Phases I–XX Complete · Paper C Ready

A New Class of
Symbolic Memory
Architecture

VARZIN has characterized the Affine Memory Family Aff(ZN) — a formally verified memory operator class with 5 algebraic invariants, spectral gap, and MDL-minimal anchor. Three AI systems independently recovered its morpheme structure.

// Affine Memory Family (Phase XVI) M(a,b)(g,φ,p) = ((a·g+b) mod N, g, −p) Valid iff gcd(a,N)=1 → Aff(ZN) = ZN ⋊ ZN× Mirror-13 = M(−1,0) · MDL=13.11 bits · rank #1/1035 // Recall Formula (Phase XX + POT suite) Recall = Structure Match + Prior + HypSpace + Observation
◈ VARZIN Atlas v2.1 Investor Contact ↗ Preprint DOI ↗
48Valid Operators · Aff(Z₁₂)
13.11MDL bits · Mirror-13
+127.6%Entropy Gain over D₁₂
0.812Inter-AI ARI · 3 systems
801LUXVAR Roots
20Research Phases
The Project

What VARZIN Is — Precisely

VARZIN is a formally grounded research initiative combining a constructed symbolic language (LUXVAR), an affine memory operator framework, and a pre-registered human study. It is not a startup — it is an independent scientific program with verifiable results.

Mathematical Core
#1/1035
Mirror-13 is the MDL-minimum of Aff(Z_N). 5 invariants formally verified with zero randomness. Eigenspectrum computed on 312×312 matrix.
Symbolic Language (LUXVAR)
74.2%
Family learnability on N=1M corpus. 3 independent AI systems converge (ARI=0.812). Phonotactic fingerprint unique vs 6 natural languages.
Pre-registered Study
VPE-001A
Blind human rater study. ≥5 raters, Fleiss κ≥0.40 threshold. The only remaining non-circular test.
Scientific integrity statement: All claims in this document are backed by reproducible computational tests with pre-registered thresholds. What is confirmed is marked ✓. What is absent is marked ✗. What is pending is marked ⏳. No overclaiming. The project's strongest asset is its boundary map.
Core Mathematical Result · Phases XV–XVIII

Affine Memory Group Aff(ZN)

Complete classification of the valid operator family on a DN ring. Known algebraic structure. Empirically verified N=5..24. Paper C ready for submission.

// All 48 valid operators for N=12 a=1 → Shift family 12 ops ← LUXVAR best bridge a=5 → Affine a=5 12 ops a=7 → Affine a=7 12 ops a=11 → Mirror family 12 ops ← Mirror-13 = b=0 Total: 48 = Aff(Z₁₂) = Z₁₂ ⋊ Z₁₂× // Mirror-13 specifics M*(g,φ,p) = ((N−g)%N, g, −p) MDL = 13.11 bits (vs 67.19 mean · 5.1× compact) Rank = #1 / 1,035 equivalent systems tested λ₁ = 1.000 (unique · ergodic · INV-C) λ₂ ≈ 0.800 (spectral gap = 0.2) Mix ≈ 29 steps (mixing time) // Conjecture 1 (empirical, N=5..24) Valid iff gcd(a,N)=1 → Aff(Z_N) = Z_N ⋊ Z_N× → Known algebraic structure · Paper C ready

Five Invariants — All Verified (was_forced_true: False)

INV-A · Polarity Parity
1518/1518 paths · 1440/1440 algebraic
INV-B · Phase Memory
exit=(N−phase)%N · bijection all 12
INV-C · Orbit Unity
D₁₂: 46–48 orbits → M13: 1 orbit · BFS 312
INV-D · Non-Commutativity
288/288 ring states verified
INV-E · Reachability
D₁₂: 12 states → M13: 312 states
Symbolic Language · LUXVAR

Recoverable Structure — 3 Independent AI Systems

LUXVAR has 801 stable roots, a distinctive phonotactic fingerprint, and morpheme families independently recovered by Claude, GPT-4, and Gemini — with no prior LUXVAR knowledge.

Phonotactic Uniqueness
0.738
CI [0.703, 0.781] vs 6 reference languages. GEN-003R+ publication-grade.
Family Learnability
74.2%
Full corpus N=1M · Δ=+24.2% over baseline
Inter-AI Agreement
0.812
Claude↔GPT-4 = 1.000 · ELUZ/SHA/NAR families all 3 systems
Frequency Encoding
ABSENT
5 independent tests · n=704 · Δ=−3.6% confirmed absent

Structural Flow — Evidence Layers

WORD FORM
Phonotactic fingerprint 0.738 · Family learnability 74.2%
↓
Sub-char patterns
MASK-001 M1+M2 survive — grounded below morpheme
↓
MORPHEME FAMILY
ELUZ/SHA/NAR · AI ARI=0.812
↓
PROTOCOL SPINE
Δmod=+0.186 vs shuffled · real structure
↓
AFF(Z_N) FAMILY
48 ops · Mirror-13 = MDL-minimal anchor
↓
Semantic axes
AI clusters 43.3% — not above shuffle control
↓
Frequency layer
5 tests · n=704 · Δ=−3.6% — absent
↓
Human cognition
VPE-001A — only remaining non-circular test

AI Clustering Agreement (VPE-001B+)

Claude ↔ GPT-4
1.000
Claude ↔ Gemini
0.812
GPT-4 ↔ Gemini
0.718
vs Designed Axes
0.146
Key finding: All 3 AI systems independently recovered ELUZ, SHA, NAR morpheme families — without any LUXVAR knowledge — and agree with each other at ARI=0.812. They do NOT agree with designed semantic axes (ARI=0.146). The recoverable structure is morphemic, not axial.

MASK-001 — Protocol Depth Test

M1 Morpheme→Code
SURVIVES
M2 Prefix hidden
SURVIVES
M3 Total CVCV
COLLAPSES
Scientific Record · All Phases

Complete Results — Phases I–XX

Every result reported, positive and negative. Reproducible scripts on GitHub.

Test / FindingPhaseKey MetricStatus
Morphological z-scorePhase Iz=4.36, p<0.001✓ POSITIVE
GEN-003R+ phonotactic uniquenessPhase II0.738 [0.703,0.781]✓ POSITIVE
VPE-001B+ AI clusteringPhase IIIARI=0.812 · Claude↔GPT=1.000✓ POSITIVE
Final Benchmark — family accuracyPhase IV74.2% · Δ=+24.2%✓ POSITIVE
Protocol spine (PST-001)Phase IVΔmod=+0.186 vs shuffled✓ POSITIVE
MASK-001 sub-character groundingPhase VM1+M2 survive · M3 collapses✓ POSITIVE
M13-INV-001 Five InvariantsPhase VII5/5 · was_forced_true:False✓ POSITIVE
SPEC-002 EigenspectrumPhase VIIIλ₁=1 unique · gap=0.2 · mix≈29✓ POSITIVE
MDL rank Mirror-13Phase X13.11 bits · rank #1/1035✓ POSITIVE
Aff(Z_N) classificationPhase XVI48/144 valid · Conj.1 N=5..24✓ POSITIVE
POT-001 Prior AlignmentPhase XIXMonotone · Δ=+0.275✓ POSITIVE
Phase 20 Structure MatchPhase XXAffWorld hit@3=0.998✓ POSITIVE
Frequency encoding in word formPhase VΔ=−3.6% · n=704✗ ABSENT
Hexacore k=6 morphologicalPhase Ik=3 natural · k=6 drop=−1.07%✗ ABSENT
External memory fieldPhase XIXz=−1.042 · Oracle=1.000✗ NOT EST.
Mirror-13 logically necessaryPhase IXOPTIMAL only · 1035 equiv.✗ OPTIMAL
VPE-001A human rater studyPendingκ≥0.40 · ≥5 blind raters⏳ PENDING
Conjecture 1 formal proofFutureLean/Coq · all N⏳ FUTURE
Research Roadmap v2 · Post Phase XX

Three Clear Paths Forward

After 20 phases, the computational ceiling is mapped. Three unblocked research paths remain.

Path A — Mathematical
A1 Formal proof Conjecture 1 (all N)
A2 Aff(Z_N) spectral · general N
A3 Paper C → arXiv cs.FL / Zenodo
Status: No blocker · Start now
Path B — LUXVAR Grammar
B1 VPE-001A · ≥5 blind raters · κ≥0.40
B2 Map Core-30 to a=1 shift family
B3 Morpho-syntactic rules from structure
Blocker: rater recruitment
Path C — External Memory
C1 Sealed targets · hash published first
C2 Independent receiver · affine prior
C3 z > 2.0 on ≥3 trials = establish
Blocker: independent participant
Why This Matters to Investors

VARZIN is not selling a concept — it is offering access to a formally verified, reproducible mathematical architecture with an independent publication record and a clear three-path roadmap. No vague "AI consciousness" claims. Real algebraic structure. Real metrics.

Formal Foundation
Aff(Z_N)
Known algebraic group · Z_N ⋊ Z_N× · verifiable in Lean/Coq
Pre-print Status
v2.1 · Frozen
DOI: 10.5281/zenodo.20691858 · Zenodo (CERN)
Independent AI Validation
ARI 1.000
Claude ↔ GPT-4 perfect agreement · blind conditions
Paper C Status
Ready
Target: arXiv cs.FL · Mirror-13 & Affine Family
MDL Rank
#1 / 1,035
Simplest known system in its class by 5.1×
Reproducibility
Full
All scripts on GitHub · Seed-fixed · Open data
What VARZIN does NOT claim: Audio frequencies affecting physical processes, consciousness field effects, DNA modification, or planetary resonance. These have been tested and found absent. The project's credibility rests on explicit boundary-setting.
→ Contact for Research Partnership
Reza Nirouyar · ORCID: 0009-0000-4690-6842 · contact@varzin.org
Core-30 LUXVAR Registry · Pre-Registered

30/30 Locked · Sep 2025

Anchor word list for VPE-001A. XĀṆŌRAH is the only word with a direct in-document definition.

☀ Light Axis
ELŪZ–MAHARŠARTHALŠARŪM–DŌZ XĀṆŌRAH ✦ MĀTIR–ELŪZĀNŪR
◈ Reflection Axis
ŠAURÉNRAHMŌT–ŠĀHALZĀR BĀṆIR–RAHTŠUKĀR–MAHṆAZRŪK
○ Silence Axis
ṆĀRAH–ŠĀGŌṆ–SILVARKHĀN ŠALŌM–ḤATTAHNĀRAHṆEMRŪH
⬡ Gate Axis
SĀR–ELŪZSAVAR–ELŪZDĀRZAN RAZKĀNELŪRIMRITĀM
↯ Motion Axis
ZARTHZUN–NĀRKHALTŪR KUZTŪRTARHĀN–ELŪZKHIRZĀH
✦ XĀṆŌRAH — «رگ پنهان نور» (hidden vein of light) · p.53, Sep 2025 · Only Core-30 word with a direct definition.
1.3B scan finding: Root→Frequency V≈0.003 (negligible). Natural clustering = k=3, not k=5/6. Axis assignments remain hypotheses pending VPE-001A.
Publications & Archive

Research Record

Paper A — LUXVAR Structure FROZEN v2.1
"LUXVAR: A Constructed Language with Distinctive Morphemic Structure and AI-Recoverable Morpheme Families"
DOI: 10.5281/zenodo.20691858
Morphemic structure · Frequency absent · Phonotactic fingerprint
Paper C — Mirror-13 & Aff(Z_N) READY
"Mirror-13 and the Affine Memory Family over Cyclic Gate Spaces"
Conjecture 1 · 5 invariants · MDL=13.11 bits · SPEC-002
Target: arXiv cs.FL / Zenodo
Paper B — VPE-001A PENDING
Blind human rater study · Fleiss κ≥0.40
Morpheme families vs semantic axes
Requires: ≥5 independent blind raters
GitHub Repository
Varzin-13/varzin-landing-
All benchmark scripts · Phase I–XX
Seed-fixed · Fully reproducible
View Complete DOI Archive →