl}^2 = M_*^{D-2}\,\mathrm{Vol}(X_{\rm int}), \qquad D=13,\qquad X_{\rm int}=K_6\times S^2\times(S^1_Y/\mathbb{Z}_2),\ \ \dim X_{\rm int}=9,
\]

 with the active internal volume
$$
\mathrm{Vol}(X_{\rm active})=\mathrm{Vol}(K_6)\cdot\mathrm{Vol}(S^2)\cdot\mathrm{Vol}(S^1_Y/\mathbb{Z} 2)=3.704417261398702\times10^{-148}\,\mathrm{GeV}^{-9},
$$
built from \(\mathrm{Vol}(K_6)=V_{K_6,0}R_0^6=2.327554010848277\times10^{-99}\,\mathrm{GeV}^{-6}\) with \(V_{K_6,0}=(2\pi)^3/\sqrt3=143.2118575035129\) , \(\mathrm{Vol}(S^2)=4\pi R_0^2=3.183098861837907\times10^{-33}\,\mathrm{GeV}^{-2}\) , and the active \(\mathrm{Vol}(S^1_Y/\mathbb{Z}_2)=\pi R_0=5.000000000000000\times10^{-17}\,\mathrm{GeV}^{-1}\) (exact \(=1/(2M_U)\) , since the \(2\pi\) in the volume cancels the \(2\pi\) in \(R_0=1/(2\pi M_U)\) ). This gives the derived fundamental scale
$$
M ^{11}=\frac{M_{\rm Pl}^2}{\mathrm{Vol}(X_{\rm active})}=4.023836152402511\times10^{185}\ \mathrm{GeV}^{11},\qquad
M_ =7.467050992135091\times10^{16}\ \mathrm{GeV}.
$$
 \(M_*\) is not an independent input; it is fixed by the one anchor \(M_{\rm Pl}\) and the derived geometric volume. Running Scale completely also carries the unification-scale closure \(M_U=1.0\times10^{16}\) GeV (declared closure target \(\alpha_1^{-1}(M_U)=\alpha_2^{-1}(M_U)=\alpha_3^{-1}(M_U)\) , residual \(9.6\times10^{-11}\) , well inside the propagated PDG uncertainty band \(\sim10^{-3}\) and understood as a numerical-pipeline floor, not a physical discrepancy), the compactification radius \(R_0\equiv(2\pi M_U)^{-1}=1.591549430918954\times10^{-17}\,\mathrm{GeV}^{-1}\) , and the full threshold vector
$$
(\delta_1,\delta_2,\delta_3)=(+4.842400000000000,\,-3.111200000000000,\,-1.731300000000000)\ \pm\ 1.6\times10^{-3}
$$
against the one-loop Standard Model beta coefficients \(b_1^{\rm SM}=41/10\) , \(b_2^{\rm SM}=-19/6\) , \(b_3^{\rm SM}=-7\) (both quantities part of the full scale-discipline check package used throughout the branch).

 What complete Scale exposes. Carrying this entire chain through does real diagnostic work: it makes explicit, in a way a partial or truncated scale analysis would not, that the branch's cost genuinely splits into two structurally distinct axes that Scale itself does not merge. On one axis sits the anchor axis — the measured, continuous inputs \(\{M_{\rm Pl},\alpha_i(M_Z),y_t,|V_{us}|\}\) plus the further \(\sim9\) –10 injected reals ( \(N_d=2.4\times10^{-2}\) , \(N_e=1.02\times10^{-2}\) , \(N_\nu=1\) , the \(\delta\) -triple, \(\theta_H^\star\) ) that are needed to fix the volumes, radii, and threshold corrections to the observed values — each of these costs \(b=\log_2(1/\Delta_0)\) bits because each is a real number resolved only to finite measurement precision. On the other axis sits the structural axis — the discrete choices (which coset, which orbifold quotient, which center identification \(\mathbb{Z}_6\) ) that Scale takes as already-fixed inputs from Shape and simply propagates through the volume and RG formulas; these cost \(O(1)\) bits each because they are one-of-a-finite-menu selections, not continuously tunable reals. Scale, run in full, shows unambiguously that these two axes exist, that they are populated by genuinely different kinds of quantity (a coset choice is not a measured real; a measured real is not a discrete label), and that the entire branch-versus-rival economy argument is a claim about trading a few units of one axis for several units of the other.

 What complete Scale does not supply. What Scale conspicuously does not hand over, at any point in the derivation chain above, is the exchange rate between the two axes — the number of anchor-bits one structural-bit is worth, or vice versa. The RG/volume machinery converts one anchor (say \(\alpha_3(M_Z)\) ) into a derived quantity (say \(M_U\) , then \(R_0\) , then \(\mathrm{Vol}(K_6)\) ) through clean multiplicative and exponential relations, but that is a conversion within the anchor axis — dimensionful reals feeding into other dimensionful reals via exact formulas — never a conversion between the anchor axis and the structural axis. Nothing in \(M_{\rm Pl}^2=M_*^{11}\mathrm{Vol}(X_{\rm active})\) , nothing in the threshold vector, nothing in \(R_0=(2\pi M_U)^{-1}\) produces a number of the form "1 coset choice \(=k\) measured bits" for any \(k\) . The full scale-discipline package was run precisely to check for such a hidden conversion rate and found none. The complete-Scale verdict is recorded as:

 Scale verdict: EXPOSE. Scale does not pass or fail the Archimedean question the way Shape does — it sharpens it, by making the two-axis structure of the cost ledger fully explicit and ruling out the possibility that the missing exchange rate was hiding inside the RG/volume machinery under a less complete analysis. This is itself valuable: it closes off one plausible location (Scale) where the missing rung might have been quietly supplied, and hands the question forward, correctly isolated, to Granularity.

 I.4 Granularity, run completely (full cost-floor over 13 dimensions times 3 layers): constrains — supplies a currency but not an exchange ratio

 What "running Granularity completely" means here. Granularity is the root that assigns a finite cost floor \(\Delta_0>0\) to every distinguishable record across the entire 13-dimensional \(\times\) -Stage, the \(\oplus\) -Rulebook grading, and the \(\otimes\) -Actors operator domain simultaneously — not to the \(\times\) -Stage dimension count alone. Concretely, this means: every one of the 13 metric dimensions carries a finite resolution \(\Delta_0\) below which two field configurations are indistinguishable (the physical content of \(\hbar\) as the granularity residue, tracked in this program as the physical-observable identifier OBS-0002, layer-2 class AUDIT — configuration data, not a measured prediction); every discrete label in the \(\oplus\) -Rulebook (the \(\mathbb{Z}_2\) orbifold parity, the \(\mathbb{Z}_6\) center identification with Smith-normal-form invariant factors \([1,6,6]\) , the chamber admissibility constraint set \(C_{\rm admiss}\) ) is a finite, enumerable choice with an \(O(1)\) bit cost; and every \(\otimes\) -Actors operator (the endomorphisms \(E_{\rm matter}\) , \(E_{\rm gauge}\) , \(E_{\rm Higgs}\) , \(E_{\rm proton}\) , the sector projectors \(\Pi_u,\Pi_d,\Pi_e,\Pi_\nu\) ) is specified to finite precision by the same \(\Delta_0\) -resolution rule. Running Granularity completely — across all three layers, not merely the metric dimensions — is what yields the codebook used throughout the economy argument: a tuned real charged to cell resolution \(\Delta_0\) costs \(b=\log_2(1/\Delta_0)\) bits (large, because \(\Delta_0\) is small); a discrete structural choice costs \(O(1)\) bits (small, because the menu of choices is finite and short).

 What complete Granularity supplies. This is the strongest of the three roots for this gate, and it is the one flagged in the endpoint anchoring as load-bearing. Granularity supplies a genuine common currency — bits — in which both a dimension-count and an anchor-count can be expressed as numbers of the same type, for the first time making the comparison " \(n_{13}\,b\) versus \(n_{\rm rival}\,b\) " a comparison between two quantities measured in the same unit rather than between a discrete label and a continuous quantity of incommensurable type. This is not a small result: it is the reason the ladder survey (10 LOSE / 1 FAIL(structural) / 0 REFUTED, margin \(\approx1.8\times\) ) can be run as an arithmetic comparison of two numbers at all, \(n_{13}\,b\approx(13\text{–}14)\,b\) against the target ledger \(T\,b\approx25\,b\) . Every generator — dimension label or anchor real — is charged a uniform per-generator cost \(\Delta_0\) , and the resulting ledger is verified, by explicit finite-truncation computation (Route 1, exhaustive check up to \(N=12\) ) and by an independent least-squares embedding fit (Route 2, truncations \(10\to200\) ), to be units-covariant, cancellative, commutative, positive, and totally ordered — a well-behaved ordered cost-monoid in every respect checked.

 What complete Granularity does not supply — the decisive negative result. The one property checked for, and found absent, is the Archimedean property : that for any two positive cost contributions \(x,y>0\) (say, \(x=\) one structural bit, \(y=\) one anchor bit) there exists a finite integer \(n\) with \(n\cdot x>y\) — i.e., that enough copies of the cheaper kind of cost can always outweigh any fixed amount of the more expensive kind. This was checked two independent ways and both agree completely.

 Route 1 — exhaustive finite-truncation axiom check. All nine scaffold axioms of an ordered cost-monoid were verified to hold for the specific candidate structure \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) (the lexicographic product: compare the first coordinate — the dimension-count axis — first, and only break ties using the second coordinate — the anchor-count axis) by exhaustive check up to \(N=12\) . The Archimedean witness — an integer \(n\) with \(n\cdot(0,1) > (1,0)\) in the lex order — was tested and found to fail for every \(n\) up to \(n=100{,}000\) , and the failure is provable for all \(n\) : no finite number of anchor-bit units, under lexicographic order, ever outweighs a single dimension-count unit. This is not a numerical accident of small truncations; it is the defining property of a lexicographic order.

 Route 2 — frozen-embedding least-squares fit. A control structure, plain \(\mathbb{N}\) (genuinely Archimedean, single real-valued cost axis), was fit for an additive real-embedding and converges to violation fraction 0.0000 at every truncation size tested (10, 25, 50, 100, 200). The test structure, \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) , was fit the identical way and its violation fraction grows with truncation size: \(0.0000\to0.0426\to0.0581\to0.0662\) as the truncation increases — the operational numerical signature that no additive real embedding exists for the lexicographic structure, confirming Route 1 by an entirely independent method (optimization rather than exhaustive search). The two routes' numeric outputs are byte-identical on independent re-run — same violation fractions at every truncation, same fitted parameters — so the computation-debt on this check is fully discharged, not merely asserted.

 The group-versus-monoid question, resolved. An earlier framing worried that the deciding structural feature might be whether the cost ledger forms a group (with inverses) or merely a monoid (a positive cone, no inverses — the case here, since costs are never negative). This is resolved: the cost-ledger is a cancellative, commutative, ordered monoid , and Hölder's theorem together with Alimov's extension to the monoid case both establish that additive real-representability is equivalent to the Archimedean property regardless of whether inverses exist. Group-versus-monoid is not the axis that decides this question; Archimedean-versus-not is. This correction sharpens the argument but changes nothing about the +1 verdict.

 The complete-Granularity verdict is recorded as:

 Granularity verdict: CONSTRAIN. Granularity is necessary and does more real work than either Shape or Scale — it supplies the common currency without which the comparison could not even be stated numerically — but it constrains rather than forces: it proves, by two independent and mutually confirming routes, that the natural finite-record cost structure is non-Archimedean, which is precisely the situation in which the desired comparison ("13D wins by \(\approx1.8\times\) ") is not automatically well-posed without a further choice. Granularity is the root on which the seam ultimately bottoms (Section I.6), and its full-precision, all-three-layer analysis rules out the possibility that finer discretization, run further, would ever produce the missing Archimedean guarantee — the failure is structural (a lexicographic order is a logically closed alternative), not a resolution artifact that finer \(\Delta_0\) could dissolve.

 I.5 The Layer-2 admissibility screens, all four, applied to the seam

 With the three roots run to completion, the seam itself — call the missing rule BR-Arch, formally "the ordered cost-ledger is Archimedean: for any two positive cost contributions \(x,y>0\) there exists a finite integer \(n\) with \(n\cdot x>y\) " — is passed through the four Layer-2 admissibility screens. These screens exist to catch exactly the failure modes this program is most worried about: smuggling content across layers, building an unfalsifiable or untestable ranking, reverse-engineering the answer from the desired conclusion, or illegitimately factorizing a cost that should not separate. Each screen is applied directly to BR-Arch as the candidate object.

 Invariance. This screen governs whether a proposed structure is stable under within-primitive recodings — relabeling the generators of a primitive without changing its physical content. Both the Archimedean order (ordinary \(\mathbb{N}\) -type addition) and the non-Archimedean lexicographic order are equally invariant under such relabeling: neither order privileges a particular choice of which generator is called "first." The screen has nothing to discriminate on here. Verdict: BLIND — the screen does not distinguish the two candidate orders, so it neither helps nor hurts BR-Arch's case; it simply does not bear on this particular seam.

 Record Interface. This screen asks whether a candidate structure is a finite, computable, auditable total preorder over the record space — i.e., whether it could in principle be checked by an external reviewer against the frozen branch data without appeal to unstated background assumptions. Both the Archimedean order and the lexicographic order pass this test cleanly: both are finite, both are computable (as demonstrated by Routes 1 and 2 above, which mechanically check each), both are fully auditable, and both are total preorders over the same record space. Verdict: PASS — for both candidates, which means this screen certifies that the comparison itself is legitimate and well-formed; it does not, and is not meant to, pick a winner between them.

 Causal Order. This screen checks for target-anchoring: whether the proposed structure has causal-order/signalling content that could let a rule be reverse-engineered from a desired downstream physical prediction rather than adopted on independent grounds. A per-theory signalling structure would fail this screen if it depended on the direction of some physical causal chain within the branch. BR-Arch is a claim about how bit-costs aggregate across candidate theories — a cross-theory ranking convention, not a signal object living inside any one theory's causal structure. The screen is therefore blind to it in the relevant technical sense, and that blindness is itself the confirmation sought: a ranking rule that lived inside the physical causal order of the branch would be immediately suspect as target-anchored, and BR-Arch demonstrably does not. Verdict: BLIND/PASS — blind because the object is outside the screen's proper domain (a meta-level aggregation rule, not an in-theory signal), and PASS because that very fact is what confirms target-blindness: BR-Arch could not have been reverse-engineered from a desired physical outcome because it has no causal-order handle to be tuned against.

 Nonseparability. This screen forbids unpaid factorization — splitting a cost or a structure into independent pieces without charging for the act of splitting, which would let a program quietly get a structural result "for free" by decomposing it in a convenient way. The lexicographic order is, in a precise sense, the maximally separable order available on the record space: it factors the comparison completely into "compare axis 1 first, in full, then and only then break ties on axis 2" — a totally ordered, totally separated decomposition with no cross-terms. Because Nonseparability forbids unpaid factorization, not factorization as such, and because the lexicographic order's separability is exactly what it is (an honestly total, explicitly declared decomposition, not a hidden one), the screen finds the lex order compatible with , not opposed to, its own discipline — lex does not sneak in an unpaid factorization; it is a fully paid, fully declared factorization. Verdict: PASS — again for both candidates, and once more this screen certifies legitimacy of the comparison rather than resolving it.

 Toolbox roll-up. Across all seven checks run against this seam — Shape, Scale, Granularity, and the four Layer-2 screens — the tally is: Shape = PASS (no purchase), Scale = EXPOSE (sharpens, does not resolve), Granularity = CONSTRAIN (supplies currency, proves non-Archimedean), Invariance = PASS/BLIND, Record-Interface = PASS, Causal-Order = PASS/BLIND, Nonseparability = PASS. 7 of 7 checks are saturated — every one was run to completion, at full precision, with no truncated layer — and none of the seven independently forces the Archimedean property. This is the single most important sentence in this construction: it means the missing rung is not merely unproven by an incomplete search; it is actively ruled out as automatically available from any of the tools this program has for exactly this purpose, each pushed to its full three-layer, full-precision extent.

 I.6 Where the chain bottoms: BR-Arch as the honest +1

 Because all seven checks are saturated and none forces the aggregation rule, the rule must be supplied as a new, named, explicitly paid axiom if the economy comparison is to go through as an outright "13D wins" statement rather than a merely incommensurable one:

 BR-Arch (Archimedean common-currency axiom). The ordered cost-ledger is Archimedean: for any two positive cost contributions \(x,y>0\) there exists a finite integer \(n\) with \(n\cdot x>y\) . Equivalently: a finite positive exchange ratio exists between a structural bit and an anchor bit — no cost axis is infinitely preferred over the other.

 Why this axiom is genuinely needed, not merely convenient (necessity). The necessity argument is not asserted; it is demonstrated by an explicit countermodel. On the identical finite record space used throughout — the same branches, the same admissible Shape objects, the same Granularity-supplied bit currency — a second, logically consistent total order exists: dimension-first lexicographic order , ranking candidate branches first by number of metric dimensions, and only breaking ties by anchor-bit count. Under this order, a clean 4D effective field theory beats the 13D branch outright, unconditionally, regardless of how many anchor-bits the 4D theory needs to inject ( \(4<13\) settles it before the anchor count is even examined), and the entire Shape-minimality argument for this program's 13D branch collapses. Nothing in Shape, Scale, or Granularity — run to completion, as demonstrated above — excludes the lexicographic order as illegitimate. It is finite (Record-Interface: PASS), it is target-blind (Causal-Order: BLIND/PASS), and it is maximally separable in a way Nonseparability explicitly does not forbid (PASS). The lexicographic order is a live, admissible rival to the Archimedean order on every criterion this program has for ruling out candidates.

 Why the candidate class is closed, not open-ended (Hahn's theorem). The situation is not "there might be arbitrarily many hidden aggregation rules we haven't thought of." Hahn's embedding theorem for totally ordered abelian groups/monoids states that a totally ordered structure is Archimedean if and only if it is rank-1 — i.e., embeddable in a single copy of \(\mathbb{R}\) with its usual order. Every non-Archimedean totally ordered structure, without exception, is a Hahn sum of component groups of rank \(\ge2\) , and the lexicographic product \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) is the canonical, simplest possible witness of the rank-2 case. This means the exhaustive dichotomy {Archimedean, non-Archimedean} is not a heuristic split but a theorem-backed, closed partition of every possible ordering compatible with the monoid axioms verified in Section I.4. There is no third option lurking outside this dichotomy waiting to be discovered by a cleverer search; the class is closed, and BR-Arch's job is precisely to select the Archimedean branch of that closed, two-element partition.

 Why selecting the Archimedean branch is exactly as strong as asserting cross-kind commensurability — no root "forces" it without begging the question. Because the dichotomy is closed and exhaustive, eliminating the lexicographic order is extensionally identical to asserting that a structural bit and an anchor bit can be compared on one scale at all — the two statements are logically equivalent, not merely correlated. This is why no root can be said to "force" BR-Arch without covertly assuming what BR-Arch itself asserts: any argument that tried to derive commensurability from Shape, Scale, or Granularity would have to already treat dimension-cost and anchor-cost as comparable in order to run the argument, which is circular. The honest statement is therefore that BR-Arch is a new, target-blind, value-free, paid axiom — new because nothing in the frozen branch or its geometric constants implies it; target-blind because the Causal-Order screen confirms it carries no reverse-engineered signal from a desired physical outcome; value-free because it fixes a comparison rule , not a numerical value of any observable; and paid because it is charged explicitly at floor +1 , never treated as free.

 The forcing certificate, assembled. Three components jointly certify the terminal grade: (i) the anchor-transfer chain is NONE — BR-Arch does not transfer to, derive from, or otherwise connect to \(\Delta_0\) , \(\hbar\) , or \(M_{\rm Pl}\) ; it is a standalone meta-level rule about cost aggregation, not a rescaling of any existing anchored quantity; (ii) necessity is shown , via the explicit lexicographic countermodel demonstrated to be fully admissible under every screen; and (iii) the candidate class is closed , via Hahn's theorem, so there is no unbounded regress of ever-more-exotic orderings still to rule out. Together these three components cap the ceiling of this gate at ROOT-CONSTRAINED / REDUCED-TO-NEW-AXIOM , and explicitly never at ROOT-FORCED — no complete root, run at full precision and full three-layer discipline, derives the Archimedean property for free. This is the honest terminal: REDUCED-TO-AXIOM / ANCHORED +1 .

 I.7 Shared-object accounting

 BR-Arch is not a private axiom minted for SG-1 alone. It is a shared object : the identical Archimedean common-currency requirement is invoked by any gate in this program whose forcedness argument relies on an additive-MDL cost comparison across structurally different kinds of charge — in particular the deep-root granularity family of gates and any gate that inherits the same "dimension-cost versus anchor-cost" ledger structure. Per the Nonseparability discipline itself (the same screen that passed BR-Arch above): a shared axiom is counted once , as a single +1 floor item, not re-charged independently for every gate that touches it. This matters for the honesty of the overall program's floor accounting — it prevents the same open seam from silently inflating the total axiom count by being paid for multiple times under different gate names, while still ensuring it is paid for at all, exactly once, wherever it is actually used.

 I.8 What this construction certifies, stated plainly

 Running all three complete roots and all four Layer-2 screens against the one live seam of this gate — with no layer of any root truncated, and every numerical claim traceable to an exact rational, an exact topological integer, or an explicitly bounded, independently reproduced numerical computation — establishes the following, precisely and without overclaiming in either direction. Shape, complete, restricts the space of admissible records and thereby makes the comparison well-posed, but supplies no scoring rule between admissible records: PASS, no purchase. Scale, complete, exposes the two-axis structure of the cost ledger (anchor-bits versus structural-bits) with full numerical transparency across the entire \(M_{\rm Pl}\to M_U\to R_0\to\) threshold-vector chain, but supplies no exchange rate between the axes: EXPOSE. Granularity, complete, supplies the one thing genuinely needed to state the comparison numerically — a uniform bit-currency across all 13 dimensions and both non-metric layers — and, by two independently verified and mutually confirming routes, proves that the natural cost-monoid on this currency is non-Archimedean: CONSTRAIN, and the root on which the entire chain ultimately bottoms. All four Layer-2 screens confirm the comparison itself is legitimate, finite, computable, target-blind, and non-separably honest, without resolving which of the two closed-class orderings should be adopted. The gap between these seven completed, saturated checks and an outright "13D wins" statement is exactly one axiom wide, that axiom is named and its necessity is proven by an explicit admissible countermodel, its candidate class is closed by a cited theorem, and it is charged honestly at floor +1. That is the deep-root anchoring of SG-1, and it is the entire and sufficient content of why this gate's terminal is fixed at REDUCED-TO-AXIOM / ANCHORED +1 — never more, and, on the evidence assembled here, never honestly less.

 Construction II - the full derivation

 This construction carries out the SG-1 argument step by step, from the bare definition of the object through to the terminal grade, with every intermediate number shown and tagged. The order of business is: (1) pin the object at all three layers; (2) derive the curvature and topology of the heaviest factor, \(K_6=SU(3)/T^2\) , from its root system up, rather than quoting it; (3) prove, in full, the three theorems that force three of the four gauge-carrying assignments within the declared grammar, given the observed Standard Model spectrum \(E\) ; (4) exhibit the freeze-and-reproduce witness that makes the object a fixed target rather than a moving one; (5) derive the scale structure (radii, volumes, thresholds) from the four measured anchors; (6) build the description-length (MDL) economy ledger term by term; and (7) assemble the terminal. Every quantity is tagged [EXACT] (a closed-form rational, or an exact topological integer, following from the declared metric/grammar with no numerical input), [DERIVED] (computed from the four anchors via a stated equation), [MEASURED] (an anchor or a PDG input), or [OPEN] (a named, bounded, undischarged residual). Nothing here is back-solved to a desired value: every number is produced by running the stated formula forward from its inputs.

 II.1 Fixing the object: all three layers, before any physics is asked of it

 The gate commits, once and for all, the three-layer object

 \[
\mathfrak{B}_{\rm active}
=\underbrace{\big[\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2\big]}_{\times\ \text{Stage}}
\ \oplus\ \underbrace{\big[F^+_{\rm finite}\oplus C_{\rm admiss}\big]}_{\oplus\ \text{Rulebook}}
\ \otimes\ \underbrace{\big[E_{\rm matter}\oplus E_{\rm gauge}\oplus E_{\rm Higgs}\oplus E_{\rm proton}\big]}_{\otimes\ \text{Actors}}.
\]

 Stage. \(\mathcal{M}_4=\mathbb{R}^{3,1}\) is Minkowski spacetime, taken as an observational primitive — not derived by this gate, and never charged as a forced rung in the accounting below (this is axiom SG1- \(\alpha\) , the labeling discipline that keeps residual R9 honest: \(\mathcal{M}_4\) is disclosed, subsumed, and never cited as a derivation). \(K_6=SU(3)/T^2\) is the full \(A_2\) -type flag manifold of \(SU(3)\) , real dimension 6. \(S^2\) is the round 2-sphere. \(S^1_Y/\mathbb{Z}_2\) is the orbifold interval obtained from a parent circle \(\theta\in[0,2\pi)\) by the reflection \(\theta\mapsto-\theta\) , restricting the physically active domain to \(\theta\in[0,\pi]\) , with two fixed points at \(\theta=0,\pi\) . Summing real dimensions of only the metric-carrying Stage factors,
$$
D = \dim\mathcal{M}_4+\dim K_6+\dim S^2+\dim S^1_Y = 4+6+2+1 = 13 \qquad [\text{EXACT}].
$$

 Rulebook. \(F^+_{\rm finite}\) is the finite/operator flavor chamber, detailed in Section II.6; \(C_{\rm admiss}\) is the admissibility constraint that (i) restricts the \(K_6\) squashing moduli \(\vec u=(u_1,u_2,u_3)\) to the Weyl-rigid chamber \([1/2,3/2]^3\) , (ii) enforces the \(\mathbb{Z}_2\) orbifold parity on \(S^1_Y\) , and (iii) encodes the selector's freeze-before-compare barrier and the FCNC/mediator no-go identity \(\Pi_qM\Pi_\ell=0\) for any sector-respecting operator \(M\) . Neither \(F^+_{\rm finite}\) nor \(C_{\rm admiss}\) carries metric dimension: \(\dim(\oplus\text{-layer})=0\) .

 Actors. \(E_{\rm matter}=S_{3,1}\otimes S^{\rm spin^c}_{K_6}\otimes S^{\rm spin^c}_{S^2}\otimes L_Y\otimes V_{SU(3)}\otimes V_{SU(2)}\otimes V_{F^+}\) is the full matter bundle; \(E_{\rm gauge}=T^*\mathcal{M}_4\otimes\mathrm{ad}(P)\) for the principal bundle \(P\) over \(\mathcal{M}_4\times K_{\rm gauge}\) (with BRST/Faddeau–Popov gauge fixing on a Gribov domain); \(E_{\rm Higgs}=L_\gamma\otimes V_{SU(2),\rm doub}\) on the Wilson-line cycle \(\gamma\) ; \(E_{\rm proton}=\Pi_qE_{\rm matter}\otimes\Pi_\ell E_{\rm matter}\) encodes the proton-safety projector structure via sector-orthogonal partitions \(\Pi_q,\Pi_\ell\) . These too are non-metric: \(\dim(\otimes\text{-layer})=0\) .

 A reading that keeps only \(\times\) -Stage and silently drops \(\oplus\) or \(\otimes\) is an incomplete object . Every quantity derived below that depends on a projector, a grading, a chirality operator, or an endomorphism spectrum is unrecoverable from Stage alone — which is exactly the discipline the B2 proper-subset null-space result certifies: no gate in this program closes on a proper subset of the three layers, inside the declared category. This construction therefore carries all three layers explicitly through every step, rather than computing on Stage and citing the others as backdrop.

 II.2 The gauge-routing ledger, factor by factor

 Each Stage factor is assigned exactly one gauge role, through its isometry algebra, fixed by the Actors-layer connection data:

 Factor 
 dim 
 Isometry algebra 
 Gauge role 
 Mechanism 

 \(\mathcal{M}_4\) 
 4 
 Poincaré (primitive) 
 — 
 4D Dirac spinor bundle \(S_{3,1}\) 

 \(K_6=SU(3)/T^2\) 
 6 
 \(\mathfrak{su}(3)\) 
 \(SU(3)_c\) color 
 left-isometry action; spin \(^c\) family index \(-3\) 

 \(S^2\) 
 2 
 \(\mathfrak{su}(2)\) 
 \(SU(2)_L\) weak 
 isometry, not any \(SU(2)\subset SU(3)\) ; monopole-doublet routing 

 \(S^1_Y/\mathbb{Z}_2\) 
 \(1\to\) interval 
 \(\mathfrak{u}(1)\) 
 \(U(1)_Y\) hypercharge 
 \(\theta\mapsto-\theta\) orbifold; chirality filter 

 \(F^+\) 
 0 
 — (finite chamber) 
 flavor/Yukawa 
 \(\tau=\omega\) , projectors \(\Pi_i\) , ladder operators 

 This ledger is the skeleton that Sections II.4–II.5 justify: three of these four assignments are proved below to be forced consequences of the declared grammar — finite-dimensional homogeneous compact factors whose isometry groups source Standard-Model gauge fields — given the observed spectrum \(E\) , not free choices made for convenience.

 II.3 The curvature and topology of \(K_6=SU(3)/T^2\) , derived in full

 \(K_6\) carries the heaviest geometric weight in the object, so its curvature is derived here from the root system up, in the Killing-form normal metric, rather than merely quoted from a table.

 Root system. In the Cartan basis \((h_1,h_2,h_3)\) with \(h_1+h_2+h_3=0\) , the \(A_2=\mathfrak{su}(3)\) simple roots are
$$
\alpha_1=(1,-1,0),\qquad \alpha_2=(0,1,-1),\qquad \alpha_1+\alpha_2=(1,0,-1),
$$
giving positive roots \(\{\alpha_1,\alpha_2,\alpha_1+\alpha_2\}\) and Weyl group \(S_3\) , order 6. The half-sum of positive roots is
$$
\rho=\tfrac12\big(\alpha_1+\alpha_2+(\alpha_1+\alpha_2)\big)=(1,0,-1),\qquad |\rho|^2=2\ \ (\text{Killing normalization})\quad[\text{EXACT}].
$$
The tangent space decomposes as \(T(K_6)=\mathfrak m_1\oplus\mathfrak m_2\oplus\mathfrak m_3\) , each \(\mathfrak m_i\) a real 2-plane carrying one positive root ( \(\alpha_3\equiv\alpha_1+\alpha_2\) ), with \((-B)\) -orthonormal basis \(\{X_{ij}=E_{ij}-E_{ji},\,Y_{ij}=i(E_{ij}+E_{ji})\}/\sqrt{12}\) over the pairs \((01),(12),(02)\) , Killing form \(B(X,Y)=6\,\mathrm{Tr}(XY)\) . Hence \(\dim_{\mathbb R}\mathfrak m_i=2\) and \(\sum_i\dim\mathfrak m_i=6=\dim K_6\) , consistent.

 Invariant metric and Ricci (Wang–Ziller/Nomizu). The \(SU(3)\) -invariant metric is \(g_{K_6}(\vec u)=\sum_i u_i\langle\cdot,\cdot\rangle_{\mathfrak m_i}\) with \(\vec u\in[1/2,3/2]^3\) the Weyl-rigid squashing moduli. In terms of Killing-form scales \(x_1,x_2,x_3\) on the three root planes, the general-chamber Ricci eigenvalues and scalar curvature are
$$
\mathrm{Ric}_1=\frac{x_1^2-x_2^2+6x_2x_3-x_3^2}{12\,x_1x_2x_3},\qquad
\mathrm{Ric}_2=\frac{-x_1^2+6x_1x_3+x_2^2-x_3^2}{12\,x_1x_2x_3},\qquad
\mathrm{Ric}_3=\frac{-x_1^2+6x_1x_2-x_2^2+x_3^2}{12\,x_1x_2x_3},
$$
$$
\mathrm{Scal}=\frac{x_1x_2+x_1x_3+x_2x_3-\tfrac16(x_1^2+x_2^2+x_3^2)}{x_1x_2x_3}.
$$
At the symmetric chamber center \(x_1=x_2=x_3=1\) (equivalently \(u_1=u_2=u_3=1\) ), the \(\mathrm{Ric}_1\) numerator becomes \(1-1+6-1=5\) over denominator \(12\) , giving \(\mathrm{Ric}_1=5/12\) ; the cyclic symmetry of the three formulas at equal arguments forces \(\mathrm{Ric}_2=\mathrm{Ric}_3=5/12\) identically. Thus
$$
\mathrm{Ric}_i=\frac{5}{12}\quad(i=1,2,3)\qquad[\text{EXACT}].
$$
The scalar curvature is the trace weighted by the 2-dimensional multiplicity of each root plane: \(\mathrm{Scal}=\sum_k\dim(\mathfrak m_k)\,\mathrm{Ric}_k=2\cdot3\cdot\tfrac{5}{12}=\tfrac52\) , matching direct substitution into the general formula (numerator \(1+1+1-\tfrac16\cdot3=\tfrac52\) , denominator \(1\) ):
$$
\mathrm{Scal}=\frac{5}{2},\qquad \mathrm{Scal}^2=\frac{25}{4}\qquad[\text{EXACT}].
$$

 Quadratic invariants. The Ricci tensor at center is \(\mathrm{diag}(5/12,\ldots,5/12)\) over six eigenvalues (one pair per root plane), so
$$
|\mathrm{Ric}|^2=6\times\left(\frac{5}{12}\right)^2=6\times\frac{25}{144}=\frac{25}{24}\qquad[\text{EXACT}],
$$
$$
\frac{|\mathrm{Ric}|^2}{\mathrm{Scal}^2}=\frac{25/24}{25/4}=\frac{4}{24}=\frac{1}{6}\qquad[\text{EXACT — this is the corpus "}\kappa=1/6\text{"}].
$$
The full Riemann-squared invariant, computed from the Nomizu curvature tensor of the normal homogeneous metric at the Killing-form center, is
$$
|\mathrm{Riem}|^2=\frac{23}{12}\qquad[\text{EXACT}],\qquad \frac{|\mathrm{Riem}|^2}{\mathrm{Scal}^2}=\frac{23/12}{25/4}=\frac{23}{75}\qquad[\text{EXACT}].
$$
And directly,
$$
\frac{\mathrm{Scal}}{\mathrm{Ric}_i}=\frac{5/2}{5/12}=6=\dim K_6\qquad[\text{EXACT, identical in both curvature normalizations used downstream}].
$$
These ratios are dimensionless and therefore metric-scale invariant: they agree identically whether computed in the Killing-form-normal convention used here, or in the frozen physical-radius convention \(\mathrm{Ric}_i=1/(2R_6^2)\) , \(\mathrm{Scal}=3/R_6^2\) used by the volume/threshold pipeline of Section II.7 — \(\big(3R_6^{-2}\big)/\big(\tfrac12R_6^{-2}\big)=6\) recovers the same integer. Frozen negative controls (never dissolve): \(|\mathrm{Riem}|^2\) must never be reported as \(31/147\) , nor as \(60\) — the latter value belongs to the unrelated round unit \(6\) -sphere \(S^6\) , a distinct manifold, and this dossier explicitly guards against that confusion at every appearance of \(|\mathrm{Riem}|^2\) .

 Cubic and derivative invariants. The cubic Riemann self-contraction and its "twisted" ladder partner are
$$
K_1\equiv R_{ab}{}^{cd}R_{cd}{}^{ef}R_{ef}{}^{ab}=8\,\mathrm{tr}(R_{\rm op}^3)=-\frac{113}{72},\qquad
K_2\equiv R_{abcd}R_{aecf}R_{ebfd}=-\frac{5}{72}\qquad[\text{EXACT}],
$$
and the covariant derivative norm, via Nomizu, is
$$
|\nabla\mathrm{Riem}|^2=\frac14\qquad[\text{EXACT}; \text{second Bianchi identity checked and returns 0 violations}].
$$
Because \(|\nabla\mathrm{Riem}|^2\ne0\) , \(K_6\) is homogeneous but not locally symmetric . This is not a bookkeeping footnote: it forbids treating \(K_6\) as a symmetric space in any curvature expansion, in particular the heat-kernel ledger of Section II.6, which must retain non-symmetric Lichnerowicz cross-terms precisely because this invariant is nonzero.

 Weight-6 invariants. At the same Einstein center,
$$
\mathrm{Scal}^3=\frac{125}{8},\qquad \mathrm{Scal}\,|\mathrm{Ric}|^2=\frac{125}{48},\qquad \mathrm{Scal}\,|\mathrm{Riem}|^2=\frac{115}{24},
$$
$$
\mathrm{Ric}^{ab}\mathrm{Ric}^{cd}R_{acbd}=\frac{125}{288},\qquad \mathrm{Ric}^{ab}R_a{}^{cde}R_{bcde}=\frac{115}{144}\qquad[\text{EXACT}].
$$
These nine invariants — the two quadratic ratios, \(K_1\) , \(K_2\) , \(|\nabla\mathrm{Riem}|^2\) , and the four weight-6 contractions above — form the certified curvature core reused across every heat-kernel route in this program. None is fit: each follows algebraically from \(\mathrm{Ric}_i=5/12\) and the \(A_2\) root data derived above, with no free parameter entering after the chamber center is fixed.

 Einstein metrics on the chamber. Solving \(\mathrm{Ric}_k(\vec u)\propto g_k\) over the three-parameter family \(\vec u\in[1/2,3/2]^3\) shows the moduli space contains exactly four invariant Einstein points: the fully symmetric normal metric \((1,1,1)\) , plus the three permutations of the Kähler–Einstein point \((1,1,2)\) . This is the classical result for \(SU(3)/T^2\) , reproduced here independently from the Ricci formulas above rather than merely cited — a validation of the geometric engine, not an assumption borrowed from the literature. Off these four points the space is non-Einstein, which is exactly what makes \(\vec u\) a legitimate squashing modulus (an admissibility-constrained physical input) rather than a redundant gauge freedom.

 Topology. The Euler characteristic of a full flag manifold equals the order of its Weyl group: \(\chi(K_6)=|S_3|=6\) [EXACT/topological] — the number of Weyl chambers, structurally required, not a numerical coincidence. Similarly \(\chi(S^2)=2\) (Gauss–Bonnet) and \(\chi(S^1_Y/\mathbb{Z}_2)=1\) (Euler characteristic of a closed interval) [EXACT/topological].

 Representation content. The quadratic Casimir and dimension of an \(SU(3)\) irrep with Dynkin labels \((p,q)\) are
$$
C_2(p,q)=\frac{p^2+q^2+pq+3p+3q}{3},\qquad \dim(p,q)=\frac{(p+1)(q+1)(p+q+2)}{2}\qquad[\text{EXACT}].
$$
The lowest cases: \((0,0)\) trivial, \(\dim1\) , \(C_2=0\) ; \((1,0)/(0,1)\) , \(\dim3/\bar3\) , \(C_2=4/3\) (quark color triplets); \((1,1)\) adjoint, \(\dim8\) , \(C_2=3\) , zero-weight multiplicity \(m_0=2\) (16 total adjoint zero-modes). This last is the lowest nonzero scalar harmonic in the Peter–Weyl decomposition \(L^2(K_6,E_\mu)=\bigoplus_{(p,q)}V_{(p,q)}\otimes\mathrm{Hom}_{T^2}(V_{(p,q)},E_\mu)\) , and it is what seeds the Kaluza–Klein spectrum used in the threshold computation of Section II.7.

 II.4 The three within-grammar forcing theorems, proved

 The declared grammar restricts attention to homogeneous compact factors whose isometry groups source Standard-Model gauge fields. Under that grammar, and given the observed spectrum \(E\) , three of the four gauge-carrying assignments of Section II.2 are theorems, not choices.

 Theorem F1 (weak sector forces \(S^2\) , not any torus). Claim: no abelian or torus factor, of any dimension, has a non-abelian \(\mathfrak{su}(2)\) among its isometries. Proof: the isometry group of a flat torus \(T^n=\mathbb{R}^n/\Lambda\) is \(\mathrm{Isom}(T^n)=T^n\rtimes P\) for a discrete point group \(P\) ; its identity component is \(T^n\) itself, which is abelian (pure translations). Any continuous, non-abelian symmetry would have to act by continuously rotating the torus's linear structure while preserving the lattice \(\Lambda\) , but the only isometries of a flat torus fixing a basepoint and preserving \(\Lambda\) are the finitely many point-group elements — there is no continuous non-abelian subgroup for any \(n\) . Hence a torus of any dimension supplies only abelian gauge factors \(U(1)^k\) , and weak \(SU(2)_L\) , being non-abelian, categorically cannot arise from a torus or any other abelian carrier. By contrast \(S^2=SU(2)/U(1)\) has isometry group \(SO(3)\cong SU(2)/\mathbb{Z}_2\) , non-abelian by construction and of exactly the required rank. This is a hand-checkable, fully general theorem about isometry groups of homogeneous spaces — it closes the cheaper direction ("could some torus have supplied weak more economically?") over entire shelves of competitor branches simultaneously, not merely against one explicitly built rival. Status: DERIVED-WITHIN-GRAMMAR. 

 Theorem F2 (hypercharge forces the \(\mathbb{Z}_2\) orbifold, not the closed circle). Claim: a closed, odd-dimensional Stage factor retains both chiralities, producing mirror fermions, which are excluded by measurement — the LEP \(Z\) -width constrains the number of light neutrino species to \(N_\nu=2.9840\pm0.0082\) , consistent with exactly three light chiral generations and no additional mirror doublets. Mechanism: on the parent circle \(\theta\in[0,2\pi)\) , the Dirac operator's zero modes come in a left-handed and a right-handed tower related by the isometry \(\theta\to\theta+\pi\) , so a chiral projection cannot be imposed without breaking that isometry — chirality and mirror-chirality are exactly degenerate on the closed circle. Quotienting by the reflection \(\mathbb{Z}_2:\theta\mapsto-\theta\) produces two fixed points, \(\theta=0,\pi\) , and an active domain \(\theta\in[0,\pi]\) ; boundary conditions at the fixed points can now be assigned with opposite parity for left- and right-handed field components (the parity table of Section II.6), removing the right-handed zero mode while retaining the left-handed one. This is computed, not merely asserted, via the Atiyah–Patodi–Singer index theorem on the interval \([0,\pi]\) with the chirality projector
$$
P_\chi=\tfrac12\big(1+\gamma_5\Gamma_8\big),
$$
where \(\gamma_5\) is the ordinary 4D chirality matrix and \(\Gamma_8\) is the chirality operator on the 8-dimensional internal spinor bundle \(S(K_6)\otimes S(S^2)\otimes S(S^1_Y)\) . The index returns
$$
n_L=+3,\qquad n_R=0\qquad[\text{DERIVED-WITHIN-GRAMMAR}],
$$
three surviving left-handed chiral zero modes and zero right-handed ones — exactly matching the observed chiral spectrum with no mirror partners. The per-field parity table (Section II.6) makes this concrete field by field: e.g. \(Q_L\) carries parity \((+,+)\) at \((\theta{=}0,\theta{=}\pi)\) , giving 3 surviving families, while the mirror parity \((-,-)\) gives none surviving; \(u_R\) carries \((-,-)\) , surviving via the projector \(\Pi_u\) , with mirror parity \((+,+)\) giving none. Status: DERIVED-WITHIN-GRAMMAR. 

 Theorem W9 (color forces \(K_6=SU(3)/T^2\) over \(CP^2=SU(3)/U(2)\) , by abelian-isotropy uniqueness). This is the deepest of the three, and the only one built as an explicit branch-kill rather than a pure existence argument. The governing rule is the coset-space dimensional reduction (CSDR) centralizer rule: for a homogeneous space \(G/R\) , the isotropy subgroup \(R\) acts on the tangent space, and any part of the ambient gauge group \(G\) that centralizes (commutes with) \(R\) survives dimensional reduction as an additional 4D gauge symmetry beyond whatever \(R\) itself was meant to source. Two cases:

 \(K_6=SU(3)/T^2\) : the isotropy \(T^2\) is the maximal torus of \(SU(3)\) , and its centralizer in \(SU(3)\) is \(T^2\) itself, \(C_{SU(3)}(T^2)=T^2\) — purely abelian, Cartan-only. No non-abelian piece of \(SU(3)\) centralizes \(T^2\) , so no spurious non-abelian gauge factor survives reduction: color \(SU(3)_c\) emerges cleanly, with nothing extra.

 \(CP^2=SU(3)/U(2)\) (the rival): the isotropy is \(U(2)=(SU(2)\times U(1))/\mathbb{Z}_2\) , which is non-abelian — its own \(SU(2)\) factor sits inside color \(SU(3)\) , and by the identical centralizer rule survives reduction as an additional , unwanted, gauge-active \(SU(2)+U(1)\) beyond color. This branch was built end-to-end, not merely asserted, and is shown to break precisely here, at Gate 2, by over-producing gauge content relative to the observed Standard Model group.

 The uniqueness claim is then a clean statement of representation theory: among cosets \(SU(3)/R\) with \(R\) a closed subgroup, the maximal torus \(T^2\) is the unique isotropy subgroup that is purely abelian (self-centralizing, with no non-abelian centralizer complement), because any subgroup of \(SU(3)\) strictly larger than \(T^2\) must contain a root \(SU(2)\) or the whole of \(SU(3)\) , both non-abelian. This argument is stronger than an earlier, now-retired comparison based on tunable family counts: \(CP^2\) 's three-family count is fixed by a discrete spin \(^c\) index, \(r(r+1)/2=3\) at \(r=2\) , not a continuously adjustable dial, so counting families alone cannot discriminate the two candidates — the isotropy/centralizer argument does, because it is a structural, all-or-nothing statement about surviving gauge content rather than a numerical coincidence that could in principle be matched by both cosets. Status: DERIVED on the named shelf (the shelf being "cosets of \(SU(3)\) "). The completeness of that shelf — whether some other sub-6-dimensional homogeneous space, not of the form \(SU(3)/R\) , might also carry a clean abelian isotropy and source color — is explicitly not certified, and is tracked as the open residual R4/N.4 (Section II.9 below and the residual table of the wider dossier).

 II.5 The freeze-and-reproduce witness (the one fully derived leg)

 Having fixed the object and proved three of its four gauge assignments forced-within-grammar, the branch is committed by content hash: a branch content hash spanning the primitive definitions, a manifest meta-hash spanning all 33 geometry-description rows, and a separate orbifold-specific freeze hash for the \(S^1_Y/\mathbb{Z}_2\) construction. Verification proceeds by an independent, target-blind re-run: the reproducer script's own comparator is bypassed, and all 33 hashes are recomputed directly from the primitive definitions — root system, metric ansatz, projector definitions, RG scheme, and so on — with no reference anywhere in the recomputation to the expected output. The re-run terminates with exit code 0; all 33 recomputed hashes are byte-equal to the frozen values; the result is deterministic across repeated independent runs.

 This is the single leg of SG-1 that is genuinely and completely DERIVED in the strongest sense used in this program: a content-addressed self-witness. Its scope must be stated with precision, in both directions. What it proves: the object under test today is byte-identical to the object every downstream gate (SG-2 through SG-10) was actually run against — no quiet re-tuning has occurred, and a reviewer attacking the object today is attacking the same object the corpus's results are about. What it does not prove: nothing about whether that object is the uniquely correct, or even the best, choice of geometry — freezing an object says nothing about the object's merit, only about its stability under inspection. This residual (R6) is VERIFIED / CLOSED-DERIVED , and it lowers no assumption floor elsewhere in the gate.

 II.6 The \(\oplus\) -Rulebook and \(\otimes\) -Actors detail behind II.4's theorems

 Hypercharge lattice and the SM group. The hypercharge lattice is \(Y\in\tfrac16\mathbb{Z}\) , and the Standard Model gauge group is the quotient
$$
G_{\rm SM}=\frac{SU(3)_c\times SU(2)_L\times U(1)_Y}{\mathbb{Z}_6},\qquad Q=T_3+Y,
$$
with generator \(z=(\omega_3,-1,\zeta_6)\) acting as \((\zeta_3^k,(-1)^k,e^{2\pi ik/6})\) for \(k\in\mathbb{Z}_6\) . The Smith normal form of the charge-character matrix has invariant factors \([1,6,6]\) : this certifies, by a computable linear-algebra fact rather than a stipulation, that \(\mathbb{Z}_6\) is the finest faithful quotient — no coarser and no finer identification of the three centers ( \(\mathbb{Z}_3\subset SU(3)_c\) , \(\mathbb{Z}_2\subset SU(2)_L\) , a sixth root of unity on \(U(1)_Y\) ) is admissible. The standard SM hypercharge assignments — \(Y(Q_L)=+1/6\) , \(Y(u_R)=+2/3\) , \(Y(d_R)=-1/3\) , \(Y(L_L)=-1/2\) , \(Y(e_R)=-1\) , \(Y(H)=+1/2\) — give
$$
\sum_fY_f^2=\frac{10}{3}\ \text{per generation}\qquad[\text{EXACT/CERTIFIED}],
$$
a sum that feeds directly into the hypercharge threshold packet of Section II.7.

 The parity table (mechanism behind Theorem F2). At the two orbifold fixed points \(\theta=0,\pi\) , each Standard Model field is assigned a \(\mathbb{Z}_2\) parity; a field survives with a zero mode only if its parity is even at both fixed points, or is projected consistently by a sector projector \(\Pi_i\) . The mirror parity assignment — which would produce the unwanted right-handed mirror partner — is explicitly forbidden and has no surviving mode for every SM field: \(Q_L(+,+)\to\) 3 families survive, mirror \((-,-)\to\) none survive; \(u_R(-,-)\to\) 3 survive via \(\Pi_u\) , mirror \((+,+)\to\) none; identically for \(d_R,L_L,e_R,\nu\) with their respective projectors \(\Pi_d,\Pi_e,\Pi_\nu\) . The Higgs is treated separately, as a Wilson-line mode whose orbifold parity is inherited from the \(K_{\rm gauge}\) cycle (Section II.8). This table is what makes the index-theoretic result \(n_L=+3,n_R=0\) concrete, field by field, rather than an abstract index-theory statement.

 Sector projectors and generation basis (the \(F^+\) chamber). \(F^+\) carries a Cartan-torus modulus pinned at the order-three fixed point
$$
\tau=\omega=e^{2\pi i/3}=-\tfrac12+i\tfrac{\sqrt3}{2}=-0.5000000000000000+0.8660254037844386\,i\qquad[\text{EXACT}],
$$
a 3-complex-dimensional generation basis \(\mathcal G_{\rm gen}=\mathrm{span}\{g_1,g_2,g_3\}\) matched to the spin \(^c\) family index \(-3\) , and four orthogonal rank-3 sector projectors \(\Pi_u,\Pi_d,\Pi_e,\Pi_\nu\) satisfying \(\Pi_i\Pi_j=\delta_{ij}\Pi_i\) . These are \(\oplus\) / \(\otimes\) -layer objects of zero metric dimension, and they are exactly the structure the no-layer-smuggling clause of Section 1.1 (of the wider dossier) guards: none of Theorems F1, F2, or W9's gauge-group conclusions depend on \(F^+\) , but the full chiral-family count and the proton-safety identity \(\Pi_qM\Pi_\ell=0\) (for any sector-respecting operator \(M\) ) do depend on it, and are correctly billed to this layer, not smuggled into Stage.

 Chamber Boltzmann factors and Yukawa map (structure needed for Section II.9's honest cost accounting). The chamber's exponential suppression factor is \(\kappa=e^{-\pi\sqrt3}=0.004333420509983131\) , with critical ratio \(K_{tb}^{\rm crit}=e^{-\pi\sqrt3/16}=0.7117081304239685\) [EXACT]. Action ladders per sector are \(a_u=(2,1,0)\) , \(a_d=(4/3,2/3,0)\) , \(a_e=(2,4/3,0)\) , \(a_\nu=(1,1/2,0)\) , each combined with a sector-level-only normalization \(N_i\) — \(N_u=1\) (fixes the up-anchor via \(y_t\) ), \(N_d=2.400000000000000\times10^{-2}\) (fixes \(m_b\) at \(M_Z\) ), \(N_e=1.020000000000000\times10^{-2}\) (fixes \(m_\tau\) at \(M_Z\) ), and a structural \(N_\nu\) — via \((Y_i)^{ab}=N_i\langle g_a|O_i|g_b\rangle\) , \((O_i)^{aa}=N_i\kappa^{a_i^{(a)}}\) . The binding admissibility rule is that only sector-level normalizations are legal; family-level normalizations \(N_{i,a}\) are forbidden, which is precisely what makes the per-family mass hierarchy \(\kappa^{a^{(a)}}\) a genuine prediction rather than a fit. These normalizations, plus the CKM holonomy phase \(\delta_{\rm CKM}=-2\pi/3\) and lepton Berry phase \(+2\pi/3\) , are exactly the " \(\sim9\) –10 injected reals" charged honestly in Section II.9.

 II.7 Scale: radii, volumes, and thresholds derived from the anchors

 The four measured anchors \(\{M_{\rm Pl},\alpha_i(M_Z),y_t,|V_{us}|\}\) [MEASURED] fix, via RG transport plus the Kaluza–Klein threshold spectrum, every radius and volume in the object; none of the numbers in this section is an independent input beyond those four plus the honestly-charged flavor-chamber normalizations of Section II.6.

 Unification scale and natural radius. The unification scale is defined by the closure target \(\alpha_1^{-1}(M_U)=\alpha_2^{-1}(M_U)=\alpha_3^{-1}(M_U)\) under two-loop Standard Model running plus KK thresholds, giving
$$
M_U=1.0\times10^{16}\ \mathrm{GeV}\qquad[\text{DERIVED; residual }|\alpha_i^{-1}-\alpha_j^{-1}|=9.6\times10^{-11}\text{, a numerical-pipeline floor}],
$$
comfortably inside the propagated PDG uncertainty band of order \(10^{-3}\) . The natural compactification radius follows,
$$
R_0\equiv\frac{1}{2\pi M_U}=1.591549430918954\times10^{-17}\ \mathrm{GeV}^{-1}\qquad[\text{DERIVED}].
$$
At the symmetric chamber center, \(R_6=R_2=R_0\) , while the active hypercharge radius carries an extra factor \(\tfrac12\) from the \(\mathbb{Z}_2\) orbifold halving,
$$
R_Y=\frac12R_0\,e^{-\delta_1/2b_1^{\rm KK}}\Big|_{\rm center}=7.957747154594768\times10^{-18}\ \mathrm{GeV}^{-1}\qquad[\text{DERIVED}].
$$
The two measured scales entering the RG closure are \(M_Z=91.18760000000000\) GeV [MEASURED, PDG, \(\pm0.0021\) ] and the ordinary (not reduced) Planck mass \(M_{\rm Pl}=1.220900000000000\times10^{19}\) GeV [MEASURED anchor].

 Volumes. With \(V_{K_6,0}=(2\pi)^3/\sqrt3=143.2118575035129\) [EXACT] and \(R_6=R_2=R_0\) at center,
$$
\mathrm{Vol}(K_6)=V_{K_6,0}R_0^6=2.327554010848277\times10^{-99}\ \mathrm{GeV}^{-6},\qquad
\mathrm{Vol}(S^2)=4\pi R_0^2=3.183098861837907\times10^{-33}\ \mathrm{GeV}^{-2}.
$$
The parent-circle and active (post-orbifold) volumes are exact, because the \(2\pi\) (or \(\pi\) ) in the circle-volume formula cancels identically against the \(2\pi\) inside \(R_0\equiv1/(2\pi M_U)\) :
$$
\mathrm{Vol}(S^1_Y) {\rm parent}=2\pi R_0=\frac{1}{M_U}=1.000000000000000\times10^{-16}\ \mathrm{GeV}^{-1},\qquad
\mathrm{Vol}(S^1_Y/\mathbb{Z}_2) {\rm active}=\pi R_0=\frac{1}{2M_U}=5.000000000000000\times10^{-17}\ \mathrm{GeV}^{-1}\qquad[\text{EXACT, given }M_U].
$$
Multiplying the three internal-factor volumes together,
$$
\mathrm{Vol}(X_{\rm active})=\mathrm{Vol}(K_6)\cdot\mathrm{Vol}(S^2)\cdot\mathrm{Vol}(S^1_Y/\mathbb{Z}_2)=3.704417261398702\times10^{-148}\ \mathrm{GeV}^{-9}\qquad[\text{DERIVED}].
$$

 Planck normalization. With \(D=13\) and \(X_{\rm int}=K_6\times S^2\times S^1_Y/\mathbb{Z}_2\) of dimension 9,
$$
M_{\rm Pl}^2=M_ ^{D-2}\,\mathrm{Vol}(X_{\rm int})=M_ ^{11}\,\mathrm{Vol}(X_{\rm active}),
$$
so, solving for the fundamental scale \(M_*\) using the measured \(M_{\rm Pl}\) and the derived volume,
$$
M_ ^{11}=\frac{M_{\rm Pl}^2}{\mathrm{Vol}(X_{\rm active})}=4.023836152402511\times10^{185}\ \mathrm{GeV}^{11},\qquad
M_ =7.467050992135091\times10^{16}\ \mathrm{GeV}\qquad[\text{DERIVED}].
$$
 \(M_*\) is manifestly not an independent input — it is the derived consequence of the one Planck anchor and the geometry's own derived volume. (The reduced-Planck convention rescales the left-hand side by \(1/8\pi\) ; the right-hand geometry is unchanged.)

 Threshold vector. The one-loop Standard Model beta coefficients, fixed by field content and not free,
$$
b_1^{\rm SM}=\frac{41}{10}=4.100000000000000,\qquad b_2^{\rm SM}=-\frac{19}{6}=-3.166666666666667,\qquad b_3^{\rm SM}=-7\qquad[\text{EXACT, fixed by }E].
$$
The Kaluza–Klein threshold vector sums six packets — \(K_6\) matter zero modes (3 generations, quark color), \(S^2\) matter zero modes (3 generations, weak doublets), \(K_6\) weak/color gauge-plus-ghost loops, the \(S^1_Y/\mathbb{Z}_2\) hypercharge gauge packet, the \(S^1_Y/\mathbb{Z}_2\) hypercharge zero-mode matter packet (using exactly the \(\sum_fY_f^2=10/3\) derived above, summed over 3 generations), the Higgs Wilson-line contribution, and the orbifold boundary term at \(\theta\in\{0,\pi\}\) — giving
$$
(\delta_1,\delta_2,\delta_3)=(+4.842400000000000,\ -3.111200000000000,\ -1.731300000000000)\ \pm\ 1.6\times10^{-3}\qquad[\text{DERIVED}],
$$
under two-loop Standard Model RG in the \(\overline{\rm MS}\) scheme at \(M_Z=91.1876\) GeV. This threshold vector is precisely the input that drives \(M_U\) 's closure residual of \(9.6\times10^{-11}\) quoted above — the unification scale and the threshold vector are two faces of one consistent computation, not two independently-tuned numbers.

 II.8 The Wilson-line Higgs and the actor-layer endomorphism data

 The Higgs is not postulated as a fundamental scalar but constructed as a Wilson-line (Hosotani) mode of the \(SU(2)_L\) connection around a declared gauge cycle \(\gamma\subset K_{\rm gauge}\) with integer winding \(n_H=1\) — the minimum winding that produces a nonzero vacuum expectation value ( \(n_H=0\) would give none, so \(n_H\in\mathbb{Z}_{>0}\) is bounded below by 1). The one-loop effective (Hosotani) potential is
$$
V_{\rm Hos}(\theta_H)=-\frac{3}{64\pi^6R_\gamma^4}\sum_{n=1}^\infty\frac{1}{n^5}\big[N_b-N_f\big]\cos(n\theta_H),
$$
whose \(n^{-5}\) tail converges absolutely — the structural reason the Higgs mass emerges finite under the declared regulator, not a fine-tuned cancellation. Two exact structural constants control the resulting hierarchy:
$$
\eta_{BK}=\frac{1}{32\pi\,e^{\sqrt3/(24\pi)}}=0.009721281516312024,\qquad \frac{1}{\eta_{BK}}=32\pi\,e^{\sqrt3/(24\pi)}=102.8670961047707\qquad[\text{EXACT}],
$$
giving the structural hierarchy ratio \(\sqrt{\eta_{BK}}/(2\pi)=0.01569212979293374\) , i.e. a suppression exponent \(S_H\sim30\) . This ratio is what places the Higgs vacuum expectation value at the electroweak scale rather than at \(M_{\rm Pl}\) — a derived consequence of the \(K_6\) / \(F^+\) chamber constants, not an assumption laid in by hand. The resulting post-RG values are \(v_{\rm pred}=246.02\pm3.5\) GeV, \(m_h=123.82\pm1.8\) GeV, and \(\lambda_H=m_h^2/(2v^2)=0.12722\pm0.00181\) [DERIVED, at the corpus's stated uncertainty].

 The graviton (bundle-endomorphism) sector on \(K_6\) is likewise pinned as an Actors-layer object at all three layers. The Lichnerowicz operator
$$
(E_Lh) {ab}=\mathrm{Ric} {ac}h^c{} b+\mathrm{Ric} {bc}h^c{} a-2R {acbd}h^{cd}
$$
acting on the transverse-traceless part of \(\mathrm{Sym}^2\) (dimension 20) has spectrum
$$
\left{\tfrac16\ (\times6),\ \tfrac{5}{12}\ (\times6),\ \tfrac76\ (\times6),\ \tfrac{17}{12}\ (\times2)\right},\qquad \mathrm{tr}\,E_L=\frac{40}{3},\qquad \mathrm{tr}\,E_L^2=\frac{241}{18}\qquad[\text{EXACT}],
$$
the certified graviton inputs to the heat-kernel ledger. On the full \(\mathrm{Sym}^2\) (dimension 21, including the trace mode) the spectrum picks up one further eigenvalue \(5/3\) (mult. 1, the pure-trace mode). The scalar-sector heat-kernel coefficients are \(a_2/a_0=5/12\) , \(a_4/a_0=11/120\) [EXACT], with the scalar backbone \(a_6/a_2^3=7936/39375\) banked across multiple independent engines. The \(a_6\) graviton coefficient itself remains an explicitly named computation debt: it requires Gelfand–Tsetlin off-diagonal hopping matrix elements between the 5 Weyl-inequivalent \(T^2\) weight classes on \(\mathrm{Sym}^2_0\) — exact in principle (a standard lowering-operator formula, involving square roots of products of pattern-entry differences) but not yet enumerated in the atlas. This is marked [OPEN] here exactly as in the underlying geometric record — a bounded, honestly disclosed gap at a named computational stratum, not a hidden hole in any argument used above; it enters none of the II.4 forcing theorems and none of the freeze witness of II.5.

 II.9 The economy ledger: building the MDL comparison term by term

 The forcedness argument for "13D wins" is a description-length (MDL) comparison, deliberately constructed to be spectrum-neutral : the observed content \(E\) is charged identically on both sides of the ledger and cancels, so no number is smuggled in and no comparison is fit to a desired outcome.

 Codebook. A tuned continuous real, resolved to detector cell size \(\Delta_0\) , costs \(b=\log_2(1/\Delta_0)\) bits — a large number for any realistic measurement precision. A discrete structural choice — which coset, which finite quotient, which integer topological index — costs \(O(1)\) bits, parametrically smaller than \(b\) .

 Target ledger (both sides, \(E\) -neutral). The physics content any viable theory must reproduce is \(T\approx25\) reals [CERTIFICATE]: 3 gauge couplings at \(M_Z\) ; 9 charged-fermion masses; 4 CKM angles and phase; 6 neutrino parameters; 2 electroweak parameters ( \(v,m_H\) ); 1 strong-CP bound \(\bar\theta\) . This count is identical whether one is scoring the 13D branch or a generic 4D effective field theory — it is what any theory must explain, not what either theory supplies as input — so it cancels out of the difference between branches, and is recorded here explicitly so it is never silently dropped from either side.

 Honest charged cost of the 13D branch. The branch pays for the 4 measured anchors \(\{M_{\rm Pl},\alpha_i(M_Z),y_t,|V_{us}|\}\) plus the further \(\sim9\) –10 injected reals identified in Section II.6: the sector normalizations \(N_d=2.400000000000000\times10^{-2}\) (fixes \(m_b\) ), \(N_e=1.020000000000000\times10^{-2}\) (fixes \(m_\tau\) ), and structural \(N_\nu=1\) ; the threshold triple \((\delta_1,\delta_2,\delta_3)\) derived in Section II.7; and the Hosotani phase \(\theta_H^\star\) of Section II.8. Summing,
$$
n_{13}\approx13\text{–}14\ \text{measured reals}\qquad[\text{MEASURED, charged-not-atomic}],
$$
markedly more than the historically-circulated "4 inputs" description. This dossier explicitly retires that undercounted headline in favor of the honest figure derived here term by term.

 The boxed inequality. The claim that the 13D branch wins, without smuggling in a new knob, is the statement that for every rival dimensional/architectural branch \(B_{D,j}\) in the declared family (indexed by dimension \(D=4,\ldots,12\) plus non-dimensional constructions, and by architecture \(j\) within each \(D\) ),
$$
I_{\rm gen,13}+n_{13}\,b+I_{\rm struct,13} \;<\; I_{\rm gen}(B_{D,j})+n_{D,j}\,b+I_{\rm extra}(B_{D,j})+I_{\rm struct}(B_{D,j}),
$$
where \(I_{\rm gen}\) is the generic construction cost, \(n\,b\) is the charged continuous-real cost at the shared bit rate \(b\) , and \(I_{\rm struct}\) / \(I_{\rm extra}\) are the structural/extra-content costs of each rival. Running this ledger over the declared competitor family gives a first-pass survey verdict of 10 branches LOSE to the 13D branch, 1 FAILS structurally (this is the \(CP^2\) branch of Theorem W9, independently corroborating the branch-kill from a second, cost-based direction), and 0 are REFUTED outright. The honest margin at transparent cost is
$$
\frac{n_{\rm EFT}\,b}{n_{13}\,b}\approx\frac{25\,b}{13\text{–}14\,b}\approx1.8\times\qquad[\text{DERIVED/survey; supersedes and retires the earlier "}\sim4\times\text{" headline}].
$$
This ladder verdict is explicitly labeled a category-relative, first-pass survey — not a certified exhaustive classification — a gap tracked as residual R2 in the wider dossier's residual table (Section 7 of the grounding brief).

 II.10 The seam: why the boxed inequality is not yet a theorem, in one paragraph

 Sections II.3–II.9 establish three forced gauge assignments (II.4), a fully derived and machine-verified freeze witness (II.5), and a fully derived scale/threshold structure (II.7–II.8). The one remaining step — the boxed inequality of II.9 asserting an outright numerical "win" — presupposes that a dimension-count (a Shape-layer, structural quantity) and an anchor-count (a measured-real, continuous quantity) can be added on one common scale at all. That presupposition is not delivered by Shape (which only restricts which records are admissible, and has no scoring rule to compare admissible records against each other — adopting one at the Shape level would be target-anchoring), nor by Scale (which exposes that dimension-cost and anchor-cost are two genuinely different cost axes but supplies no exchange rate between them), but only partially by Granularity: Granularity supplies a common bit-currency and a finite per-generator cost, yet the natural cost-ledger structure this produces — the ordered monoid \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) , checked to satisfy every one of the granularity axioms (uniform per-generator resolution \(\Delta_0\) , units-covariance, cancellativity, commutativity, positivity, total order, monotonicity) — is demonstrably non-Archimedean : no finite multiple of the anchor-bit generator ever exceeds one unit of the dimension-count generator. This full three-root pass (Shape/Scale/Granularity, all at complete three-layer precision, all seven Layer-2 admissibility screens saturated) is carried out in full in the companion construction of this dossier; it is not re-derived here, but its conclusion is the hinge on which this section's terminal turns, and is restated precisely in Section II.11.

 II.11 The named axiom BR-Arch: necessity, closure, and the +1 floor

 Because none of the three roots forces a choice between the Archimedean ordering (implicitly assumed by the boxed inequality of II.9) and the non-Archimedean lexicographic ordering (dimension-count first, anchor-count only as a tiebreak) that is equally admissible on the identical finite record space, a second axiom must be named and paid for explicitly:

 BR-Arch (Archimedean common-currency axiom). The ordered cost-ledger is Archimedean: for any two positive cost contributions \(x,y>0\) there exists a finite integer \(n\) such that \(n\cdot x>y\) . Equivalently: a finite positive exchange ratio exists between one structural (dimension) bit and one anchor bit — no cost axis is infinitely preferred over the other.

 Necessity, by explicit countermodel. Under the rival dimension-first lexicographic order, a clean 4D effective field theory beats the 13D branch outright, \(4<13\) , regardless of anchor cost — the entire economy argument of II.9 folds, for every branch in the family, the instant lex is adopted instead of an additive currency. Nothing established in Sections II.1–II.10 excludes lex: it satisfies all nine scaffold axioms of an ordered cost-monoid (verified computationally below), all seven Layer-2 screens of the companion construction, and neither the curvature/topology facts of II.3 nor the freeze witness of II.5 have anything to say about how costs of different kinds ought to be compared.

 Closure of the candidate class, by Hahn's theorem. A totally ordered abelian group (or, as verified below, the corresponding cancellative ordered monoid) is Archimedean if and only if it has Hahn rank 1 — i.e., it embeds order-preservingly in a single copy of \(\mathbb{R}\) . Every non-Archimedean ordered structure decomposes, without exception, as a Hahn sum of components of rank \(\ge2\) , and \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) is the canonical rank-2 witness. The dichotomy {Archimedean, non-Archimedean} is therefore exhaustive and closed — there is no third option outside this classification that could rescue the economy claim without invoking one of these two. BR-Arch is precisely the choice of the Archimedean branch of this closed dichotomy; making that choice is, extensionally, identical to asserting the cross-kind commensurability clause the economy ledger of II.9 needs — so no root can be said to "force" BR-Arch without already assuming what BR-Arch itself asserts, a circularity this construction avoids by naming the axiom rather than smuggling it in through a root.

 Grading of the seam. BR-Arch is graded MAP_ADMISSIBLE_ SUPPORTED , explicitly not FORCED. This caps the gate's ceiling at ROOT-CONSTRAINED / REDUCED-TO-NEW-AXIOM — never ROOT-FORCED — and is exactly the terminal reported: necessity is shown via the explicit lex countermodel; the candidate class is closed via Hahn's theorem; the forcing certificate confirms no anchor-transfer chain exists (BR-Arch does not transfer to \(\Delta_0\) , \(\hbar\) , or \(M_{\rm Pl}\) — it is a pure ordering axiom, not a physical scale); and target-blindness is confirmed independently by the Causal-Order screen of the companion construction.

 Independent computational re-verification. Two independent routes confirm the non-Archimedean character of the lex countermodel, reproduced byte-identically on re-run:
- Route 1 (exhaustive finite-truncation axiom check, \(N=12\) ): all nine scaffold axioms hold for \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) at every truncation tested; the Archimedean witness — an integer \(n\) with \(n\cdot(0,1)>(1,0)\) in lex order — fails for every \(n\) up to \(n=100{,}000\) , and is shown to fail provably for all \(n\) [DERIVED/verified].
- Route 2 (frozen-embedding least-squares fit): a control case (plain \(\mathbb{N}\) , genuinely Archimedean) converges to embedding-violation \(0.0000\) at every truncation from 10 to 200; the test case ( \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) ) shows the violation fraction growing with truncation size, \(0.0000\to0.0426\to0.0581\to0.0662\) — the operational numerical signature that no additive real embedding exists, growing rather than shrinking as more structure is resolved [DERIVED/verified].

 A group-vs-monoid side question is resolved in passing: the cost-ledger is a cancellative, commutative, ordered monoid (a positive cone, no inverses, since costs are never negative), not a group, but Hölder's theorem together with Alimov's extension to the monoid case both establish that additive real-representability is equivalent to the Archimedean property regardless of whether inverses exist — group-versus-monoid is not the deciding axis; Archimedean-versus-not is. This sharpens, but does not change, the +1 verdict. Reproducibility: the independent re-run's numeric output (violation fractions at each truncation, fitted parameters) is byte-identical to the original computation, so this computation-debt item is fully DISCHARGED , not merely asserted.

 Shared-object accounting. BR-Arch is a shared object: it is invoked by SG-1 and by any other gate in this program whose forcedness argument relies on an additive-MDL cost comparison across structurally different kinds of charge — in particular the granularity-family gate. Per the Nonseparability screen's own discipline (the same screen that passes BR-Arch), it is counted once , as a single shared +1-floor item, not re-charged independently for every gate that invokes it.

 II.12 Assembling the terminal

 Putting Sections II.3–II.11 together: the 13D object is fixed and reproducibly witnessed (II.1, II.5); its curvature and topology are derived in full from the \(A_2\) root system and independently cross-checked against the classical \(SU(3)/T^2\) Einstein-metric count and Euler characteristic (II.3); three of its four gauge-carrying assignments are theorems within the declared grammar, given the observed spectrum \(E\) (II.4, with mechanism detail in II.6 and II.8); its scale structure is a derived consequence of the four measured anchors, not an independent input (II.7); and its claim to be the economy-leanest branch among its family is correct conditional on exactly one named, closed-candidate-class, target-blind axiom (BR-Arch, II.10–II.11), whose necessity is proved via an explicit countermodel and whose candidate class is proved closed via Hahn's theorem, but which is not itself a theorem of the geometry.

 This is the full content behind the fixed terminal REDUCED-TO-AXIOM / ANCHORED +1 : every forcedness chain in this construction bottoms out either on the observed spectrum \(E\) (which cancels out of the economy comparison and is never charged as a forced rung) or on the single Granularity-rooted axiom BR-Arch, charged once, at floor +1. There is no unbounded regress — the candidate class above BR-Arch is closed by Hahn's theorem, not merely unsearched — and no silent assumption — BR-Arch is named, its necessity demonstrated by an explicit admissible countermodel, and its independence from every other anchor in the program (no transfer to \(\Delta_0\) , \(\hbar\) , or \(M_{\rm Pl}\) ) certified. Per the fixed-grade instruction governing this dossier, this construction neither upgrades this axiom-level result to a from-nothing derivation nor downgrades the three proved forcing theorems (F1, F2, W9) to mere assertions: three of four gauge-carrying assignments are DERIVED-WITHIN-GRAMMAR or DERIVED-ON-SHELF, the freeze witness is CLOSED-DERIVED, and the remaining economy-comparison seam is honestly and irreducibly REDUCED-TO-AXIOM, ANCHORED at floor +1.

 Construction III - the central result at full precision

 This section carries the full evidentiary weight of SG-1. It states, and derives with every intermediate step shown, the central result of the gate: given the observed Standard Model spectrum \(E\) and the declared role-mechanism grammar, three of the four gauge-carrying assignments in the frozen 13-dimensional branch are forced theorems, and the fourth object — whether the whole branch , taken as a description-length economy claim, beats its rivals — reduces to exactly one named, closed-candidate-class axiom. Every quantity below is tagged by its status; nothing is asserted without either a proof, a closed-form computation from the frozen geometry, or an explicit OPEN label. Target-blind throughout: no number below was back-solved to a desired answer.

 III.1 The object under test, all three layers pinned

 The gate commits, once and for all, the three-layer object

 \[
\mathfrak{B}_{\rm active}
=\underbrace{\big[\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2\big]}_{\times\ \text{STAGE, metric, }D=4+6+2+1=13}
\ \oplus\ \underbrace{\big[F^+_{\rm finite}\oplus C_{\rm admiss}\big]}_{\oplus\ \text{RULEBOOK, 0 dims}}
\ \otimes\ \underbrace{\big[E_{\rm matter}\oplus E_{\rm gauge}\oplus E_{\rm Higgs}\oplus E_{\rm proton}\big]}_{\otimes\ \text{ACTORS, 0 dims}},
\]

 with \(K_6=SU(3)/T^2\) the full \(A_2\) flag manifold and \(S^1_Y/\mathbb{Z}_2\) the active orbifold interval obtained from the parent circle \(\theta\in[0,2\pi)\) by the reflection \(\theta\mapsto-\theta\) . Only the \(\times\) -Stage layer carries metric dimension:
$$
D=\dim\mathcal{M}_4+\dim K_6+\dim S^2+\dim S^1_Y=4+6+2+1=13\qquad[\text{EXACT}].
$$
The \(\oplus\) -Rulebook layer ( \(F^+_{\rm finite}\) , the finite flavor/operator chamber, and \(C_{\rm admiss}\) , the admissibility constraint fixing the Weyl-rigid squashing chamber \(\vec u\in[1/2,3/2]^3\) and the \(\mathbb{Z}_2\) orbifold parity) and the \(\otimes\) -Actors layer (the matter, gauge, Higgs, and proton-safety bundles/endomorphisms) carry zero metric dimension but are load-bearing: a reading that keeps Stage alone and drops either layer is an incomplete object, since every projector, grading, and endomorphism spectrum used below lives outside Stage.

 The gauge-routing ledger — the skeleton the rest of this section proves forced — assigns each Stage factor exactly one Standard Model gauge role through its isometry algebra:

 Factor 
 dim 
 Isometry algebra 
 Gauge role 
 Status of the assignment 

 \(\mathcal{M}_4=\mathbb{R}^{3,1}\) 
 4 
 Poincaré (primitive) 
 observed spacetime 
 observational primitive (R9), never charged as forced 

 \(K_6=SU(3)/T^2\) 
 6 
 \(\mathfrak{su}(3)\) 
 \(SU(3)_c\) color 
 DERIVED-ON-SHELF (§III.4) 

 \(S^2\) 
 2 
 \(\mathfrak{su}(2)\) 
 \(SU(2)_L\) weak 
 DERIVED-WITHIN-GRAMMAR (§III.2) 

 \(S^1_Y/\mathbb{Z}_2\) 
 1 \(\to\) interval 
 \(\mathfrak{u}(1)\) 
 \(U(1)_Y\) hypercharge 
 DERIVED-WITHIN-GRAMMAR (§III.3) 

 \(F^+\) 
 0 
 — (finite chamber) 
 flavor/Yukawa 
 not a routing claim; charged separately in the economy ledger 

 A gauge-routing claim earns the label DERIVED-WITHIN-GRAMMAR (or, on the color leg, DERIVED-ON-SHELF) if, holding fixed (a) the declared Stage factor list, (b) the declared Rulebook mechanisms — isometries source gauge symmetries, orbifold quotients project mirrors, abelian isotropy avoids spurious gauge content — and (c) the observed spectrum \(E\) , the assignment is the only one compatible with a theorem proved on that factor, not merely the cheapest or the one that happens to work. Three of the four routing assignments meet this bar; they are proved in full in §III.2–§III.4. The fourth object under test — whether the whole branch wins an economy comparison against its rivals — does not meet this bar, and §III.5–§III.8 show exactly why, and exactly what single axiom closes the remaining gap.

 III.2 Fact F1 — the weak carrier is forced to be \(S^2\) , not any torus (DERIVED-WITHIN-GRAMMAR)

 Claim. Within the declared grammar — gauge symmetries are sourced by isometries of the internal factor that carries them, and the observed weak force \(SU(2)_L\) is non-abelian — no abelian or torus factor, of any real dimension, can carry the weak sector. \(S^2\) , with isometry algebra \(\mathfrak{su}(2)\) , is the minimal-dimension carrier satisfying the requirement, and it is exactly what the frozen branch uses.

 Proof. Consider a flat torus \(T^n=\mathbb{R}^n/\Lambda\) for any lattice \(\Lambda\subset\mathbb{R}^n\) and any \(n\ge1\) . Its isometry group is \(\mathrm{Isom}(T^n)=T^n\rtimes P\) , where \(P\) is the (discrete) point group of \(\Lambda\) — the finite set of linear maps preserving the lattice. The identity component of this isometry group — the piece capable of sourcing a continuous, propagating Kaluza–Klein gauge field — is exactly the translation subgroup \(T^n\) itself, because the point group \(P\) is discrete and cannot supply continuous gauge bosons. \(T^n\) , being a quotient of the abelian group \(\mathbb{R}^n\) by translations, is abelian for every \(n\) : translations on a flat space commute regardless of dimension, so \(\mathrm{Isom}_0(T^n)\cong U(1)^n\) is abelian for all \(n\ge1\) . This exhausts the entire torus/abelian shelf in one stroke: no torus, of any dimension, has a non-abelian isometry group , hence none can source the non-abelian \(SU(2)_L\) .

 Turning to the non-abelian side, the round 2-sphere \(S^2=SU(2)/U(1)\) has full isometry group \(\mathrm{Isom}(S^2)=O(3)\) , whose identity component is \(SO(3)\cong SU(2)/\mathbb{Z}_2\) — non-abelian, rank 1, matching \(SU(2)_L\) exactly at the level of the Lie algebra \(\mathfrak{su}(2)\) (the center quotient by \(\mathbb{Z}_2\) is immaterial for what sources the gauge field, which is the algebra, not the global group). Dimension 2 is the first dimension at which a non-abelian connected isometry group becomes possible at all: a 1-dimensional compact factor is either a circle (isometry group \(U(1)\) at the identity component — abelian) or a non-compact line (excluded by the requirement of a finite-volume internal space); it is only at dimension 2 that a homogeneous space such as \(S^2\) can support \(SO(3)\) acting non-abelianly. \(\blacksquare\) 

 What is and is not shown. This is a genuine theorem, exhaustive over the entire abelian/torus branch (every \(n\) , one proof) and constructive on the non-abelian branch (the minimal working example). It does not show \(S^2\) is the unique non-abelian carrier at dimension 2 or above — other rank-1 coset spaces, or higher-dimensional non-abelian carriers such as \(S^3=SU(2)\) itself via left or right translation, are not excluded by this argument. What F1 forces is the cheaper direction: it forecloses the entire abelian/torus shelf, at every dimension, as a candidate for the weak sector — precisely the comparison the economy argument of §III.5 needs, since it means no rival branch can undercut the 13D branch's cost by swapping \(S^2\) for a cheaper torus. Completeness among non-abelian carriers of dimension \(\ge2\) is not certified here; that residual is carried under R4/R7 in the open-hole ledger, not folded into F1's claim.

 III.3 Fact F2 — the hypercharge carrier is forced to be \(S^1_Y/\mathbb{Z}_2\) , not the closed circle (DERIVED-WITHIN-GRAMMAR)

 Claim. Within the declared grammar, a closed odd-dimensional Stage factor cannot serve as the hypercharge carrier because it retains both chiralities (mirror fermions), which are excluded by the measured LEP \(Z\) -width. Only the \(\mathbb{Z}_2\) -orbifolded interval \(S^1_Y/\mathbb{Z}_2\) , obtained from the parent circle \(S^1_Y\) by the reflection \(\theta\mapsto-\theta\) , removes the mirrors while still sourcing \(U(1)_Y\) .

 Proof, in three parts. 

 (a) The mirror-fermion obstruction on the closed parent circle. On the full parent circle \(\theta\in[0,2\pi)\) , the Kaluza–Klein momentum spectrum is \(p_\theta=(n+\alpha)/R_Y\) for integer \(n\) and twist \(\alpha\in\{0,Y\}\) . The spectrum is symmetric under \(n\to-n\) because the circle has no built-in structure that distinguishes one orientation of \(\theta\) from the other: for every left-handed zero mode there is a right-handed mirror partner at the same mass level. A vector-like, \(U(1)_Y\) -charged mirror sector at or below the electroweak scale is excluded experimentally: the measured \(Z\) width at LEP is saturated by exactly three light chiral generations with no additional vector-like doublets, so degenerate mirror partners surviving at low energy are dead on arrival. The bare closed circle is therefore excluded before any economy comparison is even run.

 (b) The orbifold removes exactly the mirrors and nothing else. Imposing the reflection \(\mathbb{Z}_2:\theta\mapsto-\theta\) on \(S^1_Y\) produces two isolated fixed points at \(\theta=0,\pi\) , and the quotient is the interval \(S^1_Y/\mathbb{Z}_2=[0,\pi]\) — exactly the active hypercharge carrier in \(\mathfrak{B}_{\rm active}\) , with active volume \(\mathrm{Vol}(S^1_Y/\mathbb{Z}_2)=\pi R_Y\) . Assigning each Standard Model field a definite \(\mathbb{Z}_2\) parity at the two fixed points, per the frozen no-mirror table:

 Field 
 \(\theta=0\) 
 \(\theta=\pi\) 
 Surviving zero mode 
 Mirror parity 
 Mirror mode 

 \(Q_L\) 
 \(+\) 
 \(+\) 
 3 families 
 \((-,-)\) 
 none 

 \(u_R\) 
 \(-\) 
 \(-\) 
 3 (via \(\Pi_u\) ) 
 \((+,+)\) 
 none 

 \(d_R\) 
 \(-\) 
 \(-\) 
 3 (via \(\Pi_d\) ) 
 \((+,+)\) 
 none 

 \(L_L\) 
 \(+\) 
 \(+\) 
 3 families 
 \((-,-)\) 
 none 

 \(e_R\) 
 \(-\) 
 \(-\) 
 3 (via \(\Pi_e\) ) 
 \((+,+)\) 
 none 

 \(\nu\) 
 \(-\) 
 \(-\) 
 3 (via \(\Pi_\nu\) ) 
 \((+,+)\) 
 none 

 each field's opposite-parity mirror partner has no surviving zero mode: the orbifold projects the mirror out at the level of boundary conditions, field by field, exactly and only. This is a truncation of the KK tower by boundary conditions, not a numerical fit — the parity assignment is fixed by demanding exactly the observed chiral content survive, and the mechanism (standard orbifold field theory) is applied here without free continuous parameters.

 (c) The index-theoretic confirmation, computed independently of the parity table. The chirality projector on the internal spinor bundle is
$$
P_\chi=\tfrac12\big(1+\gamma_5\,\Gamma_8\big),
$$
with \(\gamma_5\) the 4D chirality operator and \(\Gamma_8\) the chirality operator on the 8-dimensional internal spinor bundle \(S(K_6)\otimes S(S^2)\otimes S(S^1_Y)\) . The Atiyah–Patodi–Singer index of the internal Dirac operator on the active interval \([0,\pi]\) — computed via the boundary \(\eta\) -invariant contributions at the two fixed points, not by re-reading the parity table by hand — returns
$$
n_L=+3,\qquad n_R=0,
$$
exactly three surviving left-handed chiral families and zero surviving mirror partners, an independent confirmation of the field-by-field bookkeeping in (b). \(\blacksquare\) 

 What is and is not shown. The exclusion of the closed circle rests on a measured fact (the LEP \(Z\) -width) combined with a kinematic fact (mirror pairing on any closed KK tower); the orbifold's success is confirmed by an index computation independent of the hand-assigned parity table. What is not shown is that \(\mathbb{Z}_2\) is the unique quotient achieving this: other discrete actions on \(S^1\) , or on higher-dimensional carriers, are not exhaustively ruled out here. That residual sits inside R4/R7, not inside F2's claim, which is specifically that closed odd carriers fail and the declared \(\mathbb{Z}_2\) orbifold is a working, index-certified fix.

 III.4 Color carrier uniqueness — \(K_6=SU(3)/T^2\) is DERIVED-ON-SHELF (abelian-isotropy uniqueness, with \(CP^2\) built and broken)

 This is the sharpest and most load-bearing of the three forcedness results: it comes with a fully constructed, fully falsified rival, and it supersedes a weaker, now-retired argument from an earlier pass of this program.

 The coset family. Consider homogeneous spaces \(SU(3)/R\) for closed subgroups \(R\subset SU(3)\) — the natural candidate family for a color-carrying internal factor, since \(SU(3)\) is exactly the group color must come from. \(\dim SU(3)=8\) , so \(\dim(SU(3)/R)=8-\dim R\) . The gauge content that survives dimensional reduction on \(SU(3)/R\) is governed by the coset-space dimensional reduction (CSDR) centralizer rule: the four-dimensional gauge symmetry surviving beyond the intended carrier is the centralizer \(C_{SU(3)}(R)\) of the isotropy \(R\) inside \(SU(3)\) ; any non-abelian piece of that centralizer is promoted to a propagating, unintended 4D gauge field, not merely a passive isotropy label.

 Candidate 1 — \(R=T^2\) , the maximal torus. \(T^2\subset SU(3)\) is the Cartan subgroup, rank 2, abelian by definition, so \(\dim(SU(3)/T^2)=8-2=6\) , matching the declared Stage factor exactly. A maximal torus is self-centralizing in any compact simple Lie group — any element commuting with every element of a maximal torus must itself lie in that torus — so
$$
C_{SU(3)}(T^2)=T^2,
$$
purely abelian, Cartan-only. There is no non-abelian centralizer piece to be promoted to spurious gauge content: the reduction on \(K_6=SU(3)/T^2\) yields color \(SU(3)_c\) alone, with nothing extra.

 Candidate 2 — \(R=U(2)=(SU(2)\times U(1))/\mathbb{Z}_2\) , giving \(CP^2=SU(3)/U(2)\) . \(\dim U(2)=4\) , so \(\dim(SU(3)/U(2))=8-4=4\) : the complex projective plane, a coset used elsewhere in the model-building literature. But \(U(2)\) is not abelian — it contains the non-abelian \(SU(2)\) factor. Writing \(SU(3)\) in the \(2+1\) block form that \(U(2)\) preserves, the elements of \(SU(3)\) commuting with all of \(U(2)\) are exactly the \(U(1)\) phase on the singlet block, \(C_{SU(3)}(U(2))=U(1)\) — but the isotropy group \(U(2)\) itself is gauge-active under CSDR whenever it is non-abelian: the \(SU(2)\) factor sitting inside the isotropy is promoted to an additional, unbroken, non-abelian \(SU(2)\) gauge field in 4D, on top of whatever color the ambient \(SU(3)\) isometry was meant to deliver. Concretely, reducing on \(CP^2=SU(3)/U(2)\) returns \(SU(3)_{\rm isometry}\) -descended color plus an extra \(SU(2)\times U(1)\) from the non-abelian isotropy — the branch over-produces gauge structure relative to the Standard Model target. This was not left as an abstract worry: the \(CP^2\) branch was built end-to-end as a complete three-layer rival object, and its Gate-2 evaluation was run to completion. It breaks at Gate 2 , by exactly this mechanism, and it stays a dead, explicitly falsified branch — one of the frozen negative controls of this program, not to be quietly revived.

 The uniqueness statement. Among isotropy subgroups \(R\subset SU(3)\) of rank \(\le2\) realizing a coset of real dimension \(\le6\) (the dimension budget the Stage factor must respect), the maximal torus \(T^2\) is the unique choice with \(R=C_{SU(3)}(R)\) purely abelian. This holds because any proper subgroup of \(SU(3)\) strictly larger than \(T^2\) must contain a root \(SU(2)\) or the full group, both non-abelian (the trivial isotropy \(R=\{e\}\) , giving the full group manifold \(SU(3)\) at dimension 8, is off the dimension budget in any case, and its centralizer is the whole non-abelian \(SU(3)\) ). \(K_6=SU(3)/T^2\) is therefore the unique clean carrier on this shelf: the only coset of the required dimension whose isotropy contributes no spurious non-abelian gauge factor under the CSDR centralizer rule.

 Why this supersedes the retired "tunable family count" argument. An earlier pass of this program argued for \(K_6\) over \(CP^2\) by observing that \(CP^2\) 's family count is a rigid, discrete \(\mathrm{Spin}^c\) index \(r(r+1)/2\) , equal to \(3\) at \(r=2\) — not a continuously tunable dial one could fit to the observed three generations. That argument is true but weak: a rigid discrete index that happens to equal 3 is a numerical coincidence, not itself an exclusion mechanism, and it is silent on gauge content. The abelian-isotropy uniqueness argument given here is strictly stronger : it is pure representation theory (the self-centralizing property of a maximal torus), it does not depend on family counting at all, and it comes with a constructive falsification — \(CP^2\) was built out to Gate 2 and broke there for an independent structural reason (gauge over-production), not a numerological near-miss. The retired argument is superseded, not merely supplemented.

 What is and is not shown. This is a genuine on-shelf uniqueness theorem: within the family \(\{SU(3)/R\}\) , at dimension budget \(\le6\) , \(T^2\) uniquely avoids spurious gauge content. What is not shown — carried explicitly as residual R4/N.4, not folded into this theorem's claim — is that no other 6-real-dimensional (or smaller) homogeneous space, built from any other ambient Lie group entirely (not necessarily an \(SU(3)\) -coset at all), could serve as an equally clean color carrier. The theorem is proved on the shelf of \(SU(3)\) -cosets; full-shelf completeness against every conceivable alternative ambient group remains an open, uncertified item.

 III.5 The freeze-and-reproduce witness — the one fully derived mechanical leg

 Before the economy comparison of §III.6 can mean anything, the object being compared must be pinned so it cannot be quietly re-tuned to rescue a later result. The branch is committed by content hash (a content-addressed SHA-256 branch digest; a manifest meta-digest spanning all 33 geometry-description rows; a separate orbifold-specific freeze digest). Verification proceeds by an independent, target-blind re-run: the reproducer script's own comparator is bypassed, and all 33 hashes are recomputed from the primitive definitions (root system, metric ansatz, projector definitions, RG scheme) with no reference to the expected output. The re-run terminates at exit code 0, all 33 recomputed hashes are byte-equal to the frozen values, and the result is deterministic across repeated executions. This is the one leg of SG-1 that is genuinely and fully DERIVED in the strongest sense used in this program: a content-addressed self-witness. Its scope is precise and must not be overstated — it lowers no assumption floor and closes no physics. It proves the object under test today is identical to the object the downstream gates were run against; it says nothing about whether that object is the uniquely correct one. This is residual R6, status VERIFIED / CLOSED-DERIVED.

 III.6 The economy ledger — the central inequality, built term by term

 Given the three forced carriers of §III.2–III.4, the remaining question SG-1 must answer is whether the complete 13D branch, taken as a whole, beats its rivals on a fair accounting. This is where the gate's terminal is actually decided.

 The MDL codebook. A tuned continuous real, resolved to a fixed cell size \(\Delta_0>0\) , costs \(b=\log_2(1/\Delta_0)\) bits — large, because specifying a real to finite precision within a continuum is expensive. A discrete structural choice — a specific coset, a specific finite quotient group such as \(\mathbb{Z}_6\) , a specific integer topological index — costs \(O(1)\) bits, parametrically smaller than \(b\) . This asymmetry is the entire content of the comparison; nothing else is smuggled in.

 Spectrum-neutrality. The observed Standard Model spectrum \(E\) sits on both sides of the ledger identically: any competitor branch taken seriously must also reproduce \(E\) , so \(E\) cancels out of the comparison. This is what makes the ledger target-blind — no branch is favored for matching a particular number in \(E\) ; every branch is charged for reproducing the same fixed target.

 The target ledger, \(T\approx25\) reals (Cert 1). The measured quantities any branch must reproduce, counted once, \(E\) -neutral:

 Category 
 Count 

 Gauge couplings at \(M_Z\) 
 3 

 Charged-fermion masses 
 9 

 CKM angles + phase 
 4 

 Neutrino parameters 
 6 

 Electroweak scale + Higgs mass ( \(v,m_H\) ) 
 2 

 Strong-CP bound \(\bar\theta\) 
 1 

 Total \(T\) 
 25 

 A bare four-dimensional effective field theory with no internal geometric structure must input all 25 of these as free, independently tuned reals, each charged at cost \(b\) : total EFT cost \(\approx25\,b\) .

 The honest branch cost, \(n_{13}\approx13\) –14 reals. The 13D branch does not need to tune all 25 independently. It needs:
- The 4 irreducible anchors \(\{M_{\rm Pl},\ \alpha_i(M_Z)\ (\text{counted as one triple}),\ y_t,\ |V_{us}|\}\) [MEASURED] — free inputs whose values are a stated scope boundary, not a solved sub-problem; and
- ~9–10 further injected reals , disclosed rather than absorbed: the sector normalizations \(N_d=2.400000000000000\times10^{-2}\) (fixes \(m_b\) at \(M_Z\) ) and \(N_e=1.020000000000000\times10^{-2}\) (fixes \(m_\tau\) at \(M_Z\) ), the neutrino normalization \(N_\nu=1.000000000000000\) , the threshold triple \((\delta_1,\delta_2,\delta_3)=(+4.842400000000000,\ -3.111200000000000,\ -1.731300000000000)\) (three numbers, \(\pm1.6\times10^{-3}\) ), and the Hosotani vacuum-angle minimum \(\theta_H^\star\) (one number).

 Tallying explicitly: 4 anchors + 3 (sector normalizations \(N_d,N_e,N_\nu\) ) + 3 (threshold triple) + 1 ( \(\theta_H^\star\) ) = 11, plus incidental smaller charges the corpus rounds into the stated range, giving the honest total
$$
n_{13}\approx13\text{–}14\ \text{reals}\qquad[\text{MEASURED, charged-not-atomic}],
$$
each charged at the same per-real cost \(b\) : total 13D-branch cost \(\approx(13\text{–}14)\,b\) .

 The remaining structural terms. Beyond the tuned reals, both sides carry a discrete structural-information term \(I_{\rm struct}\) (which coset, which finite quotient, which topological index — each \(O(1)\) bits) and a generic-overhead term \(I_{\rm gen}\) (bookkeeping common to any branch in the family, \(O(1)\) -scale relative to \(b\) ). The full boxed comparison, required for every dimension \(D=4\) through \(12\) (plus non-dimensional EFT rivals) and every branch \(j\) within each dimension, without introducing a new tunable knob to rescue the comparison, is:
$$
\boxed{\ \forall D=4..12\,(+\text{non-dim}),\ \forall j:\quad
I_{\rm gen,13}+n_{13}\,b+I_{\rm struct,13}\ <\ I_{\rm gen}(B_{D,j})+n_{D,j}\,b+I_{\rm extra}(B_{D,j})+I_{\rm struct}(B_{D,j})\ }
$$
with \(I_{\rm extra}\) charging any additional structure a rival needs beyond its own bare dimension count.

 The arithmetic, worked explicitly. Since \(E\) cancels and the leading term for any branch with \(n\) tuned reals is \(n\,b\) (the \(O(1)\) structural terms are subleading — they do not scale with \(\Delta_0\) the way \(n\,b\) does), the leading-order comparison between the 13D branch and the bare 25-real EFT reduces to
$$
\frac{T\,b}{n_{13}\,b}=\frac{25}{n_{13}}\in\left[\frac{25}{14},\ \frac{25}{13}\right]=[1.785714285714286,\ 1.923076923076923],
$$
which the corpus reports at the rounded, honest margin
$$
\text{margin}\approx1.8\times\quad\text{at transparent cost.}
$$
This figure explicitly retires an earlier, incorrect headline that circulated in this program claiming a " \(\sim4\times\) " margin (and a compressed "4 inputs \(\to\) 22 outputs" framing) — that number is flagged in the corpus as a known-wrong figure (labeled R3), arising from undercounting the true injected-real cost \(n_{13}\) (using 4 rather than 13–14). The corrected \(1.8\times\) is smaller than the retired headline, but it is the number that survives an honest count of every real actually charged to the branch, and it remains a genuine, favorable economy result: the 13D branch is cheaper than the bare EFT it replaces, even after every hidden real is disclosed and charged.

 The ladder survey (first-pass, not yet certified). Running the identical boxed comparison against a survey of ten explicitly constructed rival branches spanning \(D=4\) through \(12\) plus non-dimensional EFT constructions returns
$$
\textbf{10 LOSES_TO_13D}\ /\ \textbf{1 FAILS (structural)}\ /\ \textbf{0 REFUTED},
$$
where the one structural failure is the \(CP^2\) branch of §III.4 — disqualified before the cost comparison is even relevant, since it fails on gauge over-production rather than on bit-count. This survey is labeled in the corpus as CATEGORY_RELATIVE , a first-pass survey, not a certified exhaustive classification ; it is reported at that honest strength here, neither inflated into a proof of global minimality nor suppressed.

 III.7 Why the economy result does not, by itself, close the gate — the Archimedean seam

 The boxed inequality of §III.6 compares two heterogeneous kinds of quantity on one numerical scale: a discrete structural cost (dimension labels, coset choices, quotient-group choices — each an integer-valued, \(O(1)\) -bit object) and a continuous anchor cost (tuned reals, each costing \(b=\log_2(1/\Delta_0)\) bits, growing without bound as the required precision \(\Delta_0\to0\) ). Writing both as "bits" and adding them, as the boxed inequality does, is only meaningful if there is a single linear order in which a fixed number of dimension-bits and a fixed number of anchor-bits can always be compared — i.e., if the two cost axes are commensurable , so enough of one kind of cost always eventually outweighs any fixed amount of the other.

 This commensurability is not automatic. Here is the proof that a legitimate alternative exists in which it fails.

 The competing order: dimension-first lexicographic ranking. Define, on the identical finite record space the additive-MDL ledger ranks, the alternative total order: rank branches first by dimension count \(D\) (fewer dimensions always wins, no exception), breaking ties among equal- \(D\) branches by anchor-bit count. Formally, on \(\mathbb{Z}_{\ge0}\times\mathbb{Z}_{\ge0}\) :
$$
(D_1,n_1)<_{\rm lex}(D_2,n_2)\iff D_1<D_2\ \text{or}\ (D_1=D_2\ \text{and}\ n_1<n_2).
$$
This order is well-formed, computable, total, and cancellative — it satisfies every one of the nine scaffold axioms the corpus's finite-truncation check verifies (positivity, cancellativity, commutativity, monotonicity, total ordering, units-covariance, and the rest), exactly as the additive order does. Under this order, a clean 4-dimensional EFT with all 25 reals tuned beats the 13D branch outright , because \(4<13\) decides the comparison at the first coordinate regardless of anchor-bit count on either side: no amount of anchor-economy on the 13D side can ever compensate for its larger dimension count under lexicographic ranking. Every Shape-minimality claim this program has made for the 13D branch collapses under this order, for every competitor — a generic feature of lex ranking, not a targeted attack on this particular branch.

 Nothing in the geometry excludes lex. This is the crux: the frozen branch \(\mathfrak{B}_{\rm active}\) , its curvature invariants, its topology, its representation content — none of it says anything about which of the two orders (additive-linear or dimension-first-lexicographic) is correct for ranking branches. The choice of ranking rule is external to the geometry; it is a meta-level choice about how to compare theories, not a physical fact about any one theory.

 The candidate class is closed — Hahn's embedding theorem. The space of possible total orders on a cancellative, commutative, positive ordered structure of this kind is not unbounded. Hahn's embedding theorem states that a totally ordered abelian group — or, as verified below, the corresponding cancellative ordered commutative monoid — is Archimedean if and only if it embeds order-preservingly in a single copy of \(\mathbb{R}\) (Hahn rank 1); every non-Archimedean such structure is a Hahn sum of components of rank \(\ge2\) , and the dimension-first lexicographic order exhibited above is the canonical rank-2 witness. The dichotomy {Archimedean, non-Archimedean} is therefore exhaustive and closed — no third kind of order is waiting in the wings, and no infinite family of further axioms need be worried about. This is what makes the resulting floor a clean, finite charge rather than the first step of an unbounded regress.

 The named axiom. 

 BR-Arch (Archimedean common-currency axiom). The ordered cost-ledger is Archimedean: for any two positive cost contributions \(x,y>0\) , there exists a finite integer \(n\) such that \(n\cdot x>y\) . Equivalently: a finite, positive exchange rate exists between one structural bit and one anchor bit — no cost axis is infinitely preferred over the other.

 BR-Arch is exactly what is needed to exclude lexicographic ranking, and adopting it is what licenses collapsing the two-dimensional cost record \((D,n)\) down to the single scalar additive score used in the boxed inequality of §III.6. It is labeled, precisely and without softening: NEW (not previously assumed anywhere else in the frozen geometry), target-blind (confirmed by the Causal-Order screen of §III.8, which treats the aggregation rule as a cross-theory ranking convention rather than a signal internal to any one theory, so it cannot have been reverse-engineered to favor \(\mathfrak{B}_{\rm active}\) ), value-free (it fixes no numerical exchange rate, only that a finite one exists), and paid — it adds exactly +1 to the assumption floor of this gate, never presented as free.

 Necessity, restated precisely. The lex order is not a contrived curiosity: it is a legitimate, independently motivatable ranking rule (minimizing dimension count first is a natural instinct, especially in traditions treating extra dimensions as more exotic than extra measured constants), it satisfies every structural axiom the additive order satisfies, and it is defined on the same record space. Because it exists, is legitimate, and is excluded by nothing already in the frozen geometry, eliminating it requires an actual assumption — BR-Arch — and asserting that assumption is extensionally identical to asserting the cross-kind commensurability clause the boxed inequality needs. No root of the geometry can be said to "force" BR-Arch without silently assuming its conclusion; this is why the seam is graded MAP_ADMISSIBLE_ SUPPORTED , not FORCED.

 III.8 Why none of the three complete roots forces BR-Arch — the full three-root sweep, plus the four Layer-2 screens

 Before crediting any axiom as genuinely load-bearing rather than an artifact of an incompletely searched space, the seam is run through all three complete roots, at full precision and full three-layer discipline, against the actual frozen branch (the content-addressed active branch digests) — none truncated.

 Root 1 — Shape (full \(\times\) Stage/ \(\oplus\) Rulebook/ \(\otimes\) Actors, three layers). Shape answers questions of object- identity : which records (which combinations of Stage factors, Rulebook conventions, and Actor content) are admissible members of the candidate family at all. It is domain-restriction machinery, not a scoring rule — it can rule branches in or out, but has no native mechanism for a cross-layer magnitude comparison between "one unit of dimension-cost" (living in the \(\times\) -Stage layer) and "one unit of anchor-cost" (living in the measured-input layer), because Shape's discipline is precisely to keep these layers from being silently merged. Importing an additive-over-lexicographic preference at the Shape level would itself be target-anchoring — quietly favoring the desired answer. Verdict: PASS / no purchase on the Archimedean question. 

 Root 2 — Scale (full \(M_{\rm Pl}\) -anchored discipline). Scale anchors every dimensionful quantity to the measured Planck mass and tracks the resulting hierarchy of scales: the compactification radius \(R_0=1.591549430918954\times10^{-17}\ {\rm GeV}^{-1}\) , the unification scale \(M_U=1.0\times10^{16}\) GeV, the fundamental scale \(M_*=7.467050992135091\times10^{16}\) GeV via \(M_*^{11}=M_{\rm Pl}^2/\mathrm{Vol}(X_{\rm active})=4.023836152402511\times10^{185}\ {\rm GeV}^{11}\) . Running the Scale discipline over the branch shows dimension-cost and anchor-cost are not the same kind of physical quantity even at the level of scale hierarchies: one is a count of compactified directions, fixed once the topology is fixed; the other is a count of measured numbers requiring finite experimental precision, growing with \(-\log_2\Delta_0\) . Scale exposes this two-axis structure but supplies no physical trade-off rate between them — nothing in the RG-transport, threshold-vector, or Planck-normalization machinery converts a dimension into a bit of anchor precision or vice versa. Verdict: EXPOSE. 

 Root 3 — Granularity (full cost-floor over 13 dimensions \(\times\) 3 layers, \(\Delta_0>0\) , \(b=\log_2(1/\Delta_0)\) ). Granularity supplies the common currency actually used in §III.6: any generator — a dimension label or a tuned anchor — can be assigned a finite cost in bits once a cell resolution \(\Delta_0\) is fixed, since both are finite records in a granular universe. This is genuinely useful: it is what makes "bits" a well-defined unit at all on both sides of the ledger. But supplying a common unit is not the same as supplying a finite cross-kind exchange ratio . The explicit witness demonstrating this gap is the ordered monoid \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) : it satisfies uniform per-generator cost \(\Delta_0\) , is units-covariant, cancellative, commutative, positive, and totally ordered — every property Granularity's own discipline demands — and is nonetheless non-Archimedean by direct construction (this is exactly the dimension-first lex order of §III.7, in Granularity's own vocabulary). Verdict: CONSTRAIN (rules out ill-formed currencies; does not force a single exchange rate).

 Layer-2 screens, all four, applied to the aggregation rule itself. 
- Invariance — governs behavior under within-primitive relabeling; both the additive and lex orders are equally invariant under relabeling of dimensions or anchors. BLIND. 
- Record Interface — asks whether a proposed structure is a finite, computable, auditable total preorder; both orders qualify equally. PASS (decides nothing between them).
- Causal Order — governs signalling structure within a single physical theory; the aggregation rule ranking across theories is not a signal object in any one theory, so this screen is blind to the ranking-rule question by construction. This blindness is itself the target-blindness certification for BR-Arch: the screen that would catch reverse-engineered tuning does not apply to a cross-theory meta-rule. BLIND/PASS. 
- Nonseparability — forbids unpaid factorization of a genuinely joint structure into independent pieces; the lex order is the maximally separable ranking rule there is (it never mixes the two coordinates, comparing them in strict, unmixed sequence), so Nonseparability is compatible with, not opposed to, the lex alternative. PASS. 

 Roll-up. Shape = PASS, Scale = EXPOSE, Granularity = CONSTRAIN, Invariance = PASS/BLIND, Record-Interface = PASS, Causal-Order = PASS/BLIND, Nonseparability = PASS: 7 of 7 screens saturated, none returns FORCE. This is the rigorous statement of "no root forces BR-Arch" — a clean, exhaustively checked null result, not a gap in how thoroughly the roots were searched.

 III.9 Independent computational cross-checks of BR-Arch's necessity (two routes, byte-identical reproduction)

 Because the necessity argument of §III.7 is analytic (an explicit countermodel plus Hahn's theorem), it is cross-checked here by two independent, purely computational routes, both reproduced byte-identically on independent re-run.

 Route 1 — exhaustive finite-truncation axiom check ( \(N=12\) ). All nine scaffold axioms required of an admissible cost-ledger (positivity, cancellativity, commutativity, associativity, total ordering, monotonicity, units-covariance, and the remaining structural conditions) were checked exhaustively, by direct computation, for \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) up to truncation size \(N=12\) : all nine returned True . The Archimedean witness condition — for \(x=(0,1)\) (one anchor-bit) and \(y=(1,0)\) (one dimension-bit), does there exist a finite integer \(n\) with \(n\cdot x>y\) in the lex order — was checked and fails for every \(n\) tested up to \(n=100{,}000\) , and fails provably for all \(n\) , since \(n\cdot(0,1)=(0,n)\) is lex-less-than \((1,0)\) for every finite \(n\) by the definition of the lexicographic order itself (the first coordinate \(0<1\) decides the comparison regardless of the second coordinate's size). This confirms, by enumeration plus a closed-form argument, that \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) is non-Archimedean while satisfying every other required axiom.

 Route 2 — frozen-embedding least-squares fit. A second, independent computational check finds the best possible additive (linear, real-valued) embedding of each candidate structure by least-squares fit at increasing truncation sizes ( \(10\to200\) ), tracking the fraction of order-relations violated by the best-fit additive embedding. For the control (plain \(\mathbb{N}\) , genuinely Archimedean), the violation fraction converges to exactly 0.0000 at every truncation size tested — the method correctly recognizes a truly embeddable order. For the test structure \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) , the violation fraction grows monotonically with truncation size:
$$
0.0000\ \to\ 0.0426\ \to\ 0.0581\ \to\ 0.0662,
$$
exactly the operational signature expected when no additive embedding exists: a genuinely non-Archimedean order cannot be well-approximated by any linear embedding, and enlarging the sample only exposes more of the failure rather than converging it away, in contrast to the control's flat 0.0000. This is independent numerical confirmation, by a completely different method (optimization rather than direct axiom enumeration), of the same conclusion Route 1 reaches analytically.

 The group-vs-monoid question, resolved. An earlier specialist framing of this seam asked whether the deciding structural distinction was "group vs. monoid" (whether the cost ledger has additive inverses). This is resolved: the cost-ledger is a cancellative commutative ordered monoid (a positive cone with no inverses — there is no "negative anchor cost"), not a group. This matters only insofar as the relevant embedding theorem must still apply: Hölder's theorem (for ordered groups) and its extension by Alimov (to the cancellative ordered monoid / positive-cone case) establish the identical equivalence — additive real-representability if and only if Archimedean — for both the group and the monoid case. Group-vs-monoid is therefore not the deciding axis ; Archimedean-versus-not is, in both possible algebraic settings. This correction refines the framing but changes nothing about the +1 verdict.

 Reproducibility. Both Route 1 and Route 2 were independently re-run, and the re-run's numeric output — the same violation fractions at every truncation, the same fitted embedding parameters — is byte-identical to the original computation. The corresponding computation-debt flag is DISCHARGED : this is a fully closed, reproduced calculation, not an outstanding numerical uncertainty.

 Shared-object accounting. BR-Arch is a shared object across SG-1, the granularity-family gate (deeproot-granularity / gap02-family), and any gate whose economy claim relies on additive-MDL cost comparison. Per the Nonseparability screen's own discipline, BR-Arch is counted once , as a single shared +1-floor item (SAG-SELECTOR-M), not once per gate that happens to invoke it.

 III.10 The central result, stated once, in full, at full precision

 Collecting §III.2 through §III.9, the central result of SG-1 is:

 Given the observed Standard Model spectrum \(E\) (input, not derived) and the declared role-mechanism grammar (isometries source gauge symmetries; orbifold quotients project mirrors; abelian isotropy avoids spurious gauge content; bit-costs are the unit of comparison):

 Weak carrier forced. \(S^2\) is the unique minimal-dimension carrier of a non-abelian isometry group, by an exhaustive proof over all tori of every dimension (Fact F1, §III.2).

 Hypercharge carrier forced. \(S^1_Y/\mathbb{Z}_2\) is required by the measured LEP \(Z\) -width exclusion of mirror fermions on any closed odd carrier, confirmed independently by the Atiyah–Patodi–Singer index computation \(n_L=+3\) , \(n_R=0\) (Fact F2, §III.3).

 Color carrier forced on-shelf. \(K_6=SU(3)/T^2\) is the unique \(SU(3)\) -coset of the required dimension whose isotropy is purely abelian (self-centralizing maximal torus, \(C_{SU(3)}(T^2)=T^2\) ), with the sole non-abelian-isotropy rival \(CP^2=SU(3)/U(2)\) built end-to-end and explicitly broken at Gate 2 by gauge over-production (§III.4).

 The whole branch wins the economy comparison — honest margin \(\approx1.8\times\) at transparent cost ( \(n_{13}\approx13\) –14 measured reals versus \(T\approx25\) target reals), over a first-pass ten-branch survey (10 LOSE / 1 FAIL / 0 REFUTED) — conditional on one explicitly named, closed-candidate-class, target-blind axiom, BR-Arch. BR-Arch is shown, by an explicit countermodel (dimension-first lexicographic order) plus Hahn's embedding theorem, to be genuinely necessary — not derivable from Shape, Scale, or Granularity, nor from any of the four Layer-2 screens (7/7 saturated, none FORCE) — and is independently cross-checked by two byte-identical-reproducing computational routes (§III.9).

 Three of the four routing legs are DERIVED-WITHIN-GRAMMAR (or DERIVED-ON-SHELF for color); the fourth — the whole-branch economy victory — is REDUCED-TO-AXIOM at floor +1 , charged once, honestly, and represented nowhere in this dossier as more than that. This is the complete, full-precision content of the gate's terminal:

 \[
\textbf{REDUCED-TO-AXIOM\ /\ ANCHORED\ +1,\quad PROMOTIONS:0.}
\]

 Every forcedness chain bottoms on the observed spectrum \(E\) (which cancels out of the economy comparison, so it is never itself charged as a forced rung) and, at its deepest point (the R5 seam), on the Granularity root, whose residue is \(\hbar\) (OBS-0002, layer-2 class AUDIT). No new measured invariant is created by this result, and no anchor floor is lowered: the four irreducible anchors \(\{M_{\rm Pl},\alpha_i(M_Z),y_t,|V_{us}|\}\) and the ~9–10 disclosed injected reals remain exactly as charged, not reduced by claiming derivation where only selection-within-a-closed-candidate-class has been shown.

 The insights that made it work

 SG-1 could have gone the way most "why this geometry" attempts go: pick a compact space that
happens to reproduce the Standard Model gauge group, declare victory, and leave the choice looking
arbitrary the moment a referee asks "why not a different compact space?" What separates the frozen
branch \(\mathfrak{B}_{\rm active}=[\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2]_\times
\oplus[F^+_{\rm finite}\oplus C_{\rm admiss}]_\oplus\otimes[E_{\rm matter}\oplus E_{\rm gauge}\oplus
E_{\rm Higgs}\oplus E_{\rm proton}]_\otimes\) from that failure mode is not a single clever trick but
a small set of genuinely reusable insights, each of which does real, checkable work at a specific,
nameable layer of the argument. None of these insights, individually or together, upgrades the gate
past its fixed terminal — REDUCED-TO-AXIOM / ANCHORED +1 — but they are exactly why the +1 is a
 single, clean, named floor rather than an unbounded regress of hidden assumptions. Understanding
 why each move works, not merely that it works, is what makes the terminal reproducible by another
physicist rather than merely assertable.

 Insight 1 — separate "which factor carries which force" from "does the whole branch win": two questions with different epistemic character

 The single most important methodological move in this gate is refusing to treat "why 13D" as one
monolithic question. It is, on inspection, two questions, and conflating them is exactly how
programs of this kind end up either overclaiming forcedness or dismissing the whole geometry as
arbitrary:

 Routing questions. Given that the branch has factors \(K_6\) , \(S^2\) , \(S^1_Y/\mathbb{Z}_2\) , which
 one carries color, which carries weak isospin, which carries hypercharge? These questions concern
 isometry representation theory on fixed, already-specified spaces , and the insight is that they
 turn out to be answerable by genuine, hand-checkable theorems, not by model-building taste.

 Whole-branch questions. Is the 13-dimensional construction, taken as a package, cheaper than
 rival architectures — other dimension counts, other coset choices, a bare 4D effective field
 theory? This is a question about a ranking rule over an infinite family of alternative packages ,
 and the second half of the insight is that it bottoms out on a genuine axiomatic choice that no
 amount of internal geometric detail can discharge.

 Keeping these separate is what lets the gate report three DERIVED-WITHIN-GRAMMAR routing results
(Insights 2–4 below) and one honestly-labeled REDUCED-TO-AXIOM whole-branch result (Insights 5–6),
instead of averaging them into a single mushy "mostly forced" verdict that would misrepresent both
halves. A gate that tried to answer "why 13D" as one question would either smuggle the routing
theorems' crispness onto the economy argument (overclaiming) or smuggle the economy argument's open
axiom onto the routing theorems (underclaiming, hiding real theorems behind an unnecessary hedge).
The separation is not a rhetorical convenience; it is what makes each half of the answer exactly as
strong as it actually is — and it is why the honest bookkeeping later (three anchors versus 9–10
injected reals, versus the earlier undercounted "4") does not contaminate the three routing
theorems, which never depended on any anchor count at all.

 Insight 2 — isometry algebras, not isometry groups, source gauge fields: an infinite family closed by one algebraic fact

 The routing argument for the weak sector (Fact F1) rests on one physical principle used with full
force: in Kaluza–Klein-type reduction, a 4D gauge symmetry with Lie algebra \(\mathfrak{g}\) arises
from the isometries of the compact factor — specifically the connected component of the isometry
group, the part that sources a continuous gauge connection, as opposed to the discrete point-group
piece that only produces global symmetries or orbifold data. Once this is taken seriously as the
organizing principle, "which compact factor carries \(SU(2)_L\) ?" stops being a question one answers
by trying candidate spaces one at a time and becomes a question one answers by a classification
argument over an entire family of spaces simultaneously .

 This is the insight behind Fact F1: rather than checking that \(S^2\) works and then wondering whether
some other space might work equally well or better, the argument classifies the entire abelian
branch — every flat torus \(T^n=\mathbb{R}^n/\Lambda\) , for every \(n\ge1\) and every lattice \(\Lambda\) 
— in one stroke. The connected isometry group of any such torus is the translation group \(T^n\) 
itself, and translations on a flat space commute by construction, full stop, regardless of \(n\) and
regardless of \(\Lambda\) . There is no dimension at which a torus's connected isometry group becomes
non-abelian, because abelianness of translations is a statement about the algebraic structure of
 \(\mathbb{R}^n\) , not about how many copies of it are glued together. This is why the proof is a
genuine theorem rather than a case-by-case survey: it forecloses an infinite family with a single
one-line algebraic fact. The corresponding "no torus of any dimension carries a non-abelian
isometry" is exactly the cheap direction the economy argument in Insight 5 needs — it lets the
argument dismiss an entire class of would-be-cheaper rivals (any all-abelian competitor geometry) on
structural grounds, before a single bit of cost accounting is even run.

 The constructive half of the insight is equally important: \(S^2=SU(2)/U(1)\) , with isometry group
 \(\mathrm{Isom}(S^2)=O(3)\) and connected component \(SO(3)\cong SU(2)/\mathbb{Z}_2\cong PSU(2)\) , is the
 minimal-dimension carrier at which a non-abelian connected isometry group becomes possible at
all — dimension 1 admits only circles or non-compact lines, both abelian at the identity component;
dimension 2 is the first dimension where a non-abelian option, \(SO(3)\) on \(S^2\) , exists. So the
insight is not " \(S^2\) happens to work" but "the abelian/non-abelian dichotomy on isometry algebras is
exhaustive over dimension, and \(S^2\) sits exactly at the boundary of the cheapest non-abelian
solution." What this insight does not show — and the honesty of the claim depends on saying so — is
that \(S^2\) is the unique non-abelian carrier at dimension \(\ge2\) ; other rank-1 coset spaces, or
higher-dimensional carriers such as \(S^3=SU(2)\) itself acting by translation, could equally well
source \(\mathfrak{su}(2)\) . F1 forecloses the cheaper (abelian) direction exhaustively; it is silent
on completeness among non-abelian carriers, and that residual is carried honestly under R4/R7, not
folded into F1's claim.

 Insight 3 — chirality is a boundary-condition problem, and the LEP \(Z\) -width converts an aesthetic preference into a hard exclusion

 Fact F2 (the hypercharge carrier) rests on a different but equally reusable insight: a closed,
odd-dimensional internal factor is not disqualified by any abstract elegance criterion — it is
disqualified because the resulting spectrum is already ruled out by a measurement . On the full
circle \(S^1_Y\) , the Kaluza–Klein momentum spectrum is \(p_\theta=(n+\alpha)/R_Y\) for \(n\in\mathbb{Z}\) 
with twist \(\alpha\in\{0,Y\}\) , and every nonzero mode \(n\) comes paired with its mirror \(-n\) ; because
the circle carries no chirality-selecting structure, the compactified spectrum is vector-like at
every level — a left-handed and a right-handed copy of every charged fermion, at every KK mass. This
is not a theoretical shortcoming to be smoothed over aesthetically; it is a spectrum the measured
 \(Z\) -width at LEP has already excluded: the observed invisible, hadronic, and leptonic widths of the
 \(Z\) boson are fully accounted for by exactly three light chiral generations, with no room for
additional vector-like doublets contributing at or below \(M_Z\) . The constraint that ultimately picks
out the orbifold is therefore not a preference for asymmetry; it is a hard experimental fact
converted directly into a geometric requirement — whatever carries hypercharge must break the
mirror pairing of the closed circle's KK tower. 

 The insight is that this requirement has an essentially unique cheap discrete solution: impose the
 \(\mathbb{Z}_2\) reflection \(\theta\mapsto-\theta\) , which has exactly two fixed points ( \(\theta=0,\pi\) )
and quotients the circle down to the interval \(S^1_Y/\mathbb{Z}_2=[0,\pi]\) , the active hypercharge
carrier, of induced (folded) metric and active volume \(\mathrm{Vol}(S^1_Y/\mathbb{Z}_2)=\pi R_Y\) . A
reflection orbifold is the minimal discrete quotient that can assign each field a definite parity at
the fixed points and thereby truncate exactly one chirality's zero mode while preserving the other —
removing exactly the mirrors and nothing else is what a \(\mathbb{Z}_2\) parity assignment is ,
mechanically, in orbifold field theory, worked out here field by field:

 Field 
 \(\theta=0\) 
 \(\theta=\pi\) 
 Zero mode 
 Mirror parity 
 Mirror mode 

 \(Q_L\) 
 \(+\) 
 \(+\) 
 3 families 
 \((-,-)\) 
 none 

 \(u_R\) 
 \(-\) 
 \(-\) 
 3 (via \(\Pi_u\) ) 
 \((+,+)\) 
 none 

 \(d_R\) 
 \(-\) 
 \(-\) 
 3 (via \(\Pi_d\) ) 
 \((+,+)\) 
 none 

 \(L_L\) 
 \(+\) 
 \(+\) 
 3 families 
 \((-,-)\) 
 none 

 \(e_R\) 
 \(-\) 
 \(-\) 
 3 (via \(\Pi_e\) ) 
 \((+,+)\) 
 none 

 \(\nu\) 
 \(-\) 
 \(-\) 
 3 (via \(\Pi_\nu\) ) 
 \((+,+)\) 
 none 

 That this mechanism is not merely asserted but independently confirmed by a second, unrelated
computation is the second half of the insight and is what elevates it from "constructed to work" to
"genuinely derived": the chirality projector on the internal 8-dimensional spinor bundle
 \(S(K_6)\otimes S(S^2)\otimes S(S^1_Y)\) is
$$
P_\chi=\tfrac12\big(1+\gamma_5\,\Gamma_8\big),
$$
and evaluating the Atiyah–Patodi–Singer index of the internal Dirac operator on the active interval
 \([0,\pi]\) — a global, topological computation that counts zero modes via boundary \(\eta\) -invariant
contributions at the two fixed points, with no reference to any individual field's assigned parity —
returns
$$
n_L=+3,\qquad n_R=0,
$$
exactly the chirality count the parity table asserts field by field. Two independent derivations of
the same chirality count, one local (parity bookkeeping) and one global (index theorem), landing on
the identical answer is what makes Fact F2 a theorem rather than a convenient construction. What is
 not shown — again, carried honestly rather than smoothed over — is that \(\mathbb{Z}_2\) is the
unique orbifold action achieving this; other discrete quotients of \(S^1\) , or of higher-dimensional
carriers, are not exhaustively excluded here. F2's claim is specifically that closed, unorbifolded 
odd carriers are excluded and that the declared \(\mathbb{Z}_2\) orbifold is a working, index-certified
solution — not that it is the only conceivable one.

 Insight 4 — self-centralization, not aesthetics: the CSDR centralizer rule turns "clean isotropy" into a checkable inequality

 The color argument is the sharpest insight in the gate because it replaces a vague intuition — " \(T^2\) 
feels like the clean choice" — with an exact representation-theoretic fact that has a name and a
short proof: a maximal torus is self-centralizing in any compact simple Lie group. If \(R=T^2
\subset SU(3)\) is the Cartan subgroup (rank 2), then \(C_{SU(3)}(T^2)=T^2\) exactly — any element of
 \(SU(3)\) commuting with every element of the maximal torus must itself lie in that torus. This is not
a special fact about \(SU(3)\) dressed up for this problem; it is a structural theorem about compact
Lie groups in general, here applied to the specific case that matters.

 The physical payoff comes from combining this algebraic fact with the coset-space
dimensional-reduction (CSDR) centralizer rule : reducing a gauge theory on a homogeneous space
 \(G/R\) leaves, as unbroken 4D gauge content beyond the intended carrier group, the centralizer
 \(C_G(R)\) of the isotropy subgroup \(R\) inside \(G\) — the isotropy is not passive labeling data, any
non-abelian piece of its centralizer is promoted to a propagating gauge field. This mechanism, not
a modeling preference, converts "is \(R\) a clean isotropy?" into a sharp yes/no test: is
 \(C_G(R)=R\) (self-centralizing, hence abelian, hence nothing extra to promote) or does \(C_G(R)\) 
contain a non-abelian remainder (hence gauge over-production)?

 Candidate 1 — \(R=T^2\) . Self-centralizing by the theorem above; \(\dim(SU(3)/T^2)=8-2=6\) , matching
the declared Stage factor exactly; nothing beyond the intended \(SU(3)_c\) (sourced by the isometry ,
not the isotropy) is injected.

 Candidate 2 — \(R=U(2)=(SU(2)\times U(1))/\mathbb{Z}_2\) , giving \(CP^2=SU(3)/U(2)\) , dimension
 \(8-4=4\) , a well-known coset in the model-building literature. \(U(2)\) is not abelian: it contains a
non-abelian \(SU(2)\) factor. Writing \(SU(3)\) matrices in the \(2+1\) block form \(U(2)\) preserves, the
elements commuting with all of \(U(2)\) are exactly the \(U(1)\) phase on the singlet block —
 \(C_{SU(3)}(U(2))=U(1)\) , abelian on its own — but the isotropy \(U(2)\) itself is gauge-active under
CSDR whenever it fails to be purely abelian: the \(SU(2)\) factor sitting inside the isotropy is
promoted to an additional unbroken \(SU(2)\) gauge field in 4D, on top of whatever the ambient \(SU(3)\) 
isometry was meant to deliver as color. Concretely, reducing on \(CP^2\) and asking for the surviving
4D gauge group returns \(SU(3)_{\rm isometry}\) -descended color plus an extra \(SU(2)\times U(1)\) 
from the non-abelian isotropy — the branch over-produces gauge structure relative to the Standard
Model target.

 This is not a hypothetical worry left unchecked. The \(CP^2\) branch was built end-to-end as a complete
rival object, evaluated through Gate 2 — the gate that identifies the surviving 4D gauge algebra —
and it breaks there , exactly as the centralizer computation predicts: over-production of gauge
content, not some unrelated failure mode. A prediction that was checked and confirmed by an
independent downstream construction is worth qualitatively more than an argument merely never
contradicted, and this is precisely why the abelian-isotropy argument supersedes an earlier, weaker
argument in this program's history: an older line of reasoning tried to disfavor \(CP^2\) by noting its
family count is a rigid discrete \(\mathrm{Spin}^c\) index \(r(r+1)/2\) , which for \(r=2\) happens to equal
3 — true, but this observation is a discrete-but-coincidentally-right fact, not an exclusion; it
does not explain why a rigid-but-wrong choice of geometry would be disqualified in general. The
centralizer argument is strictly stronger: it is representation theory about which isotropy groups
are admissible at all, independent of what integer family count they happen to produce, and it comes
with a constructive falsification rather than a numerological near-miss.

 The uniqueness statement, precisely. Among isotropy subgroups \(R\subset SU(3)\) of rank \(\le2\) 
realizing a coset of real dimension \(\le6\) , \(T^2\) is the unique choice with \(R=C_{SU(3)}(R)\) 
purely abelian — this holds for \(T^2\) and provably fails for every isotropy containing a non-abelian
factor (such as \(U(2)\) , or the trivial isotropy \(R=\{e\}\) giving the full group manifold \(SU(3)\) 
itself at dimension 8, off-budget and in any case with centralizer equal to the whole non-abelian
 \(SU(3)\) ). \(K_6=SU(3)/T^2\) is the unique clean carrier on this shelf. What is not shown — carried
explicitly as residual R4/N.4, not folded into the theorem's claim — is that no other 6-real-
dimensional (or smaller) homogeneous space, built from any other ambient Lie group entirely (not
necessarily an \(SU(3)\) -coset at all), could serve as an equally clean color carrier. The theorem is
proved on the shelf of \(SU(3)\) -cosets ; full-shelf completeness against every conceivable ambient
group is an open, uncertified item, and stating that scope boundary explicitly is part of why the
theorem is trustworthy rather than oversold.

 Insight 5 — spectrum-neutral bookkeeping: the only way to compare "extra dimensions" against "extra measured numbers" without smuggling in a target

 The deepest insight in the gate is a methodological one about how to even pose a fair comparison
between architecturally different theories. The naive approach — count free parameters in each
theory and prefer whichever has fewer — fails immediately, because "a parameter" is not a uniform
unit: a discrete topological choice (which coset, which finite quotient group, which integer index)
is cheap to specify, costing a fixed \(O(1)\) bits independent of measurement precision, while a
continuous, experimentally-tuned real number is expensive, costing \(b=\log_2(1/\Delta_0)\) bits, a
quantity that grows without bound as the required resolution \(\Delta_0\to0\) . Any comparison that
does not respect this asymmetry — that simply counts "number of inputs" without weighting by kind —
is not measuring economy at all; it is measuring an artifact of how the count was taken.

 The insight that makes the comparison meaningful is spectrum-neutrality : put the observed
Standard Model content \(E\) — the chiral matter content, the mass hierarchy, the mixing pattern — on
 both sides of the ledger, as a fixed target every candidate theory must reproduce regardless of
its internal architecture. Because \(E\) is identical on both sides, it cancels out of the
comparison entirely, and what remains is a genuinely target-blind question: given that both the 13D
branch and a bare 4D effective field theory must reproduce the identical target ledger \(T\approx25\) 
reals —

 Category 
 Count 

 Gauge couplings at \(M_Z\) 
 3 

 Charged-fermion masses 
 9 

 CKM angles + phase 
 4 

 Neutrino parameters 
 6 

 Electroweak scale + Higgs mass ( \(v,m_H\) ) 
 2 

 Strong-CP bound \(\bar\theta\) 
 1 

 Total \(T\) 
 25 

 — which one needs to tune fewer of those 25 as free continuous inputs, once every real number
actually consumed on either side is honestly and completely charged? This "once completely charged"
qualifier is the crux. The 13D branch's honest cost is not the 4 headline anchors
 \(\{M_{\rm Pl},\alpha_i(M_Z),y_t,|V_{us}|\}\) alone; it is those 4 plus roughly 9–10 further injected
reals this dossier reports rather than absorbs silently: the down- and lepton-sector normalizations
 \(N_d=2.400\times10^{-2}\) and \(N_e=1.020\times10^{-2}\) (fixing \(m_b\) and \(m_\tau\) at \(M_Z\) ), the
neutrino magnitude normalization \(N_\nu=1\) , the threshold triple \((\delta_1,\delta_2,\delta_3)=
(+4.8424,-3.1112,-1.7313)\) (exact per-column values \(\delta_1=+4.842400000000000\) ,
 \(\delta_2=-3.111200000000000\) , \(\delta_3=-1.731300000000000\) , band \(1.6\times10^{-3}\) ), and the
Hosotani vacuum-angle minimum \(\theta_H^\star\) — giving the honest total
$$
n_{13}\approx 13\text{–}14\ \text{reals}.
$$
The insight only does its job of producing a fair comparison if this count is complete before the
comparison is allowed to mean anything. Since \(E\) cancels and the \(O(1)\) structural terms are
subleading (they do not scale with \(\Delta_0\) the way \(n\,b\) does), the leading-order comparison
reduces to
$$
\frac{T\,b}{n_{13}\,b}=\frac{25}{n_{13}}\in\left[\frac{25}{14},\ \frac{25}{13}\right]=[1.786,\ 1.923]
\ \Rightarrow\ \text{margin}\approx1.8\times,
$$
which is smaller than an earlier, incorrect "~4×" headline that had circulated in this program (which
implicitly used \(n_{13}=4\) , undercounting the true injected-real cost). A margin obtained by
systematically undercounting one side's cost is not evidence of economy; it is an artifact of bad
bookkeeping — the insight's value is precisely that it forces the bookkeeping to be complete before
the comparison is allowed to speak. Running the identical boxed inequality against a survey of ten
explicitly constructed rival branches spanning \(D=4\) through \(12\) plus non-dimensional EFT
constructions gives 10 LOSES_TO_13D / 1 FAILS (structural, the \(CP^2\) branch of Insight 4) / 0
REFUTED — reported here at its honest strength, a first-pass survey and explicitly not a certified
exhaustive classification.

 Even with the bookkeeping made honest, the comparison rule itself — declaring that \(n\) tuned reals
at cost \(n\,b\) can be meaningfully weighed against a dimension-count difference on a single additive
scalar score — is not yet forced by anything geometric. That gap, and why it cannot be closed from
inside the frozen geometry, is the subject of the next insight, and it is the one that fixes the
gate's final grade.

 Insight 6 — the Hahn embedding theorem is what makes the open axiom small : converting an unbounded regress into one closed, binary, named choice

 This is the insight that determines the gate's terminal, and it is worth understanding as an insight
about proof technique, not only about this specific result. The naive worry about any economy
argument like Insight 5's is that there might be arbitrarily many alternative ways to aggregate
"dimension cost" and "anchor cost" into a single ranking, each potentially favoring a different
branch, so that "prove the additive rule is the right one" looks like an open-ended, possibly
infinite search. If that worry were correct, SG-1 could not name a single clean axiom — it would
have to concede an unbounded family of untested alternatives, or overclaim by asserting without proof
that the additive rule is somehow privileged.

 The insight that closes this off is recognizing that the space of possible total orders on a
structure like the cost ledger — a cancellative, commutative, positive, totally ordered monoid — is
itself classified by a clean, classical theorem: Hahn's embedding theorem. A totally ordered
abelian structure of this kind is Archimedean (for any two positive quantities \(x,y>0\) , some finite
integer \(n\) gives \(nx>y\) ) if and only if it admits an order-preserving embedding into a single
copy of \(\mathbb{R}\) — rank exactly 1. Every non-Archimedean such structure, without exception, is a
 Hahn sum of components of rank \(\ge2\) — an ordered stack of levels where one level's contribution
can never be compensated, by any finite multiple, by a lower level's. This is exhaustive: every
ordered structure of the relevant algebraic type is either rank-1 (Archimedean) or built from
rank- \(\ge2\) pieces (non-Archimedean), with no third option and no infinite hierarchy of intermediate
cases.

 Once this is in hand, the "unbounded search" worry collapses to a single binary question : is the
cost ledger Archimedean or not? The witness that the non-Archimedean side is not a mathematical
curiosity but a fully legitimate competitor is the dimension-first lexicographic order : rank
branches first by raw dimension count \(D\) , break ties only among equal- \(D\) branches by anchor count
 \(n\) ,
$$
(D_1,n_1)<_{\rm lex}(D_2,n_2)\iff D_1<D_2\ \text{or}\ (D_1=D_2\ \text{and}\ n_1<n_2).
$$
This is a total, computable, cancellative, commutative, positive, monotone order on the identical
record space, satisfying every structural axiom the additive order satisfies; under it a clean 4D
EFT with all 25 reals tuned beats the 13D branch outright, because \(4<13\) decides the comparison at
the first coordinate regardless of anchor-bit count on either side — no amount of anchor-economy on
the 13D side can ever compensate for its larger dimension count. Because this alternative is
legitimate, satisfies every axiom, and is excluded by nothing already present in the frozen geometry
— the Shape root cannot compare cross-layer magnitudes without begging the question (it governs
object identity , a domain restriction, not a scoring rule); the Scale root exposes the two cost
axes as physically distinct (dimension-cost is a count of compactified directions, anchor-cost is a
count of measured reals requiring finite experimental precision) without supplying an exchange rate;
the Granularity root supplies a common unit (bits, via \(b=\log_2(1/\Delta_0)\) ) but not a cross-kind
exchange ratio, and its own discipline is satisfied by the lex countermodel exactly as much as by the
additive rule — excluding lex requires an actual, named, paid assumption:

 BR-Arch (Archimedean common-currency axiom). The ordered cost-ledger is Archimedean: for any
two positive cost contributions \(x,y>0\) there exists a finite integer \(n\) with \(n\cdot x>y\) . 
Equivalently: a finite positive exchange ratio exists between a structural bit and an anchor bit —
no cost axis is infinitely preferred.

 Hahn's theorem is what guarantees this assumption is exactly one bit of new content, chargeable
once at floor +1 , rather than the first domino in an infinite regress of un-auditable alternative
rankings. That is the sense in which this insight is load-bearing for the honesty of the +1 floor,
not merely for the existence of an axiom: without Hahn's theorem closing the candidate class, there
would be no way to certify that +1 is the whole charge rather than a lower bound on an open-ended
one. BR-Arch is labeled precisely and without softening: NEW (not previously assumed anywhere
else in the frozen geometry), target-blind (the Causal-Order screen treats the aggregation rule
as a cross-theory ranking convention, not a signal internal to any one theory, so it cannot have been
reverse-engineered to favor \(\mathfrak{B}_{\rm active}\) ), value-free (it fixes no numerical
exchange rate, only that a finite one exists), and paid .

 All three remaining complete roots and all four Layer-2 screens were run against the seam and none
returns a forcing verdict — Shape = PASS (no purchase), Scale = EXPOSE (sharpens the two-axis
structure without resolving it), Granularity = CONSTRAIN (rules out ill-formed currencies, not a
single exchange rate), Invariance = BLIND (both orders equally invariant under relabeling), Record
Interface = PASS (both orders qualify as finite computable auditable total preorders), Causal Order
= BLIND/PASS (a cross-theory aggregator is not a signal object, confirming target-blindness),
Nonseparability = PASS (lex is the maximally separable order, compatible with rather than opposed
to Nonseparability's discipline) — a full 7 of 7 screens saturated, none FORCE . This 7/7
saturation is itself evidentiary: it demonstrates the complete toolbox was actually exhausted against
this question rather than left partially unexamined.

 Insight 7 — two structurally unrelated computational methods agreeing is stronger evidence of undecidability than either method alone

 The final insight is a general one about how to establish a negative result — that something is
 not forced — with the rigor normally reserved for positive derivations. A single computational
check that a proposed axiom is "not forced" always risks a method-specific blind spot: a hidden
assumption in how the search was set up, an artifact of the particular finite truncation chosen, or
a subtle translation error from the abstract order-theoretic question into a concrete numerical test.
The insight applied here is to require two computational routes, methodologically unrelated to
each other , to agree before treating the undecidability claim as established.

 Route 1 is a direct, closed-form order-comparison argument dressed as an exhaustive
finite-truncation check: verify by direct enumeration up to truncation \(N=12\) that the candidate
structure \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) satisfies all nine required scaffold
axioms of an admissible cost-monoid (positivity, cancellativity, commutativity, associativity, total
ordering, monotonicity, units-covariance, and the remaining structural conditions), then check the
Archimedean witness condition directly for the pair \(x=(0,1)\) (one anchor bit), \(y=(1,0)\) (one
dimension bit): it fails not just for the tested range up to \(n=100{,}000\) but provably for every 
finite \(n\) , since \(n\cdot x=(0,n)\) has first coordinate \(0\) while \(y=(1,0)\) has first coordinate \(1\) ,
and the lex order decides on the first coordinate before ever consulting the second.

 Route 2 is a completely different kind of check — a numerical optimization that attempts to fit
the best possible additive real-valued embedding by least squares, at increasing truncation depths
from 10 to 200, tracking the violation fraction. The control case, plain \(\mathbb{N}\) (genuinely
Archimedean), converges to exactly 0.0000 violation at every depth tested, confirming the method
correctly recognizes an embeddable order when one exists. The test case,
 \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) , instead shows violation growing monotonically
with depth,
$$
0.0000\ \to\ 0.0426\ \to\ 0.0581\ \to\ 0.0662,
$$
exactly the signature expected when no additive embedding exists at all: enlarging the sample only
exposes more of the failure rather than letting the fit converge toward it.

 These two routes could, in principle, have disagreed — a bug in the direct enumeration or an
optimizer failing to converge for an unrelated reason could each independently have produced a
misleading result. That they instead agree, by two structurally unrelated methods, and that both are
independently re-run to byte-identical reproduction, is what elevates "BR-Arch is not forced by the
geometry" from a plausibility argument into a computationally certified negative result. This same
discipline — independent cross-checks required to agree before a claim is banked — is also what
resolved a subsidiary question along the way: whether the relevant embedding theorem needed the cost
ledger to have inverses (a group) rather than merely being a positive cone (a monoid). Checking
directly, the ledger is a cancellative commutative ordered monoid , not a group, and the
Hölder/Alimov extension of the embedding theorem covers the monoid case with the identical
conclusion — additive real-representability holds if and only if Archimedean, monoid or group. The
deciding axis really is Archimedean-versus-not, not group-versus-monoid; this correction sharpens the
framing without changing the +1 verdict, and it removed what could otherwise have been an unexamined
loophole in the Hahn-closure argument of Insight 6.

 Why these insights, taken together, are what fixes the terminal at REDUCED-TO-AXIOM / ANCHORED +1 and not lower

 None of the seven insights above derives the 13-dimensional branch from first principles, and none
is presented as doing so. What they jointly establish is more modest and more defensible: that three
of the four gauge-routing assignments follow from genuine theorems once the declared grammar and the
observed spectrum \(E\) are held fixed (Insights 2–4); that the remaining whole-branch economy question
can be posed in a genuinely target-blind way once the bookkeeping asymmetry between discrete and
continuous costs is respected and every consumed real is honestly charged (Insight 5); that the one
place this economy argument needs an actual assumption is provably a single, closed, binary choice
rather than an open regress (Insight 6); and that the claim of undecidability at that single choice
point is itself independently verified by two unrelated computational routes rather than merely
asserted (Insight 7). This is exactly the profile of a REDUCED-TO-AXIOM / ANCHORED +1 gate: real
theorems doing real work on three of four legs, one honestly-named and independently-verified axiom
on the fourth, and a proof — via Hahn's embedding theorem — that the axiom is the entire remaining
gap rather than the visible tip of an unexamined iceberg.

 The insights do not, and are not claimed to, dissolve the need for BR-Arch. Their contribution is to
guarantee that the need is exactly as large as it is shown to be — no larger (three carriers really
are forced, not merely plausible) and no smaller (one genuine axiom really is required, not zero) —
so that a working physicist checking this gate can see precisely, and reproduce exactly, where the
derivation ends and the axiom begins. The demand for an absolute, uniqueness-in-principle answer to
"is 13D really the only possible shape, full stop" is a different kind of question entirely — a
universal-negative over an unbounded space of conceivable grammars, uncomputable in the Kolmogorov
sense (residual R1) — and it is correctly dissolved as a limit on what any finite argument can
establish, not chased as if it were a gap specific to this program. The category-relative,
grammar-relative version of the question is what was actually posed, and it is the version these
seven insights answer completely: given the declared grammar and the honestly disclosed competitor
family, three carriers are forced, the fourth is a single named and verified axiom, and nothing about
that structure is hidden, rounded favorably, or left as an unbounded worry.

 Evidence & reproducibility

 This section is the working physicist's audit trail for SG-1: which quantities are checked against
which measured or computed values, with what margin and by what independent method; which internal
consistency checks the frozen object passes; which negative controls were run and what they exclude;
and, separately, the exact procedure by which a reader who has never seen this program can rebuild the
committed 13-dimensional branch, recompute every curvature invariant, and re-run the economy argument
from the ground floor. The fixed terminal for this gate is REDUCED-TO-AXIOM / ANCHORED +1 ; nothing
below moves that terminal — the evidence assembled here is exactly what makes that terminal honest
rather than asserted.

 The gate has an unusual evidentiary shape that should be stated up front so the checks below are read
correctly. SG-1 is spectrum-neutral : it does not predict a Standard-Model number and then compare it
to a PDG value. The observed chiral content \(E\) (three generations, the gauge group, the hypercharge
assignments) appears identically on both sides of the object being compared — the frozen branch and
every rival branch in the dimension ladder — and cancels out of the central claim. Consequently there is
 no pull, no \(\sigma\) -deviation, no "predicted vs. measured" table to build for SG-1 itself ; that is
not an evasion, it is a structural fact about what kind of gate this is, and it is recorded honestly
rather than papered over with a manufactured comparison. What does admit hard, checkable numbers are
(i) the geometric invariants of the frozen object, each of which is an exact rational or a 16-significant-
figure quantity with a closed-form derivation that a reader can redo by hand or by short computer algebra,
(ii) the reproducibility of the freeze itself, which is a machine-checkable byte-identity claim, and
(iii) the two-route numerical demonstration underlying the one live axiom this gate pays for (BR-Arch).
Each of these is walked through below with enough detail to redo independently.

 1. Numerical checks: model vs. measured, with honest pulls

 1.1 There is no SM-observable pull at the SG-1 level, and this is stated rather than hidden. 
The physical-observable identifier on file for this gate is OBS-0002 = \(\hbar\) — the granularity residue,
i.e. the finite bit-cost per tuned quantity that the whole economy argument runs on — and its layer-2
class is AUDIT , meaning it is a configuration object (it fixes what a "unit of structural cost" means)
rather than a measured prediction to be checked against data. No independent Standard-Model falsifier is
wired to SG-1; the falsifier for the deeper claim (whether nature's theory-selection economy behaves
additively at all) lives one level up, at the unresolved seam R5 discussed in §3 below. A reader
expecting a \(\chi^2\) table at this point will not find one, and should not: manufacturing one would be
target-anchoring exactly the sin the gate's own guardrails forbid.

 1.2 The quantities that are numerically checked are the frozen geometric constants themselves ,
checked not against experiment but against independent closed-form recomputation — the correct notion of
"prediction vs. measurement" for a piece of declared, frozen mathematics. Four such checks are load-bearing:

 The Einstein-metric count on \(K_6 = SU(3)/T^2\) . The Wang–Ziller/Nomizu Ricci formula at general
 chamber \(\vec u = (u_1,u_2,u_3) \in [1/2,3/2]^3\) ,
 $ \(\mathrm{Ric}_k(\vec u) = \frac{(u_k-u_i+u_j)(u_k+u_i-u_j)}{2R_6^2\,u_iu_ju_k}\quad(i,j,k)\text{ cyclic},\) $
 has exactly 4 solutions to \(\mathrm{Ric}_1=\mathrm{Ric}_2=\mathrm{Ric}_3\) (equivalently four
 \(\vec u\) -classes making the space Einstein): the fully symmetric normal metric \((1,1,1)\) and the
 Kähler–Einstein metric \((1,1,2)\) together with its three permutations. This is not a number invented
 for the framework — it reproduces a classical fact about the flag manifold \(SU(3)/T^2\) known independently in
 the mathematics literature, so getting "4" here is an external cross-check the frozen object passes,
 not a free parameter tuned to get 4.

 The curvature invariants at the symmetric center \(\vec u=(1,1,1)\) , Killing-form normalization 
 ( \(g = (-B)|_{\mathfrak m}\) , \(B(X,Y)=6\,\mathrm{Tr}(XY)\) on \(\mathfrak{su}(3)\) ): \(\mathrm{Ric}_i = 5/12\) ,
 \(\mathrm{Scal}=5/2\) , \(\mathrm{Scal}^2=25/4\) , \(\lVert\mathrm{Ric}\rVert^2=25/24\) ,
 \(\lVert\mathrm{Riem}\rVert^2=23/12\) , ratio \(\lVert\mathrm{Riem}\rVert^2/\mathrm{Scal}^2=23/75\) ,
 \(\lVert\mathrm{Ric}\rVert^2/\mathrm{Scal}^2=1/6\) (the " \(\kappa=1/6\) " figure used elsewhere in the corpus),
 and \(\mathrm{Scal}/\mathrm{Ric}_i=6=\dim K_6\) — this last ratio checked and found identical in both
 the physical \(R_6\) -normalization and the dimensionless Killing normalization , exactly as required
 because curvature ratios are metric-scale invariant. Reproducing this identity in both
 normalizations independently is itself a consistency check: an error in either normalization's
 bookkeeping would show up as a mismatch here, and none is found.

 The cubic (weight-6) invariants : \(K_1=R_{ab}{}^{cd}R_{cd}{}^{ef}R_{ef}{}^{ab}=-113/72\) ,
 \(K_2=R_{abcd}R_{aecf}R_{ebfd}=-5/72\) , and \(\lVert\nabla\mathrm{Riem}\rVert^2=1/4\neq0\) . The last of
 these is a checked structural fact, not a free number: it certifies that \(K_6\) is homogeneous but
 not locally symmetric, and it is cross-checked by verifying the second Bianchi identity is
 satisfied with zero violations on the computed Nomizu tensor — a nontrivial internal consistency
 requirement that a mis-specified curvature tensor would fail.

 Topological invariants : \(\chi(K_6)=6\) , matching independently the order of the Weyl group
 \(S_3\) ( \(|S_3|=6\) ) for a full \(A_2\) flag manifold — a second, independent cross-check of the same
 underlying root-system data (Weyl order and Euler characteristic are computed by different routes
 and agree); \(\chi(S^2)=2\) ; \(\chi(S^1_Y/\mathbb{Z}_2)=1\) (an interval, contractible).

 1.3 Frozen negative controls (the closest thing this gate has to a falsifiable numeric bet). 
Three specific wrong values are named and checked against, precisely so that a reviewer has something
concrete to try to break:
- \(\lVert\mathrm{Riem}\rVert^2\) must equal \(23/12\) and must never equal \(31/147\) — a value that
 appears in an earlier miscomputation elsewhere in the corpus and is retained here as a tripwire.
- \(\lVert\mathrm{Riem}\rVert^2\) must never equal \(60\) — that is the value for the round unit
 six-sphere \(S^6\) , a manifold with a different isotropy structure (isotropy \(SO(6)\) , not \(T^2\) ); getting
 60 would mean the calculation had silently substituted the wrong homogeneous space. The \(S^6\) heat-kernel
 row ( \(a_2/a_0=5\) , \(a_4/a_0=12\) , \(a_6/a_0=1139/63\) ) is kept in the geometry pack precisely as this
 calibration control — the \(a_4\) formula is checked to reproduce exactly \(12\) on \(S^6\) , confirming the
 heat-kernel machinery is correctly wired before it is trusted on \(K_6\) .
- The Smith normal form of the charge-character matrix for \((\mathbb{Z}_3,\mathbb{Z}_2,\mathbb{Z}_6)\) 
 must return invariant factors \([1,6,6]\) — checked by direct computation on the \(3\times3\) generator
 matrix — certifying \(\mathbb{Z}_6\) as the finest faithful quotient, neither coarser nor finer. A
 different invariant-factor list would mean the declared gauge group
 \(G_{\rm SM}=(SU(3)_c\times SU(2)_L\times U(1)_Y)/\mathbb{Z}_6\) was wrong.

 None of these three controls is a "prediction" in the sense of forecasting an undetermined experimental
number; they are arithmetic identities on a fully specified frozen object , and the check is that
independent recomputation returns the frozen value and not a plausible-looking wrong one. This is the
correct and only honest notion of "numerical check" available to a gate that is, by its own construction,
spectrum-neutral.

 1.4 The economy-ledger numbers, checked at face value (with the correction applied). The MDL
comparison scores the frozen branch's honest bit-cost against a target ledger of \(T\approx25\) reals
(3 gauge couplings at \(M_Z\) , 9 charged-fermion masses, 4 CKM angles+phase, 6 neutrino parameters, 2
electroweak parameters \(v,m_H\) , and the 1 strong-CP bound \(\bar\theta\) — an \(E\) -neutral list, identical
on both sides of the ledger) against the branch's own honest charged cost of \(n_{13}\approx13\) –14 reals
(the 4 headline anchors \(\{M_{\rm Pl},\alpha_i(M_Z),y_t,\lvert V_{us}\rvert\}\) plus the \(\sim\) 9–10
injected reals \(N_d=2.4\times10^{-2}\) , \(N_e=1.02\times10^{-2}\) , \(N_\nu=1\) , the threshold \(\delta\) -triple,
and \(\theta_H^\star\) ). The resulting margin is checked to be \(\approx25/14\approx1.8\times\) , not the
 \(\sim4\times\) figure that circulated earlier in the corpus and has since been retired as a known-wrong
headline. This correction is itself a check: recomputing \(T/n_{13}\) from the disclosed reals, rather
than from the earlier under-counted \(n\approx4\) , is the arithmetic that catches the error, and the
corrected \(\approx1.8\times\) is the number that should be quoted going forward. The dimension-ladder
survey against ten named rival geometries returns 10 LOSE / 1 FAIL(structural) / 0 REFUTED under this
corrected accounting — reported explicitly as a first-pass survey, not a certified exhaustive
classification (see negative controls, §2, and residual R2 in §5).

 2. Internal consistency cross-checks and negative controls

 Seven independent cross-checks converge on the same terminal, each probing a different failure mode:

 Freeze reproducibility (the mechanical leg, fully closed). The branch's content and manifest are
 fixed by a 33-row description covering every frozen object in this pack (radii, volumes, curvature
 rationals, chamber operators, threshold vector, topological data). An independent re-run recomputes
 all 33 rows from the declared closed-form equations without consulting the original comparator 
 (target-blind) and checks the recomputed values against the frozen ones. The check returns exit
 code 0 — every one of the 33 recomputed values matches the frozen value exactly — and the run is
 deterministic across repetitions (rerunning it a second time reproduces the identical 33 values,
 not merely values within tolerance). This is the one leg of SG-1 that is genuinely derived rather
 than selected or axiomatized: it is a self-witness that the declared object is what the gate claims
 it is, and it closes no physics by itself — a passed reproducibility check pins an object, it does not
 validate a law of nature.

 The abelian-isotropy uniqueness check on the color carrier, cross-checked against a live
 counter-example. The claim that \(T^2\) is the unique purely abelian isotropy among \(SU(3)/R\) cosets
 rests on the centralizer computation \(C_{SU(3)}(T^2)=T^2\) (Cartan-only, no extra commuting non-abelian
 piece). The check that this claim is not vacuous is the \(CP^2=SU(3)/U(2)\) counter-example: \(U(2)=
 (SU(2)\times U(1))/\mathbb{Z}_2\) is manifestly non-abelian, and the coset-space-dimensional-reduction
 (CSDR) centralizer rule predicts that this non-abelian isotropy is gauge-active and over-produces 
 an extra \(SU(2)+U(1)\) beyond the color group at the gauge-fixing gate. This was built end-to-end 
 (not merely argued) and independently confirmed to BREAK at that downstream gate exactly as
 predicted — a genuine structural exclusion with a real failure mode observed, not merely asserted.
 \(CP^2\) is retained in the record as a permanently dead branch precisely so this negative control stays
 checkable by a future reader.

 The Hahn-embedding closure of the axiom's candidate class. The claim that BR-Arch is needed 
 (not merely convenient) is checked against the existence of a genuine rival total order on the same
 finite record space: dimension-first lexicographic ordering, under which a 4-dimensional EFT beats the
 13-dimensional branch outright ( \(4<13\) ) regardless of anchor cost. Hahn's embedding theorem is used as
 an independent classification check: it says Archimedean orders are exactly the rank-1 orders
 (embeddable in a single copy of \(\mathbb{R}\) ), and every non-Archimedean order is a Hahn sum of rank
 \(\geq2\) , with the lexicographic order the canonical rank-2 witness. Because {Archimedean,
 non-Archimedean} is an exhaustive and closed dichotomy, this check certifies that there is no third
 alternative hiding outside the two cases already analyzed — the classification is complete, not a
 partial survey.

 Two independent numerical routes confirming the lexicographic order is non-Archimedean , run and
 cross-checked against each other:
 - Route 1 (exhaustive finite-truncation axiom check, truncation size \(N=12\) ): all nine scaffold
 axioms for an ordered cancellative commutative monoid are verified True on
 \(\mathbb{Z}_{\geq0}\times_{\rm lex}\mathbb{Z}_{\geq0}\) by direct enumeration; the Archimedean witness
 (existence of finite \(n\) with \(n\cdot x>y\) for arbitrary positive \(x,y\) ) is checked and found to
 fail for \(n\) up to \(100{,}000\) , and the failure is shown to persist for all \(n\) by the structure
 of the lexicographic order itself (any first-coordinate deficit can never be closed by increasing
 the second coordinate, however large \(n\) is).
 - Route 2 (frozen-embedding least-squares fit): a control case, plain \(\mathbb{N}\) (genuinely
 Archimedean), is fit for an additive real embedding and the fit converges to 0.0000 violation at
 every truncation size tested (10 through 200) — confirming the fitting procedure correctly reports
 "no violation" when none exists. The test case, \(\mathbb{Z}_{\geq0}\times_{\rm lex}\mathbb{Z}_{\geq0}\) ,
 is fit with the same procedure and the violation fraction grows monotonically with truncation
 size: \(0.0000\to0.0426\to0.0581\to0.0662\) . A growing (not shrinking, not plateauing) violation
 fraction as the truncation is refined is the correct operational signature of "no additive embedding
 exists at all," which is exactly what Route 1's exact argument predicts — the two routes, one exact
 and combinatorial, one numerical and asymptotic, independently agree .
 - The re-run of both routes is checked to reproduce the original computation's numeric output exactly:
 identical violation-fraction values at every truncation, identical fitted parameters. This closes
 the computation-debt on this seam.

 The group-vs-monoid side-check (a potential objection, checked and resolved). A natural worry is
 that the cost ledger might secretly need to be a group (with inverses) for the Hölder/Alimov
 Archimedean-representability theorem to apply, and that a mere monoid (the actual structure here — a
 positive cone of costs, no inverses, since a bit-cost cannot be negative) might evade the theorem. This
 was checked explicitly: Hölder's and Alimov's theorems cover the cancellative commutative ordered
 monoid case identically to the group case (additive real-representability if and only if
 Archimedean), so the group-vs-monoid distinction is not the deciding axis; Archimedean-vs-not is.
 This check corrects an earlier looser framing without changing the +1 verdict it supports.

 The four-screen Layer-2 toolbox, run to saturation as a negative control on "did some other
 mechanism secretly force the axiom for free." Each of the four generic screens is checked against
 the two competing orders (Archimedean and lexicographic) and none is found to force a decision:
 Invariance is blind (both orders are equally invariant under within-primitive relabeling);
 Record Interface passes both (both are finite, computable, auditable total preorders); Causal
 Order is blind/pass (a cross-theory aggregation rule is not itself a signalling object, which
 additionally confirms target-blindness of the whole construction); Nonseparability passes the
 lexicographic order rather than opposing it (lex is in fact the maximally separable order, so the
 nonseparability screen cannot be used to exclude it). The toolbox is therefore checked to be
 saturated 7/7 (three roots plus four screens) with zero forcing hits — a genuine negative
 result, reported as such, that is precisely what licenses calling BR-Arch a paid axiom rather than a
 derived theorem.

 Ledger convergence across independently maintained records. Three separately maintained corpus
 documents — one on the order-theory analysis, one on the descent/attack-sequence result, and one on
 the floor/credit registry — are checked against each other and found to agree that BR-Arch is graded
 #4 REDUCED-TO-AXIOM (+1 floor) with zero floor-delta between them. This is convergence from
 independent bookkeeping, not a single self-report being counted three times; per the Nonseparability
 accounting rule, BR-Arch is nonetheless counted once as a single shared floor item (tagged
 SAG-SELECTOR-M) across every gate whose economy claim depends on it, precisely so this convergence
 cannot be double-banked into a bigger number than it is.

 Two additional negative controls worth stating explicitly because they bound what SG-1 does not 
claim. First, the family index \(\chi(K_6,E)=-3\) is deliberately not re-derived or re-checked inside
SG-1's own evidence base — it is recorded as a given-E import from a separate gate (SG-3), and treating
it as an SG-1 output would be a layer-smuggling error the ×/⊕/⊗ discipline is built to catch. Second, the
"4 inputs \(\to\) 22 outputs" framing is deliberately not used as supporting evidence anywhere in this
section; it is a disclosed-corrected overclaim (residual R3), and the honest cost used throughout is the
 \(n_{13}\approx13\) –14-real figure.

 3. Reproducing the result from scratch: the exact procedure

 A reader with no access to this corpus and only the equations printed in this dossier can rebuild the
committed object and re-run every check above. The procedure has five stages.

 Stage 1 — write down the arena and count dimensions. Declare
 \(\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2\) with \(K_6=SU(3)/T^2\) the full flag manifold
of \(A_2=\mathfrak{su}(3)\) and \(S^1_Y/\mathbb{Z}_2\) the orbifold of a parent hypercharge circle under
 \(\theta\mapsto-\theta\) . Count: \(\dim\mathcal M_4=4\) , \(\dim K_6=6\) , \(\dim S^2=2\) , and the orbifolded circle
contributes \(1\) (it is a 1-manifold quotiented to an interval, not a dimension-reducing operation) —
total \(D=4+6+2+1=13\) , an exact integer count requiring no numerical input. Separately declare the
non-metric layers: the rulebook \(F^+_{\rm finite}\oplus C_{\rm admiss}\) and the actor bundles
 \(E_{\rm matter}\oplus E_{\rm gauge}\oplus E_{\rm Higgs}\oplus E_{\rm proton}\) ; verify by inspection that
neither contributes to \(D\) (both are 0-dimensional data on top of the 13-dimensional stage) — this is the
check that guards against silently dropping the \(\oplus/\otimes\) content and reading a "13D" claim as
metric-dimension-only, which the brief flags as an incomplete-object error.

 Stage 2 — recompute the \(K_6\) curvature from the root system, by hand. Fix the Cartan basis
 \((h_1,h_2,h_3)\) with \(h_1+h_2+h_3=0\) and the simple roots \(\alpha_1=(1,-1,0)\) , \(\alpha_2=(0,1,-1)\) ,
 \(\alpha_1+\alpha_2=(1,0,-1)\) ; the half-sum of positive roots is \(\rho=(1,0,-1)\) with
 \(\lVert\rho\rVert^2=2\) in the Killing normalization, and the Weyl group is \(S_3\) of order 6. Build the
tangent decomposition \(T(K_6)=\mathfrak m_1\oplus\mathfrak m_2\oplus\mathfrak m_3\) (each a real 2-plane
carrying one root) with the \((-B)\) -orthonormal basis given in the geometry pack. Substitute the isotropic
scales \(x_1=x_2=x_3=1\) (the symmetric chamber center) into the general Wang–Ziller Ricci formula and the
scalar-curvature formula quoted above; this reproduces \(\mathrm{Ric}_i=5/12\) and \(\mathrm{Scal}=5/2\) by
direct substitution — a five-minute hand calculation, not a black box. From there, \(\mathrm{Scal}^2=25/4\) ,
 \(\lVert\mathrm{Ric}\rVert^2=6\times(5/12)^2=25/24\) (six equal eigenvalues, each squared and summed), and
 \(\lVert\mathrm{Riem}\rVert^2=23/12\) follow from the same Nomizu-tensor bookkeeping (the cubic invariants
 \(K_1=-113/72\) , \(K_2=-5/72\) , and \(\lVert\nabla\mathrm{Riem}\rVert^2=1/4\) require carrying the calculation
one order further, contracting the Riemann tensor with itself twice more, but use no additional
assumptions). Check the answer against the frozen negative controls: it must not equal \(31/147\) or \(60\) .

 Stage 3 — recompute the radii and volumes from the two anchors \(M_U\) and \(M_{\rm Pl}\) . Take
 \(M_U=1.0\times10^{16}\) GeV as the RG-transport closure target (obtained by requiring
 \(\alpha_1^{-1}(M_U)=\alpha_2^{-1}(M_U)=\alpha_3^{-1}(M_U)\) using the one-loop SM beta coefficients
 \(b_1=41/10\) , \(b_2=-19/6\) , \(b_3=-7\) together with the threshold vector
 \((\delta_1,\delta_2,\delta_3)=(+4.8424,-3.1112,-1.7313)\) ; a reader can verify the closure residual is
 \(9.6\times10^{-11}\) , i.e. essentially exact given a competent two-loop \(\overline{\rm MS}\) RG solver).
From \(M_U\) , compute \(R_0=(2\pi M_U)^{-1}=1.591549430918954\times10^{-17}\,\mathrm{GeV}^{-1}\) by direct
division — this is pure arithmetic, no fit. At chamber center, \(R_6=R_2=R_0\) and
 \(R_Y=\tfrac12R_0=7.957747154594768\times10^{-18}\,\mathrm{GeV}^{-1}\) (the factor \(\tfrac12\) is the
 \(\mathbb{Z}_2\) orbifold halving, itself checkable: the active interval \([0,\pi]\) is half the parent
circle's circumference). Compute the volumes: \(V_{K_6,0}=(2\pi)^3/\sqrt3=143.2118575035129\) ,
 \(\mathrm{Vol}(K_6)=V_{K_6,0}R_0^6=2.327554010848277\times10^{-99}\,\mathrm{GeV}^{-6}\) ,
 \(\mathrm{Vol}(S^2)=4\pi R_0^2=3.183098861837907\times10^{-33}\,\mathrm{GeV}^{-2}\) ,
 \(\mathrm{Vol}(S^1_Y/\mathbb{Z}_2)=\pi R_0=5.0\times10^{-17}\,\mathrm{GeV}^{-1}\) exactly (the \(2\pi\) in the
volume integral cancels the \(2\pi\) in the definition of \(R_0\) , leaving exactly \(1/(2M_U)\) — check this
symbolic cancellation directly rather than trusting the decimal). Multiply the three volumes to get
 \(\mathrm{Vol}(X_{\rm active})=3.704417261398702\times10^{-148}\,\mathrm{GeV}^{-9}\) . Finally take the
measured \(M_{\rm Pl}=1.2209\times10^{19}\) GeV (ordinary, not reduced) and solve
 \(M_*^{11}=M_{\rm Pl}^2/\mathrm{Vol}(X_{\rm active})\) for
 \(M_*^{11}=4.023836152402511\times10^{185}\,\mathrm{GeV}^{11}\) , hence
 \(M_*=7.467050992135091\times10^{16}\,\mathrm{GeV}\) — a single algebraic division, checkable on a
calculator once the volume is in hand.

 Stage 4 — recompute the discrete/topological data. Verify \(\chi(K_6)=6\) by counting Weyl chambers
( \(|S_3|=6\) for the full \(A_2\) flag manifold) and cross-check against the general Gauss–Bonnet-type
relation for homogeneous spaces (both routes must agree, as noted in §1.2 above). Build the
 \(3\times3\) integer matrix encoding the centers \(\mathbb{Z}_3\subset SU(3)_c\) , \(\mathbb{Z}_2\subset
SU(2)_L\) , \(\mathbb{Z}_6\) on \(U(1)_Y\) with generator \(z=(\omega_3,-1,\zeta_6)=(1,1,1)\) , and run the
standard Smith-normal-form algorithm (row/column integer operations) to recover invariant factors
 \([1,6,6]\) — a mechanical linear-algebra exercise with a unique correct answer, not a judgment call.
Recompute \(\sum_fY_f^2=10/3\) per generation directly from the six SM hypercharge assignments
 \(Y(Q_L)=+1/6\) , \(Y(u_R)=+2/3\) , \(Y(d_R)=-1/3\) , \(Y(L_L)=-1/2\) , \(Y(e_R)=-1\) (each squared, summed with
multiplicity, and checked to total \(10/3\) ).

 Stage 5 — re-run the economy comparison and the BR-Arch check. Assemble the target ledger
 \(T\approx25\) (count: 3 + 9 + 4 + 6 + 2 + 1, itemized in §1.4) and the honest branch cost
 \(n_{13}\approx13\) –14 (4 anchors + \(\sim\) 9–10 injected reals, itemized in the same place); divide to get
the margin \(\approx1.8\times\) — this arithmetic is the entire "forcedness" claim at the numerical level
and takes one line. Then, separately, construct the two candidate total orders on the same finite record
space (plain additive/Archimedean cost vs. dimension-first lexicographic cost) exactly as specified in
§2 item 3, and either (a) run the exhaustive finite-truncation axiom check (Route 1) up to some chosen
truncation \(N\) and confirm the Archimedean witness fails, or (b) run a least-squares additive-embedding
fit (Route 2) at increasing truncation sizes and confirm the violation fraction grows rather than
shrinks. Either route independently reproduces the conclusion that no forced common currency exists
between "one dimension label" and "one anchor bit," which is precisely the content of BR-Arch being a
 paid, not free , axiom — completing the reconstruction of the REDUCED-TO-AXIOM / ANCHORED +1 terminal
from nothing but the equations in this dossier.

 What this procedure does not, and cannot, reproduce. It cannot produce a proof that the 13D branch
is the unique or absolutely minimal geometry (residual R1, the uncomputable Kolmogorov-complexity
question — a dissolved universal-negative, not a gap in this procedure). It cannot certify that no other
sub-6-dimensional \(SU(3)\) -homogeneous space with clean abelian isotropy exists beyond \(\{K_6,CP^2\}\) 
(residual R4, the shelf-completeness question). And it cannot discharge the granularity-to-MDL bridge
itself (residual R5) — that is the one open, decisive seam this whole evidence base surrounds without
closing, and the reproduction procedure above is precisely what keeps that seam honestly located rather
than quietly smuggled into an unstated step.

 4. Summary of what the evidence base establishes and what it leaves open

 Every exact rational and every 16-significant-figure quantity used in this gate's central claim has been
walked through above with its derivation reproducible by hand or short computation; three independent
convergent checks (Einstein-metric count, \(S^6\) heat-kernel calibration, Smith-normal-form) confirm the
frozen curvature and topology data against externally known mathematics rather than against a
self-referential comparator; the freeze itself is machine-verified to reproduce byte-identical values
under an independent target-blind re-run; and the one live axiomatic seam (BR-Arch) is supported by two
independently constructed and mutually agreeing numerical demonstrations plus a closed (Hahn)
classification of the candidate order-theoretic alternatives. No experimental pull is reported because
none is structurally available at this spectrum-neutral gate — this is stated plainly rather than
manufactured. The evidence supports exactly the fixed terminal: REDUCED-TO-AXIOM / ANCHORED +1 ,
resting on measured anchors \(\{M_{\rm Pl},\alpha_i(M_Z),y_t,\lvert V_{us}\rvert\}\) plus \(\sim\) 9–10
charged reals, three grammar-forced structural carriers (weak \(S^2\) , hypercharge \(S^1_Y/\mathbb{Z}_2\) ,
color \(K_6\) ), and one named, paid, target-blind axiom — with the absolute-uniqueness question correctly
dissolved as a universal-negative limit on all knowledge, not banked as a solved derivation.

 Open gaps & the specialist closure path

 SG-1 is fixed at REDUCED-TO-AXIOM / ANCHORED +1 . That terminal is reached honestly and is not
in question in what follows. What follows is the residual family — ten named objects, R1 through
R10 — that remain after the terminal is reached, together with the two structural seams (SEAM-1,
SEAM-2) they cluster into, and the concrete, target-blind machinery that would move each one.
None of these residuals can re-open the gate: the terminal is a floor. Closing any of them can only
 strengthen SG-1 — upgrading the verb from "selected" toward "realization-minimal given \(E\) ," or in
one case (R5) resolving the single remaining seam that keeps "selected" from becoming "forced." The
governing discipline throughout is target-blindness: every closure criterion below is stated so that
a negative result (the geometry loses, the axiom is refuted, a competitor survives) is as
reportable and as valuable as a positive one. A hostile referee reading this section should be able
to verify that no criterion was chosen to make the pre-existing 13D branch win.

 Two meta-facts orient the whole family. First, every one of the ten residuals bottoms, when traced
to its root, on exactly one of two kinds of object: an unproven theorem-target with a stated proof
strategy and no artifact yet on disk (SEAM-1: R2, R5), or an uncomputable universal negative whose
proof would require checking every member of an unbounded family (SEAM-2: R4, R7, R8), with the
remaining residuals (R1, R3, R6, R9, R10) being disclosure/labeling items or an already-closed
mechanical leg. Second, the single highest-leverage move in the entire residual family is R5: it is
the one object whose resolution changes the grade , not merely the confidence interval around the
grade, because it is the one axiom the terminal is explicitly reduced to.

 R5 — the decisive seam: does granularity force an additive (Archimedean) cost currency?

 (a) The precise open object. SG-1's economy claim — that the frozen branch \(\mathfrak{B}_{\rm
active}\) is the leanest-bit-cost object over its declared competitor family — requires comparing two
structurally different kinds of cost on one numerical scale: a discrete structural cost (one more
compact dimension, one more coset factor, one more discrete quotient) and a continuous anchor
cost (one more measured real number, charged at \(b=\log_2(1/\Delta_0)\) bits per unit cell of
resolution \(\Delta_0\) ). The object that must exist for "13D wins" to be a theorem rather than an
assertion is a single positive real-valued exchange rate between these two cost kinds — equivalently,
the statement that the ordered cost-ledger is Archimedean : for any two positive cost
contributions \(x,y>0\) there is a finite integer \(n\) with \(n\cdot x>y\) . This statement is named
 BR-Arch in the ledger. It is currently charged as a new, paid, target-blind, value-free axiom 
at floor +1 — REDUCED-TO-AXIOM, not derived from the geometry, not derived from any of the three
complete roots.

 (b) Why it is hard, and the specific traps. The difficulty is structural, not computational: it
is a foundations-of-measurement question (in the sense of Krantz–Luce–Suppes–Tversky), not a
physics calculation, and it sits exactly at the boundary the corpus calls Granularity. The trap that
must be named explicitly is target-anchoring : it would be illegitimate to simply declare 
additive aggregation because the 13D branch needs it to win. The dossier's own audit already shows
why this trap is real and not hypothetical — a second, fully consistent total order exists on the
identical finite record space, dimension-first lexicographic order (rank candidate branches
first by raw dimension count, and only break ties by anchor-bit count). Under lex ordering, any
four-dimensional effective field theory beats the 13-dimensional branch outright , \(4<13\) ,
regardless of how many measured anchors the 4D theory needs to charge — the anchor cost never even
enters the comparison. Nothing intrinsic to the frozen geometry, and nothing in the three complete
roots run at full precision, excludes lex. A second trap is conflating "no counterexample found"
with "proven." The two independent computational checks already performed (Route 1: exhaustive
finite-truncation axiom check on \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) out to \(N=12\) ,
with the Archimedean witness failing for \(n\) up to \(100{,}000\) ; Route 2: frozen-embedding
least-squares fit, where the lex-ordered test space shows an additive-embedding violation fraction
growing \(0.0000\to0.0426\to0.0581\to0.0662\) as truncation size increases from 10 to 200, while the
plain- \(\mathbb{N}\) control stays at exactly \(0.0000\) throughout) are strong numerical evidence for 
non-Archimedean behavior of the lex order, not a proof that BR-Arch is false for the physical 
cost ledger — they establish the necessity argument (a consistent non-Archimedean alternative
exists) but not a resolution of which ledger nature actually uses. A third trap, already flagged and
corrected once in this program, is the group-vs-monoid framing : an earlier pass suspected that
whether the cost ledger is a group (with inverses) versus a monoid (positive cone only) was the
deciding axis. It is not — the cost ledger is a cancellative, commutative, positive, totally ordered
monoid (no inverses; you cannot have negative bit-cost), and Hölder's theorem together with its
monoid extension (Alimov) shows additive real-representability is equivalent to the Archimedean
property in either case. The specialist closing R5 should not re-litigate group-vs-monoid; that
axis is settled and orthogonal to the open question.

 (c) What closes it, target-blind, with success and refutation criteria. The closing object is a
proof or disproof of the following prove-or-refute pair , stated with three sub-claims that must
all be discharged for a positive closure:

 Sub-claim A (granularity \(\Rightarrow\) MDL): does the existence of a finite cost-floor
 \(\Delta_0>0\) over the 13 metric dimensions and 3 layers (the Granularity root, already shown to
 supply a common currency — bits — and a finite cost per single generator) force the stronger
 claim that the total cost functional is additive across generators of different kind (structural
 vs. anchor)? The already-completed Layer-2 audit shows Granularity's verdict is CONSTRAIN , not
 FORCE : it rules out some cost structures (e.g., it forbids a cost function that assigns zero
 bits to an unbounded family of distinct discrete choices) but does not by itself rule out
 lexicographic aggregation, since lex also respects a uniform per-generator \(\Delta_0\) , is
 units-covariant, cancellative, commutative, positive, totally ordered, and monotone — it merely
 fails to be Archimedean. Closing sub-claim A means either exhibiting an additional physical
 principle beyond raw granularity that rules out lex specifically (and showing that principle is
 itself not a second smuggled axiom), or accepting that granularity alone cannot decide it.

 Sub-claim B (anchors dominate dimension under a stated information measure \(\mathcal{I}\) ): is
 there an operational, target-blind information-theoretic reason (e.g., minimum description length
 under a fixed, pre-registered universal Turing machine or a fixed physically motivated encoding of
 "one more compact dimension" versus "one more measured constant") why anchor-bits and
 dimension-bits should be commensurable at all, prior to and independent of which specific
 branches are being compared? This is the sub-claim most likely to yield a genuine theorem, because
 it converts "is the order Archimedean" into "is there a physically motivated common encoding," a
 narrower and more tractable question.

 Sub-claim C (symmetric economy-win with the generator map fully charged): with the
 geometry-to-observables generator map — the full chain from \(\{M_{\rm Pl},\alpha_i(M_Z),y_t,
 |V_{us}|\}\) plus the ~9–10 injected reals through to the ~25-real target ledger \(T\) — charged in
 both directions (i.e., the same fairness standard applied to every rival branch on the ladder,
 not just to the 13D branch), does the ~1.8 \(\times\) margin survive under whichever aggregation rule
 is eventually justified? This guards against the specific failure mode where a resolution of A and
 B is achieved abstractly but silently re-introduces asymmetric bookkeeping when applied to the
 concrete ladder (guards labeled G1/G2/G5 in the program's fairness audit).

 Success criterion: a theorem (not a numerical experiment) establishing that the physical
 cost-ledger relevant to comparing compactification branches is Archimedean — equivalently, that no
 consistent non-Archimedean total order compatible with the Granularity, Shape, and Scale
 constraints already established can be constructed — would discharge BR-Arch and promote the +1
 floor to a genuinely forced (+0, RESOLVED) result. This must be a general argument applicable to
 the whole competitor family, not a case-by-case exclusion of lex alone (excluding one
 non-Archimedean order while leaving the (closed, per Hahn) class of rank- \(\ge2\) orders otherwise
 unaddressed would not be a proof).

 What a refuting result looks like, and why it must be reported with equal confidence: a
 demonstration that the physical cost ledger is not Archimedean — for instance, an explicit
 physical argument that "one more propagating dimension" carries an ontological cost that no finite
 number of "one more measured constant" charges can ever outweigh (which is exactly the
 lexicographic intuition, stated as physics rather than as bookkeeping) — would flip SG-1's economy
 verdict: a clean four-dimensional effective theory would then win the comparison outright, family
 by family, regardless of how many free parameters it charges. This outcome, labeled
 REFUTED-ECONOMY in the ledger, is explicitly flagged in this program as the "equally-valuable
 honest twin" of a positive closure — it must be written up with the same rigor and prominence as a
 positive result, not suppressed or minimized, because target-blindness cuts both ways: a program
 that only publishes the branch of the prove-or-refute pair that favors its own geometry is not
 target-blind.

 (d) The machinery to start from, described not cited. The natural starting point is the
mathematical theory of ordered algebraic structures and additive representability : Hölder's
theorem (1901) for Archimedean ordered groups, and its extension by Alimov and later authors to
positive cancellative ordered monoids, both state precisely "Archimedean \(\iff\) order-embeds in
 \((\mathbb{R},+,<)\) ." The complementary tool is Hahn's embedding theorem (1907), which
classifies every totally ordered abelian group (or monoid) as a "Hahn sum" — a generalized power
series with exponents in a totally ordered index set (the "rank") and coefficients in
 \(\mathbb{R}\) -valued Archimedean pieces; rank 1 recovers the ordinary reals, and lexicographic order
on \(\mathbb{Z}\times\mathbb{Z}\) is the canonical minimal example of rank 2. Because {Archimedean,
non-Archimedean} is an exhaustive and closed dichotomy under Hahn (there is no third alternative;
every non-Archimedean order is some higher-rank Hahn sum, and lex is merely the simplest witness),
the closing theorem does not need to defeat every conceivable non-Archimedean order one at a time —
it needs to either (i) prove a rank bound (show the physical cost ledger's Hahn rank is forced to be
exactly 1 by some independent physical principle), or (ii) find the physical principle of sub-claim B
that renders the rank question moot by fixing a specific , pre-registered encoding under which the
comparison is well-posed regardless of rank. A second body of machinery worth invoking is
 information-theoretic minimum description length (MDL) in the Rissanen sense, particularly the
two-part code / universal-code framework, which is the natural home for making sub-claim B precise:
one would need to fix a reference universal Turing machine (or an equivalence class of them up to an
additive constant, per the invariance theorem) and show that the additive two-part code length
 is the right cost functional for comparing physical theories — which is itself a substantive,
non-trivial claim in the philosophy-of-science / algorithmic-information-theory literature (related
to, but distinct from, Solomonoff induction's use of MDL as a prior over hypotheses rather than as a
comparison rule between two already-specified theories). A specialist attacking R5 should treat this
as primarily an ordered-algebra and foundations-of-measurement problem with an MDL gloss, not a
physics calculation — the physical content (which branch has which raw bit-cost) is already computed
and frozen; what is missing is the aggregation theorem.

 (e) Leverage — what else closes if this closes. R5 is explicitly the single highest-value
target in the entire SG-1 residual family, and its leverage extends well beyond SG-1: BR-Arch is
recorded as a shared object (SAG-SELECTOR-M) across SG-1, the deeproot-granularity family of
gates (the gap-02 cluster), and any other gate in the program whose forcedness or economy argument
relies on an additive-MDL cost comparison across heterogeneous cost kinds. Discharging it once
therefore discharges it everywhere it is used — per the Nonseparability screen, it is counted once,
not once per gate, and a proof would retire the +1 floor across the whole shared-axiom ledger
simultaneously, not just here. Within SG-1 specifically, closing R5 positively would be the single
step that converts the terminal from REDUCED-TO-AXIOM (+1) to a forced RESOLVED (+0) result — the
only such step in the entire residual family — because R5 is the only residual whose object is
literally the axiom the current terminal is reduced to. R2 (below) is downstream of R5 in the sense
that a full dimension-ladder lower-bound theorem presupposes a settled aggregation rule; closing R5
first makes R2 tractable rather than the reverse.

 R2 — the dimension-ladder lower-bound matrix (Cert 3/4): first-pass survey, not a certified classification

 (a) The precise open object. The claim "13D wins by \(\sim1.8\times\) " currently rests on a
 first-pass ladder survey : ten named rival branches at other dimensionalities (4 through 12, plus
non-standard/non-dimensional constructions) were scored against the 13D branch under the (still
axiom-dependent) additive cost rule, returning 10 LOSE / 1 FAIL(structural) / 0 REFUTED . This is
explicitly flagged in the program's own ledger as CATEGORY_RELATIVE , a survey and not a certified
exhaustive classification. The open object is the general theorem that would upgrade this from a
finite, hand-picked sample to a proof that covers the whole family: for every admissible branch \(B\) 
compatible with the declared role-mechanism grammar \(\mathcal{G}\) (three-role: color-carrier,
weak-carrier, hypercharge-carrier, each realized by some geometric factor), the cost of \(B\) is
bounded below by a function of its role-mechanism decomposition, with the 13D branch's honest charged
cost \(n_{13}\approx13\) –14 measured reals sitting at or near that lower bound.

 (b) Why it is hard, and the traps. This is a combinatorial/classification problem layered on top
of an unresolved foundational one (R5): until the aggregation rule is fixed, "lower bound" does not
even have a well-defined numerical meaning across dimension kinds, so R2's proof is gated behind
R2's statement being well-posed, which is gated behind R5. The trap is attempting to brute-force
the ladder wider (adding more rival branches by hand) without first proving the normal-form /
exhaustion theorem (labeled O1) that every model \(B\) satisfying the grammar's axioms reduces,
cost-non-increasingly, to some role-mechanism tuple drawn from a finite or at least well-ordered
menu of geometric primitives — without that reduction theorem, no finite ladder survey, however
large, can be honestly reported as exhaustive, because an adversary can always propose one more
exotic branch outside the sampled family. A second trap is silently changing the cost-counting
convention between rungs of the ladder (e.g., charging one rival branch's Yukawa sector at the
finite-chamber \(F^+\) level while charging another's at the full continuum level) — this is exactly
the asymmetric-bookkeeping failure mode guarded against by fairness guards G1/G2/G5 in sub-claim C of
R5 above, and it applies with equal force here.

 (c) What closes it, target-blind, with success/refutation criteria. Closure requires proving
theorem O1 — every \(B\models\mathcal{G}\) (every model of the declared grammar's axioms: three
independent gauge-carrying factors realizing an \(SU(3)\times SU(2)\times U(1)\) -compatible isometry
algebra, plus a finite Yukawa chamber) reduces cost-non-increasingly to a role-mechanism tuple — and
then computing, not surveying, the cost lower bound for each equivalence class of tuples, comparing
each honestly to \(n_{13}\approx13\) –14. Success criterion: the theorem holds and the 13D branch's
tuple achieves the lower bound (or comes within a stated, explained gap of it) across the entire 
class, not merely the ten sampled rungs. What a refuting result looks like: discovery of an
admissible role-mechanism tuple, satisfying the identical grammar \(\mathcal{G}\) , whose honestly
charged cost is lower than \(n_{13}\) — this would not merely dent the margin, it would overturn the
"13D wins" headline entirely for that branch of the family, and per the target-blindness discipline
such a discovery must be reported as a LOSES or REFUTED verdict against the 13D branch, not
quietly excluded from the ladder on ad hoc grounds.

 (d) The machinery to start from. The natural framework is coset-space dimensional reduction
(CSDR) classification theory, which already provides a systematic (if not fully enumerated) map
from isotropy subgroups \(R\subset G\) of candidate cosets \(G/R\) to the surviving 4D gauge algebra via
the centralizer rule \(\mathfrak{g}_{\rm surviving}=\mathfrak{g}\ominus\mathfrak{r}\) -orbit content —
this is exactly the rule already used once, decisively, to kill the \(CP^2=SU(3)/U(2)\) rival (its
isotropy \(U(2)=(SU(2)\times U(1))/\mathbb{Z}_2\) is non-abelian, so the centralizer rule predicts it
over-produces gauge content, and building it end-to-end confirmed the break at Gate 2). The
exhaustion theorem O1 should be attempted as a classification result over compact homogeneous spaces
 \(G/R\) with \(\mathrm{rank}(G)\le\) some small bound, using the standard Lie-theoretic classification of
maximal-rank and non-maximal-rank subgroups (Borel–de Siebenthal theory for the former). The
lower-bound computation itself is a direct extension of the MDL codebook already defined in Cert 2
(structural choices cost \(O(1)\) bits, tuned reals cost \(b=\log_2(1/\Delta_0)\) bits) — mechanically
straightforward once O1 fixes what "all admissible tuples" means and R5 fixes how to add the two
kinds of bits.

 (e) Leverage. A positive closure of R2 would retire the "first-pass survey" caveat entirely and
convert the \(\sim1.8\times\) margin into a proven, not merely surveyed, global minimum over the whole
grammar-compatible family — the natural complement to R5: R5 fixes how to compare costs, R2 proves
the comparison has been made exhaustively . Together they would jointly discharge SEAM-1.

 R4 — search-category completeness on the color shelf (the SU(3)-carrier "no other coset" question)

 (a) The precise open object. Color is currently DERIVED-ON-SHELF: among cosets \(SU(3)/R\) , the
maximal torus \(T^2\) is proven to be the unique isotropy subgroup with purely abelian centralizer
( \(C_{SU(3)}(T^2)=T^2\) , Cartan-only), so \(K_6=SU(3)/T^2\) is the unique clean carrier within that named
shelf , and the leading rival \(CP^2=SU(3)/U(2)\) is built end-to-end and shown to break. What remains
open, named N.4 in the ledger, is full-shelf completeness : has every sub-six-real-dimensional,
 \(SU(3)\) -homogeneous space with a clean (centralizer-surviving-abelian) isotropy been enumerated, or
could there be some other coset \(SU(3)/R'\) , or some non-coset \(SU(3)\) -homogeneous or even
non-homogeneous carrier, with a clean isotropy that has simply not yet been checked?

 (b) Why it is hard, and the traps. This is a universal-negative claim — "no other carrier
exists" — over a family that is large but not obviously exhausted by the two cases considered so far
( \(T^2\) and \(U(2)\) are the two isotropy subgroups of maximal and next-to-maximal rank inside \(SU(3)\) ,
but \(SU(3)\) has other closed subgroups, e.g. \(SO(3)\) , finite subgroups, \(U(1)\times U(1)\) realized
non-maximally, or products embedded reducibly). The specific trap is conflating "the two most
natural candidates were checked" with "the shelf is exhausted." A second trap, sharper here than
elsewhere, is scope creep: R4 is scoped to sub-6-real-dimensional carriers of \(SU(3)\) specifically ,
because that is the dimension budget \(K_6\) occupies in the frozen 13D branch; a specialist must
resist the temptation to solve the unbounded, dimension-agnostic version of the question (which is
R1's uncomputable territory, not R4's bounded one).

 (c) What closes it, target-blind, with success/refutation criteria. The closing move is a
 bounded coset-classification plus CSDR centralizer check : enumerate every closed subgroup
 \(R\subset SU(3)\) with \(\dim(SU(3)/R)\le6\) (a finite, group-theoretically tractable list, since
 \(\dim SU(3)=8\) bounds \(\dim R\ge2\) , and closed subgroups of a rank-2 compact simple Lie group of
dimension \(\ge2\) fall into a short, classifiable list: the maximal torus \(T^2\) , \(U(2)\) and its
conjugates, \(SO(3)\) , and a small number of others up to conjugacy), and for each, apply the same
centralizer rule already used to kill \(CP^2\) : does \(C_{SU(3)}(R)/R\) contain a non-abelian factor? If
so, the CSDR rule predicts spurious gauge over-production, matching the \(CP^2\) precedent.
 Success criterion: every subgroup on the finite list other than \(T^2\) either fails the
clean-isotropy test (over-produces gauge content, as \(U(2)\) does) or is excluded on an independent,
equally rigorous physical ground (e.g., wrong isometry rank to carry \(SU(3)_c\times\) nothing else),
leaving \(T^2\) the unique survivor and closing N.4 completely. What a refuting result looks like: a
subgroup \(R'\ne T^2\) is found with clean abelian centralizer and dimension \(\le6\) — this would
produce a second candidate color carrier at the same or lower cost, directly undermining the
"unique clean carrier" language in the current forcedness argument (though it would not necessarily
overturn the specific frozen \(K_6\) branch, since \(K_6\) could still be independently selected by the
R5/R2 economy argument even if not uniquely forced on representation-theoretic grounds alone) — this
distinction must be stated explicitly in any closure writeup so a second clean carrier is not
mistaken for a refutation of SG-1's terminal.

 (d) The machinery to start from. Standard compact Lie group subgroup classification : the
closed subgroups of \(SU(3)\) up to conjugacy are known from the classical theory of maximal subgroups
of simple Lie groups (Dynkin's classification of maximal subalgebras, refined for \(SU(3)\) 
specifically in the physics GUT-model-building literature, where the coset spaces \(SU(3)/T^2\) ,
 \(SU(3)/U(2)\) , and \(SU(3)/SO(3)\) are the three classically studied homogeneous spaces of low
codimension). The centralizer computation for each is a direct Lie-algebra exercise: compute
 \(C_{\mathfrak{su}(3)}(\mathfrak r)\) for each candidate subalgebra \(\mathfrak r\) and check whether the
quotient contains a non-abelian summand — mechanically the same computation already performed for
 \(T^2\) (abelian, clean) and \(U(2)\) (non-abelian, dirty).

 (e) Leverage. Closing R4 would upgrade the color-carrier forcedness claim from "DERIVED-ON-SHELF
(rep-theory theorem on a named, but not certified-complete, shelf)" to a genuinely complete
representation-theoretic uniqueness result, removing one of the two AXIOM-OPEN items filed under
SEAM-2 and tightening the non-claim in section 1.3 of the brief (" \(K_6\) is the unique SU(3) carrier"
— currently qualified as unique only among clean carriers). It does not, by itself, touch R5/R2 or
change the +1 floor, since R4 concerns representation-theoretic forcedness of one factor, not the
cross-kind cost-aggregation axiom the terminal is reduced to.

 R7 — functional three-role necessity and realization-minimality (B2 / C1–C10)

 (a) The precise open object. The B2 result establishes that the three-role grammar (a
color-role, a weak-role, a hypercharge-role, each needing its own geometric carrier) is
 category-relative : within the declared architecture, no proper subset of the three \(\times/\oplus/
\otimes\) layers closes the scoped gates — a null-space result , proven. What remains open is
whether the three-role decomposition itself is a necessary feature of any admissible
architecture whatsoever, not merely of the one declared here — i.e., is \(k_{\rm role}\ge3\) a
theorem about gauge theories with the observed group \(G_{\rm SM}\) , or a feature of this particular
grammar's bookkeeping? A companion, harder question is realization-minimality : even granting
three roles are necessary, is the specific realization chosen for each role (one coset factor per
role, rather than e.g. a single higher-dimensional factor realizing two roles simultaneously via a
more exotic isometry group) itself forced?

 (b) Why it is hard, and the traps. The necessity direction ( \(k_{\rm role}\ge3\) ) is a claim about
 every admissible architecture, which makes it another bounded-but-nontrivial universal-negative:
one must show that no admissible architecture can realize two of the three gauge factors from a
single geometric role without violating some other already-certified constraint (e.g., without
re-merging color and weak into a shared non-abelian isometry group, which would violate the observed
independence of \(\alpha_2\) and \(\alpha_3\) running, or without losing the clean chirality-filter
mechanism that currently lives specifically on the odd-dimensional orbifold factor). The trap is
 "defeat F6 per role" — F6 is the program's name for the general hazard of silently smuggling
governance/bookkeeping choices into what looks like a physics constraint (the same hazard flagged
separately as R10); a specialist must show role-necessity using only physics content (measured gauge
independence, chirality data) and not by definitional fiat within the grammar itself. Realization-
minimality is harder still and is explicitly flagged as only "sharper-OPEN," not attackable with
currently available tools — it is closer in character to R1's uncomputability than to R4's bounded
classification.

 (c) What closes it, target-blind, with success/refutation criteria. The tractable first step is
to upgrade B2 to a functional three-role necessity theorem : for any admissible architecture
consistent with the measured independence of the three SM gauge couplings' running (i.e., three
functionally distinct beta-function sectors) and the measured absence of mirror fermions (LEP
 \(Z\) -width), show that no single geometric factor can supply two of {color, weak, hypercharge} without
contradicting one of those two measured facts. Success criterion: a general argument, not
restricted to the specific \(K_6\times S^2\times S^1_Y/\mathbb{Z}_2\) split, that three independent
roles are needed for any geometric realization of \(G_{\rm SM}\) with three independently running
couplings. What a refuting result looks like: an explicit construction of a two-role (or
one-role) architecture that reproduces \(G_{\rm SM}\) and the observed chirality structure without a
third independent geometric factor — this would show the current three-factor split, while
sufficient, is not necessary, demoting "forced within the grammar" language for the overall
three-role structure (though the individual F1/F2/abelian-isotropy arguments for why each named
role needs the specific carrier it has would be unaffected).

 (d) The machinery to start from. This sits at the interface of CSDR gauge-symmetry-breaking
classification and renormalization-group scheme independence arguments : the tool is to treat
"functionally independent running" as the operational definition of "independent role," and ask
whether the one-loop beta-function structure ( \(b_1^{\rm SM}=41/10\) , \(b_2^{\rm SM}=-19/6\) ,
 \(b_3^{\rm SM}=-7\) , already fixed by SM matter content and not free) forces three independent
UV-completions of the three couplings under any grammar consistent with the RG data — this is a
question in the mathematics of Lie algebra decompositions applied to the specific representation
content of the SM fermions, not a new physics input.

 (e) Leverage. Because R7 concerns the necessity of the number of roles (three) rather than the
specific realization of any one role, closing it positively would strengthen the overall SG-1
architecture-neutrality claim broadly — it is the piece of SEAM-2 most directly connected to the
brief's honest non-claim that "13D is the minimal/forced geometry" is only selector-minimal and
category-relative; closing R7's necessity half would narrow that gap without yet closing it (the
realization-minimality half stays sharper-OPEN regardless).

 R8 — actor-layer minimality and the un-forced chiral content \(E\) 

 (a) The precise open object. The discrete \(\mathbb{Z}_6\) charge-quantization kernel is computed
 given the chiral content \(E\) : the hypercharge assignment obeys \(q\equiv3z_2-2z_3\pmod6\) once \(E\) is
fixed, and this computation is solid. What is open is twofold: (i) whether a no-competitor
universal negative at the level of the \(\otimes\) -Actors layer holds — i.e., is there any other
finite operator/endomorphism structure (a different choice of \(E_{\rm matter}\oplus E_{\rm
gauge}\oplus E_{\rm Higgs}\oplus E_{\rm proton}\) , or a wholly different actor-layer architecture such
as a noncommutative-geometry finite Dirac operator à la Connes–Chamseddine spectral-triple model
building) that reproduces the same observed spectrum at lower actor-layer cost \(k_{\rm actor}\) ; and
(ii) the standing, explicitly disclosed fact that \(E\) itself — the chiral matter content — is not
derived anywhere in this gate or in the program at large; it is an input, consumed, not produced.

 (b) Why it is hard, and the traps. This residual was explicitly demoted on skeptic-verify in
an earlier pass of this program — an initial claim of "AXIOM_CLOSED" for the actor layer was found to
be premature, because the no-competitor claim had not actually been checked against the most
serious rival architecture (NCG finite spectral triples, which are a mature, independently developed
framework for deriving exactly this kind of finite discrete data from an operator-algebraic actor
layer, with its own extensive literature on classifying finite geometries compatible with the SM).
The trap, already fallen into once, is banking a universal-negative closure without an explicit
competitor scoring pass — the corpus is explicit that this must not be repeated: "Do NOT bank
AXIOM_CLOSED — campaign already corrected that."

 (c) What closes it, target-blind, with success/refutation criteria. The closing move is to
 score the actor-competitor matrix target-blind , with the NCG finite Dirac operator as the leading
named rival: build the NCG finite spectral triple that reproduces the same \(G_{\rm SM}\) representation
content and chirality structure, count its actor-layer cost \(k_{\rm actor}^{\rm NCG}\) under the
identical MDL codebook used for the frozen branch, and compare honestly to the frozen branch's
 \(k_{\rm actor}\) . Success criterion: the frozen branch's actor cost is shown to be at or below
every scored competitor's, for a stated, non-exhaustive-but-representative competitor set, with the
scoring methodology identical on both sides (same fairness guards as R5/R2). What a refuting
result looks like: the NCG (or another) competitor reproduces the identical physical content at
strictly lower actor-layer cost — this would not break the frozen geometry's Stage-layer forcedness
arguments (F1, F2, abelian-isotropy) but would show the Actors layer specifically is not minimal,
narrowing the economy margin or possibly flipping it if the actor-layer saving is large enough to
outweigh the Stage-layer advantage.

 (d) The machinery to start from. The primary machinery is the Connes–Chamseddine spectral
action / finite-geometry classification program : a finite spectral triple \((\mathcal{A}_F,
\mathcal{H}_F,D_F)\) consists of a finite-dimensional \(*\) -algebra, a Hilbert space carrying a
representation of it, and a finite Dirac operator \(D_F\) encoding Yukawa-type couplings; the existing
NCG literature already contains explicit classifications of which finite algebras reproduce
 \(G_{\rm SM}\) (the algebra \(\mathbb{C}\oplus\mathbb{H}\oplus M_3(\mathbb{C})\) being the best-known
example), and the cost of that classification (dimension of \(\mathcal{A}_F\) , rank of \(D_F\) , number of
free Yukawa-type parameters in \(D_F\) ) is directly commensurable with the MDL codebook already defined
for the frozen branch's \(F^+\) chamber. A specialist should treat this as primarily a literature-
comparison-and-recount exercise — the NCG classification already exists; what is missing is
charging it under the same cost accounting used here.

 (e) Leverage. Closing R8 positively (frozen branch wins or ties on actor-layer cost) would remove
the last AXIOM-OPEN item under SEAM-2 tied to a genuine, mature, named rival framework, materially
strengthening the "no serious competitor beats this on economy grounds" component of the overall
economy argument that R5/R2 are trying to make rigorous at the aggregation-rule level. It does not
touch the un-forced status of \(E\) itself, which is disclosed as a standing, permanent, unclosable
scope boundary (R9-adjacent), not a gap awaiting a specialist.

 R1, R3, R6, R9, R10 — disclosure, labeling, and already-discharged items

 These five residuals require no new physics and are listed for completeness because they remain in
the ledger, but none of them is a live target for a specialist in the sense R2/R4/R5/R7/R8 are.

 R1 (absolute irreducibility, Kolmogorov complexity \(K(T)\) ) is a dissolved unicorn , not an
open research question: demanding a proof that the 13D branch has the shortest possible description
among every logically conceivable encoding is an appeal to an uncomputable universal negative
(Kolmogorov complexity is provably non-computable in general), and no amount of specialist effort
closes an uncomputable quantity. The correct and complete disposition, already executed, is to state
minimality as grammar-relative ( \(\mathcal{G}\) -relative) and explicitly living/extensible —
i.e., to declare the grammar \(\mathcal{G}\) openly (as this dossier and the brief do) and to
acknowledge that absolute, grammar-independent minimality is a limit on all knowledge, not a defect
of this program. Any future specialist work here should extend the grammar explicitly (adding named
axioms as R4/R7 close) rather than chase the unicorn directly.

 R3 (the "~9–10 injected reals beyond the 4 anchors") is a disclosed-and-corrected labeling
item, not a physics gap: the historical "4 inputs \(\to\) 22 outputs" framing overstated the true input
count, and the honest accounting — 4 measured anchors \(\{M_{\rm Pl},\alpha_i(M_Z),y_t,|V_{us}|\}\) plus
roughly 9–10 further injected reals ( \(N_d=2.4\times10^{-2}\) , \(N_e=1.02\times10^{-2}\) , \(N_\nu=1\) from
the sector normalizations, the threshold \(\delta\) -triple, and the Hosotani phase \(\theta_H^\star\) ),
for a total honest charged cost of ~13–14 reals , not 4 — is already the number used consistently
throughout this dossier. The remaining task is purely one of institutional hygiene: an owner
countersign to propagate the corrected "~22 outputs from ~13–14 effective inputs" phrasing everywhere
the old "4 \(\to\) 22" language might still appear, retiring both the ~4 \(\times\) overclaim and any
over-corrected ~1.6 \(\times\) undershoot in the other direction. No specialist physics work is owed.

 R6 (the reproducer machine-witness) is already CLOSED, fully DERIVED , and is listed only for
ledger completeness: the branch's content-addressed identity — the full 33-row geometry-description
manifest covering the metric factors, the orbifold quotient, and the finite chamber — was
independently regenerated in a target-blind re-run (the verification script's own comparator
bypassed) with exit code 0 and every recomputed digest byte-equal to the frozen original,
deterministic across reruns. This is the one leg of SG-1 that is genuinely, unconditionally derived
— a content-addressed self-witness — and it is done. It lowers no assumption floor (a reproducibility
check proves fidelity to a declaration, not necessity of the declaration) but nothing further is
owed here.

 R9 ( \(\mathcal{M}_4\) as a declared axiom) is a disclosed, permanent, non-defective scope
boundary: observed 3+1 spacetime is consumed as an observational primitive , not derived, and this
is stated as such throughout (subsumed under the label AXIOM SG1- \(\alpha\) ). The only owed action is a
 labeling sweep — ensuring \(\mathcal{M}_4\) is never inadvertently counted as a forced rung in a
future summary, and never cited as though it were a derivation. This is editorial discipline, not
physics.

 R10 (Rulebook physics-vs-governance separation, the F6 hazard) is a methodology guard with zero
forcedness weight : it flags the general risk that the \(\oplus\) -Rulebook layer's admissibility
constraints could, if stated carelessly, smuggle in governance/bookkeeping choices (e.g., how the
freeze-and-reproduce process itself is administered) as though they were physical constraints. The
closing action is to restate the admissibility chamber \(\mathcal{C}_{\rm phys}\) in
architecture-neutral language and charge only the genuine physics burden of the Rulebook layer, never
the freeze/gate-status governance overhead. This residual is explicitly tied to a parallel weakest-
link item in the flavor-sector gate (SG-8's R1, concerning the \(F^+\) chamber) and should be closed in
coordination with that gate, but again carries no forcedness weight of its own — it cannot move
SG-1's grade in either direction.

 The two seams, restated, and the honest priority order

 Collapsing the ten residuals onto their structural roots gives exactly two live seams. SEAM-1 
(granularity \(\Rightarrow\) MDL: R2 and R5) is an unproven theorem-target with a stated, tractable
proof strategy (ordered-algebra representability theory plus a bounded coset-classification
exercise) and no discharging artifact yet on disk. SEAM-2 (universal negatives: R4, R7, R8) is a
family of bounded-but-nontrivial completeness questions, each requiring an explicit, target-blind
competitor-scoring pass rather than a proof from first principles, and each already correctly
resisting the temptation to bank a premature "AXIOM_CLOSED" verdict. R1 is a genuine dissolved
unicorn and should never be re-opened as a target. R3, R6, R9, R10 are disclosure/labeling items,
with R6 already fully discharged.

 The honest priority order for a specialist team allocating effort is: attack R5 first — it is the
only residual that changes the grade itself, its proof strategy is the most mathematically concrete
of the family (ordered-algebra representability plus Hahn-rank arguments, both established
machinery), and a resolution in either direction (BR-Arch proven, or REFUTED-ECONOMY established)
retires the shared +1 floor across every other gate in the program that leans on the same additive-
MDL comparison. R2 is the natural second target , since its own statement is not fully well-posed
until R5 fixes the aggregation rule, but its classification half (the CSDR-based exhaustion theorem
O1) can be developed in parallel. R4, R7, and R8 form a coherent third package — all three are
bounded competitor-search exercises using the same CSDR-centralizer or NCG-classification machinery,
and a single specialist team fluent in coset-space gauge-symmetry-breaking classification and
noncommutative-geometry model building could reasonably attack all three together, since they share
both method and the same "score the named competitor, don't assume the negative" discipline. None of
R1, R3, R6, R9, or R10 requires further specialist physics effort.

 Honest ceiling, scope & the endpoint

 This closing section does one job: it draws the boundary of SG-1 with enough precision that no reader — friendly or hostile — can mistake what has been shown for something larger, and then it states, in the fixed and load-bearing terminal language of this program, exactly where the gate's reasoning bottoms out. The fixed grade is REDUCED-TO-AXIOM / ANCHORED +1 . Nothing below moves that grade in either direction; the purpose of an honest-ceiling section is to make the grade legible, not to renegotiate it. PROMOTIONS:0 throughout.

 10.1 What SG-1 does not claim — dissolved ≠ solved, selection ≠ derivation, given-E ≠ derivation of E

 The single most important discipline in this gate is refusing to let a dissolution masquerade as a solution , a selection masquerade as a derivation , or a result obtained given the observed spectrum \(E\) masquerade as an explanation of \(E\) . SG-1 dissolves one genuine unicorn (§10.1.1 below) and it forces three structural sub-claims within a declared grammar (§10.2), but none of that earns the sentence "the geometry is derived." The material below restates each non-claim on its own line, deliberately, because each is a place a careless summary could quietly inflate the result — and this dossier's job is the opposite of inflation.

 "13D is the minimal/forced geometry" is not claimed. What is shown is selector-minimal, category-relative forcedness: the 13D branch
$$
\mathfrak{B} {\rm active}=\underbrace{[\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2]} {\times\ \text{Stage, }D=13}\ \oplus\ \underbrace{[F^+ {\rm finite}\oplus C {\rm admiss}]} {\oplus\ \text{Rulebook}}\ \otimes\ \underbrace{[E {\rm matter}\oplus E_{\rm gauge}\oplus E_{\rm Higgs}\oplus E_{\rm proton}]}_{\otimes\ \text{Actors}}
$$
wins an explicit bit-cost economy comparison against a named, declared family of rival branches — a first-pass ladder that returns 10 LOSE / 1 FAIL(structural) / 0 REFUTED , with the honest margin corrected to ≈1.8× at transparent cost (≈13–14· \(b\) for the 13D branch against ≈25· \(b\) for the standard effective-field-theory target ledger of \(T\approx25\) reals), not the retired ≈4× headline. That is a real result inside a stated category. It is not, and is never represented here as, an architecture-neutral proof that no admissible geometry outside the surveyed family could do better. The ladder is explicitly flagged in its own record as a first-pass survey , not a certified exhaustive classification (residual R2, disposition OPEN/GAP with bounded computation-debt).

 "The geometry is derived" is not claimed. SG-1 is, by its own charter, a freeze-and-reproduce certificate , not a derivation. The branch content commitment, its manifest, and the orbifold-freeze record were independently regenerated in a target-blind re-run: the verifier's own comparator was bypassed, all 33 geometry-description rows were recomputed from first principles, exit code 0, byte-equal, deterministic across reruns. That is residual R6, and it is the one leg of this entire gate that is genuinely, fully DERIVED — but what it derives is that the object under test is the object claimed , not that the object is the only possible object . A reproducibility check is an audit anchor, not a physics validation. Declaring a geometry once and then proving you have not silently re-tuned it afterward is a necessary discipline for everything downstream to mean anything; it is not itself a derivation of the geometry from more primitive assumptions, and this dossier never lets R6's genuine "DERIVED" status bleed into the gate's headline claim.

 "4 inputs → 22 outputs" is not claimed as historically stated. That headline overstates the honest cost. The correct accounting, used throughout this dossier, is 4 anchors \(\{M_{\rm Pl},\,\alpha_i(M_Z),\,y_t,\,|V_{us}|\}\) + ~9–10 injected reals ≈ 13–14 reals total — the injected reals being the sector normalizations \(N_d=2.400000000000000\times10^{-2}\) , \(N_e=1.020000000000000\times10^{-2}\) , \(N_\nu=1\) (fixed downstream at the flavor gate), the threshold triple \((\delta_1,\delta_2,\delta_3)=(+4.8424,\,-3.1112,\,-1.7313)\pm1.6\times10^{-3}\) (fixed at the threshold/unification gate), and the Hosotani phase \(\theta_H^\star\) (fixed at the electroweak-symmetry-breaking gate) — charged against a target ledger of \(T\approx25\) reals: 3 gauge couplings at \(M_Z\) , 9 charged-fermion masses, 4 CKM angles+phase, 6 neutrino parameters, 2 electroweak parameters ( \(v,m_H\) ), and 1 strong-CP bound ( \(\bar\theta\) ). This dossier uses the corrected ~13–14-real accounting everywhere and explicitly retires the ~4× overstatement (residual R3, disposition DISCLOSED-CORRECTED ) rather than allowing it to recirculate silently.

 " \(K_6\) is the unique \(SU(3)\) carrier" is not claimed in full generality. What is shown is that \(K_6=SU(3)/T^2\) is the unique clean, purely abelian-isotropy carrier: among cosets \(SU(3)/R\) , the maximal torus \(T^2\) is the only isotropy subgroup with abelian centralizer, \(C_{SU(3)}(T^2)=T^2\) (Cartan only), so it alone injects no spurious non-abelian gauge factor into the low-energy algebra. This is representation theory, not conjecture, and it comes with a genuine, built-and-broken negative control: the rival coset \(CP^2=SU(3)/U(2)\) has non-abelian isotropy \(U(2)=(SU(2)\times U(1))/\mathbb{Z}_2\subset SU(3)_c\) , was constructed end-to-end as a complete rival branch, and demonstrably BREAKS at the very next gate by over-producing gauge content beyond \(SU(3)\times SU(2)\times U(1)\) — a real structural exclusion, not an assumption dressed up as one. What is not shown is full-shelf completeness: whether some other sub-6-dimensional \(SU(3)\) -homogeneous space, not yet enumerated on the shelf of candidates, might also carry a clean abelian isotropy and so tie with \(K_6\) . That is residual R4 (the "N.4" full-shelf-completeness question), and it stays honestly OPEN rather than being swept in as "obviously none exist."

 "The gauge-group outcome discriminates the framework from its competitors" is not claimed. Recovering \(G_{\rm SM}=(SU(3)_c\times SU(2)_L\times U(1)_Y)/\mathbb{Z}_6\) from a compact internal geometry is a filter that string theory, M-theory, F-theory, noncommutative geometry, and lattice constructions all also pass — the community-wide "landscape" / "which coset" / "which spectral triple" degeneracy is shared by every framework that tries this move, not a defect unique to this one. This is stated plainly as a TIE across frameworks, not dressed up as a discriminating victory.

 Two further non-claims sharpen the same boundary from a different angle, and both matter because they are the two most tempting places to over-claim. First, SG-1 does not derive the chiral matter content \(E\) of the Standard Model — three generations of quarks and leptons in the representations they occupy — nor does it derive \(\mathcal{M}_4=\mathbb{R}^{3,1}\) itself; both are declared, observational inputs , carried forward honestly as primitives rather than smuggled in as outputs of the geometric construction (residual R9: \(\mathcal{M}_4\) is subsumed as a labeled observational-primitive axiom, never counted as a forced rung). Second, the family index \(\chi(K_6,E)=-3\) , which fixes the count of left-handed generations at three, is a given-E result computed downstream once the chiral content \(E\) is already assumed; SG-1 only records it as part of the frozen ledger and takes no credit for producing it. And \(CP^2\) stays a dead branch throughout — built out completely, broken completely, never quietly reopened anywhere in this document as a live possibility.

 10.1.1 The one genuine dissolution — and why it is not a gap

 The corpus contains exactly one legitimate unicorn at this gate: the demand for absolute, architecture-neutral irreducibility — the Kolmogorov-complexity question "is this the single shortest possible description of a universe with this Standard Model, evaluated against every logically conceivable grammar, not merely the one declared here?" (residual R1, the \(K(T)\) question). This is not merely a hard open problem; it is uncomputable in principle . Kolmogorov complexity admits no general algorithm for any object, and "every logically conceivable grammar" is not a well-defined, enumerable search space to begin with — there is no procedure, even in principle, that could certify a description as absolutely shortest across all possible formal systems for describing physics. Demanding that SG-1 answer this question is demanding a universal negative that no physical theory, in this framework or any competing one, could ever supply; it is a limit on all knowledge, not a hole peculiar to this program's homework.

 The correct treatment — and the one this dossier follows without softening it into an open technical debt — is to dissolve the demand rather than leave it dangling: state the declared grammar \(\mathcal{G}\) explicitly (the role-mechanism family: an internal geometry that sources exactly three propagating gauge factors plus a zero-dimensional finite flavor chamber, with rival branches compared by an additive bit-cost ledger), and assert minimality as \(\mathcal{G}\) -relative and extensible — living, not final, open to a sharper or larger grammar being declared later, but never claimed absolute. This is the single dissolution in the gate, and it is worth being explicit that "dissolved" and "solved" are different words used differently on purpose in this dossier: R1 is dissolved — the demand itself is retired as malformed, because no finite construction could ever discharge it — while R5 (the Archimedean seam, discussed next) is emphatically not dissolved. R5 is a live, well-posed, in-principle-decidable mathematical question that simply has not yet been decided either way, and treating it as a second unicorn would be a category error in the opposite, dishonest direction — quietly discarding an open, answerable question by mislabeling it uncomputable.

 10.2 Selection ≠ derivation, made precise at the level of each forced piece

 Three of the geometry's four gauge-bearing carriers are forced within the declared grammar, given the observed spectrum \(E\) — and it matters exactly what "forced-within-grammar" buys and does not buy, because this is the precise boundary between the gate's genuine strength and its remaining seam.

 Weak \(S^2\) (Fact F1, DERIVED-WITHIN-GRAMMAR). No abelian or torus carrier, of any dimension whatsoever, has a non-abelian \(SU(2)\) among its isometries. This is a hand-checkable general theorem about isometry groups of compact homogeneous spaces, and it closes the cheaper direction over entire shelves of competitor geometries in one stroke: any branch that tries to source weak \(SU(2)_L\) from a torus factor — however many extra dimensions it spends doing so — fails outright, independent of dimension count. Binding qualifier: this is a torus source specifically, \(S^2\) supplies \(SU(2)_L\) via its isometry algebra \(\mathfrak{su}(2)\) , and this is explicitly not any \(SU(2)\subset SU(3)\) sitting inside \(K_6\) — the two carriers are structurally distinct and the routing is not interchangeable. It is forced given the requirement "source \(SU(2)_L\) from the isometry group of a compact internal factor" — a requirement that is itself part of the declared grammar, not a deeper theorem about what nature must do.

 Hypercharge \(S^1_Y/\mathbb{Z}_2\) (Fact F2, DERIVED-WITHIN-GRAMMAR). A closed odd-dimensional factor retains both spacetime handednesses, producing mirror fermions that are already excluded by the measured LEP \(Z\) -width. The \(\mathbb{Z}_2\) orbifold action \(\theta\mapsto-\theta\) (two isolated fixed points at \(\theta=0,\pi\) ) projects the mirror states out and simultaneously sources \(U(1)_Y\) ; the Atiyah–Patodi–Singer index computed on the active interval \([0,\pi]\) returns \(n_L=+3\) , \(n_R=0\) exactly, with per-field \(\mathbb{Z}_2\) parities assigning zero modes to \(Q_L,L_L\) at \((+,+)\) and to \(u_R,d_R,e_R,\nu\) at \((-,-)\) , and forbidding all mirror zero modes. This argument is forced given the declared grammar's requirement that mirror fermions be excluded — itself a measured-physics input (the LEP electroweak precision bound), not a geometric necessity that would hold in a universe without that measurement.

 Color \(K_6=SU(3)/T^2\) (abelian-isotropy uniqueness, DERIVED-ON-SHELF). Among cosets \(SU(3)/R\) , the maximal torus \(T^2\) is the unique isotropy subgroup with abelian centralizer; this is forced given the declared shelf of \(SU(3)\) -homogeneous coset candidates, under the coset-space-dimensional-reduction centralizer rule, with the explicit, built-and-broken negative control \(CP^2\) demonstrating that the rule has teeth rather than being asserted by fiat. The shelf itself — residual R4/N.4, are there other sub-6-dimensional candidates not yet enumerated? — is not certified complete.

 Each of these three results is real, each is independently checkable against the stated grammar, and each earns the word DERIVED with the honest qualifier "within-grammar" permanently attached — it is never dropped later in this dossier or elsewhere in the corpus. But "forced within a declared grammar, given the observed spectrum" is precisely selection , in the exact technical sense this program uses that word: the grammar itself — three propagating internal gauge-bearing factors sourcing \(SU(3)\times SU(2)\times U(1)\) , plus a zero-dimensional flavor chamber, compared to rivals via an additive bit-cost ledger — is a declared starting point, chosen and frozen, not a theorem proved from something more primitive than itself. Given- \(E\) is not derivation-of- \(E\) : none of these three arguments explains why the Standard Model has the chiral content it has; each explains, given that content, which internal factor must carry which gauge group. And selection is not derivation in the fuller sense that matters for the gate's grade: proving that \(K_6\) beats its named rivals under a declared cost metric is a different and weaker claim than proving that no other cost metric, no other grammar, and no other family of competitor geometries could ever yield a different winner. Both distinctions are load-bearing, and neither is relaxed anywhere in this dossier — including in the summary sentences below, where it would be easiest to let them slip.

 10.3 The anchors paid — the honest ledger, itemized

 SG-1 is deliberately spectrum-neutral : the observed content \(E\) sits on both sides of the economy comparison — both the 13D branch and every rival in the ladder must reproduce the same target physics — and so \(E\) cancels in the "who wins" verdict. The gate consumes no measured number as a numeric input to its central economy claim. What it does charge , honestly and completely, is the following.

 The four headline anchors \(\{M_{\rm Pl},\ \alpha_i(M_Z),\ y_t,\ |V_{us}|\}\) [MEASURED] are declared free inputs to the wider program this frozen geometry supports downstream: \(M_{\rm Pl}=1.220900000000000\times10^{19}\) GeV (ordinary, not reduced, Planck mass), the three gauge couplings \(\alpha_i(M_Z)\) at \(M_Z=91.1876\) GeV, the top Yukawa \(y_t\) , and the Cabibbo magnitude \(|V_{us}|\) . Why these four specific numerical values obtain — as opposed to why the slots for four such anchors exist — is a stated scope boundary , a question about the values of fundamental constants that lies outside this gate's remit, not a physics gap SG-1 is answerable for.

 The ~9–10 injected reals beyond the four anchors , bringing the honest total charged cost to ~13–14 reals , not 4: the sector normalizations \(N_d=2.400000000000000\times10^{-2}\) (fixes \(m_b\) at \(M_Z\) ), \(N_e=1.020000000000000\times10^{-2}\) (fixes \(m_\tau\) at \(M_Z\) ), and \(N_\nu\) (structural, fixing neutrino-sector magnitudes together with a Berry phase of \(2\pi/3\) at diagonalization); the threshold triple \((\delta_1,\delta_2,\delta_3)=(+4.8424,\,-3.1112,\,-1.7313)\pm1.6\times10^{-3}\) ; and the Hosotani vacuum-angle \(\theta_H^\star\) that fixes the electroweak vacuum expectation value \(v_{\rm pred}=246.02\pm3.5\) GeV and Higgs mass \(m_h=123.82\pm1.8\) GeV. These are not hidden costs discovered after the fact — they are the explicit subject of residual R3, disposition DISCLOSED-CORRECTED , with an instruction on file that the "~13–14 effective inputs" phrasing propagate everywhere the older, overstated "4→22" language previously appeared, and that both the retired ~4× overclaim and any over-correction below the honest ~1.8× be retired together.

 The freeze/reproducibility cost. The frozen branch carries a content commitment, a manifest record (33 rows), and an orbifold-freeze record that together pin the object under test; regenerating and verifying these against an independent, target-blind re-run is an audit cost paid once, not a recurring physics cost, and not itself a source of any of the 13–14 charged reals.

 The one named axiom — BR-Arch — at floor +1. This is the actual price of the gate's terminal, and it is restated here in full because it is the entire content of "ANCHORED +1." The claim "the 13D branch is the economy winner" requires comparing two structurally different cost axes — a discrete dimension/structure count and a continuous measured-anchor count — on one common numerical scale. All three complete geometric roots were run against this seam at full precision, and none of them supplies the needed aggregation rule for free:

 Shape (the full three-layer \(\times\) Stage/ \(\oplus\) Rulebook/ \(\otimes\) Actors object) delivers object- identity — which records are admissible — a domain restriction, not a scoring rule; it cannot express a cross-layer magnitude inequality between "one dimension label" and "one anchor bit." Adopting an additive-over-lexicographic comparison at this level would itself be target-anchoring. Verdict: PASS / no purchase on the question.

 Scale (the full \(M_{\rm Pl}\) -anchored discipline, screens SCL-A through SCL-J) exposes that dimension-cost and anchor-cost are two competing cost axes, without supplying their exchange rate. Verdict: EXPOSE .

 Granularity (the full cost-floor over all 13 dimensions × 3 layers, finite cell size \(\Delta_0>0\) , bit-cost \(b=\log_2(1/\Delta_0)\) per tuned quantity) supplies a common currency — bits — and a finite cost per single generator, but not a finite cross-kind exchange ratio. The explicit witness \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) (lexicographic order on pairs of non-negative integers: rank first by dimension count, break ties by anchor-bit count) satisfies every other structural axiom in the toolbox — uniform per-generator \(\Delta_0\) , units-covariance, cancellativity, commutativity, positivity, total order, monotonicity — while remaining non-Archimedean . Verdict: CONSTRAIN .

 Because a second, logically consistent total order genuinely exists on the identical finite record space — dimension-first lexicographic order, under which a lean 4D effective theory beats the 13D branch outright ( \(4<13\) ) regardless of anchor cost, and under which Shape-minimality folds for every competitor simultaneously — nothing in the frozen geometry excludes that order. Selecting the Archimedean branch over the lexicographic one is therefore a genuine, paid, non-vacuous new axiom:

 BR-Arch (Archimedean common-currency axiom): the ordered cost-ledger is Archimedean — for any two positive cost contributions \(x,y>0\) there exists a finite integer \(n\) with \(n\cdot x>y\) ; equivalently, a finite positive exchange ratio exists between a structural bit and an anchor bit, and no cost axis is infinitely preferred over the other.

 This axiom is labeled, precisely, NEW, target-blind, value-free, paid (+1 floor) — never free. Its necessity is demonstrated: the lexicographic countermodel is real and satisfies every other axiom in the toolbox, so BR-Arch is not a vacuous formality but does genuine work. Its candidate class is closed by Hahn's embedding theorem: an ordered structure is Archimedean if and only if it is rank-1 (embeds in a single copy of \(\mathbb{R}\) ); every non-Archimedean order is a Hahn sum of rank \(\ge2\) , of which the lexicographic order is the canonical rank-2 witness; the dichotomy {Archimedean, non-Archimedean} is jointly exhaustive, so there is no third possibility hiding in the wings that could rescue or replace BR-Arch. Its target-blindness is independently confirmed: the Causal-Order screen treats the aggregation rule as a cross-theory ranking convention rather than a signal object internal to any one theory, so it could not have been reverse-engineered to produce a preferred answer — corroborating a broader Layer-2 screen saturation of Invariance: BLIND, Record-Interface: PASS, Causal-Order: PASS/BLIND, Nonseparability: PASS, alongside the three roots' Shape: PASS, Scale: EXPOSE, Granularity: CONSTRAIN. All seven screens were run to saturation and none of the seven forces the Archimedean branch.

 This is corroborated computationally by two independent routes that agree byte-for-byte on rerun. Route 1, an exhaustive finite-truncation axiom check: all nine scaffold axioms hold for the lexicographic monoid up to truncation \(N=12\) , and the Archimedean witness fails to find a bounding integer \(n\) through \(n=100{,}000\) (and provably for all \(n\) ). Route 2, a frozen-embedding least-squares fit: the genuinely Archimedean control case (plain \(\mathbb{N}\) ) converges to violation \(0.0000\) at every truncation from 10 to 200, while the lexicographic test case's violation fraction grows monotonically with truncation depth — \(0.0000\to0.0426\to0.0581\to0.0662\) — the operational numerical signature that no additive embedding exists. A group-versus-monoid side question was checked and resolved along the way: the cost-ledger is a cancellative, commutative, ordered monoid (a positive cone with no inverses), not a group, but Hölder/Alimov's theorem covers the monoid case identically to the group case — additive real-representability is equivalent to the Archimedean property regardless of which — so this is not a second deciding axis, only a clarification that leaves the +1 verdict unchanged.

 BR-Arch is graded MAP_ADMISSIBLE_SUPPORTED , not FORCED — the forcing certificate shows an empty anchor-transfer chain (BR-Arch does not transfer to or derive from \(\Delta_0\) , \(\hbar\) , or \(M_{\rm Pl}\) ), demonstrated necessity, and a closed candidate class, which together cap the ceiling at exactly REDUCED-TO-AXIOM / ANCHORED +1 and never at ROOT-FORCED. BR-Arch is also a shared object : it is the identical seam behind SG-1's economy claim and behind every other gate in this program whose closure argument relies on additive-MDL cost comparison across cost kinds (the granularity-family gates in particular). Per the Nonseparability screen, it is charged once , as a single shared +1 floor item, not duplicated as an independent axiom at every gate that happens to touch it.

 10.4 What is genuinely still owed (named plainly, not rolled into a hedge)

 Two structurally different kinds of open item remain, and this dossier keeps them apart rather than blending them into one undifferentiated "still open" gray zone, because they close by different means and neither is required to sustain the present terminal.

 SEAM-1 — the decisive, in-principle-decidable seam: R5, granularity ⇒ MDL / BR-Arch itself. This is the single object that could flip the gate. It is a live theorem-target — prove that granularity genuinely forces an additive MDL comparison (sub-claim A: granularity ⇒ MDL), that anchors dominate dimension count under the natural information measure (sub-claim B), and that the economy-win survives with the full geometry-to-observables generator map charged, honoring guards G1/G2/G5 (sub-claim C) — or refute it. The honest, equally reportable twin outcome is named explicitly in the corpus and not suppressed: a REFUTED-ECONOMY finding, in which a clean 4D effective theory genuinely wins the bit-cost comparison once BR-Arch is dropped in favor of the lexicographic order, is just as valuable a result as the current one, and this dossier commits to reporting it with equal confidence if it is ever obtained. There is no discharging artifact for R5 on disk today; it is the single highest-value next move on this gate, named, bounded, and falsifiable — not a vague gesture at "more work needed."

 SEAM-2 — universal-negative completeness questions: R4 (color-shelf completeness), R7 (functional three-role necessity / realization-minimality), R8 (actor-layer minimality with \(E\) left un-forced). These ask, in three different guises, "is there no other admissible competitor?" — a family of bounded-but-currently-uncomputed universal negatives, each with a concrete, finite closure path rather than an infinite search. R4 asks whether any sub-6-dimensional \(SU(3)\) -homogeneous space beyond \(\{K_6,\,CP^2\}\) has a clean, centralizer-surviving abelian isotropy; it closes by a bounded coset-classification plus a centralizer check — a finite, nameable calculation, just not yet performed to certification. R7 asks whether the three-role functional necessity (color, weak, and hypercharge each needing their own carrier) can be upgraded from "necessary, not yet shown sufficient" to a proven floor \(k_{\rm role}\ge3\) for any admissible architecture, by defeating each remaining role-collapsing alternative in turn. R8 asks whether some competitor actor-layer construction — noncommutative geometry's finite Dirac operator is named explicitly as the sharpest rival to score — could achieve a lower actor-layer cost than the one this branch uses, given the identical \(E\) ; this item was explicitly demoted on a prior skeptic-verify pass specifically because the no-competitor claim was found undischarged, and this dossier does not re-inflate it back to closed.

 Two smaller, already-resolved bookkeeping items round out the ledger and are recorded here only to confirm they carry zero forcedness weight and never re-open anything. R9 : the observed spacetime \(\mathcal{M}_4=\mathbb{R}^{3,1}\) is a declared observational primitive, subsumed as a labeled axiom — never to be counted as a forced rung or cited anywhere as a derivation. R10 : the boundary between physics content and freeze/gate-status governance inside the Rulebook layer is a methodology guard with zero forcedness weight, restated architecture-neutrally so that only genuine physics burden is ever charged. Neither R9 nor R10 is a physics gap; both are labeling discipline, and both are already satisfied by the way this dossier has been written.

 None of R1 (dissolved), R4, R5, R7, R8, R9, or R10 is required to hold the current terminal in place, and none of them, if left unresolved indefinitely, degrades REDUCED-TO-AXIOM / ANCHORED +1 to something weaker — they are the map of where the gate could in principle be strengthened further (toward a sharper "realization-minimal given \(E\) " result), not a list of debts that undermine what has already been shown. Conversely, resolving R5 in the forcing direction is the only listed item that could ever change the terminal — and even then, only by discharging the one paid axiom in the upgrading direction, never by exposing a hidden flaw that would downgrade what is already certified.

 10.5 The closing endpoint statement

 Every forcedness chain in this gate has now been traced to where it stops. The observed spectrum \(E\) is the floor for the three within-grammar forcing arguments (F1 for weak \(S^2\) , F2 for hypercharge \(S^1_Y/\mathbb{Z}_2\) , the abelian-isotropy uniqueness result for color \(K_6\) ) and is explicitly not re-derived by SG-1; it is consumed as a given, and in the economy comparison itself \(E\) cancels between the 13D branch and every rival, so no measured number is smuggled into the "who wins" verdict. The one remaining forcedness gap — the cross-kind cost-aggregation rule between a structural bit and an anchor bit — bottoms out on a single, named, closed-candidate-class axiom charged against the Granularity root, at floor +1, counted once as a shared object across every gate that relies on the same additive-MDL comparison. The absolute-uniqueness demand that could have made this an infinite regress is dissolved, correctly, as an uncomputable universal negative rather than left dangling as an unfinished derivation. Nothing here is rolled up into a hedge, and nothing here is inflated past what is shown.

 Nothing left. Anchored on: Shape: the frozen 13D carrier \(\mathcal{M}_4\times K_6(=SU(3)/T^2)\times S^2\times S^1_Y/\mathbb{Z}_2\) , all three layers ( \(\times\) Stage/ \(\oplus\) Rulebook/ \(\otimes\) Actors) pinned and machine-reproduced byte-identical under an independent target-blind re-run (33 geometry-description rows recomputed from first principles, exit code 0, deterministic across reruns); Granularity: the load-bearing root here — finite records impose a bit-cost \(b=\log_2(1/\Delta_0)\) per tuned quantity, which is what makes the shape comparison decidable at all, residue \(\hbar\) (the granularity residue, physical-observable id on file, audit-class); Scale: the UV/boundary package \(M_U\sim1.0\times10^{16}\) GeV, \(R_0=1.591549430918954\times10^{-17}\ \mathrm{GeV}^{-1}\) , threshold vector \((\delta_1,\delta_2,\delta_3)=(+4.8424,\,-3.1112,\,-1.7313)\pm1.6\times10^{-3}\) — present but not the deciding root at this gate; Observables: none consumed as numeric input (the gate is spectrum-neutral, \(E\) cancels on both sides of the economy comparison), while referencing without deriving \(E\) and the ~13–14 honestly charged measured reals (the four anchors \(\{M_{\rm Pl},\alpha_i(M_Z),y_t,|V_{us}|\}\) plus \(N_d=2.4\times10^{-2}\) , \(N_e=1.02\times10^{-2}\) , \(N_\nu\) , the threshold \(\delta\) -triple, and the Hosotani angle \(\theta_H^\star\) ), tested against the granularity residue \(\hbar\) with no independent Standard-Model falsifier wired at this gate; Dissolution: the demand for absolute, architecture-neutral, Kolmogorov-style irreducibility ( \(K(T)\) ) is a genuine unicorn — dissolved as a limit on all knowledge rather than a gap in this program's derivation, with minimality correctly reframed as grammar-relative and extensible. The one paid item standing between this result and a from-nothing derivation is BR-Arch, the Archimedean common-currency axiom, REDUCED-TO-AXIOM at floor +1 — named, demonstrated necessary, target-blind, and candidate-class-closed by Hahn's embedding theorem, but not itself proven from anything more primitive. That is the whole and complete honest ceiling of SG-1: REDUCED-TO-AXIOM / ANCHORED +1 , PROMOTIONS:0.

 Closure ledger — SG-1 — geometry / shape selection

 Status (fixed): REDUCED-TO-AXIOM · ANCHORED +1

 The technical closure LEDGER (separate document)

 Gate: SG-1 — geometry / shape selection. Fixed grade (do not change): REDUCED-TO-AXIOM / ANCHORED +1. PROMOTIONS:0. 

 This is the auditor's record: every object, number, and forcing step SG-1 rests on, tagged by kind and graded on the credit ladder. It is written to be checked line by line, not read for narrative. All quantities are quoted at full precision from the frozen geometric arena; nothing here is fitted or back-solved.

 1. Layer-0 wall identity

 SG-1 is the object-commitment wall : the gate that freezes, before any downstream gate runs, the exact three-layer branch every later gate (SG-2…SG-10) is evaluated against. Its wall identity is specificity + no-layer-smuggling + reproducibility — a freeze-and-reproduce certificate, not a derivation. It closes no physics by itself; it pins the object physics is then done on.

 The committed object, all three layers pinned:

 \[
\mathfrak{B}_{\rm active}
=\underbrace{[\mathcal{M}_4\times K_6\times S^2\times S^1_Y/\mathbb{Z}_2]}_{\times\ {\rm STAGE\ (metric,\ 13\ dims)}}
\ \oplus\ \underbrace{[F^+_{\rm finite}\oplus C_{\rm admiss}]}_{\oplus\ {\rm RULEBOOK\ (0\ dims)}}
\ \otimes\ \underbrace{[E_{\rm matter}\oplus E_{\rm gauge}\oplus E_{\rm Higgs}\oplus E_{\rm proton}]}_{\otimes\ {\rm ACTORS\ (0\ dims)}}
\]

 × Stage (metric, carries the only dimension count): \(\mathcal{M}_4=\mathbb{R}^{3,1}\) (4, observational primitive) \(\times\ K_6=SU(3)/T^2\) (6, the full \(A_2\) flag manifold, color source) \(\times\ S^2\) (2, weak source) \(\times\ S^1_Y/\mathbb{Z}_2\) (1→interval, hypercharge + chirality filter). \(D=4+6+2+1=13\) [EXACT] .

 ⊕ Rulebook (0-dim, non-metric, load-bearing): \(F^+_{\rm finite}=\{\tau=\omega,\ \mathcal{G}_{\rm gen},\ \Pi_u,\Pi_d,\Pi_e,\Pi_\nu,\ O_u,O_d,O_e,O_\nu,\ \phi_i,\ N_i,\ \mathcal{N}_i,\ {\rm RG}\}\) (flavor chamber) \(\oplus\ C_{\rm admiss}=\{\) selector v3, C1–C14, freeze-before-compare barrier, anomaly conditions, no-mirror parity table, Wilson-line winding rule, FCNC/mediator no-go \(\}\) (anti-fitting firewall).

 ⊗ Actors (0-dim): \(E_{\rm matter}\oplus E_{\rm gauge}\oplus E_{\rm Higgs}\oplus E_{\rm proton}\) , with \(E_{\rm matter}=S_{3,1}\otimes S^{\rm spin^c}_{K_6}\otimes S^{\rm spin^c}_{S^2}\otimes L_Y\otimes V_{SU(3)}\otimes V_{SU(2)}\otimes V_{F^+}\) .

 A ×-only reading is an incomplete object — the ⊕ and ⊗ layers are frozen and cannot be silently dropped from any gate that cites \(\mathfrak{B}_{\rm active}\) .

 Gauge-routing ledger (which factor sources which force — isometries of the internal metric factors, not embeddings of one group in another):

 Factor 
 dim 
 Force sourced 
 Mechanism 

 \(\mathcal{M}_4=\mathbb{R}^{3,1}\) 
 4 
 — (observed spacetime) 
 observational primitive (R9) 

 \(K_6=SU(3)/T^2\) 
 6 
 \(SU(3)_c\) 
 left-isometry \(\mathfrak{su}(3)\) 

 \(S^2\) 
 2 
 \(SU(2)_L\) 
 isometry \(\mathfrak{su}(2)\) — not any \(SU(2)\subset SU(3)\) 

 \(S^1_Y/\mathbb{Z}_2\) 
 1→interval 
 \(U(1)_Y\) + chirality filter 
 isometry + \(\mathbb{Z}_2\) orbifold (kills mirrors) 

 \(F^+\) 
 0 
 flavor/Yukawa 
 finite-operator chamber 

 Freeze witnesses (content-addressed, the mechanical leg): a content-addressed SHA-256 branch digest; a manifest meta-digest (33 rows); a separate orbifold-freeze digest. R6 is machine-verified by an independent re-run: target-blind exit 0; all 33 geometry-description hashes recomputed without the script's own comparator, both frozen hashes byte-equal, deterministic across reruns. This is a self-witness that lowers no assumption floor and closes no physics — an audit anchor, not a physics validation.

 2. Layer-1 endpoint anchor

 The endpoint SG-1 bottoms on is not a measured number — the gate is spectrum-neutral by construction (the SM content \(E\) appears on both sides of the economy comparison and cancels). The Layer-1 anchor is instead a named axiom :

 BR-Arch (Archimedean common-currency axiom). The ordered cost-ledger is Archimedean: for any two positive cost contributions \(x,y>0\) there exists a finite integer \(n\) with \(n\cdot x>y\) . Equivalently: a finite positive exchange ratio exists between a structural (dimension) bit and an anchor (measured-real) bit — no cost axis is infinitely preferred over the other.

 This is the +1 floor of the REDUCED-TO-AXIOM terminal. It is labeled NEW, target-blind, value-free, and paid (not free) — adopting it is a genuine addition to the theory's assumption base, not a discharge of an existing one.

 Endpoint bottoming, root by root:
- Shape bottoms on the frozen content-addressed active branch digests (object identity).
- Granularity — the load-bearing root for this gate — bottoms on the finite-record bit-cost \(b=\log_2(1/\Delta_0)\) , residue \(\hbar\) (physical-observable id OBS-0002, AUDIT class).
- Scale bottoms on the UV package ( \(M_U\) , \(R_0\) , threshold vector \(\delta\) ) but is not the deciding root here.
- The absolute -uniqueness demand ( \(K(T)\) , Kolmogorov complexity) is a universal-negative unicorn , dissolved as uncomputable-in-principle — a limit on all knowledge, not a gap in this reconstruction — and reframed as grammar-relative forcing.

 3. Layer-2 root stack

 3.1 Tier A — Shape / Scale / Granularity, full precision, all three complete roots

 The gate's one live forcedness seam (R5, the additive-MDL / common-currency bridge) was driven through all three complete roots at full precision on the frozen branch (the content-addressed active branch digests) — none truncated.

 Root: Shape (full 3-layer ×Stage/⊕Rulebook/⊗Actors). Delivers object- identity — which records are admissible — a domain restriction, not a scoring rule. It cannot express a cross-layer magnitude inequality between "one dimension label" and "one anchor bit": doing so would require Shape to carry a metric it does not have. Adopting additive-over-lexicographic cost aggregation at the Shape level would be target-anchoring (importing the desired answer as a structural axiom). Verdict: PASS / no purchase on the Archimedean question. 

 Full-precision Shape data used elsewhere in the ledger but load-bearing for what Shape can fix (object identity, not the economy question):
- \(D=4+6+2+1=13\) [EXACT] .
- \(K_6=SU(3)/T^2\) curvature at the symmetric center \(\vec u=(1,1,1)\) (Killing-norm, exact rationals): \(\mathrm{Ric}_i=5/12\) ; \(\mathrm{Scal}=5/2\) ; \(\mathrm{Scal}^2=25/4\) ; \(|\mathrm{Ric}|^2=25/24\) ; \(|\mathrm{Riem}|^2=23/12\) ; \(|\mathrm{Riem}|^2/\mathrm{Scal}^2=23/75\) ; \(|\mathrm{Ric}|^2/\mathrm{Scal}^2=1/6\) (" \(\kappa=1/6\) "); \(\mathrm{Scal}/\mathrm{Ric}_i=6=\dim K_6\) in both the Killing-norm and R₆-norm conventions. Cubic invariants: \(K_1=-113/72\) , \(K_2=-5/72\) ; \(|\nabla\mathrm{Riem}|^2=1/4\ne0\Rightarrow K_6\) homogeneous but not locally symmetric (0 Bianchi violations). Exactly 4 invariant Einstein metrics on \(SU(3)/T^2\) : normal \((1,1,1)\) plus the 3 Kähler–Einstein \((1,1,2)\) permutations; off-center the space is non-Einstein. Frozen negative controls (never dissolve): \(|\mathrm{Riem}|^2=23/12\) is never \(31/147\) and never \(60\) (the latter is \(S^6\) , a distinct manifold).
- Topology [EXACT/topological]: \(\chi(K_6)=6\) ( \(=|S_3|\) , the Weyl-chamber count, as expected for a full flag manifold); \(\chi(S^2)=2\) ; \(\chi(S^1_Y/\mathbb{Z}_2)=1\) .
- Volumes [EXACT]: \(V_{K_6,0}=(2\pi)^3/\sqrt3=143.2118575035129\) ; \(\mathrm{Vol}(K_6)=V_{K_6,0}R_0^6=2.327554010848277\times10^{-99}\ \mathrm{GeV}^{-6}\) ; \(\mathrm{Vol}(S^2)=4\pi R_0^2=3.183098861837907\times10^{-33}\ \mathrm{GeV}^{-2}\) ; \(\mathrm{Vol}(S^1_Y/\mathbb{Z}_2)=\pi R_0=5.0\times10^{-17}\ \mathrm{GeV}^{-1}\) (exact \(=1/2M_U\) ); \(\mathrm{Vol}(X_{\rm active})=3.704417261398702\times10^{-148}\ \mathrm{GeV}^{-9}\) .
- \(\mathbb{Z}_6\) center [EXACT/CERTIFIED]: \(G_{\rm SM}=(SU(3)_c\times SU(2)_L\times U(1)_Y)/\mathbb{Z}_6\) ; Smith normal form of the charge-character matrix has invariant factors \([1,6,6]\) ⇒ \(\mathbb{Z}_6\) is the finest faithful quotient (no coarser or finer identification admissible); \(\sum_f Y_f^2=10/3\) per generation.

 Root: Scale (full \(M_{\rm Pl}\) -anchored discipline, screens SCL-A..J). Exposes that dimension-cost and anchor-cost are two competing cost axes but does not itself supply the exchange rate between them. Verdict: EXPOSE (a genuine finding — Scale shows the tension exists — but not a resolution).

 Full-precision Scale data:
- \(M_U=1.0\times10^{16}\) GeV [DERIVED] — closure target \(\alpha_i^{-1}(M_U)=\alpha_j^{-1}(M_U)\) ; residual \(9.6\times10^{-11}\) (numerical-pipeline floor).
- \(R_0\equiv(2\pi M_U)^{-1}=1.591549430918954\times10^{-17}\ \mathrm{GeV}^{-1}\) [DERIVED] ; \(R_6=R_2=R_0\) at chamber-center; \(R_Y=7.957747154594768\times10^{-18}\ \mathrm{GeV}^{-1}\) (the \(1/2\) factor is the \(\mathbb{Z}_2\) halving) [DERIVED] .
- \(M_Z=91.1876\) GeV [MEASURED, PDG, \(\pm0.0021\) ] ; \(M_{\rm Pl}=1.2209\times10^{19}\) GeV [MEASURED anchor, ordinary not reduced] .
- Planck normalization [DERIVED, fixed by geometry + \(M_{\rm Pl}\) , not independent]: \(M_{\rm Pl}^2=M_*^{D-2}\,\mathrm{Vol}(X_{\rm active})\) , \(D=13\) ⇒ \(M_*^{11}=M_{\rm Pl}^2/\mathrm{Vol}(X_{\rm active})=4.023836152402511\times10^{185}\ \mathrm{GeV}^{11}\) , \(M_*=7.467050992135091\times10^{16}\) GeV.
- Threshold vector [DERIVED]: \((\delta_1,\delta_2,\delta_3)=(+4.8424,-3.1112,-1.7313)\pm1.6\times10^{-3}\) ; SM one-loop \(b_1=41/10\) , \(b_2=-19/6\) , \(b_3=-7\) .

 Root: Granularity (full cost-floor over 13 dims × 3 layers, \(\Delta_0>0\) , \(b=\log_2(1/\Delta_0)\) ). Supplies a common currency (bits) and a finite cost per single generator; does not supply a finite cross-kind exchange ratio between a structural bit and an anchor bit. Witness: \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) satisfies uniform per-generator \(\Delta_0\) , is units-covariant, cancellative, commutative, positive, and totally ordered, monotone — and is still non-Archimedean . Verdict: CONSTRAIN — this is the root the seam actually lives on; Granularity is the load-bearing root for SG-1's residual forcedness question.

 3.2 Tier B — the four Layer-2 screens (toolbox saturated 7/7, none forces Archimedean)

 Screen 
 Test 
 Verdict 

 Invariance 
 governs within-primitive recodings; both the Archimedean order and the dimension-first lexicographic order are equally invariant under relabeling 
 BLIND 

 Record Interface 
 admits both orders as finite, computable, auditable total preorders 
 PASS 

 Causal Order 
 per-theory signalling constraint; blind to a cross-theory aggregator since a ranking rule is not a signal object — confirms target-blindness 
 BLIND/PASS 

 Nonseparability 
 forbids unpaid factorization; lex is the maximally separable order, hence compatible with (not opposed to) lex 
 PASS 

 Roll-up across all seven tools: Shape=PASS, Scale=EXPOSE, Granularity=CONSTRAIN, Invariance=PASS/BLIND, Record-Interface=PASS, Causal-Order=PASS/BLIND, Nonseparability=PASS. 7/7 saturated, 0/7 FORCE. No root or screen forces the Archimedean choice; BR-Arch must be adopted as an axiom, which is exactly what the REDUCED-TO-AXIOM terminal certifies.

 4. Measured anchors and their role

 SG-1 is unusual among the gates in that it consumes no measured number as a numeric input to its own central claim — the economy comparison is spectrum-neutral, with \(E\) (the SM chiral content) appearing identically on both sides of the ledger and cancelling. The table below distinguishes consumed (used as a numeric input inside the SG-1 derivation), referenced (charged to the cost ledger but not driving the comparison), and tested-against (compared to a prediction — there is none at this gate).

 Anchor / quantity 
 Value 
 Role at SG-1 
 Kind 

 \(M_{\rm Pl}\) 
 \(1.2209\times10^{19}\) GeV 
 referenced (fixes \(M_*\) via Planck normalization; not a numeric input to the economy comparison) 
 MEASURED 

 \(\alpha_i(M_Z)\) (3 couplings) 
 GUT-normalized, PDG values at \(M_Z=91.1876\) GeV 
 referenced (declared free inputs; charged into \(n_{13}\) ) 
 MEASURED 

 \(y_t\) 
 top Yukawa 
 referenced (declared free input; charged into \(n_{13}\) ) 
 MEASURED 

 $ 
 V_{us} 
 $ 
 Cabibbo angle magnitude 

 \(N_d\) 
 \(2.4\times10^{-2}\) 
 injected real, charged cost (SG-8 sector normalization, fixes \(m_b\) at \(M_Z\) ) 
 MEASURED, charged-not-atomic 

 \(N_e\) 
 \(1.02\times10^{-2}\) 
 injected real, charged cost (fixes \(m_\tau\) at \(M_Z\) ) 
 MEASURED, charged-not-atomic 

 \(N_\nu\) 
 \(1\) (structural) 
 injected real, charged cost 
 MEASURED, charged-not-atomic 

 \(\delta\) -triple 
 \((+4.8424,-3.1112,-1.7313)\) 
 injected reals (SG-7), charged cost 
 DERIVED but charged as cost 

 \(\theta_H^\star\) 
 Hosotani vacuum angle 
 injected real (SG-6), charged cost 
 DERIVED but charged as cost 

 Honest total charged cost: ~13–14 measured/injected reals , NOT the headline "4 anchors → 22 outputs." This correction (R3 in the residual ledger) is load-bearing: the four headline anchors \(\{M_{\rm Pl}, \alpha_i(M_Z), y_t, |V_{us}|\}\) are necessary but not sufficient to specify \(\mathfrak{B}_{\rm active}\) 's cost; the sector normalizations and phase data add ~9–10 more reals.

 Tested-against: the only physical-observable id on file for SG-1 is OBS-0002 \(=\hbar\) (the granularity residue), layer-2 class AUDIT (configuration data, not a measured prediction). No SM-number falsifier is wired for SG-1 — the falsifier lives one level up, at whether the eventual theory-selection economy behaves additively at all (i.e., at BR-Arch itself). This absence of a pulls table is a declared scope boundary, not a suppressed result.

 5. The full derivation chain — numbered ledger

 Each step is graded on the credit ladder immediately after its value. Steps 1–7 are the object-commitment chain (Shape); step 8 is the sole DERIVED self-witness (Granularity/mechanical); steps 9–13 are the three-carrier forcing arguments (given-E); steps 14–20 are the economy/MDL chain culminating in the axiom.

 Dimension count. \(D=4(\mathcal{M}_4)+6(K_6)+2(S^2)+1(S^1_Y/\mathbb{Z}_2)=13\) . [EXACT] — CLOSED-NEGATIVE-adjacent identity; grade: REDUCED-TO-AXIOM at the level of "these are the declared factors" (the factor list is a Shape declaration, not derived from a deeper principle).

 \(K_6=SU(3)/T^2\) chosen as color carrier. Abelian-isotropy uniqueness (W9): among cosets \(SU(3)/R\) , the maximal torus \(T^2\) is the unique isotropy subgroup that is purely abelian, \(C_{SU(3)}(T^2)=T^2\) (Cartan only) — it injects no spurious non-abelian gauge factor. Grade: DERIVED-ON-SHELF (rep-theory theorem on the named shelf of \(SU(3)/R\) cosets; full-shelf completeness beyond this named shelf is R4, OPEN).

 \(CP^2=SU(3)/U(2)\) branch-kill. \(U(2)=(SU(2)\times U(1))/\mathbb{Z}_2\) is non-abelian (contains \(SU(2)\subset\) color \(SU(3)\) ); by the CSDR centralizer rule it is gauge-active and over-produces \(SU(2)+U(1)\) at Gate 2. Built end-to-end as a complete rival branch, BREAKS . Grade: CLOSED-NEGATIVE — a real structural exclusion, not a tuned-away possibility.

 Weak \(S^2\) forced within grammar (Fact F1). No abelian/torus carrier of any dimension has a non-abelian \(SU(2)\) among its isometries ⇒ weak needs a genuinely non-abelian-isometry carrier, and \(S^2\) is the minimal one. Hand-checkable general theorem, closes the cheaper direction over whole shelves. Grade: DERIVED-GIVEN-grammar (given the declared role-mechanism grammar; not grammar-independent).

 Hypercharge \(S^1_Y/\mathbb{Z}_2\) forced within grammar (Fact F2). A closed odd-dimensional factor retains both handednesses ⇒ mirror fermions ⇒ excluded by the LEP \(Z\) -width measurement; the \(\mathbb{Z}_2\) orbifold ( \(\theta\mapsto-\theta\) ) projects them out and sources \(U(1)_Y\) . Atiyah–Patodi–Singer index on \([0,\pi]\) returns \(n_L=+3\) , \(n_R=0\) . Grade: DERIVED-GIVEN-grammar .

 Charge quantization. Smith normal form of the charge-character matrix on \(\mathbb{Z}_3\times\mathbb{Z}_2\times\mathbb{Z}_6\) : invariant factors \([1,6,6]\) ⇒ \(\mathbb{Z}_6\) is the finest faithful quotient, \(G_{\rm SM}=(SU(3)_c\times SU(2)_L\times U(1)_Y)/\mathbb{Z}_6\) . Grade: EXACT/CERTIFIED — a clean integer/topological computation, no continuous input.

 Freeze commitment. A content-addressed SHA-256 branch digest, a manifest meta-digest (33 rows), and a separate orbifold-freeze digest. Grade: the object commitment itself — not physics, the precondition for physics.

 Freeze reproducibility (R6). Independent target-blind re-run, all 33 hashes recomputed without the original comparator, byte-equal, deterministic. Grade: DERIVED self-witness — the one leg that is genuinely and completely derived, but it lowers no assumption floor (a hash check validates bookkeeping, not physics).

 \(K_6\) curvature invariants at Einstein center (Killing-norm). \(\mathrm{Ric}_i=5/12\) , \(\mathrm{Scal}=5/2\) , \(|\mathrm{Ric}|^2=25/24\) , \(|\mathrm{Riem}|^2=23/12\) , ratio \(23/75\) , \(|\mathrm{Ric}|^2/\mathrm{Scal}^2=1/6\) , \(K_1=-113/72\) , \(K_2=-5/72\) , \(|\nabla\mathrm{Riem}|^2=1/4\) . Grade: EXACT — closed-form evaluation of the Wang–Ziller/Nomizu formulas at the declared chamber center; not a fit.

 4 invariant Einstein metrics on \(SU(3)/T^2\) . Classic result (normal \((1,1,1)\) + 3 permutations of Kähler–Einstein \((1,1,2)\) ), independently reproduced. Grade: EXACT — a validation of the geometry engine against known mathematics, not a new physics claim.

 Target ledger \(T\approx25\) reals (Cert 1). The \(E\) -neutral comparison target: 3 gauge couplings at \(M_Z\) + 9 charged-fermion masses + 4 CKM angles/phase + 6 neutrino params + 2 electroweak ( \(v,m_H\) ) + 1 strong-CP bound \(=25\) . Grade: CERTIFICATE — an accounting object, not a derived physical quantity.

 Honest branch cost \(n_{13}\approx13\) –14. 4 headline anchors + ~9–10 injected reals (enumerated in §4 above). Grade: DISCLOSED-CORRECTED (R3) — supersedes the retired "4 inputs" headline.

 MDL codebook (Cert 2). A tuned real costs \(b=\log_2(1/\Delta_0)\) bits (large, cell-resolution-dependent); a discrete structural choice (a coset, a \(\mathbb{Z}_6\) , an integer index) costs \(O(1)\) bits (small). Grade: CERTIFICATE — the costing convention, not itself a physics result.

 Boxed inequality. \(\forall D=4..12(+{\rm non\text{-}dim}),\forall j:\ I_{\rm gen,13}+n_{13}b+I_{\rm struct,13} < I_{\rm gen}(B_{D,j})+n_{D,j}b+I_{\rm extra}(B_{D,j})+I_{\rm struct}(B_{D,j})\) — the formal statement the ladder tests. Grade: structural claim , tested next.

 Ladder verdict. 10 LOSES_TO_13D / 1 FAILS(structural) / 0 REFUTED — first-pass survey, explicitly not a certified classification. Grade: DERIVED/survey — a real computation, but honestly bounded in scope (not exhaustive over all conceivable \(D,j\) ).

 Honest margin. ~1.8× at transparent cost (≈13–14 \(\cdot b\) for the 13D branch vs ≈25 \(\cdot b\) for the standard EFT). The public "~4× / 4→22" headline is retired as known-wrong (R3). Grade: DERIVED/survey , corrected.

 The lex countermodel. Dimension-first lexicographic order (rank by # dimensions, then break ties by anchor-bit count) is a second logically consistent total order on the identical finite record space; under lex a clean 4D EFT beats 13D outright ( \(4<13\) ) regardless of anchor cost. Grade: DERIVED (an explicit, checkable countermodel — not hand-waved).

 Hahn embedding closes the candidate class. Archimedean \(\Leftrightarrow\) rank-1 (embeds in a single copy of \(\mathbb{R}\) ); every non-Archimedean order is a Hahn sum of rank \(\ge2\) ; lex is the canonical rank-2 witness. \(\{\) Archimedean, non-Archimedean \(\}\) is exhaustive and closed. Grade: EXACT (classical theorem, correctly applied) — this is what makes BR-Arch's necessity argument airtight rather than merely plausible.

 BR-Arch adoption. Since nothing in the root set (Shape PASS, Scale EXPOSE, Granularity CONSTRAIN) nor any of the four Layer-2 screens (7/7 saturated, 0 FORCE) excludes the lex countermodel, selecting the Archimedean branch is the added axiom — eliminating lex is extensionally identical to asserting the cross-kind commensurability clause. Grade: REDUCED-TO-AXIOM, +1 floor — the terminal grade of the entire gate.

 Two independent numerical/order-theoretic confirmations of non-Archimedean-ness (Route 1: exhaustive finite-truncation axiom check to \(N=12\) , all 9 scaffold axioms True for \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) , Archimedean witness fails to \(n=100000\) and provably for all \(n\) ; Route 2: frozen-embedding least-squares fit, control \(\mathbb{N}\) converges to violation \(0.0000\) at all truncations 10→200, test \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) violation fraction grows \(0.0000\to0.0426\to0.0581\to0.0662\) ). Byte-identical reproduction between original and independent re-run. Grade: DERIVED/verified — this is genuine computation confirming the necessity of BR-Arch (that the axiom is doing real work), not a discharge of the axiom itself.

 6. Credit-ladder grading summary (every leg, one table)

 # 
 Leg 
 Value / result 
 Ladder grade 

 1 
 Dimension count 
 \(D=13\) 
 REDUCED-TO-AXIOM (declared factor list) 

 2 
 \(K_6\) = unique abelian-isotropy carrier 
 \(SU(3)/T^2\) 
 DERIVED-ON-SHELF 

 3 
 \(CP^2\) branch-kill 
 over-produces gauge at Gate 2 
 CLOSED-NEGATIVE 

 4 
 Weak \(S^2\) 
 no abelian carrier has \(SU(2)\) isometry 
 DERIVED-GIVEN-grammar 

 5 
 Hypercharge \(S^1_Y/\mathbb{Z}_2\) 
 mirror-fermion exclusion + orbifold 
 DERIVED-GIVEN-grammar 

 6 
 \(\mathbb{Z}_6\) finestness 
 SNF \([1,6,6]\) 
 EXACT/CERTIFIED 

 7 
 Freeze commitment 
 content-addressed branch + manifest + orbifold-freeze digests 
 object commitment (precondition) 

 8 
 Freeze reproducibility 
 byte-equal, target-blind, exit 0 
 DERIVED (self-witness; floor-neutral) 

 9 
 \(K_6\) curvature invariants 
 $ 
 \mathrm{Riem} 

 10 
 4 Einstein metrics 
 \((1,1,1)\) +3× \((1,1,2)\) 
 EXACT (validation) 

 11 
 Target ledger \(T\) 
 \(\approx25\) reals 
 CERTIFICATE 

 12 
 Honest branch cost 
 \(n_{13}\approx13\) –14 
 DISCLOSED-CORRECTED 

 13 
 MDL codebook 
 \(b=\log_2(1/\Delta_0)\) vs \(O(1)\) 
 CERTIFICATE 

 14–15 
 Boxed inequality + ladder 
 10 LOSE/1 FAIL/0 REFUTED 
 DERIVED/survey (bounded) 

 16 
 Honest margin 
 ~1.8× (not ~4×) 
 DERIVED/survey, corrected 

 17 
 Lex countermodel 
 4D beats 13D under lex 
 DERIVED (explicit countermodel) 

 18 
 Hahn embedding 
 Archimedean \(\Leftrightarrow\) rank-1 
 EXACT (classical theorem) 

 19 
 BR-Arch adoption 
 Archimedean common-currency axiom 
 REDUCED-TO-AXIOM, +1 (gate terminal) 

 20 
 Non-Archimedean confirmation (2 routes) 
 violation fractions, witness failure to \(n=10^5\) 
 DERIVED/verified (necessity, not discharge) 

 R1 
 Absolute irreducibility \(K(T)\) 
 uncomputable universal negative 
 DISSOLVED (unicorn) 

 R9 
 \(\mathcal{M}_4\) 
 observed spacetime 
 DISCLOSED / subsumed as declared axiom (observational primitive) 

 Reading the ladder: legs 2, 4, 5 are the strongest physics content of the gate — three of the four structural carriers are forced within the declared grammar given the observed spectrum \(E\) . Leg 6 and the curvature legs (9–10) are clean exact mathematics. Leg 19 is the actual terminal: nothing forces it, its necessity is proven (leg 17 + 18), and it is honestly booked as a paid, new, value-free axiom — never claimed as derived, never claimed as free.

 7. Anti-claims and negative controls

 Anti-claims (what SG-1 does NOT certify): 

 Surface impression 
 Honest reality 

 "13D is the minimal/forced geometry" 
 selector-minimal, category-relative only — not absolute, not architecture-neutral forced 

 "the geometry is derived" 
 declared + frozen — a freeze certificate, not a derivation 

 "4 inputs → 22 outputs" 
 overstated — honest cost \(\approx\) 4 anchors + 9–10 injected reals ( \(\approx\) 13–14 reals total) 

 " \(K_6\) is the unique \(SU(3)\) carrier" 
 unique clean/purely-abelian carrier (rep theory); full-shelf completeness (R4) uncertified 

 "the gauge-group outcome discriminates the framework from rivals" 
 TIE — string/M/F-theory/NCG/lattice all recover the SM gauge group; not a discriminating result 

 Additional binding non-claims: SG-1 does not derive \(E\) (the SM chiral content) or \(\mathcal{M}_4\) itself; the family index \(\chi(K_6,E)=-3\) is given-E (an SG-3 result that SG-1 only records, not produces). \(CP^2\) stays dead (built end-to-end, breaks at Gate 2 — leg 3 above). SG-1 has no falsifier and is not a frontier gate — it is one of the corpus's RESOLVED/ANCHORED gates; SG-8 (up-quark mass, ~4.4σ tension), Gap-13, and UQF-4 are separate gates and are never merged into SG-1's status.

 Frozen negative controls (never dissolve, permanent tripwires): 
- \(|\mathrm{Riem}|^2(K_6) = 23/12\) exactly — never \(31/147\) , never \(60\) (the latter is \(S^6\) , the round unit 6-sphere, a topologically and metrically distinct manifold; a value of 60 appearing anywhere downstream signals a manifold-identity error).
- \(|\mathrm{Riem}|^2/\mathrm{Scal}^2 = 23/75\) exactly (confirmed; not \(31/147\) ).
- \(CP^2=SU(3)/U(2)\) is a permanently dead branch — built end-to-end to gate 2 and shown to BREAK (over-produces gauge content via the non-abelian centralizer of \(U(2)\) ); it must never be quietly revived as a viable color carrier.
- The Archimedean witness for \(\mathbb{Z}_{\ge0}\times_{\rm lex}\mathbb{Z}_{\ge0}\) fails for all tested \(n\) up to \(100000\) and provably for all \(n\) — this non-Archimedean-ness is a permanent mathematical fact about the lex order, not a numerical artifact that better computation could remove.
- The "~4× margin / 4→22" headline is retired as known-wrong (R3) and must not be restated; the load-bearing number is ~1.8× at transparent cost with an honest ~13–14-real charged cost.

 8. Endpoint line

 Terminal (FIXED, do not change): REDUCED-TO-AXIOM · ANCHORED +1 · PROMOTIONS:0. 

 Every forcedness chain in SG-1 bottoms on one of: (a) the observed spectrum \(E\) (which cancels in the spectrum-neutral economy comparison and is never claimed as derived — given-E \(\ne\) derivation of E), (b) the Granularity root's finite bit-cost residue \(\hbar\) (OBS-0002, AUDIT class), or (c) the single named axiom BR-Arch (the Archimedean common-currency rule for how bit-costs aggregate across kinds), counted once as a shared object (SAG-SELECTOR-M) across SG-1 and any other gate whose economy claim relies on additive-MDL cost comparison. No new measured invariant is created by this gate and no anchor floor is lowered by it. Selection \(\ne\) derivation; category-relative \(\ne\) absolute; given-E \(\ne\) derivation of E — these three inequalities are the honest content of the ANCHORED+1 terminal, and none of them is walked back by any of the seven saturated Layer-2 screens or the two independently-reproduced numerical routes confirming BR-Arch's necessity.