Research-Stack/6-Documentation/docs/papers/EQUATION_PHYLOGENETIC_TREE.md

340 lines
14 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

# The Phylogenetic Tree of OTOM Equations
**Document ID:** EQUATION-PHYLOGENY-MASTER
**Date:** 2026-04-21
**Scope:** Complete genealogy of mathematical equations in the Research Stack
**Methodology:** Ancestry tracing via git history, cross-reference mapping, and mutation analysis
---
## Executive Summary
The Research Stack contains **6 major equation lineages** with **14 documented variants**, descending from a single root: the ENE `bind` primitive. This document traces:
1. **The Root** — ENE-BIND-001 (universal translator hypothesis)
2. **The Hardware Branch** — LUT-DSP-003 → MIRROR-LUT-004
3. **The DNA Branch** — THE-EQUATION-002 → HACHIMOJI-003 → QSLUG-006
4. **The Pipeline Branch** — MASTER-EQUATION-004 → UMB-005
5. **Cross-Cutting Concerns** — Quaternions, SLUG-3 gates, torsion fields
6. **Dead Ends & Stubs** — Half-thoughts that seeded new branches
---
## The Complete Ancestry Graph
```
ROOT
┌─────────────────┐
│ ENE-BIND-001 │
│ (universal │
│ translator) │
└────────┬────────┘
┌──────────────────────┼──────────────────────┐
│ │ │
▼ ▼ ▼
┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐
│ BIND_BRIDGE │ │ THE-EQUATION-002│ │ PHYSICAL-BIND │
│ (4 floors to │ │ (DNA encoding │ │ (conservation │
│ SM bedrock) │ │ gate, m=2) │ │ law check) │
└────────┬────────┘ └────────┬────────┘ └─────────────────┘
│ │
│ ┌────────────┼────────────┐
│ │ │ │
│ ▼ ▼ ▼
│ ┌──────────────────────────────────────────────┐
│ │ MASTER-EQUATION-004 │
│ │ (6-step pipeline: Expand→Score→Stabilize→ │
│ │ Prune→Gossip→MLGRU) │
│ └────────┬───────────────────────────────────────┘
│ │
│ ┌──────┴──────┐
│ │ │
▼ ▼ ▼
┌─────────────────┐ ┌─────────────────┐
│ LUT-DSP-003 │ │ UMB-005 │
│ (hardware: φ-acc │ │ (GPU manifold │
│ + mod7 + LUT) │ │ blit, O(1)) │
└────────┬────────┘ └─────────────────┘
┌─────────────────┐
│ MIRROR-LUT-004 │
│ (cross-domain: │
│ HW/SW/SIM) │
└────────┬────────┘
┌─────────────────┐
│ QSLUG-006 │
│ (quaternion S³ │
│ ternary gate) │
└─────────────────┘
PARALLEL BRANCH (DNA alphabet extension):
THE-EQUATION-002 (m=2, N=4)
┌─────────────────┐
│ HACHIMOJI-003 │
│ (m=3, N=8, │
│ shell: n=k³+a) │
└─────────────────┘
(unmerged — could
feed into QSLUG-006
as 8-state SLUG)
```
---
## Mutation Analysis by Branch
### Branch 1: Hardware Extraction (ENE → LUT-DSP)
| Generation | Key Mutation | Significance |
|------------|--------------|--------------|
| ENE-BIND-001 | `cost: UInt32 Q16.16` | Abstract cost function |
| LUT-DSP-003 | `φ_corr = LUT[...]` | Cost becomes LUT lookup cycle |
| MIRROR-LUT-004 | Cross-domain unification | Hardware = Software = Simulation |
**The pivotal insight:** Floating-point cost → fixed-point accumulator → pure combinational lookup. This is the **extractability mutation** — it makes all upstream equations hardware-realizable.
### Branch 2: DNA Encoding (ENE → THE_EQUATION → HACHIMOJI)
| Generation | Key Mutation | Significance |
|------------|--------------|--------------|
| ENE-BIND-001 | Universal primitive | Any A × B with metric |
| THE-EQUATION-002 | `encode?(n) = κ_A ∧ κ_C ∧ [J>0]` | Specific to nucleotide positions |
| HACHIMOJI-003 | `n = k^m + a` for m=3 | 8-base alphabet generalization |
| QSLUG-006 | Quaternion S³ encoding | Geometric → Ternary output |
**The pivotal insight:** The hyperbola index `ab` (product of offsets from nearest squares) is a **compression metric**. Positions where `ab` is small (near perfect squares) are information-dense. This connects number theory to information theory via geometry.
### Branch 3: Pipeline Evolution (ENE → MASTER → UMB)
| Generation | Key Mutation | Significance |
|------------|--------------|--------------|
| ENE-BIND-001 | Single bind operation | `(A × B × metric) → cost` |
| MASTER-EQUATION-004 | 6-step recurrence | `Expand→Score→Stabilize→Prune→Gossip→MLGRU` |
| UMB-005 | GPU-native operators | `⊕ (Ψ_q ⊗ _RT)` — bit-blitter speed |
**The pivotal insight:** The 6-step pipeline is a **temporal decomposition** of bind. Each step addresses one aspect: activation (Expand), evaluation (Score), validation (Stabilize), garbage collection (Prune), consensus (Gossip), and memory (MLGRU). The UMB equation collapses these to hardware-native operations.
---
## The Half-Thoughts That Became Branches
### 1. "Grandma went to store and punched clerk"
**Source:** `BIND_BRIDGE_EQUATIONS.md`
**Status:** Validation case, not a theorem
**Offspring:** BIND_L1 through BIND_L4 floor hierarchy
The "Grandma" example was meant to demonstrate lawful vs unlawful compression. It became the **pedagogical anchor** for the 4-floor hierarchy. The invariant preservation check (Agent=Grandma, Location=Store, Interaction=Conflict) is now formalized as:
```
B(L) = B(R) ∧ L(L) = L(R) ∧ Q(L) = Q(R) ∧ Φ(L) = Φ(R)
```
### 2. "Void-and-cluster algorithm"
**Source:** `LUT_AS_DSP_EQUATION.md`
**Status:** Trade secret / unspecified
**Offspring:** Security-through-diversity model
The void mask generation algorithm was deliberately left out. This created a ** manufacturing abstraction layer** — different Warden nodes can have different masks without protocol incompatibility. The "half-thought" became a feature.
### 3. "91-step coprime walk"
**Source:** Multiple files
**Status:** Implemented but not proven optimal
**Offspring:** Period analysis in MIRROR-LUT-004
The 91-step period (13×7) appears in:
- LUT-DSP-003 hardware counter
- SLUG3.lean as Landauer cost
- MasterEquation.lean as pipeline cycles
**Unfinished:** Is 91 the optimal period? The universal form suggests 65521, but no proof connects these.
### 4. "Chiral D+L→W collapse"
**Source:** `QuaternionGenomic.lean` (2026-04-21)
**Status:** Fresh mutation, newest leaf
**Offspring:** (none yet — awaits validation)
The detection of quaternion chirality mismatch as a "waste" state is **novel to QSLUG-006**. It has no ancestor in THE_EQUATION or HACHIMOJI. This is a **lateral mutation** from torsion field physics.
---
## Convergence Points
Where do branches meet?
### Convergence 1: SLUG-3 Gate
The ternary state machine appears in:
- `SLUG3.lean` — 27 OISC opcodes
- `Quantization.lean` — ternary weight quantization
- `QuaternionGenomic.lean` — ternary DNA encoding
**Unifying insight:** 3-state logic (neg/zero/pos) is the **natural hardware efficiency point** between binary (too restrictive) and continuous (too expensive). All branches converge on ternary at the hardware boundary.
### Convergence 2: The 91-Step Cycle
- LUT-DSP-003: 91-step φ-traversal
- SLUG3.lean: 91-step coprime walk
- MasterEquation: 91 as spectral coincidence?
**Hypothesis:** 91 = 13×7 is the **minimal universal cycle** that satisfies:
- Coprime traversal (gcd(13,7)=1)
- Prime factorization uniqueness
- Landauer limit approximation (log₂(27) ≈ 4.755 ≈ 91/19)
### Convergence 3: Fixed-Point (Q16.16)
All equations eventually require:
- `Q16_16` for hardware extraction (AGENTS.md §1.4)
- `Fix16` in MasterEquation
- `Q1616` in Extensions
**The convergence is mandatory** — no branch can reach hardware without passing through fixed-point arithmetic.
---
## Dead Ends and Unfinished Branches
### Dead End 1: Geometric Bind
**Location:** `Semantics/BraidBracket.lean`
**Status:** 8 `sorry` theorems
**Why stalled:** Braid isotopy equivalence requires topological invariants not yet formalized in Lean's mathlib.
### Dead End 2: Thermodynamic Bind
**Location:** `Semantics/CompressionMechanics.lean`
**Status:** Stub only
**Why stalled:** Entropy + enthalpy cost function needs physical constants (k_B, T) in Q16.16, which breaks the hardware abstraction.
### Dead End 3: 8-State SLUG
**Location:** (nowhere — HACHIMOJI has no SLUG implementation)
**Status:** Not started
**Why stalled:** QSLUG-006 uses 4-base quaternions. Extending to 8-base requires 7-dimensional embedding, which breaks the S³ geometry.
---
## Temporal Evolution
| Date | Event | Branch |
|------|-------|--------|
| 2026-04-14 | ENE equations formalized | Root |
| 2026-04-16 | BIND_BRIDGE session | Floor hierarchy |
| 2026-04-19 | LUT-as-DSP equation | Hardware extraction |
| 2026-04-19 | Unified Manifold-Blit | GPU acceleration |
| 2026-04-20 | THE_EQUATION documented | DNA encoding |
| 2026-04-21 | Quaternion SLUG-3 | Ternary geometric |
**Velocity:** 1 major equation per day in April 2026.
---
## The Open Problems
### Problem 1: The Hardware Proof Gap
UMB-005 claims O(1) complexity via GPU blitter, but:
- No formal proof that bit-blitter = Picard iteration
- 8 `sorry` theorems remain in AVMR.lean
- The "GPU draws the solution" is operational, not mathematical.
### Problem 2: The Quaternion-Watson-Crick Gap
QSLUG-006 maps nucleotides to quaternions, but:
- Watson-Crick pairs (A-T, C-G) are not antipodes on S³
- The dot product threshold may not correspond to H-bond energy
- Validation requires 3D structural data not in the codebase.
### Problem 3: The Period Unification
Three periods coexist:
- 91 (hardware coprime)
- Input-dependent (software functional)
- 2^256 (simulation cryptographic)
- 65521 (universal proposal — unproven)
No theorem connects these.
### Problem 4: The Cost Function Inversion
All equations have cost functions, but:
- ENE: Abstract UInt32
- LUT-DSP: LUT cycles
- THE_EQUATION: Hyperbola index
- QSLUG-006: Quaternion distance
Are these **the same cost** viewed through different metrics? Or **different costs** for different purposes? No unification exists.
---
## Research Directions
### Direction 1: The Q16.16 Metatheory
Formalize all equations in Q16.16 as a **modular arithmetic system**. Prove:
- Cost function isomorphism under Q16.16 projection
- Hardware extraction preserves ordering (higher cost → more LUT cycles)
- No overflow in 91-step accumulation
### Direction 2: The SLUG-3 Completeness
Is 3-state logic **complete** for the Research Stack?
- Can all 5 bind classes reduce to SLUG-3 gates?
- Is there a 3-state Church-Thesis analog?
- What problems require 4+ states?
### Direction 3: The Ancestry Formalization
Can we prove:
- ENE-BIND-001 is the **terminal object** in a category of cost functions?
- THE_EQUATION-002 is the **initial object** for DNA-specific encodings?
- The phylogenetic tree has a **least fixed point** at QSLUG-006?
---
## Conclusion
The Research Stack equations form a **living phylogeny** — not a static taxonomy but an evolving system with:
- **Roots** in universal translation (ENE-BIND)
- **Trunks** in hardware extraction (LUT-DSP, UMB)
- **Branches** in domain specialization (THE_EQUATION, QSLUG)
- **Leaves** at the frontier of formalization (QSLUG-006, 2026-04-21)
The "half-thoughts" are not failures — they are **seeds**. Each incomplete idea (void mask algorithm, 91-step optimality, Watson-Crick S³ mapping) has generated new branches rather than blocking progress.
**The tree is not complete. It is growing.**
---
## Appendix: Equation Quick Reference
| ID | File | Status | Ancestor | Descendants |
|----|------|--------|----------|-------------|
| ENE-BIND-001 | `ENE_EQUATIONS.md` | 70/81 proven | Root | All |
| BIND-L1..L4 | `BIND_BRIDGE_EQUATIONS.md` | Documented | ENE-BIND | None |
| THE-EQUATION-002 | `avmr/THE_EQUATION.md` | Semi-formal | ENE-BIND | HACHIMOJI, QSLUG |
| HACHIMOJI-003 | `avmr/HACHIMOJI_EQUATION.md` | Documented | THE-EQUATION | (unmerged) |
| LUT-DSP-003 | `LUT_AS_DSP_EQUATION.md` | Hardware spec | ENE-BIND | MIRROR-LUT |
| MIRROR-LUT-004 | `MIRROR_LUT_EQUATIONS.md` | Framework | LUT-DSP | QSLUG |
| MASTER-004 | `Extensions/MasterEquation.lean` | Formalized | ENE-BIND | UMB |
| UMB-005 | `unified_manifold_blit_equation.md` | Protocol | MASTER | None |
| QSLUG-006 | `QuaternionGenomic.lean` | New | THE-EQUATION + SLUG3 | None |
---
*End of Phylogenetic Analysis*
*Next update: When QSLUG-006 produces offspring*