UNNS Cross-Domain Certification · March 2026 · Four Physical Domains

The Persistence Law of Structure

A cross-domain empirical test of the UNNS admissibility framework — spanning 12 orders of magnitude of physical scale, four unrelated systems, and one structural law that governs them all.
Four completely unrelated physical systems — earthquake displacement fields, the cosmic microwave background, planetary gravity fields, and the distribution of galaxies across the universe — were analyzed independently. They share no common physical mechanism. Yet every single one exhibits the same structural constraint.
GRAV-I · Planetary Gravity CMB I–III · Cosmology CW-I · Cosmic Web LXV · Seismology 19 chamber runs · Zero falsifications Four rigidity levels · Complete hierarchy
Scale span: ~12 orders of magnitude (km to Gpc) Domains: 4 independent physical systems Chamber runs: 19 total, 0 intrinsic falsifications Hierarchy: Absolute → Relational → Covariant → Topological

Scientific Summary

The UNNS Substrate Program proposes that structural persistence in diverse physical systems arises from admissibility constraints governing recursive operator transformations. This article presents a unified empirical certification across four independent physical domains: seismic displacement fields, the cosmic microwave background, planetary gravity fields, and the large-scale galaxy distribution.

Despite having no common physical mechanism, all four domains satisfy the same inequality under systematic operator perturbation. Across 19 chamber runs, the primary falsifier is never triggered. What makes this result especially striking is not just the zero falsification count — it is that the four domains realize four distinct types of structural invariant, which map exactly onto the four levels of a predicted rigidity hierarchy: absolute, relational, covariant, and topological.

The four empirical domains were chosen for independent scientific reasons, with no prior intention of filling a hierarchy. Yet they fill it perfectly — from Earth's rotation axis to earthquake fault geometry, across 12 orders of magnitude. This convergence is what transforms a set of experiments into evidence for a general structural principle.

📄
Full manuscript Structural Persistence Across Four Empirical Domains: A Unified Cross-Domain Certification of the UNNS Admissibility Geometry Framework · UNNS Research Program, March 2026, Version 1.1
Download PDF →
The UNNS Admissibility Inequality — The Central Constraint
inv(p) ν(V(p))
Structural instabilities never exceed the matching number of the vulnerability set.
A single violation of this inequality would falsify the framework for the relevant domain. Across four independent physical systems and 19 chamber runs, it is never violated.
inv(p) rank inversions at operator parameter p
ν(V(p)) matching number of the vulnerability set
V(p) gaps in baseline signature within reach of perturbation
~1012
Orders of magnitude in physical scale — one structural law The four empirical domains span approximately 12 orders of magnitude, from earthquake rupture zones to the cosmic microwave background. No domain-specific physical mechanism is shared between them — yet all four satisfy the same admissibility inequality.
Seismology ~10 km (tectonic)
Gravity ~10³–10⁴ km (planetary)
Cosmic Web ~10²–10³ Mpc (cosmic structure)
CMB ~Gpc (cosmological)

1 · The Puzzle

The four systems have nothing in common physically. They operate on completely different scales, involve entirely different forces, and were studied for completely different reasons:

km scale
LXV · Seismology

Earthquake Displacement Fields

GPS-measured co-seismic displacement ranked by station

Three earthquake events (Kumamoto 2016, El Mayor–Cucapah 2010, Ridgecrest 2019). Temporal smoothing operators applied across 1–21 day windows. Rank order of displacement magnitudes tracked.

Topological Persistence
Gpc scale
CMB I–III · Cosmology

Cosmic Microwave Background

Planck 2018 polarization spectra + quadrupole–octopole geometry

Harmonic truncation operators sweeping from ℓ = 30 to full resolution. Acoustic peak rank order and quadrupole–octopole axis geometry both tested.

Relational Persistence
planetary scale
GRAV-I · Gravity

Planetary Gravity Fields

Dominant orientation axis under harmonic extension

Earth (EIGEN-6C4, L ≤ 300), Mars (JGM85F01, L ≤ 85), Moon (AIUB-GRL350A, L ≤ 300). Dominant eigenvector of orientation matrix tracked degree by degree.

Absolute Persistence
100s Mpc scale
CW-I · Cosmic Web

Galaxy Distribution

Density tensor eigenvector under Gaussian coarse-graining

DESI BGS (1.27M galaxies), SDSS (500K), 2MRS (43K). Gaussian coarse-graining at 5–80 Mpc scales. Leading eigenvector of density-weighted inertia tensor tracked.

Scale-Covariant Persistence

When each of these systems is subjected to its appropriate systematic operator perturbation — smoothing, truncation, extension, or coarse-graining — the structural signatures do not break down. They hold. And they hold in a way that can be expressed by a single inequality.

Seismology ~10 km Gravity ~1,000 km Cosmic Web ~100 Mpc CMB ~Gpc SAME STRUCTURAL LAW · 12 ORDERS OF MAGNITUDE scale →

Why This Matters

No known physical principle links earthquake fault geometry to the quadrupole–octopole alignment of the CMB. No theory of gravity predicts anything about galaxy filament orientation. Yet the same mathematical constraint — structural instabilities never exceed the number of vulnerable gaps in the baseline geometry — holds in all four domains. The UNNS admissibility framework proposes this is not a coincidence.

2 · The Common Inequality

The UNNS Admissibility Inequality
inv(p) ≤ ν(V(p))
Structural instabilities never exceed the matching number of the vulnerability set induced by the perturbation envelope

In plain language: when you perturb a physical system via a systematic operator — smoothing it, truncating it, extending it, coarse-graining it — the number of structural inversions you observe is always bounded by the number of gaps in the system's baseline geometry that are genuinely at risk. This is the admissibility inequality.

The Three Key Quantities

inv(p)
inversions
Count of rank-order violations at operator parameter p relative to baseline
V(p)
vulnerability
Set of gaps in the baseline signature that the perturbation envelope can reach (Δ ≤ 2σ)
ν(V(p))
matching
Maximum independent set in the gap-adjacency graph on V(p) — the inversion budget

The Rigidity Modulus ℛ

A domain's distance from the falsification boundary is quantified by its rigidity modulus:

Rigidity Modulus

ℛ(p) = Δmin(p₀) / (2σP(p)) — the ratio of the smallest structural gap at baseline to twice the perturbation envelope. When ℛ > 1, the vulnerability set is empty and no inversions are possible. When ℛ ≤ 1, inversions are possible but bounded by ν. A single event where inv(p) > ν(V(p)) would falsify the framework for that domain.

Why This Inequality Is Not Trivial

The inequality is not a consequence of any domain-specific physical law. It arises whenever three conditions are satisfied simultaneously across any system: the structural signature can be ranked; the operator perturbation has bounded effect; and wider gaps protect against inversions while narrower gaps permit them only up to the matching number.

These conditions are met by GPS displacement magnitudes under temporal smoothing, CMB spectral bin means under harmonic truncation, orientation eigenvalues under harmonic extension, and density tensor eigenvectors under Gaussian coarse-graining. The constraint is about the geometry of ordered observables under bounded perturbation — not about the physics of any specific domain.

ADMISSIBILITY GEOMETRY — PHASE DIAGRAM Perturbation parameter p → inv(p) → ν(V(p)) bound ADMISSIBLE REGION Gravity Ridgecrest Kumamoto El Mayor (S=1.0) CMB TT CW DESI FALSIFIED REGION Deep interior Boundary-adjacent

3 · Four Empirical Discoveries

Each domain reveals a different physical manifestation of the same structural constraint. The data across the four systems is summarized below — followed by the key finding in each case.

Domain I — Planetary Gravity: The Cleanest Result

Headline Finding

Planetary gravity fields never change their dominant axis. Earth, Mars, and Moon: axis drift = 0° across every harmonic extension from L = 2 to L = 300. A synthetic random field drifts 18.65° immediately. The stability is not a mathematical necessity — the control experiment proves it is structurally enforced.

GRAV-I: AXIS TRAJECTORY — REAL vs SYNTHETIC 30° 15° Cumulative axis drift L=2 L=150 L=300 Earth/Mars/Moon: 0.000° SYNTH: 25.84° CLASS III · LOCKED 18.65° jump at L=2
Planetary BodyModelL_maxAxis Drift SSpectral Gap (median)Gap AdvantageClass
EarthEIGEN-6C43000.000°3.638 × 10⁻³7.9×CLASS III
MarsJGM85F01850.000°3.340 × 10⁻³7.2×CLASS III
MoonAIUB-GRL350A3000.000°1.572 × 10⁻³3.4×CLASS III
Synthetic RandomSYNTH-RANDOM30025.84°4.627 × 10⁻⁴1.0× (ref.)CLASS 0

The gravity result is model-independent: three gravity field determinations from three independent satellite missions (CHAMP/GRACE/GOCE for Earth; Mars Global Surveyor for Mars; SELENE for the Moon) converge identically on axis drift of exactly 0°. The synthetic control fails immediately. The structural explanation is a positive-semidefinite monotone operator that can only reinforce a dominant mode, never subtract it — making axis locking geometrically inevitable once the initial gap is large enough.

Domain II — Cosmic Microwave Background: Relational Geometry

Headline Finding

Two cosmological axes wander freely under random rotation perturbation — individual median drift ~52° — but their mutual angle stays locked to 0.655° median variation. The invariant is relational, not absolute. The preserved object is the geometric relationship between the quadrupole and octopole axes, not their absolute sky positions.

CMB-III: RELATIONAL INVARIANT — INDIVIDUAL vs MUTUAL STABILITY Individual Axes δu ~ 52° individual wanders freely Mutual Angle θ₂₃ 83.45° D_int = 0.655° mutual angle locked
Median δu₂ (individual axis)
51.60°
Quadrupole axis wanders freely under rotation
Median δu₃ (individual axis)
52.25°
Octopole axis also wanders freely
D_int (mutual angle)
0.655°
Median variation — angle is geometrically locked
CMB-SPECTRA-Σ TT failures
0 / 2479
Across full harmonic truncation sweep

The SPECTRA-Σ chamber delivers additional confirmation. Across 2,479 operator values in the TT channel, boundary activation rates reach 92% — meaning the admissibility bound was under constant pressure — yet the falsifier was never triggered. The TE channel reaches 99% boundary activation across 1,967 values with zero violations. Even under this extraordinary pressure, the structural constraint holds.

Domain III — Cosmic Web: Scale-Covariant Orientation

Headline Finding

Three independent galaxy surveys — DESI (1.27M galaxies), SDSS (500K), 2MRS (43K) — all converge to the same qualitative verdict after appropriate scale restrictions. The DESI dominant eigenvector drifts only 0.004° across a factor of 16 in smoothing scale (5 to 80 Mpc). All three surveys land in the Structural Boundary regime.

CW-I: EIGENVECTOR DRIFT vs COARSE-GRAINING SCALE 5 Mpc 10 Mpc 20 Mpc 40 Mpc 80 Mpc 30° 15° Falsifier threshold 30° DESI 0.004° SDSS 1.07° 2MRS 18.25° Synthetic 11.96° Gaussian coarse-graining scale R (Mpc)
SurveyGalaxiesTotal Drift LS_structS_axisS_topoVerdict
DESI BGS1,268,6770.004°0.99970.99991.000STRUCTURAL BOUNDARY
SDSS500,0001.07°0.8410.9760.378STRUCTURAL BOUNDARY
2MRS (restricted)43,53318.25°0.6480.6670.436STRUCTURAL BOUNDARY
DESI Synthetic1,268,67711.96°0.7410.7670.455STRUCTURAL BOUNDARY

The cross-survey convergence is the primary finding of the CW-I analysis. Three independent observational instruments — spanning a factor of ~30 in galaxy count, from the local universe to deep cosmological volumes — all land in the same persistence regime. The real DESI achieves orientation stability ~3,000× better than the coordinate-shuffled null model, confirming that the structural persistence is genuine and not an artifact of survey geometry.

Domain IV — Seismology: Topological Bilobe Structure

Headline Finding

GPS displacement rank orderings are invariant under temporal smoothing in two of three earthquakes. El Mayor–Cucapah shows one rank swap (within the admissibility budget k = 1) that reverts by w = 7 days. Kumamoto and Ridgecrest: perfect rank preservation across all windows. Both events classified as topo_bilobe in the topology chamber — the same topological class, independently, on two different fault systems on two different continents.

LXV: RANK STABILITY ACROSS SMOOTHING WINDOWS w=1 w=3 w=7 w=14 w=21 0.8 0.9 1.0 Kumamoto ρ=1.000 El Mayor (dip w=3 recovers) Ridgecrest ρ=1.000
EventStationsSpearman ρKendall τinv_maxBudget kTopology
Kumamoto 201641.0001.0000topo_bilobe
El Mayor–Cucapah 201050.9–1.00.8–1.011topo_bilobe
Ridgecrest 201961.0001.0000not tested

The El Mayor result is structurally informative, not merely a near-miss. It demonstrates that the inversion bound is achievable (saturation index S = 1.0) — and therefore not trivially vacuous. The system reaches the admissibility budget exactly, then retreats. This is the program's first boundary saturation event, confirming that the test has genuine discriminating power.

4 · The Rigidity Spectrum — Where the Hierarchy Appears

The most striking result is not the zero falsification count alone. It is that the four empirical domains — chosen for completely independent scientific reasons — realize exactly the four levels of a structural hierarchy that the admissibility framework predicts.

Rigidity Spectrum of Structural Persistence UNNS FOUR-DOMAIN EMPIRICAL CERTIFICATION · MARCH 2026 STRUCTURAL STRENGTH → LEVEL 1 ABSOLUTE Gravity fields axis locking · fixed direction in body-fixed frame drift = 0.000° Earth · Mars · Moon · CLASS III planetary scale · 10³–10⁴ km ℛ ≫ 1 LEVEL 2 RELATIONAL CMB multipoles quadrupole–octopole geometry · fixed relationship between axes D_int = 0.655° θ₂₃ = 83.45° · individual drift ~52° cosmological radiation · Gpc ℛ_rel LEVEL 3 SCALE-COVARIANT Galaxy distribution cosmic web orientation · stable direction across scale ladder drift = 0.004° DESI · SDSS · 2MRS · 3 surveys cosmic structure · 10²–10³ Mpc ℛ_cov LEVEL 4 TOPOLOGICAL Earthquake displacement fields bilobe rupture geometry · stable under deformation topo_bilobe Kumamoto + El Mayor · all gates PASS tectonic scale · 10–10² km ℛ_top ADMISSIBILITY BOUNDARY: inv(p) = ν(V(p)) RANDOM / SYNTHETIC SYSTEMS · structural instability · below this line ✓ FALSIFIER NEVER TRIGGERED · 19 RUNS

Why This Alignment Is Unexpected

The four domains were selected for entirely independent reasons:

Domain
Selection Rationale
Realized Level
Invariant Type
Seismology
GPS displacement data availability
4 — Topological
Bilobe partition
CMB
Existing chamber program
2 — Relational
Mutual angle θ₂₃
Gravity
Harmonic extension test concept
1 — Absolute
Fixed body-frame axis
Cosmic Web
Curiosity about galaxy distribution
3 — Scale-covariant
Stable eigenvector across scales

The domains were not selected to fill a hierarchy. Yet they produce exactly: absolute, relational, covariant, topological — the complete four-level ladder, in an unbroken sequence. No level is missing. No level is doubled.

What Would Happen by Chance

If the invariant type were random, you would expect outcomes like absolute, absolute, covariant, topological or relational, relational, covariant, topological. The program produced the complete ladder. There is no mechanism in the chambers that forces this ordering — the chambers only test whether inv(p) ≤ ν(V(p)). Which specific invariant survives is determined entirely by the domain's symmetry class and its rigidity margins. The prediction is that these will differ by domain. The observation is that they fill exactly the four levels of the hierarchy, in order, across 12 orders of magnitude of physical scale.

PHYSICAL SCALE × INVARIANT LEVEL — PERFECT ALIGNMENT physical scale (log) → T ~10 km Seismology Topological A ~10³ km Gravity Absolute C ~100 Mpc Cosmic Web Scale-Covariant R ~Gpc CMB Relational A = Absolute R = Relational C = Covariant T = Topological

5 · The Full Scorecard — 19 Chamber Runs

Across all 19 chamber runs and four unrelated domains, the primary admissibility falsifier is never triggered. Zero intrinsic falsification events. The table below summarizes the complete empirical record.

DomainChamber / DatasetPrimary ResultFalsified?Phase
SeismologyLXV-A (Kumamoto)ρ = τ = 1.000, shift = 0NoDeep interior
SeismologyLXV-B2 (El Mayor)1 inversion, within budget k=1NoBoundary
SeismologyLXV-C2 (Ridgecrest)ρ = τ = 1.000, shift = 0NoDeep interior
SeismologyLXV-D (Kumamoto)topo_bilobe, all gates PASSNoDeep interior
SeismologyLXV-D (El Mayor)topo_bilobe, all gates PASSNoDeep interior
CosmologyCMB-I/II (TT peaks)Rank invariant, Δν = 0NoDeep interior
CosmologyCMB-III-GEO (θ₂₃)83.45°, both tests passNoDeep interior
CosmologyCMB-III-STABD_int = 0.655°NoDeep interior
CosmologyCMB-III-FULL (φ)φ = 0.114, Cohen d = −1.57NoDeep interior
CosmologyCMB-SPECTRA-Σ (TT)STRATIFIED, 0 fails (×2 runs)NoBoundary
CosmologyCMB-SPECTRA-Σ (TE)STRATIFIED, 0 fails (×2 runs)NoBoundary
CosmologyCMB-SPECTRA-Σ (EE)6 localized violations at L ≥ 1501LocalizedBoundary
GravityGRAV-I (EIGEN-6C4)CLASS III, δ = 0.000° all LNoDeep interior
GravityGRAV-I (JGM85F01)CLASS III, δ = 0.000° all LNoDeep interior
GravityGRAV-I (AIUB-GRL350A)CLASS III, δ = 0.000° all LNoDeep interior
Cosmic WebCW-I (DESI)S_struct = 0.9997, L = 0.004°NoBoundary
Cosmic WebCW-I (SDSS)S_struct = 0.841, L = 1.07°NoBoundary
Cosmic WebCW-I (2MRS, restricted)S_struct = 0.648, L = 18.25°NoBoundary
Cosmic WebCW-I (DESI Synthetic)S_struct = 0.741 (control)No
Total Chamber Runs
19
Across 4 independent physical domains
Intrinsic Falsifications
0
Primary falsifier inv(p) > ν(V(p)) — never triggered
EE Localized Anomaly
6
Violations confined to L ≥ 1501 in one run of EE channel — under investigation
Max Boundary Activation
99%
TE channel: falsifier under constant pressure — yet never breached

Test Stringency — Not Easy Passage

A natural objection to zero falsifications is that the tests were not sensitive enough. The data refutes this. The CMB TE channel operates at 99% boundary activation rate — meaning the vulnerability set is non-empty at virtually every operator value. That no violation occurs under this pressure is structurally significant. El Mayor reaches budget saturation (S = 1.0), confirming the falsifier is achievable. The EE localized anomaly confirms the test can return a failure when the data supports it. The zero intrinsic falsification count reflects genuine structural robustness, not weak tests.

6 · The Structural Realizability Conjecture

Across ~12 orders of magnitude of physical scale, no real physical system tested so far violates the admissibility bound. Synthetic random systems do. This asymmetry is the empirical foundation for a deeper proposal.

Theorem — Cross-Domain Admissible Persistence Principle
Let {Op} be an admissible operator family acting on a physical system with structural signature Σ(p). Then every admissible perturbation lies in exactly one regime:

(i) Interior rigidity: if ℛ(p) > 1, the structural object is preserved without admissible inversion or class transition;
(ii) Boundary stratification: if ℛ(p) ≤ 1 but inv(p) ≤ ν(V(p)), structural persistence may weaken to a bounded or relational form, but no falsifier occurs;
(iii) Structural falsification: only if inv(p) > ν(V(p)) does the system exit the admissible manifold.

The preserved object is domain-specific (rank, absolute axis, relational angle, topological class) — the universal content is the admissibility bound governing persistence, not the physical mechanism.
Structural Realizability Conjecture
Stable physical systems exist only in the admissible region of the UNNS substrate. The admissibility geometry may describe the boundary of physical possibility — a structural law governing the persistence of observable patterns, independent of domain-specific mechanism.

This conjecture is not a confirmed result — it is the next level of the program. It states that the admissibility manifold is not merely a descriptor of what physical systems do, but a constraint on what physical systems can be. Systems that violate the inequality would not persist long enough to be observed.

What Would Falsify This

The conjecture makes clear predictions. Any physical system that exhibits inv(p) > ν(V(p)) under an admissible operator family would constitute a direct falsification. The EE localized anomaly (six violations above L = 1501) is the one candidate finding in the current corpus — it is confined to a narrow multipole range in a single run and warrants further investigation before it can be classified as a genuine falsification event versus an instrumental/noise-floor artifact.

The Deeper Structural Analogy

Physics already contains several laws that define what cannot occur: thermodynamics forbids sustained entropy decrease, relativity forbids superluminal motion, quantum mechanics restricts states to Hilbert space. The admissibility inequality may represent another constraint of this type — not a law about energy or velocity, but a law about the persistence of structure itself under systematic transformation.

If the Pattern Continues

The four tested domains span four symmetry regimes: rotational symmetry (gravity), statistical isotropy (CMB), finite-volume anisotropy (cosmic web), and local rupture geometry (seismology). Each symmetry class maps naturally to one level of the rigidity hierarchy. If additional domains show the same pattern, the rigidity spectrum may function as a classification of structural persistence in physical systems — analogous to how symmetry classes classify phases of matter.

7 · Scientific Significance and Open Questions

The cross-domain synthesis of the first four UNNS empirical domains represents a qualitative step beyond the program's earlier single-domain results. Three structural findings combine to produce a result stronger than any of its parts.

Finding I — The Admissibility Bound Is Universal

The inequality inv(p) ≤ ν(V(p)) is satisfied across four physically unrelated systems, through 19 chamber runs, spanning 12 orders of magnitude. This is the broadest empirical test of the admissibility framework to date.

Finding II — The Invariant Type Is Domain-Specific but Hierarchically Ordered

Each domain preserves a different structural object — absolute, relational, covariant, topological. These form a complete hierarchy in decreasing order of structural strength. The hierarchy is not imposed by the chambers; it emerges from the symmetry classes of the four systems.

Finding III — The Hierarchy Is Realized Across Physical Scale

The four invariant types also order by physical scale: planetary (absolute) → cosmological radiation (relational) → cosmic structure (covariant) → tectonic (topological). This scale alignment is not predicted by any domain-specific mechanism and provides the clearest suggestion that the rigidity spectrum reflects something deeper than four independent coincidences.

Open Questions

  • Does the EE localized anomaly represent a genuine falsification event or a noise-floor artifact at high multipoles?
  • Does the DESI axis alignment at (0.9999, −0.0072, 0) reflect physical structure or survey-coordinate geometry?
  • Will the rigidity hierarchy fill predictably as additional domains are tested?
  • Is there a formal derivation of the hierarchy of invariant types from first-principles admissibility geometry?
  • Can the Structural Realizability Conjecture be connected to known structural principles in physics (symmetry breaking, topological order, universality classes)?

Null Results Are Positive Evidence

The zero-falsification scorecard is not an absence of findings — it is evidence for the universality of the admissibility constraint. That the bound holds under 99% boundary activation pressure in the TE channel, and that El Mayor reaches exactly the budget without exceeding it, are the positive findings: the framework survives contact with the strongest structural tests the UNNS program has devised.

Instruments, Data, and Manuscripts

The UNNS Substrate Program is an independent research program testing the admissibility geometry framework across physical domains. All chamber instruments are self-contained HTML/JavaScript scientific tools; all data sources are public (Nevada Geodetic Laboratory, Planck 2018, DESI, SDSS, 2MRS, ICGEM). This article reports empirical results without claiming a physical explanation for the observed structural constraints.