mirror of
https://github.com/allaunthefox/Research-Stack.git
synced 2026-07-31 03:05:21 +00:00
allaun
475f6319ea
This squashes all local history (768 commits) onto the scrubbed PR #90 baseline. Individual commits were lost during filter-repo corruption; the working tree content is preserved intact. Build: N/A (working tree state only)
286 lines
13 KiB
Markdown
286 lines
13 KiB
Markdown
# Rigorization of the 16D Rotation / Yang Column Formalism
|
||
|
||
**Status:** mathematical foundation extracted from PhiNUVMAP (`PistSimulation.lean` §8),
|
||
Goxel16D (`MeshRouting.lean`), and Law 15 (`Law15_Field.lean`).
|
||
Each informal concept is paired with the exact theorem or formula that solidifies it.
|
||
|
||
| Informal concept | Rigorous counterpart |
|
||
|---|---|
|
||
| 16D shape-potential space | Kähler vector space (ℝ¹⁶, g, J, ω) ≅ ℂ⁸ |
|
||
| "rotate through all 16D" | dense orbit on the maximal torus T⁸ ⊂ U(8) (Weyl equidistribution) |
|
||
| φ-contraction + rotation | golden spiral similarity z ↦ c + φ⁻¹e^{iθ_g}(z − c) |
|
||
| Yang column | distinguished complex line with quantized winding (U(1) flux tube) |
|
||
| Kähler gate "smooth vs fractal" | two-out-of-three theorem; ∂̄-residual |
|
||
| "smooth rotation recovers EM" | Cauchy–Riemann ⟺ J-equivariance ⟹ harmonic ⟹ vacuum Maxwell |
|
||
| integer-only 16D geometry | even unimodular lattices E₈⊕E₈, D₁₆⁺; finite exact rotation group |
|
||
| primes in the geometry | shell counts r(2n) = 480·σ₇(n); explicit formula over ζ-zeros |
|
||
| finite bulk, infinite boundary | IFS attractor, Moran dimension D = ln N / ln φ |
|
||
|
||
---
|
||
|
||
## 1. The space: ℝ¹⁶ as ℂ⁸ — making J an operator
|
||
|
||
Pair the dimensions (d₀,d₁), (d₂,d₃), …, (d₁₄,d₁₅) into complex coordinates
|
||
z_k = d_{2k} + i·d_{2k+1}, k = 0…7. The almost-complex structure is the
|
||
block-diagonal **integer matrix**
|
||
|
||
J = diag(ε, ε, …, ε) (8 blocks), ε = [ 0 −1 ]
|
||
[ 1 0 ]
|
||
|
||
satisfying J² = −I **exactly** (entries 0, ±1 — exact in Q16.16).
|
||
|
||
Metric and symplectic form:
|
||
|
||
g(X, Y) = Σ_{i=0}^{15} X_i Y_i
|
||
ω(X, Y) = g(JX, Y) = Σ_{k=0}^{7} (X_{2k} Y_{2k+1} − X_{2k+1} Y_{2k})
|
||
|
||
**Theorem (two-out-of-three).** For the groups of linear maps preserving each
|
||
structure,
|
||
|
||
O(16) ∩ Sp(16, ℝ) ∩ GL(8, ℂ) = U(8),
|
||
|
||
and any *two* of {preserves g, preserves ω, commutes with J} imply the third.
|
||
Law 15K is a membership test for U(8) inside SO(16).
|
||
|
||
**Computable gate residual** (replaces `J_squared_identity : Bool`): for a
|
||
candidate transform R,
|
||
|
||
ε_K(R) = ‖RᵀR − I‖_F + ‖RᵀJR − J‖_F, ε_K(R) = 0 ⟺ R ∈ U(8).
|
||
|
||
For *similarities* (rotation combined with φ-contraction), test conformality
|
||
instead, with scale μ = φ⁻²:
|
||
|
||
ε_CK(R) = ‖RᵀR − μI‖_F + ‖RᵀJR − μJ‖_F.
|
||
|
||
"Rotating through all 16 dimensions" is then precise: take
|
||
T = diag(e^{iθ₁}, …, e^{iθ₈}) on ℂ⁸ with rationally independent θ_k.
|
||
By **Weyl's equidistribution theorem** the orbit {Tⁿ} is dense and
|
||
equidistributed on the maximal torus T⁸ ⊂ U(8) — the orbit visits every
|
||
angular sector of all 8 planes with asymptotically uniform frequency.
|
||
|
||
---
|
||
|
||
## 2. The missing rotation operator: the golden spiral map
|
||
|
||
PhiNUVMAP currently scales (`phiContract`) but never mixes components. The
|
||
canonical completion — one formula that is simultaneously the φ-contraction,
|
||
a genuine 16D rotation, and automatically Kähler-compatible — is, on each
|
||
complex plane,
|
||
|
||
S(z) = c + λ·(z − c), λ = φ⁻¹ · e^{iθ_g}, θ_g = 2π·φ⁻² ≈ 137.5078°
|
||
|
||
(θ_g is the golden angle; φ⁻² = 1 − φ⁻¹ = 0.3819660…).
|
||
|
||
Properties, each exact:
|
||
|
||
1. **Contraction law preserved.** |λ| = φ⁻¹, so ‖Sᵗ(s) − c‖ = φ⁻ᵗ‖s − c‖ —
|
||
identical to the existing `phiContractN` law; only the argument advances.
|
||
|
||
2. **Optimal angular coverage.** The argument sequence {t·φ⁻² mod 1}
|
||
equidistributes (Weyl). The **three-distance theorem** (Steinhaus): the
|
||
first N iterates partition the circle into arcs of at most 3 distinct
|
||
lengths. **Hurwitz's theorem**: |φ − p/q| < 1/(√5 q²) has the worst-possible
|
||
constant √5 attained exactly at φ — the golden angle is the *most
|
||
resonance-free* rotation that exists. No periodic lock-in, ever.
|
||
|
||
3. **Kähler compatibility is automatic.** S acts by complex scalar
|
||
multiplication, which commutes with J by construction; equivalently S is
|
||
holomorphic. So ε_CK(S) = 0 identically: the golden spiral *passes the gate
|
||
by theorem*, while any shear or plane-mixing map that breaks the pairing
|
||
fails it with quantifiable residual.
|
||
|
||
4. **Q16.16 error bound.** Per-step rounding error δ ≤ 2⁻¹⁶ per component
|
||
obeys the recursion e_{t+1} ≤ φ⁻¹ e_t + δ, hence
|
||
|
||
e_∞ ≤ δ / (1 − φ⁻¹) = δ · φ² ≈ 2.618 · 2⁻¹⁶ ≈ 4.0 × 10⁻⁵,
|
||
|
||
using the identity 1 − φ⁻¹ = φ⁻². The contraction eats its own rounding
|
||
noise; total fixed-point drift is bounded by φ² ULP for all time.
|
||
|
||
---
|
||
|
||
## 3. The Yang column: a quantized flux line
|
||
|
||
Distinguish one complex plane (say z₀ = d₀ + i·d₁) as the column's cross-
|
||
section; the column "axis" is the remaining 14 dimensions. The rigorous
|
||
identity of the column is a **U(1) vortex / flux tube**:
|
||
|
||
- **Winding number** (must be an *integer*, not Q16.16):
|
||
|
||
Ω = (1/2π) ∮ dθ ∈ ℤ, χ = sign(Ω) (chirality).
|
||
|
||
- **Gauge potential of a straight flux line** with winding n, in the plane
|
||
transverse to the column:
|
||
|
||
A(x, y) = (n/2π) · (−y, x) / (x² + y²),
|
||
∮_C A·dl = n (any loop C encircling the column),
|
||
B = ∇×A = n·δ²(x, y) ẑ (flux concentrated on the column, quantized).
|
||
|
||
This is the corrected `projectPotential`. The current placeholder
|
||
(A₁ = A₂ = A₃ = Ω·χ) forces B ≡ 0 for every TorsionState; the vortex form
|
||
makes winding *source* the magnetic sector, which is its entire job:
|
||
|
||
A₀ = Θ (torsion potential),
|
||
A₁ = −(Ω·χ/2π) · y/(x²+y²),
|
||
A₂ = +(Ω·χ/2π) · x/(x²+y²),
|
||
A₃ = κ (helical pitch; 0 for a straight column).
|
||
|
||
**Integer-native discretization (Wilson / lattice gauge theory).** Put the
|
||
potential on lattice *edges* and curvature on *plaquettes*:
|
||
|
||
F_p = Σ_{e ∈ ∂p} A_e (oriented sum around each plaquette),
|
||
Q = (1/2π) Σ_p F_p ∈ ℤ (total topological charge).
|
||
|
||
This makes `topologicalCharge` in `GoxelFieldFrame` an actual integer
|
||
invariant computed by summation — no real analysis required, fully Q16.16/ℤ.
|
||
|
||
---
|
||
|
||
## 4. Why the Kähler gate implies Maxwell (Law 15K ⟹ 15B/15C)
|
||
|
||
The bridge is one equivalence:
|
||
|
||
**Cauchy–Riemann ⟺ J-equivariance.** A differentiable map f of the plane is
|
||
holomorphic iff its differential commutes with J: df ∘ J = J ∘ df.
|
||
"Kähler-compatible rotation" and "holomorphic motion" are the same condition.
|
||
|
||
Consequence: write the projected potential pair as f = u + iv with f
|
||
holomorphic (u = A₀, v = transverse component). Then
|
||
|
||
∇²u = ∇²v = 0 (harmonic conjugates),
|
||
|
||
and the field E = (∂_x u, −∂_y u) satisfies
|
||
|
||
div E = 0, curl E = 0 — static vacuum Maxwell in the projection.
|
||
|
||
So *smooth (holomorphic) rotation of the column recovers electromagnetism as
|
||
a theorem*, not a metaphor. The failure mode is quantified by the
|
||
**∂̄-residual**:
|
||
|
||
ε_residue = |∂f/∂z̄|², ∂/∂z̄ = ½(∂_x + i·∂_y),
|
||
|
||
which is zero iff f is holomorphic. "Fractal residue" = the L² mass of ∂̄f.
|
||
Fractally folded data has ∂̄f ≠ 0 almost everywhere → rejected, routed to
|
||
`shock/rough_geometry`. The Law 15 chain becomes a logical cascade:
|
||
|
||
∂̄f = 0 (15K) ⟹ harmonicity (15B, 15C) ⟹ quantized coupling (15D via §3).
|
||
|
||
---
|
||
|
||
## 5. Canonical integer geometry: the two 16D even unimodular lattices
|
||
|
||
Dimension 16 is not arbitrary decoration — it is the first dimension with
|
||
*two* even unimodular lattices (the only smaller case is E₈ in dim 8):
|
||
|
||
E₈ ⊕ E₈ and D₁₆⁺ = D₁₆ ∪ (D₁₆ + (½,…,½)).
|
||
|
||
**Theta series.** The space of weight-8 modular forms for SL₂(ℤ) is
|
||
one-dimensional, so both lattices share
|
||
|
||
θ_Λ(τ) = E₄(τ)² = 1 + 480 Σ_{n≥1} σ₇(n) qⁿ, σ₇(n) = Σ_{d|n} d⁷,
|
||
|
||
i.e. the number of lattice vectors of norm 2n is **exactly**
|
||
|
||
r(2n) = 480 · σ₇(n).
|
||
|
||
(Check: n = 1 gives 480 = the 2×240 roots of E₈⊕E₈. ✓ This shared θ with
|
||
non-isomorphic lattices is Milnor's 1964 isospectral-tori example.)
|
||
|
||
**Primes live in the shell counts.** For n > 1:
|
||
|
||
n is prime ⟺ σ₇(n) = 1 + n⁷ ⟺ shell 2n holds exactly 480(1 + n⁷) vectors
|
||
|
||
(composites have strictly more divisors, hence strictly larger shells).
|
||
Primality is literally a *deficiency of geometric mass* on the lattice shell —
|
||
this is the rigorous landing point of the "primes in 16D" intuition.
|
||
|
||
**Exact integer rotations.** Aut(E₈⊕E₈) = (W(E₈) × W(E₈)) ⋊ ℤ₂ with
|
||
|W(E₈)| = 696,729,600. These automorphisms are **integer matrices**: the only
|
||
16D rotations expressible in fixed-point arithmetic with *zero* rounding
|
||
error, forever. The Kähler-compatible exact rotations are Aut(Λ) ∩ U(8).
|
||
|
||
**Snap-to-lattice decoder** (Conway–Sloane, per E₈ factor, O(n) and exact):
|
||
round every coordinate to ℤ; if the coordinate-sum is odd, re-round the
|
||
coordinate with the largest rounding error the other way (gives nearest D₈
|
||
point); repeat for the coset D₈ + (½,…,½); keep the nearer of the two.
|
||
|
||
---
|
||
|
||
## 6. Dimension budget of the chaos game (the horn-fiber claim)
|
||
|
||
The φ-chaos game is an iterated function system with maps
|
||
w_i(s) = c_i + φ⁻¹(s − c_i). By Banach/Hutchinson it has a unique compact
|
||
attractor approached at rate φ⁻ᵗ. With N anchors in general position the
|
||
**Moran equation** N·(φ⁻¹)^D = 1 gives the attractor dimension
|
||
|
||
D = ln N / ln φ.
|
||
|
||
Bulk-filling threshold in 16D:
|
||
|
||
N* = φ¹⁶ ≈ 2206.9995… ≈ L₁₆ = 2207 (Lucas number; φⁿ rounds to Lₙ).
|
||
|
||
- N < 2207 → D < 16: the attractor is a measure-zero scaffold. Rotation
|
||
cannot expand the bulk — it only re-aims which boundary sectors the orbit
|
||
visits. This **proves** the horn-fiber statement quantitatively.
|
||
- N ≥ 2207 → the IFS can have positive 16D measure (overlap regime).
|
||
|
||
Finite volume with unbounded boundary is classical (Gabriel's horn: revolve
|
||
y = 1/x for x ≥ 1; V = π, surface area = ∞). The second-order boundary-sector
|
||
ODE in `bodegaflow_horn_fiber_refinements.md` is a *model postulate*, not a
|
||
derived law; its mathematical obligation is stability — the companion
|
||
polynomial of d²A/dt² = αA + β‖τ‖² + χ d‖τ‖²/dt + γ·RRM must satisfy
|
||
Routh–Hurwitz (all roots in the left half-plane) for bounded sector dynamics.
|
||
|
||
---
|
||
|
||
## 7. The prime anchor (verified numerically, session 2026-06-11)
|
||
|
||
Riemann's explicit formula — the "vector set on an imaginary curve" — is
|
||
|
||
ψ(x) = x − Σ_ρ x^ρ/ρ − log 2π − ½·log(1 − x⁻²),
|
||
|
||
sum over zeros ρ = ½ + iγ of ζ. Each zero is one wave 2·Re(x^ρ/ρ); primes are
|
||
the points of constructive interference. Truncating at 100 zeros resolves
|
||
every prime and prime power below 31 (demo: `/tmp/explicit_formula_demo.py`;
|
||
jump at x detects log p, composites collapse to ≈0). Inside the 16D system,
|
||
primes enter exactly through §5's σ₇ shell counts; per-residue-class structure
|
||
(the 6k±1 helix) is governed by the same formula over Dirichlet L-functions.
|
||
|
||
---
|
||
|
||
## 8. Implementation map (Lean, Q16.16-native)
|
||
|
||
1. **J as data.** `def J16 : Array (Array Int)` (block ε's). Replace
|
||
`KahlerState.J_squared_identity : Bool` with the computed residual
|
||
ε_K(R) / ε_CK(R) of §1 over an explicit 16×16 Q16.16 matrix R.
|
||
`native_decide` witnesses: golden spiral passes ε_CK = 0 (±ULP);
|
||
a shear `[[1,1],[0,1]]` on one plane fails.
|
||
2. **Golden spiral step.** Extend `phiContract` to
|
||
`phiSpiral (s c : Array Q16_16)` applying λ = φ⁻¹e^{iθ_g} per plane
|
||
(fixed cos θ_g, sin θ_g constants; document the φ²·ULP drift bound of §2.4).
|
||
3. **Winding is an Int.** `TorsionState.windingField : Int` (quantization is
|
||
the physics); chirality = its sign, drop the separate field or keep as
|
||
derived.
|
||
4. **Vortex projectPotential.** Implement §3's formula at a sample point;
|
||
theorem: ∃ TorsionState with nonzero discrete plaquette curl
|
||
(kills the current B ≡ 0 degeneracy); theorem: Maxwell residual = 0
|
||
away from the column core.
|
||
5. **Plaquette charge.** `def plaquetteFlux : … → Int` implementing
|
||
Q = (1/2π)Σ F_p; makes `GoxelFieldFrame.topologicalCharge` an integer.
|
||
6. **Optional lattice layer.** `def e8Decode : Array Q16_16 → Array Int`
|
||
(Conway–Sloane); shell-count primality witness
|
||
`σ₇(p) = 1 + p⁷` via `decide` for small p.
|
||
7. **Chaos-game invariant.** Assert/document anchors < 2207 ⟹ scaffold
|
||
regime (D = ln N / ln φ < 16).
|
||
|
||
## Provenance of the named theorems
|
||
|
||
Two-out-of-three: standard Kähler geometry (e.g. Huybrechts, *Complex
|
||
Geometry* §1.2). Weyl equidistribution: Weyl 1916. Three-distance: Sós/
|
||
Surányi/Świerczkowski 1957–58. Hurwitz bound: Hurwitz 1891. Vortex flux
|
||
quantization: Abrikosov 1957 / Aharonov–Bohm 1959. Wilson plaquettes:
|
||
Wilson 1974. Even unimodular classification in dim 16: Witt 1941; isospectral
|
||
consequence: Milnor 1964. E₄² coefficient identity: one-dimensionality of
|
||
M₈(SL₂(ℤ)). Moran equation: Moran 1946; IFS attractor: Hutchinson 1981.
|
||
E₈ decoder: Conway–Sloane, *SPLAG* ch. 20. Explicit formula: Riemann 1859 /
|
||
von Mangoldt 1895.
|