SG-3 — Three generations χ=−3: the gate anchor ledger

The honest one-line: SG-3 has a real, checkable result — the number of matter generations comes out of this geometry as a rigid whole number with no dial, $|\chi(K_6,E)| = 3$ — but it is computed with the observed Standard-Model content $E$ as input (given-E ≠ derivation of E), the magnitude (not the chirality sign) is what is pinned, and the bundle returning $-3$ is selected, not proven unique.

This page is the gate-specific instantiation of the anchoring method: it takes the master anchor and the bridges and applies them, object by object, to one gate. Every exact thing SG-3 touches gets its own row — its status, what it is allowed to claim, and what it is forbidden to claim. It follows the same eleven-part shape as the canonical SG-4 ledger.

1. Gate status header

• Gate-level SG-3 roll-up: DERIVED-GIVEN-E — RESOLVED +0 on the live board (ratified 2026-07-08).
• Taxonomy reconciliation (2026-07-05): under the ratified closure taxonomy (board 2026-07-08) the gate-level grading is DERIVED-GIVEN-E — RESOLVED +0, read as TERMINAL + RESIDUALS-SHOWN — a reached terminal with its residual family (R1–R8, R3 highest) listed and carried unchanged below. The facts are unchanged: any earlier “open” roll-up reflected the superseded least-closed-residual rule, not different physics. The residuals in this ledger remain listed and are neither closed nor re-graded here.
• The closed leg — a rigid integer family count: given the frozen geometry and the bundle $E$ (which is the observed chiral content), the holomorphic Euler characteristic of the carrier returns $$\chi(K_6,E) = -3, \qquad |\chi(K_6,E)| = 3 = \text{number of generations}.$$ The chiral half adds the boundary index on the orbifold fold: $$(n_L, n_R) = (+3,\,0) \quad\text{on } S^1_Y/\mathbb{Z}_2 .$$
• Status of the closed leg: DERIVED-GIVEN-E — the integer is a deformation-proof topological index, genuine and convention-independent in magnitude; the bundle whose Chern class gives $-3$ is $E$ itself, so this certifies "this geometry+bundle yields three families, given the particles we observe." It does not derive $E$, and it does not prove three is forced across all geometries.

The result is a genuine win: a family count is exactly the kind of object that should be an index — a winding number, not a tunable volume — and here it is one. What stays open is everything beyond the magnitude: the uniqueness of the bundle/weight, whether the carrier preference over the cheaper $\mathbb{CP}^2$ is target-blind, the structure label ("spin-$\mathbb{C}$" is refuted for pure SM), the chirality sign convention bit, and whether the from-scratch certificate re-derives the index.

selection ≠ derivation · given-E ≠ derivation-of-E · frozen/reproducible ≠ proven-unique.

2. Frozen inputs (what SG-3 stands on, not what it produces)

• Frozen branch hashes dcc66f1b2685 (branch) / a5b1e6f9d951 (manifest meta), with object data 0fd19c9ae0c1 ((twisted-)spin bundle on $K_6$) and ac4d2df3e708 ($\mathbb{Z}_2$ orbifold + boundary parity ledger). The branch is read-only. These hashes are audit anchors — they certify which object was tested and that it cannot be quietly retuned. They do not validate the physics.
• Upstream spectrum $E$ (the SM chiral content) is given / charged / inherited — it enters SG-3 as input, and is the bundle whose index is computed. SG-3 does not derive $E$. Using $\chi = -3$ to "force" $E$ would be circular, because $E$ is the input.
• External admissibility-boundary datum: the LEP/SLD light-neutrino count $N_\nu = 2.984 \pm 0.008$. This excludes a fourth light generation; it is a measurement, not a mechanism, and it sits at the class boundary (see R5).

3. The object anchors (given-E / upstream)

The chiral actors live in the matter bundle $E_{\rm matter}$, one generation being the SM set of $15$ Weyl fermions ($16$ with $\nu_R$): $$ Q_L=(3,2)_{+1/6},\quad u_R=(\bar 3,1)_{-2/3},\quad d_R=(\bar 3,1)_{+1/3},\quad L_L=(1,2)_{-1/2},\quad e_R=(1,1)_{+1}. $$ The generation module is $\mathcal{G}_{\rm gen}=\mathrm{span}\{g_1,g_2,g_3\}$, $\dim = 3$, matched to $|\chi|=3$ and consumed downstream by the flavor chamber. Status: GIVEN-E / upstream-inherited — not SG-3-derived.

4. Root and master-anchor traceability

Deep roots that are load-bearing for SG-3:

Scale and causal order are not primary load-bearing anchors for the SG-3 count.

Master anchors in play: finite invariant ledgers (the integer index) · no unpaid labels (the $\mathbb{Z}_6$ congruence, $m_0(p,q)$, computed target-blind) · the frozen branch · given-$E$ (the bundle is $E$) · the declared admissibility class · open-residual discipline.

5. The SG-3 anchor ledger (the universal table)

6. The arithmetic / construction — the rigid integer, in full

The count is generated by two index computations on the frozen geometry: $K_6$ counts the families; $S^1_Y/\mathbb{Z}_2$ makes them chiral by projecting out the mirror.

(1) The carrier (hand-checkable). $K_6=SU(3)/T^2$ is the complete flag manifold of $\mathbb{C}^3$: $$\dim K_6 = \dim SU(3) - \dim T^2 = 8 - 2 = 6, \qquad \chi(K_6) = |W(SU(3))| = |S_3| = 6 .$$

(2) The Borel–Weil–Bott count. On a homogeneous Kähler $G/T$, BWB gives the cohomology of a line bundle $L_\lambda$ in closed form, so the holomorphic Euler characteristic $$\chi(K_6,L_\lambda) = \sum_i (-1)^i \dim H^i(K_6,L_\lambda)$$ is integer-valued and representation-theoretic — not a numerical fit. For the frozen bundle $E$, $$\boxed{\;\chi(K_6,E) = -3\;}\qquad |\chi| = 3 \text{ generations.}$$ The integer is rigid: no continuous modulus can move it inside the declared admissibility category (Atiyah–Singer rigidity). That is what "no dial" means.

(3) The chirality fold (no-go + control). A closed internal factor is handedness-neutral (corpus fact F2): a bare circle $S^1_Y$ mirrors every fermion and the APS index reads $(n_L,n_R)=(+3,+3)$ — wrong physics. The orbifold $S^1_Y/\mathbb{Z}_2$ folds by $\theta\mapsto-\theta$ with two fixed points; under the frozen parity ledger the Atiyah–Patodi–Singer boundary index returns $$(n_L,n_R) = (+3,\,0).$$ A one-sided count is impossible on any closed factor; the bare-$S^1_Y$ control $(+3,+3)$ proves the fold is load-bearing.

Diagnostic — the result is specific, not trivial. Three independent specificity checks distinguish this from "arranged to give 3":

1. The count is the right kind of number. "Three by dial" fails as a dial even when its value is 3 — a continuous bundle modulus on the cheaper $\mathbb{CP}^2$ is not an index. Only a deformation-proof integer survives the type filter.
2. The mirror control is non-degenerate. The bare-$S^1_Y$ value $(+3,+3)\neq(+3,0)$: the fold changes the answer, so the one-sidedness is earned, not assumed.
3. The engine regenerates target-blind. $m_0(p,q)=\min(p,q)+1$ re-derives via Freudenthal's recursion and the $\mathbb{Z}_6$ congruence $q\equiv 3z_2-2z_3\ (\mathrm{mod}\ 6)$ holds field-by-field — both computed from the actual SM hypercharges, no generation number in sight.

The obstruction split. The closed leg is the magnitude: $$O_{\rm SG3,local}(E) = \big(\chi(K_6,E)+3,\ (n_L,n_R)-(+3,0)\big),\qquad O_{\rm SG3,local}(E)=0 .$$ The full gate obstruction additionally carries the bundle-uniqueness and carrier-tie-break pieces, which are not closed: $$O_{\rm SG3}(E)=\big(O_{\rm uniqueness}(E),\,O_{\rm carrier-tiebreak}(E),\,O_{\rm sign}(E),\,O_{\rm local}(E)\big).$$ We do not assert $O_{\rm SG3}(E)=0$.

7. The "spin-$\mathbb{C}$ index" phrase, split into honest objects

The single phrase "the spin-$\mathbb{C}$ Borel–Weil–Bott index" hides distinct claims with distinct statuses:

1. The magnitude $|\chi|=3$. A count, not a structure label. Status: DERIVED-GIVEN-E — it survives any structure-label correction unchanged.
2. The "spin-$\mathbb{C}$" structure label. Status: REFUTED for pure SM. A spin-$\mathbb{C}$ structure exists iff some $U(1)\subset G_{\rm SM}$ assigns all-odd charges to the $15$ Weyl fermions; an all-odd scan returns EMPTY for pure SM (reproducing Davighi–Gripaios–Lohitsiri, arXiv:1910.11277 Sec. 7). Test-the-test: inserting a gauged $B-L$ makes the search succeed, proving the scanner discriminates.
3. The genuine forced object. With $(-1)^F$ identified as the $SU(2)$ $2\pi$ rotation in the gauge centre, the global structure is the twisted $$(\mathrm{Spin}\times G_{\rm SM})/\mathbb{Z}_6,\qquad \text{twist order } n=2 .$$ Status: FORCED-GIVEN-E (it lands on $E$ + the anomaly rulebook + a Dai–Freed single-valuedness principle; it carries no generation number, so it is not a forcing of $E$).

So the value is DERIVED-GIVEN-E, the spin-$\mathbb{C}$ label is refuted, and the twisted structure is forced-given-E. The open target is to mount the absent twisted-Dirac spectrum and confirm the twisted-Dirac index $=3$ matches the BWB-on-line-bundle value without tuning to the known answer.

8. The carrier preference, split into honest objects (the flagged target-aligned tie-break)

"$K_6$ is forced" hides two different claims:

1. Carrier-for-gauge. $K_6=SU(3)/T^2$ is the unique purely-abelian $SU(3)$ isotropy giving a clean gauge sector ($\mathbb{CP}^2=SU(3)/U(2)$ over-produces gauge content at the gauge gate). Status: DERIVED / data-independent — PASS.
2. Carrier-for-the-family-count. $K_6$ (6D) beats the cheaper $\mathbb{CP}^2$ (4D) on the count only by paying $+2$ dimensions under an anti-fitting tie-break ("adjustable counts as fail") that is aligned with the 3-generation target. Status: FLAGGED — the flagged target-aligned tie-break (R2).

The legitimate gauge result must not be allowed to launder the family-count tie-break, which is separate. Naming the $+2$-dimension payment out loud — rather than hiding it — is what keeps the strong magnitude claim credible.

9. Open residuals — beyond the rigid magnitude

The magnitude is the closed face of SG-3. These distinct residuals make up the rest; none is closed by the integer count. Leverage / attack order: R3 → (R1 via R3) → R2 → R4 → R6 → R5 / R7.

• R3 — bundle admissibility / uniqueness (HIGHEST LEVERAGE). The weight on $K_6$ returning $\chi=-3$ is selected inside a declared admissibility class, not proven the unique admissible one. AXIOM-OPEN. This is the only path that could lift the given-E conditioning.
• R1 — given-E. The bundle whose index is $-3$ is $E$. AXIOM-OPEN (closeable only via R3); the gate stays DERIVED-GIVEN-E until then.
• R2 — bundle-selected vs bare-carrier-forced (the flagged target-aligned tie-break). The $+2$-dimension payment for $K_6$ over $\mathbb{CP}^2$ leans on a target-aligned tie-break. OPEN.
• R4 — "spin-$\mathbb{C}$" wording refuted for pure SM. Cleanest correctness win; wording patch reachable now, from-scratch twisted-Dirac re-index BLOCKED on the absent k6_dirac_spectrum.csv. DISCLOSED-CORRECTED now / BLOCKED_INPUTS for the value.
• R6 — the G04_chirality certificate is a consistency lint, not a from-scratch re-derivation. It checks against the declared $-3$ and $(+3,0)$. AUDIT / BLOCKED_INPUTS (same missing CSV).
• R5 — the index-changing deformation is excluded by data, not topology. The neighboring values $2,4$ are excluded by LEP $N_\nu$, an external datum; "no dial" is topological in-category + empirical at the class boundary. DISCLOSED-CONSISTENT (dependent on R3).
• R7 — the chirality sign convention bit is unpinned. The magnitude 3 and one-sidedness $n_R=0$ are convention-independent; the sign $-3$ (left- vs right-handed) rides an unpinned $\mathrm{Pin}^\pm$ bit. OPEN — and its on-disk default is the WRONG sign downstream: $\chi=-3$ ($=5\bmod 8$) gives $\arg=-135^\circ=e^{-3i\pi/4}$, the wrong sign for the BG-10 leptogenesis target (which needs $\sigma_\nu=+1\to e^{+i\pi/4}$). The axiom departs from the target rather than wearing it.
• R8 — inheritance (dependency record, not a closure target). $\dim\mathcal{G}_{\rm gen}=3$ is inherited at SG-7/8; a revision of "3" cascades. Keep the cascade wired and the given-E qualifier travelling with "3."

10. Anti-claims (what this page refuses to say)

• SG-3 does not derive $E$. The index is computed with $E$ as input; "three families are derived (from nothing)" is forbidden overclaim. The honest hook is "given the particles we observe." Formally: $E\in\ker O_{\rm SG3,local}$, not $\ker O_{\rm SG3,local}=\{E_{\rm SM}\}$.
• Three is not forced across all geometries. That is a universal negative — unprovable in principle for any theory. Only this geometry+bundle is certified.
• Anomaly-freedom does not fix the generation number. It is a filter, not a determiner; there are infinitely many anomaly-free chiral $U(1)$ extensions of the SM. The internal claim that $E$ is forced by anomaly-freedom + minimality is REFUTED.
• NOT "the spin-$\mathbb{C}$ Borel–Weil–Bott index." Pure-SM content provably forbids spin-$\mathbb{C}$; the forced object is twisted $(\mathrm{Spin}\times G_{\rm SM})/\mathbb{Z}_6$. The label is refuted on-disk while still live in the manuscript; $|\chi|=3$ survives.
• NOT "the bundle/weight is unique." Only selected inside a declared class (R3).
• NOT "left-handed, no surviving mirror" as a settled chirality output. Only the magnitude and one-sidedness are convention-independent; the sign rides an unpinned $\mathrm{Pin}$ bit whose default is on-disk wrong-sign for BG-10 (R7).
• NOT "the G04_chirality certificate establishes the index." It is a consistency lint against the declared values, not a from-scratch re-derivation (R6).
• The frozen-branch hashes are audit anchors; they do not validate the physics.
• Local magnitude closure is not whole-gate closure. $O_{\rm SG3,local}=0$ does not imply $O_{\rm SG3}=0$.

11. Specialist closure plan

Each open residual is a concrete, finite work-package, to be carried out without using the known generation number as an input. A closure path that introduces a new target-aligned criterion (e.g. "pick the weight that gives three") has relocated the input, not removed it.

1. R3 / bundle uniqueness — enumerate the $K_6$ weights compatible with the $\mathbb{Z}_6$ Tong-congruence lattice; apply BWB to each (build weight $\mapsto\chi$); test whether minimality singles out a unique weight giving $|\chi|=3$ without invoking "three" and without LEP. Named posit: AXIOM-MIN-WEIGHT-LIFT. Endpoint: sharper-OPEN → AXIOM-CLOSED if minimality yields 3 value-blind. Valid refuting close: a strictly-lower-cost weight with $\chi\neq-3$ ⇒ the bundle was 3-selected.
2. R2 / carrier tie-break — re-run the selector with the 3-generation target masked: confirm the $\mathbb{CP}^2$ count is a continuous modulus (rejected as a dial, value-blind), survey sub-6D $SU(3)$ carriers for a rigid-integer-count survivor, and test whether the $+2$-dimension payment still goes through on "rigid-integer-count" alone. Named posit: AXIOM-COUNT-MUST-BE-INDEX. Endpoint: AXIOM-CLOSED if the masked run passes; sharper-OPEN (tie-break confirmed target-aligned) if it only passes because it yields 3.
3. R4 / spin structure — replace "spin-$\mathbb{C}$ index" with "twisted $(\mathrm{Spin}\times G_{\rm SM})/\mathbb{Z}_6$ index" in the manuscript (reachable now), citing DGL + Tong (+ Hsieh–Tachikawa–Yonekura; García-Etxebarria–Montero). Named posit: AXIOM-TWISTED-SPIN-GIVEN-E. For the DERIVED-CLOSED tier, mount the absent k6_dirac_spectrum.csv and confirm the twisted-Dirac index $=3$ (BLOCKED until generated — do not fabricate it).
4. R6 / certificate — re-derive the index from the bundle data target-blind (do not read the declared $-3$), confirm it equals the frozen $-3$ and that APS re-derives $(+3,0)$, then re-hash. Gated on the same missing CSV as R4.
5. R7 / chirality sign — a spin-bordism question: is the orientation $\mathrm{Pin}$ bit on $S^1_Y/\mathbb{Z}_2$ fixed by Dai–Freed single-valuedness, or a free convention? Named posit: AXIOM-CHIRALITY-ORIENTATION. Endpoint: AXIOM-CLOSED (count unaffected); surface the wrong-sign default as the real datum, not a bland "unpinned bit."
6. R5 / data-vs-topology — keep the topological-in-category vs empirical-at-the-boundary split explicit wherever "no fourth family" appears. Named posit: AXIOM-INDEX-RIGID-IN-CATEGORY. Becomes boundary-rigid only if R3 closes.

Closing these upgrades SG-3's honesty and reach — but the given-E qualifier is not removable short of R3 closing, and R3's own realistic ceiling is AXIOM-CLOSED. No DERIVED-CLOSED is promised.

12. Completion tests for this page

Required presence (all met): gate roll-up DERIVED-GIVEN-E + RESOLVED +0 · the closed leg $\chi(K_6,E)=-3$, $|\chi|=3$ · APS $(n_L,n_R)=(+3,0)$ · $E$ not derived · frozen hashes (AUDIT ONLY) · carrier $\dim K_6=6$, $\chi(K_6)=6=|S_3|$ · the bare-$S^1_Y$ control $(+3,+3)$ · the $\mathbb{Z}_6$ Tong congruence · $m_0(p,q)=\min(p,q)+1$ · the spin-$\mathbb{C}$ refutation + twisted structure · the $+2$-dimension target-aligned tie-break · LEP $N_\nu$ measured anchor · every open residual (R1–R8) as its own item · specificity diagnostics · the gate's anti-claims.

Required absence (all held): no claim that three is derived from nothing · three forced across geometries · anomaly-freedom fixes the count · the bundle/weight is unique · "spin-$\mathbb{C}$ index" asserted as correct · the chirality sign settled · $\ker O_{\rm SG3}=\{E_{\rm SM}\}$ · the certificate establishes the index · hashes validate physics · local magnitude closure = whole-gate closure · any reader-visible build-process vocabulary.

This is the SG-3 gate anchor ledger; it follows the same eleven-part shape and universal table as the canonical SG-4 ledger.

See also: the anchoring method · the master anchor · Layer 4 — carrier-forcing & the given-E wall (why the bundle is selected and $E$ is given) · the shape-minimality challenge (the carrier-over-$\mathbb{CP}^2$ tie-break) · the full SG-3 dossier.