Topographical Group Theory and the Sedenion Crack

The Octonion Triad, Forced Democracy, and the One Load-Bearing Rung — v2.1

Ernest N. Prabhakar (Radical Centrism / ihack.us) and Bench (Claude, Anthropic)

July 6, 2026. Version 1 published July 6, 2026; same-day errata (E1–E6) published July 6 and fully incorporated here. All v1 numerical claims stand; several v1 attributions and one v1 sentence do not, and are corrected below at the point of use.

0. What this is, and what changed

Topographical Group Theory (TGT) is an experimental survey of the landscape around exact structures: pre-registered predictions, scale-relative tolerances, known-limit gates, contradiction treated as instrument diagnostic, and machine zero treated as candidate theorem. Version 1 applied the protocol to the sedenions 𝕊 = A₄ and reported the octonion triad, the automorphism group construction, and a democracy theorem. A literature pass run hours after publication found one false sentence, one misposed open problem, and a substantial attribution ledger (see the prepended Errata). This version incorporates all corrections, adds five further test campaigns (split signature, the multiplication family, isotopy, bi- and tri-composition, and the 32-dimensional rung), and restates every claim at post-literature strength.

The honest headline survives deflation: the sedenions force a 3 = 1 ⊕ 2 structure on their three octonion frames, by a mechanism that exists at exactly one rung of the Cayley–Dickson ladder, in exactly one signature. What is new here is the forcing; the objects being forced were known.

1. The landscape at dimension 16 — with corrected attributions

The alternative cone (Moreno). The alternator T_x = L_{x²} − L_x² measures the failure of alternativity elementwise. Its zero locus in 𝕊 is not the triad (v1’s error): by Moreno’s theorem, an element (a, b) is alternative iff a and b are linearly dependent, so the alternator’s vacuum is the 8-dimensional cone {a + t(a e₈)}. We verified this at machine precision (‖T‖ ≤ 7e-16 at pencil slopes 0.5, 1.0, 2.7).

The octonion triad (Chan–Đoković; Gresnigt). Inside the cone, closure under multiplication is a sharper condition. Writing candidate 7-spaces as W = {λv + μ(v·e₈) : v ∈ Im 𝕆}, generic CD identities reduce subalgebra closure to a single cubic:

μ(3λ² − μ²) = 0,

whose three roots — slopes 0, +√3, −√3 — give exactly three octonion subalgebras O₁, O₂, O₃ at mutual 60° in the pencil. The existence and count of these subalgebras follow from the Chan–Đoković conjugacy classification (2006), and their use as a generations substrate was proposed by Gresnigt, extending Furey. What is ours is the derivation: the closure cubic, the pencil geometry, and the corrected characterization — the triad is the subalgebra locus inside Moreno’s cone, the three slopes at which the cone’s fibers close.

The automorphism group (Schafer; Eakin–Sathaye; Dray–Manogue). Aut(𝕊) = G₂ × S₃. The order-3 generator τ — simultaneous rotation by 2π/3 in every plane (q, q e₈) — is stated explicitly in Dray and Manogue’s Geometry of the Exceptional Lie Groups; the explicit order-3 generator is Brown’s (1967: right multiplication of pure octonions by ζ = −½ − (√3/2)e₈), and the group structure, conjectured by Brown, is Eakin–Sathaye’s theorem (1990, Cor. 2.5; see §8). Ours: the gated numerical construction; the full-circle scan showing automorphy at exactly {0°, 120°, 240°} (defect 1e-14 vs. controls of order 40); ψτψ = τ⁻¹; and the completeness sweep of the monomial sector (2688 elements, none frame-moving), establishing the S₃ as irreducibly non-monomial.

The triskelion. Under G₂ × S₃, 𝕊 = (1 ⊗ triv) ⊕ (1 ⊗ sgn) ⊕ (7 ⊗ std): the identity, the axis e₈ (odd under ψ), and the seven-dimensional pencil carrying the standard representation three ways.

2. The static anatomy is eigentheory (Biss–Christensen–Dugger–Isaksen; Moreno)

Everything structural about single-element multiplication is owned by the eigentheory literature, and our instruments blindly reproduced it — which we report as calibration, verified both directions at ≤ 1e-15:

The operator M_a = |a|⁻²L_{a*}L_a (our “composition operator” LᵀL) is BCDI’s central object.
Quartets: every eigenspace is a module over C₄ = span{1, e₈}, hence of real dimension divisible by 4 (BCDI Prop. 3.19). The quadruple degeneracies throughout our data are this theorem.
Sum rule: eigenvalues sum to 2ⁿ = 16 (Prop. 3.16) — a normalization identity, not a conservation law.
Hulls: Eig₁(a) = ker T_a — the 8-dimensional octonion through each element, on which multiplication is exactly isometric (Lemma 3.17, Remark 3.18). Every private hull contains the axis plane.
Ceiling: all eigenvalues lie in [0, 2^(n−3)] = [0, 2] for 𝕊, sharp (Cor. 4.7) — our √2 singular-value ceiling.
Generic spectrum: {1−s, 1⁸, 1+s}, s = 2|a||b|sinθ/(|a|²+|b|²), multiplicities 4-8-4 (Cor. 7.3); the pair sums to 2 identically (v1’s “circle law”).
Zero divisors: spectrum {0⁴, 1⁸, 2⁴} with eigenspaces (x, ∓cx), c = ab/|ab| (Prop. 7.4).

One synthesis is ours, at one-sentence strength: combining Prop. 7.4 with the S₃, the killed and doubled quartets at every zero divisor are exchanged by ψ (verified: ψ(Eig₀) = Eig₂, principal overlaps 1.000000). The reflection that survived the crack is the operator that balances its books.

What remains ours in this territory is the dynamical reading, labeled as interpretation: iterating y ↦ xy concentrates amplitude in the dominant quartet at asymptotic rate ln√(1+s) ≤ ½ ln 2 per step, with equality exactly on the zero-divisor locus; continuous flows exp(tL_x) are isometries, so growth and decay exist only in discrete events (a two-ledger structure: one book spins, one settles); and the trace identity reads as a budget — annihilation funds amplification at par. The rate ceiling is packaging-invariant: sandwich U_x = xyx reaches spectral radius 2.000000 (two multiplications, same rate per product-use), ad_x reaches √2, and the Jordan half is a pure rank-2 clock (rotation e₀ ↔ x, Euler’s formula in situ). Dissipation is an interference effect between the Jordan and Lie halves of the product; neither alone produces it.

3. Theorems: what the crack forces (ours)

Theorem A (alternator factorization). For every x ∈ 𝕊, the alternator factors exactly as

T_x = A₇(x) ⊗ J₂,
a 7×7 antisymmetric block tensored with the standard symplectic 2×2 — measured residual 3.2e-16, and strengthened by Moreno’s cone (the symmetric pencil component now vanishes on all lines, not merely three). The alternator is a pure inter-frame transition generator: single elements can rotate the generation dial and do nothing else.

Theorem B (forced democracy and 3 = 1 ⊕ 2). The six inter-frame couplings of T_x are exactly equal for every x — forced by sin 60° = sin 120°, i.e., by the same cubic that created the frames. The coupling matrix on frame space is therefore the democratic matrix with spectrum

(2, −1, −1):
one frame combination singled out, two exactly degenerate. Three generations with a democratic texture splitting as 1 ⊕ 2 is not an input; it is a theorem of the sedenions. This is the paper’s core claim, and no prior statement of it was found in the literature pass. (Democratic mass matrices are a known flavor-physics ansatz — Fritzsch lineage; the derivation from an algebra is the new content.)

Theorem C (dynamical no-go). Composition N(xy) = N(x)N(y) forces L_x ∈ O(n): division-rung dynamics is clocks only. In 𝕊 the spectrum leaves the unit circle under the fixed 2ⁿ budget, with Lyapunov ceiling ½ ln 2 per product-use — attained exactly at the crack’s zero divisors. The doubling constant √2 that built the algebra is the same constant that bounds its dynamics: the crack leaks at most half of the doubling that created it, per event.

4. Test 3 — signature: the triad is compact-only

Split-doubling (γ = +1) flips the closure cubic to μ(3λ² + μ²) = 0: one real root, one frame. The triad and its S₃ do not exist in split signature (verified: division-side closure at 0, ±√3 to machine zero; split closure fails by O(1)). ψ survives; boosts are never automorphic; Der stays the compact g₂ (Killing signature (0, −14)) in both signatures — a failed pre-registration of ours, logged. The alternator’s pencil components swap exactly between signatures — (7.6e-16, 1.724) division vs. (2.145, 0.000) split — rotation generator and boost generator as two real forms of one dynamics. Hulls persist 8-dimensional as split octonions (4,4) with exact split composition; 4-cores are split quaternions (2,2). Minkowski (1,3) appears in no Cayley–Dickson norm at any rung (pre-registered and confirmed): spacetime is not derived here. It is located — the (1,3) Hessian lives in the determinant of J₂(ℂ), the complex-rung Jordan algebra sitting at the color-singlet core of J₃(𝕆), i.e., on the same ℂ that section 6 finds threaded through every sedenion structure. Located, not possessed.

5. Test 4 — isotopy: the S₃ is triality’s ghost

The autotopism (triality) algebra tri(A) = {(F,G,H) : F(xy) = G(x)y + xH(y)} measures continuous isotopy freedom. Gates: dim tri = 11 (ℍ), 30 (𝕆: Barton–Sudbery’s 28 plus two dilations). Measured for 𝕊: 18 = 14 ⊕ 2 ⊕ 2 — derivations, dilations, and a two-dimensional remnant. The Moufang nucleus (elements a whose multiplication triples remain autotopic) collapses from all of 𝕆 to exactly span{e₀, e₈}. Of the fourteen continuous triality directions the octonions enjoy, twelve die at the crack; what survives of the symmetry is its discrete skeleton — the frame permutations. The sedenion S₃ is the ghost of Spin(8) triality: 28 → 14 ⊕ 2 ⊕ 2, twelve continuous directions reborn as six group elements. This is the third instance of the campaign’s recurring law: failure sets keep the form of what failed (the norm died leaving its symmetry as the zero-divisor shape; composition died leaving its constant as the rate ceiling; triality died leaving its component group as the family symmetry).

6. Test 5 — pairs: the axis wakes, and one plane wears five hats

The bilinear associator A_{x,y}(z) = [x,y,z] splits by (x↔y) symmetry. The symmetric part is the polarized alternator — still exactly ⊗J₂, still pure frame rotation. The antisymmetric part is new structure (mixed pencil character, the first object outside the ⊗J₂ cage) and it does what nothing bilinear-diagonal can: it moves the axis (max ‖[x,y,e₈]‖ = 2.40; the symmetric part moves it by 1.7e-16). Since e₈ lies in every private hull, no single element ever subjects the axis to non-isometric dynamics: the axis is a collective-only degree of freedom — single-particle safe, two-particle alive — and its pair dynamics is carried entirely by the inherited alternating form (§7).

Hull geometry of pairs: for every generic pair, dim(O_x ∩ O_y) = 2, and the intersection is always span{e₀, e₈}. The book of private octonions is bound at a single complex spine. That plane is now characterized five independent ways: Der-fixed subspace (formally Eakin–Sathaye Cor. 1.9: the joint kernel of all derivations is ₃A, which at n = 4 is exactly this plane); carrier of triv ⊕ sgn (the ψ-odd axis); the entire surviving Moufang nucleus; the universal hull spine; the unique pair-only dynamical target. A U(1)-sized object coupling only through pairs, sitting on the ℂ where §4 located the Minkowski form — flagged, unpriced.

One failed pre-registration, logged as open: the classical formula D_{x,y} = [L_x,L_y] + [L_x,R_y] + [R_x,R_y] fails to be a derivation of 𝕊 even for octonion pairs (defect 2.19). Der(𝕊) is not internally generated from the multiplication algebra in the alternative-algebra manner; how the crack’s g₂ is generated remains open.

7. Test 6 — triples: irreducibly triadic matter

The full associator 3-tensor, decomposed under slot permutations (flexibility gates 0.0e+00 in both algebras): 𝕆 is 100% alternating; 𝕊 splits with exact integers,

‖A_𝕊‖² = 7392 = 11 × ‖A_𝕆‖², 87.9% alternating ⊕ 12.1% mixed ⊕ 0% symmetric
(the mixed fraction ≈ 4/33, suspiciously rational; exact-arithmetic check pending). The Jacobiator equals 6·Alt(A) identically, so the mixed component is invisible to every commutator-built probe — Lie brackets see only the alternating shadow. Malcev holds on frame-pure triples (3.4e-13) and fails generically (13.4). And the support census, at machine zero: the mixed part vanishes on same-frame triples, vanishes on any axis slot, and is supported exclusively on cross-frame configurations (O₁, O₂, O₃: 168.000 exact). The only genuinely new trilinear content the crack creates requires all three frames simultaneously — matter that cannot be decomposed into pairs and cannot exist without three distinct generations present. (The pair-level axis dynamics of §6 lives in the alternating part; everything coheres.) We note the resonance without pricing it: this is Peircean Thirdness realized as a measured tensor component with an exact support law.

8. Test 7 — the tower: ternary forever, load-bearing once

Three senses of “can we do four,” three answers:

Arity — never. The pentagon identity closes in every algebra (verified 4.6e-16): the binary product has exactly one primitive obstruction, at arity 3. All higher coherence defects are associator composites.

Frames — never. The closure polynomial is cubic at every rung and in both signatures; no fourth root exists.

Rungs — the mint repeats, and this corrected v1’s misposed open problem. At dimension 32 the same generic computation yields three sedenion frames at ±60° (closure defects ≤ 2.5e-16), and the 32-level pencil rotation is an automorphism (defect 1.9e-14 vs. controls 41–58; order 3; ψ₃₂τ₃₂ψ₃₂ = τ₃₂⁻¹; 3-cycles the sedenion frames). It moves e₈ and therefore lies outside the lifted sedenion group, and it commutes with the lifted τ₁₆ exactly: Aut(A₅) ⊇ G₂ × S₃ × S₃. This is in fact a theorem, not a conjecture: Eakin–Sathaye (1990, Cor. 2.5) proved Brown’s conjecture that Aut_F(A_n) ≅ Aut_F(A_{n−1}) ⊕ G for n ≥ 4, where G ≅ S₃ if √(−3γ_n⁻¹) ∈ F and ℤ/2 otherwise. Over ℝ with γ ≡ −1 the criterion is √3 ∈ ℝ, satisfied at every rung, so Aut(A_n) ≅ G₂ × S₃^(n−3) for all n ≥ 3 — the ladder is a proven ternary tree, our rung-32 measurement is the n = 5 instance, and no τ₆₄ check is needed. (Our v1 memory of a “frozen” group was a garble of this recursion into a saturation; the failed pre-registration is logged and now diagnosed.) Note the field criterion: √(−3γ⁻¹) = √3 is exactly the slope of our closure cubic’s nontrivial roots — over ℚ only the slope-0 frame is rational and the factor collapses to ⟨ψ⟩ ≅ ℤ/2. The mint exists precisely when the field can see all three frames: their arithmetic criterion and our pencil geometry are the same fact. Der(A₅) = 14: the new symmetries are purely discrete (consistent with Schafer’s derivation theorem, ES Thm. 1.6, and their Cor. 1.9).

But the machinery that makes a triad physical exists once. The level-32 frames are sedenions, hence not alternative: the alternator does not vanish on them (‖T‖ = 3.35, vs. 1e-15 at level 16). No vacuum cone organized around them, no antisymmetry theorem, no forced democracy, no 1 ⊕ 2. Higher mints are symmetries without dynamics — decoration, not generations. The selection principle, measured: generations are not “the frames”; they are the unique frames that are dynamically silent — alternative frames, which exist at one rung (16), in one signature (compact), numbering three (the cubic).

9. Physics assembly, at deflated strength

Derived: a three-fold family structure with exactly democratic couplings and forced 1 ⊕ 2 degeneracy, from the unique vacuum triad of the unique load-bearing rung. Located, not derived: a (1,3) quadratic form at the ℂ-core (J₂ row of the magic square), on the same distinguished plane that carries the pair-only U(1)-sized freedom. Compatible, not derived: the Furey–Gresnigt program (our results un-exclude it and supply forcing theorems it lacked; they supply the fermion content we lack); sector recombination, chirality, and any manifold remain absent here. Open seams, ranked by stakes: the S₃-breaking correction that splits the degenerate pair (a mass hierarchy would be the decisive test of this framework); the 364-dimensional shell decomposition under G₂ × S₃; the ℂ⊗𝕊 transplant; the exact 4/33 check; τ₆₄.

10. Ledger

Failed pre-registrations (all logged, two discovery-grade): split-form Killing prediction; e₈-mediated quartet pairing (resolved by BCDI Prop. 7.4: the pairing is ψ); classical derivation-generation for octonion pairs; the rung-32 freeze (the mint repeats). Protocol addenda earned: sample the complement of a hypothesis at interior points, not only distinguished ones; the literature pass is an instrument — run it before strong nouns, and expect it to close open problems by reading. Pending: Chan–Đoković full text (triad conjugacy presentation); 4/33 in exact arithmetic. (The Eakin–Sathaye item closed July 7 with the paper in hand; see §8.)

References

J. Baez, The Octonions, Bull. AMS 39 (2002). C. Barton, A. Sudbery, Magic squares and matrix models of Lie algebras, Adv. Math. 180 (2003). D. K. Biss, J. D. Christensen, D. Dugger, D. C. Isaksen, Eigentheory of Cayley–Dickson algebras (and companion papers [DDD], [DDDD]). R. B. Brown, On generalized Cayley–Dickson algebras, Pacific J. Math. 20 (1967). K.-C. Chan, D. Ž. Đoković, Conjugacy classes of subalgebras of the real sedenions, Canad. Math. Bull. 49 (2006) 492–507. T. Dray, C. A. Manogue, The Geometry of the Exceptional Lie Groups, §Cayley–Dickson automorphisms. P. Eakin, A. Sathaye, On automorphisms and derivations of Cayley–Dickson algebras, J. Algebra 129 (1990) 263–278. C. Furey, Three generations, two unbroken gauge symmetries, and one eight-dimensional algebra, Phys. Lett. B 785 (2018). A. Gillard, N. Gresnigt, on three generations from ℂ⊗𝕊, Adv. Appl. Clifford Algebras (2019). G. Moreno, The zero divisors of the Cayley–Dickson algebras over the real numbers, Bol. Soc. Mat. Mex. 4 (1998); Alternative elements in the Cayley–Dickson algebras, in the Plebański volume, World Scientific (2006), arXiv:math/0404395. R. D. Schafer, On algebras formed by the Cayley–Dickson process, Amer. J. Math. 76 (1954). G. P. Wilmot, Automorphisms of Sedenions, arXiv:2512.07210 (2025). Full reproduction scripts: exceptional_algebras_lab.py and session transcripts, July 5–7, 2026.

v1’s flags were load-bearing; v2 is what they were carrying. — E.N.P. & Bench