Axiom-0

Collapse is generative.
Only what returns is real.

Collapsus generativus est; solum quod redit, reale est.

The Framework

What is GCD?

Generative Collapse Dynamics is a mathematical framework that measures what survives collapse and what returns. It begins from a single axiom — Axiom-0 — and derives a complete measurement protocol that works identically across 21 scientific domains, from particle physics to consciousness studies.

The framework is implemented as UMCP (Universal Measurement Contract Protocol), a reproducible validation system with frozen parameters that are seam-derived — discovered by the mathematics, not chosen by convention.

At a Glance

20,337
Tests
23
Domains
44
Identities
47
Lemmas
245
Closures
3
Languages

Six Invariants, Three DOF

The Kernel

The kernel function K: [0,1]ⁿ × Δⁿ → (F, ω, S, C, κ, IC) maps any trace vector to six invariants — but only 3 effective degrees of freedom.

Click any invariant to learn what it measures, why it exists, and how it connects to the others.

Primitive — computed directly from the trace vector
Derived — follows algebraically from a primitive

How the Six Invariants Relate

4 are primitive (computed directly), 2 are derived (follow algebraically). Together, they have only 3 effective degrees of freedom because S is asymptotically determined by F and C.

Every Claim Follows This Path

The Spine

Every claim, validation, and discourse follows a fixed five-stop spine. Click each stop to see what happens there.

Derived from Gates, Never Asserted

Three Regimes

Regime classification is derived from gates on (ω, F, S, C) — never asserted. Click to explore each regime.

Fisher Space Partition

12.5%
24.4%
63.1%

Stability is rare — 87.5% of the manifold lies outside it. Return from collapse to stability is what the axiom measures.

Minimal Prose Vocabulary

The Five Words

The minimal prose vocabulary for telling the story of any collapse-return cycle. Click a word to see its ledger role and operational definition.

Immutable Dependency Structure

Three-Tier Architecture

Every symbol, function, and artifact belongs to exactly one tier. Click to understand each layer.

One-Way Dependency (No Back-Edges)

Tier-1 Tier-0 Tier-2

No feedback from Tier-2 to Tier-1 or Tier-0 within a frozen run.

Compounding Parametrized Sweeps

20,337 Tests

Not 20,337 separate assertions — compounding parametrized sweeps across 23 domains. ~800 base test functions expand 18× through domain × entity × configuration parametrization.

How Test Compounding Works

~800
Base test functions
Written once per pattern
~18×
Parametrization factor
Domain × entity × config
20,337
Collected test items
Every invariant, every domain

Example: One Tier-1 identity test compounds across 23 domains

# One parametrized test function:
@pytest.mark.parametrize("domain", ALL_20_DOMAINS)
def test_duality_identity(domain):
traces = domain.get_all_entities() # 12–118 entities each
for trace in traces:
F, ω = kernel(trace)
assert abs(F + ω - 1) == 0.0 # Exact to 0.0e+00
→ 1 function × 23 domains × 12–118 entities = hundreds of assertions from one test

Test Distribution by Domain

Click a tier to see its test domains.

231 test files, numbered test_000 through test_338

Proto
Core
Domains

Verified to Machine Precision

44 Structural Identities

All derived from Axiom-0 and verified to machine precision. They fall into four series. Click a series to see its identities and what they prove.

6 Connection Clusters

The 44 identities form a connected network. These clusters reveal the deep structure.

Same Formulas, Three Implementations

Three Languages, One Kernel

The same kernel function K: [0,1]ⁿ × Δⁿ → (F, ω, S, C, κ, IC) is implemented in three languages — each serving a distinct architectural role. Click to explore each layer.

Three-Layer Sandwich Architecture

Python Domain closures · CLI · Dashboard · Fleet
C++17 Types · Validation · pybind11 zero-copy
C99 Raw math · Kernel · Seam · Pipeline

Same formulas, same frozen parameters — three layers of the same truth

Formal Foundation

47 Lemmas

The formal foundation beneath the kernel. Each lemma proves a property used by the implementation — tagged as OPT-* in the code. Click a category to explore.

Lemmas in Code

Every optimization in the kernel references its proving lemma via OPT-* tags:

# OPT-1 (Lemmas 4, 10, 15): Homogeneity fast path — 40% speedup
# OPT-2 (Lemma 1): Range validation — instant bug detection
# OPT-3 (Lemmas 4, 34): Heterogeneity gap analysis
# OPT-4 (Lemmas 2, 3): Log-space κ computation — 10% speedup + stability
# OPT-12 (Lemmas 23, 30): Lipschitz error propagation

Every speedup has a proof. Every optimization has a lemma number. Nothing is ad hoc.

Jump To

Explore