How Physical Systems Find Their Place in Admissibility Space

The Corpus at a Glance

From Rankings to Geometry: The Conceptual Shift

Earlier work in the UNNS program evaluated ladder encodings one at a time, recording verdicts (FULL, GIANT, TAIL, HARD) and comparing similarity scores across domains. Those comparisons revealed something striking — helium and the CMB scored remarkably similarly — but the why remained elusive.

The Canonical Ladder Engine v2.0.0 changed this by replacing single-point evaluation with deformation-field analysis. Instead of asking "what verdict does this ladder receive at physical constants?", the pipeline now asks: "how does this structure survive as constants vary across a 20% deformation envelope?"

This reframing resolves three problems simultaneously: it explains why the similarity values are high (basin co-occupation, not physical equivalence), it explains why adversarial ladders partially overlap with real ones (both can produce identical stitching-defect geometry), and it explains why some domains appear to "fail" falsification (internal bifurcation, not adversarial penetration).

The 5D Admissibility Vector

For each ladder encoding L, the CLE v2.0.0 pipeline constructs a deformation field Φ(L) over the grid (α,μ) ∈ [0.8, 1.2]² (81 points at step 0.05). From this field, five summary statistics are extracted — together forming the coarse coordinate of persistence geometry.

The five components of v(L) capture distinct structural properties. Persistence depth (𝒫_depth) is the mean giant-component ratio across all 81 deformation grid points: a direct measure of how deeply the system sits inside ℳ_adm. Rigidity variance (σ²_GR) measures how stable that depth is — low variance means a smooth, coherent manifold. Fragmentation rate counts how often the deformed ladder loses structural coherence. Admissibility persistence is the fraction of the deformation grid that retains FULL or GIANT verdict. Anisotropic persistence probes directional vulnerability: does the system fragment preferentially under certain constant variations?

The Stratified Basin Architecture

The admissibility manifold ℳ_adm is not a flat, homogeneous region. It is stratified by the connectivity margin m(L) and a percolation threshold operator T with discrete fixed points at κ_conn ∈ {0.562, 0.750, 1.000, 2.000, 10.000}. These fixed points partition ℳ_adm into four qualitatively distinct basin types.

Five Domain Portraits

Helium: The Two-Layer Separation

Helium spectroscopy produced the most theoretically important result in the corpus. On symmetric deformation grids, all five encoding families produce identical topology: GR ≡ 0.867 ± 0.001 across all 81 grid points for all families. Topology is universal.

On anisotropic grids, persistence depths diverge — from 0.473 (zeeman_triplet, RIGID_STABLE) to 0.454 (delta_zeeman_singlet, RIGID_PLASTIC). Rigidity is representation-dependent. This proves that the topology layer and the rigidity layer are genuinely independent structural observables.

Cosmology: The Deepest Interior

The three Planck 2018 power spectra (TT, TE, EE) achieve global admissibility = 1.0 across the entire deformation grid. The TT spectrum has the highest mean GR (0.918) and lowest GR variance (0.016) of any sequence tested. Not a single stitching defect under the deformation protocol. The CMB occupies the Type I dense interior without exception — and does so across a 20% variation in both α and μ, confirming this is not an artifact of specific constant values.

Protein MSM: Observable-Dependent Geometry

The protein MSM corpus (QT45 ribozyme, Folding@home) is qualitatively different from the spectral and cosmological ladders: it encodes conformational dynamics rather than static observables. The same physical system contains both coherent and turbulent sub-manifolds depending on which observable is ladderized. Population ladders produce 𝒫_depth ≈ 0.96, τ ≈ 0.005 — Type I co-occupant with the CMB. Out-strength ladders produce 𝒫_depth ≈ 0.71, τ ≈ 0.12, with three isolated nodes (kinetic traps). Persistence geometry is observable-dependent, not system-intrinsic.

Voyager: The First Dynamic Trajectory

Voyager 1's heliopause crossing (2011–2017) provides the only time-ordered trajectory through admissibility space in the corpus. At t* = 2012, κ_conn achieves its annual minimum — confirmed across three independent window scales (3/3 scale robustness). The connectivity margin m(L) approaches zero while the giant-component ratio remains non-trivially positive: the system is compressed toward ∂ℳ_adm but does not cross it. After t*, the trajectory recovers into Type II geometry as Voyager enters the ISM. This is Margin-Confinement dynamics confirmed in a continuously tracked physical trajectory.

Neutrino Detector: Internal Manifold Ecology

The neutrino corpus (67 encodings, five families: TMVA, DeepL, raw_sig, raw_bkg, Fib) reveals that what looked like a "noisy, partially failing domain" is actually an internally structured manifold ecology spanning five distinct structural phases.

Signal encodings cluster in STABLE_ISLAND; background encodings dominate COLLAPSE_BASIN. No label information was provided to the CLE pipeline — the structural geometry discovered the signal/background separation independently. The DeepL family splits by physical function despite identical algorithmic origin. Δ-lifting recovers FULL or GIANT continuity in 100% of tested HARD-classified cases: apparent collapse is representation-induced and recoverable.

The Adversarial Locality Principle

The 650-ladder adversarial corpus (100 uniform, 100 Pareto, 100 randomwalk, 350 shuffle) was the first large-scale test of whether admissibility geometry can distinguish real physical encodings from synthetic random ones. The results are decisive — but not in the way a simple pass/fail test would suggest.

Hierarchical Falsification Results

The correct falsification criterion requires two conditions: (1) adversarial penetration rate < 5%, and (2) within-domain cosine similarity > 0 (coherence prerequisite). Domains that fail condition (2) are internally bifurcated — their canonical lift should be tested instead.

Basin Co-Occupancy: What the 0.993 Actually Means

The pairwise domain similarity matrix, computed from 5D admissibility vectors, contains one number that has driven scientific discussion throughout the program:

Helium and the CMB differ by less than 1% of the maximum inter-domain separation in normalized 5D space. A separation that spans coherent rigid manifolds on one end and plastic transition manifolds on the other — and helium and cosmology fall at the same end, nearly identically.

But what does this actually mean? The answer is now precise: both helium and the CMB are Type I co-occupants of ℳ_adm. Both reside in the dense interior, far from the admissibility boundary, with no fragmentation and near-zero GR variance under the full deformation envelope. They are similar because they occupy the same geometric region — not because they share any physics.

This is the central conceptual improvement over the previous "structural universality" framing. "Universality" suggested shared laws, shared mechanisms, shared physics. "Co-occupancy" says exactly what the data supports: unrelated systems can inhabit the same geometric neighborhood inside ℳ_adm. The language is more precise, more defensible, and more informative.

Formalizing Turbulence

The manuscript introduces the first formal quantification of manifold turbulence — converting an intuitive distinction into measurable structural observables.

Three indices characterize the turbulence structure of a manifold. The turbulence index τ(L) combines GR variance, anisotropic fragility, and fragmentation rate into a single dimensionless number — separating coherent from turbulent manifolds by over two orders of magnitude. The turbulence radius ξ(L) measures how far from the physical parameter point turbulence first emerges. The manifold roughness ℛ(L) captures the total variation of the GR field across the deformation grid.

These are not metaphors. They are measurable properties of the deformation field — the first formal quantification of structural turbulence in the UNNS program.

The Dynamics Frontier

Every domain analyzed in Parts I and II of the manuscript represents a static snapshot: one deformation field, one basin assignment, one structural position. The Voyager corpus breaks this pattern — and in doing so, opens an entirely new theoretical horizon.

A physical system moving through parameter space traces a continuous path through ℳ_adm. The Voyager trajectory from 2011 to 2017 is the first operational measurement of this path. The structural coordinates evolve year by year. At t* = 2012, the connectivity threshold κ_conn reaches its minimum — maximum structural compression, closest approach to ∂ℳ_adm — and then recovers as the spacecraft enters the ISM.

Significance: What the Program Has Gained

The Canonical Structures manuscript, combined with the CLE v2.0.0 corpus, represents the most substantial conceptual advance of the UNNS program to date. It is not a new empirical result added to an existing framework. It is a new layer of the framework itself.

The layered architecture of the UNNS program now has a stable form:

The deepest discovery is this: real physical systems may be organized less by what they are physically, and more by how their structures persist under deformation. Persistence geometry is the structural layer between raw physical description and abstract admissibility law. It was always present in the mathematics. The Canonical Structures manuscript and CLE v2.0.0 made it visible.