UNNS Research · Admissibility Geometry

From Empirical Stability to Operator-Manifold Phase Boundaries

Admissibility Geometry, Stratified Manifolds, and Observability through Invariance — a formal answer to the question: does the UNNS Substrate have a shape, and can that shape be observed?

📅 UNNS Research Collective · 2025 🧮 LXV Seismic Suite Validation 📐 Operator Manifold Theory

Overview

The UNNS Substrate has long been described as the mathematical arena inside which structural laws emerge from recursive dynamics. But what does that arena look like? This paper gives a precise, measurable answer for the first time.

The shape of the substrate is not visible as a fault line or a physical surface. It is inferred through invariance geometry — the pattern of descent stability under admissible operator families. This work formalizes that shape as a stratified manifold inside the space of bounded linear operators, proves its convexity properties, and validates the predicted phase structure empirically using three earthquake events spanning two orders of magnitude in the Rigidity Modulus R.

The core result: structural lawhood in the UNNS Substrate exists precisely in the interior of admissibility margin 𝒜 > 1. The boundary is not a failure mode — it is a structural feature.

Rigidity Moduli
Rmag, Rgeo
Kumamoto R
≈ 21.6
Ridgecrest Reff
≈ 4.7
El Mayor R
≈ 0.19
Tight Budget ν(V)
1
Strata Dim.
2D Manifold
The Central Question

🔭 Does the Substrate Have a Shape?

The answer is yes — but only indirectly, through invariance geometry. The substrate is not a geometric object in ordinary space. It is an abstract arena — the space of admissible operator families — and its shape is the geometry of the admissibility region inside that space.

The Shape Has Three Geometric Components

📏

Order Geometry

The baseline adjacent gaps {Δk} define a gap spectrum encoding separability of the station network in displacement space. Dense gaps imply fragility; sparse gaps imply robustness.

🎯

Perturbation Envelope

The pair (σP, δP) defines a two-dimensional parameter space 𝒫 = ℝ²≥0. The envelope reflects both measurement noise and the extent of the admissible operator family.

🔗

Matching Geometry

As σP increases, gaps become vulnerable: V(σP) = {k : Δk ≤ 2σP}. The matching number ν(V) governs the tight inversion budget; its adjacency structure is the degeneracy geometry.

Together, these three components determine the geometric object: the admissibility region in 𝒫, stratified by ν-level sets. The substrate shape is, precisely, the shape of admissibility.

📊 Diagram I — Phase Structure in (Rmag, Rgeo) Space

Each seismic event maps to a point in the phase plane defined by the two rigidity moduli. The lines Rmag = 1 and Rgeo = 1 are the phase boundaries — the RNP-critical surfaces at which structural descent becomes unstable.

Phase Plane: (Rmag, Rgeo) — LXV Suite
0.1 0.3 1 5 15 22 0.5 1 5 20 ≫1 R_mag (Magnitude Rigidity Modulus) R_geo (Geometric Rigidity Modulus) FULLY RIGID D(w)=0, ARI=1 MAG-DEGENERATE bounded inversions DIR-DEGENERATE FRAGMENTATION K Rc EM Kumamoto R≈21.6 Ridgecrest R_eff≈4.7 El Mayor (B2) R≈0.19 ← boundary R_mag = 1 boundary R_geo = 1 boundary Kumamoto (Rigid) Ridgecrest (Rigid, eff.) El Mayor (Boundary) Boundaries
Figure 1. Phase plane (Rmag, Rgeo) for the LXV seismic suite. Dashed lines mark the RNP-critical boundaries R = 1. Kumamoto and Ridgecrest sit in the fully rigid interior; El Mayor (B2) falls in the magnitude-degenerate quadrant — on the boundary manifold — while maintaining geometric rigidity. The two-order-of-magnitude spread in Rmag arises from station network geometry, not seismic energy.

🧩 Diagram II — The Stratified Admissibility Region

The rigid admissibility region 𝒜rig in envelope parameter space (σP, δP) is an open 2D submanifold of ℝ². As σP grows, vulnerable gaps are crossed one by one — producing a staircase stratification with discrete jumps in the matching number ν. This is not smooth collapse; it is discrete-stratified phase transition.

Stratified Admissibility in (σP, δP) Space — El Mayor B2 Instantiation
0 0.24 0.50 0.8 1.3 1.7 σ_P (mm) 34° 15° δ_P (deg) ν=0 RIGID ν=1 BOUNDARY ν=2 FRAGMENTATION ZONE DIRECTION-DEGENERATE (δ_P > Θ_min/2) σ=Δ_1/2 σ=Δ_2/2 σ_P empirical 1.3 mm El Mayor B2 (σ_P≈1.3, ν=1, D=1) 𝒞_1 (convex rect.)
Figure 2. Stratified admissibility region in (σP, δP) envelope space, instantiated on El Mayor B2. Green zone (ν=0): rigid interior — an open interval, hence convex. Amber band (ν=1): boundary stratum where one inversion is structurally forced. Red region (ν=2): fragmentation zone. Dashed boundary lines at critical values σP = Δk/2 separate strata. The teal dashed rectangle marks 𝒞1, the convex certification region for budget k=1.

📉 Diagram III — Vulnerability Graph and Gap Spectrum

The vulnerability graph plots each adjacent gap Δk against the perturbation threshold 2σP. Gaps below the threshold line are vulnerable; their adjacency pattern determines ν(V). This is the substrate's degeneracy geometry, made visible.

Gap Spectrum vs. Perturbation Threshold — Three LXV Events Compared
Kumamoto (Rigid) 2σ_P=6mm k=1 334 k=2 130 k=3 273 All gaps ≫ 2σ_P → V=∅, ν=0 Ridgecrest (Eff. Rigid) 2σ_P=12mm 1 444 2 8.9 ★ 3 57 4 125 5 11.8 V={2}, ν=1, but pair-specific R_eff=4.7 El Mayor B2 (Boundary) 2σ_P=2.6mm 1 55.8 2 41.2 3 0.48 ★ 4 0.50 ★ adjacent! V={3,4} adj., |V|=2, ν=1 ← tight budget Safe (above threshold) Vulnerable ★ (in V) Near-threshold 2σ_P threshold
Figure 3. Gap spectrum diagrams for all three LXV events. Red-filled bars (★) fall below the perturbation threshold 2σP and are vulnerable. Kumamoto: no vulnerable gaps (V=∅, ν=0). Ridgecrest: one global vulnerable gap, but pair-specific modulus for that boundary is Reff≈4.7 (still rigid). El Mayor B2: two adjacent vulnerable gaps (k=3, k=4); adjacency forces ν(V)=1, making the minimal tight budget kminallowed=1 — matching the chamber's operational choice exactly.

Why Adjacency Matters: The Matching Number

When two vulnerable gaps are adjacent (as in El Mayor B2, k=3 and k=4), only one can be independently swapped without forcing a station to shift two ranks. This is not a conservative heuristic — it is a structural theorem. The matching number ν(V) is the tight minimal budget, computable by a greedy O(|V|) scan: pick k whenever k > lastPicked+1.

Operator Manifold Generalization

🏗️ From Envelope Plane to Operator Manifold

The envelope plane (σP, δP) is a derived coordinate chart — a projection of a deeper object. The full substrate shape lives in the space of bounded linear operators ℬ(X), and the admissibility region is an open submanifold ℳrig(x) inside the admissible operator manifold ℳ.

Envelope projection: E_x : ℳ → ℝ²≥0, E_x(T) = (σ_x(T), δ_x(T))

Rigid region: ℳ_rig(x) = { T ∈ ℳ : 2σ_x(T) < Δ_min(x) ∧ 2δ_x(T) < Θ_min(x) }

Critical hypersurfaces: ℋ_k = { T ∈ ℳ : 2σ_x(T) = Δ_k(x) }
🟢

Open Submanifold of Rigidity

rig(x) is open in ℳ (operator norm topology). Rigidity is a continuous property; strict inequalities are open conditions. It is a genuine d-dimensional smooth submanifold.

🔶

Stratification Hypersurfaces

Boundary degeneracy corresponds to crossing ℋk. Between hypersurfaces, νx(T) is locally constant — phase transitions are discrete jumps, not smooth collapse.

📈

Monotone Path Structure

Convexity of ℳ is not intrinsic. The correct replacement: along any path where σx(γ(t)) is nondecreasing, νx(γ(t)) is nondecreasing. Certification regions are path-stable in the "inward" direction.

Core Theoretical Advancement

The (σP, δP) coordinates are not primitive parameters — they are event-induced projections of operator geometry. Every admissibility result in envelope-plane language is a shadow of a richer statement in operator space. The admissible operator manifold ℳ is the substrate object. This is no longer just smoothing robustness — it is a geometric theory of admissible operator action.

📐 The Convexity Trichotomy

A precise and non-trivial characterization of which aspects of the substrate shape are convex and which are not — with theorems for each case.

Theorem 1

Rigid Region is Convex

The cross-section of 𝒜rig in σP is exactly the interval [0, Δmin/2) — an interval, hence convex. The rigid interior has simple, clean shape.

✓ CONVEX (interval in σ_P)
Theorem 2

Phase Surface ν(σ_P) is NOT Convex

ν(σP) is a right-continuous step function — monotone nondecreasing but never convex (violates Jensen's inequality for nonconstant step functions).

⚠ MONOTONE, NOT CONVEX
Corollary

Certification Regions Are Convex

For any fixed budget k ≥ 0, by monotonicity of ν, the certification region 𝒞k = [0,σk) × [0, Θmin/2] is a rectangle — hence convex.

✓ CONVEX (rectangle)

Why This Trichotomy Is Novel

Prior treatments of operator robustness addressed convexity of the stability region or convexity of loss functions, but not the full geometric picture: a convex interior, a non-convex phase surface, and convex budget-constrained regions simultaneously. This trichotomy gives the first complete geometric characterization of the UNNS admissibility structure.

🌍 Empirical Validation: LXV Seismic Suite

Three earthquake events, spanning two orders of magnitude in Rigidity Modulus R, confirm the predicted phase structure. Crucially, this spread is not a function of seismic magnitude — all three events are comparable in Mw≈7. It encodes near-fault station geometry.

Event (Chamber) Δmin (mm) σP (mm) R |V| ν(V) 𝒟 Phase D(w) observed
Kumamoto 2016 (LXV-A) ≈130 ≈3.0 ≈21.6 0 0 0 RIGID 0 (all windows)
Ridgecrest 2019 (LXV-C2) 57–125 (eff.) ≈6.1 ≈4.7 (eff.) 1 (global) 1 0.2 RIGID 0 (no inversion)
El Mayor 2010 (LXV-B2) 0.484 ≈1.3 ≈0.19 2 (adjacent) 1 0.5 BOUNDARY 1 (max rank shift=1)
El Mayor Topology (LXV-D) δP<1.5° Rgeo ≫1 DIR-RIGID ARI=1 (all windows)

El Mayor B2: The Paradigm Boundary Event

El Mayor B2 demonstrates a striking decoupling: Rmag≈0.19 (magnitude channel on the boundary) while Rgeo≫1 (directional channel fully rigid). The two components of the admissibility vector can and do decouple in real data. The observed inversion count D(w)=1 matches exactly the predicted tight budget ν(V)=1 — confirming that the chamber's operational gate kallowed=1 is not merely conservative but the theoretically tight minimal budget.

RNP as a Manifold Theorem

⚡ The RNP Phase-Boundary Theorem

The Rigid–Nonrigid Transition Principle has been upgraded from a structural principle into a geometric theorem. It now reads:

RNP Phase-Boundary Theorem (Upgraded Statement)

Under perturbation-admissible smoothing and the adjacent-swap regime:

(1) Rigid descent: If 𝐑 > (1,1) componentwise, the structural signature Σ(E) is invariant on each Gadm-orbit; it descends through the quotient and is law-admissible.

(2) Boundary forcing: If 𝐑 meets a phase boundary (Rmag≤1 or Rgeo≤1), bounded degeneracy is structurally forced — there is no way to enforce a single rigid description across all admissible transforms.

(3) Nonrigid impossibility: If either modulus drops sufficiently below 1 that no bounded-degeneracy certificate holds, no orbit-invariant structural signature exists and fragmentation is unavoidable.

From Principle to Geometry

Previously, RNP stated: "Some structures descend, others don't." Now it states: Descent holds on an open submanifold of admissible operators. The substrate has a measurable geometry of structural admissibility — not metaphorical, not philosophical, but measurable through orbit flatness, phase margins, and degeneracy adjacency structure.

Structural Identification Table

Perturbation-Stability Construct UNNS Substrate Object
Smoothing window w ∈ 𝒲Admissible operator Ow
Perturbation envelope σPBounded admissible deformation radius
Rigidity modulus RAdmissibility margin in substrate phase space
Degeneracy index 𝒟Boundary proximity measure
Descent of Σ through ΠQuotient stability (law-admissibility)
Phase boundary R = 1RNP-critical surface in the substrate
Admissibility region 𝒜rigOpen submanifold of operator space
Stratification hypersurfaces ℋkCodimension-one phase boundaries

🌐 Significance and Broader Implications

What This Adds to UNNS Theory

Upgrade 1

Admissibility Is Now Geometric

Before: admissibility was a logical constraint. Now: admissibility is an open region in an operator manifold with interior, codimension-one hypersurfaces, stratified degeneracy layers, and orbit-level descent structure.

Upgrade 2

Phase Transitions Are Discrete-Stratified

The matching-number stratification reveals that UNNS phase transitions are not smooth collapse — they are discrete jumps at codimension-one surfaces. This is a deep structural property of the substrate.

Upgrade 3

Operator Nesting Is the True Variable

The admissible operator manifold generalizes beyond smoothing windows to resolution operators, filtering, coarse-graining, and representation transforms — all as points in ℳ. This future-proofs the theory across UNNS domains.

What This Adds to General Science

Most sciences rely on smoothing, aggregation, resolution change, and coarse-graining — but lack a formal stability region in operator space. This framework provides the operator manifold, envelope projection, stratified phase structure, and minimal instability budget. That is new at the methodological level.

The Deepest Implication

Structure persistence is not binary — it occupies an open manifold region. Scientific "laws" are interior points in admissibility geometry. Boundary artifacts arise at codimension-one surfaces. This is a clean unification of law detection, robustness analysis, phase transitions, and representation dependence into a single geometric framework.

A Template for Other UNNS Domains

Wherever there is nested resolution, smoothing families, model approximations, or operator deformations — quantum algorithm diagnostics, spectroscopy analysis, fundamental constants validation — the same framework applies: identify (σP, Δmin, ν(V)) for the domain, compute R, and locate the system in its admissibility phase.

👁 Three Methods for Making the Shape Visible

The substrate shape is not directly observable. But it can be inferred through three complementary empirical methods:

📍

Method A: Phase Diagram

Plot each event as a point in (Rmag, Rgeo) space and draw boundary lines at R=1. The four quadrants directly visualize phase regions. See Diagram I above.

📊

Method B: Vulnerability Graph

Plot gap index k vs. gap size Δk, mark threshold 2σP. The pattern of adjacency among vulnerable gaps is the degeneracy geometry; ν(V) reads off directly. See Diagram III above.

🔄

Method C: Operator-Orbit Plot

Map w ↦ (D(w), ARI(w)) for each window. A flat orbit signals rigid substrate; bifurcation signals proximity to a hypersurface ℋk. Orbit shape = substrate geometry.

From Invisibility to Measurability

The UNNS Substrate is not invisible abstraction. It manifests as phase margins, orbit flatness, degeneracy adjacency structure, and envelope-induced contraction. The shape is the admissibility region in operator space — and it is now formally measurable, not merely conceptually invoked.

⚠️ Falsifiability: Explicit Criteria

The framework admits distinct falsification classes, as required by scientific rigor:

Structural Falsifiers (Theorems Are Wrong)

F1: D(w)>0 while R>1 under the empirical envelope — falsifies Order Rigidity Theorem.

F2: Cluster reassignment (ARI<1) while Θ>2δP — falsifies Directional Stability.

F3: Inversion count exceeding |V|(N−1) — falsifies the Degeneracy Bound.

Parameter Falsifiers (Model Fit Is Wrong)

F4: Observed inversion while R̂(α)>1 — falsifies sub-Gaussian noise model or guard protocol.

F5: Systematic σ/√w decay violation — indicates violated independence assumptions.

F6: D(w) > |V| in boundary regime with adjacent-only perturbations — falsifies tight kallowed bound.

Correctly Scoped Non-Evidence

Perfect ARI in a TOPO_SINGLE event provides no evidence for or against the Topology Rigidity Theorem (degenerate case). Ridgecrest's global Rglobal=0.73 combined with no observed inversion is consistent with theory — the boundary regime predicts vulnerability, not necessity of inversion. These distinctions matter for honest science.

📚 Resources & Downloads

All chambers are self-contained HTML applications with no external dependencies, ensuring full reproducibility. Data exports in JSON format are provided for each chamber run. The LXV suite instantiates every logical branch of the RNP Phase-Boundary Theorem empirically: rigid descent (Kumamoto), boundary forcing (El Mayor B2), and topological rigidity (El Mayor D). No fragmentation was observed within the tested admissible family.