SCC Research Overview

Date: 2026-05-09 (W6 EOD) | Canonical version: CV-1.11 (2026-05-06) | Theorem status: 54 Category A + 14 Category B + 5 Category C + 5 Retracted (78 claims, ~69% fully proved)

1. WHAT IS THIS RESEARCH?#

Core Thesis#

Soft Cognitive Cohesion (SCC) is a formal mathematical theory of how coherent formations (proto-objects, coherent regions, meaningful structures) emerge prior to discrete objecthood.

The theory operates on a principle: objects are not the starting point. They are the end point — stabilized formations that satisfy multiple structural requirements simultaneously. What comes before objects is a continuous graded cohesion field (u_t : X_t → [0,1]) — a distribution of "degree of participation in a formation" across a relational support space.

Key Conceptual Inversion#

Standard approach (rejected):

• Start with discrete objects or pixel labels
• Apply algorithms to segment, cluster, or track them
• Objects are primitives; algorithms operate on them

SCC approach (novel):

• Start with a primitive soft cohesion field — a graded field of participation
• Define four operators that measure structural properties of that field
• Use variational energy minimization to find stable field configurations
• Objects (if they emerge) are formations that simultaneously satisfy all four structural requirements: Bind (self-support), Sep (distinction), Inside (articulation), Persist (temporal continuity)

Ontological Commitment#

"Relational structure is prior to discrete individuation. What makes a formation cohere is not an intrinsic property of isolated points but a pattern of local mutual support: sites that reinforce one another, that belong together not because of a shared label but because their relational configuration is self-sustaining." — Canonical Spec §2

2. THE PROBLEM CONSCIOUSNESS: Why This Theory?#

The Perceptual Problem (from cognitive science perspective)#

How does a visual system recognize a coherent object?

Naive answer: It detects edges, segments regions, groups pixels. But this presupposes the object is already there, waiting to be detected.

The real problem: Before any algorithm runs, there must be some structural state that makes a coherent region coherent — independent of whether we happen to label it as "object" or "background." What defines coherence at the pre-objective level?

The Mathematical Problem#

Can you define coherence formally without presupposing:

• Boundary positions?
• Fixed graph structure?
• Predetermined classes?
• Objects?

And can you do it in a way that:

• Admits degrees (graded cohesion, not binary)?
• Is self-referential (the field defines what counts as coherent)?
• Is variational (minimizes energy)?
• Connects to temporal persistence (same formation, different time)?

The Theoretical Gap#

Existing frameworks:

• Clustering/segmentation: Presuppose objects exist, partitions are the answer
• Tracking: Assumes identity across time; doesn't explain how identity emerges
• Gestalt psychology: Powerful phenomenology; lacks formal mathematics
• Deep learning: Learns representations; doesn't explain what makes a region coherent to the system
• Predictive processing: Explains prediction; doesn't explain formation of coherent representations to begin with

SCC fills this gap: Provides formal machinery to define coherence before partitions, objects, or identity.

3. FORMAL MATHEMATICAL STRUCTURE#

The Primitive Ontology#

The formal universe is a tuple:

𝔇 = (T, {X_t}, {u_t}, {Cl_t}, {N_t, D_t}, {M_{t→s}})

Key point: The soft field u_t is NOT a probability or soft label. It is an intensity of cohesive participation — the primitive entity from which all structure is derived.

Four Operators (The Core Machinery)#

Energy Functional (4-Term Variational)#

On a connected weighted graph with volume constraint Σ u = m:

E(u) = λ_cl · E_cl(u) + λ_sep · E_sep(u) + λ_bd · E_bd(u) + λ_tr · E_tr(u)

Key architectural decision: No interaction term between formations (multi-formation interaction is kinetic/barrier-based, not thermodynamic/energy-based).

Derived Diagnostic Vector#

d = (Bind, Sep, Inside, Persist) ∈ [0,1]^4

• Bind: Field self-support intensity = 1 - √(E_cl/n)
• Sep: Field distinction = Σ u_x · D_x(1-u) / Σ u_x
• Inside: Morphological articulation = Q_morph · H_0(persistence)
• Persist: Temporal inheritance under transport

A formation is "proto-coherent" when all four components are high.

4. WHAT HAS BEEN DISCOVERED AND PROVED?#

Phase-Transition Phenomenon (Core Result)#

Theorem T8-Core (Universality): On any connected graph with Fiedler eigenvalue λ₂, there exists a phase transition.

Condition: β/α > 4λ₂ / |W''(c)|

Above this threshold, uniform field (u ≡ m/n) becomes unstable → non-trivial minimizers with formation structure emerge.

Why important: Formation birth is NOT an artifact of the graph topology. It depends only on the spectral gap. This means formation behavior is topologically universal.

Closure Structure#

Theorem T6b (Contraction): The sigmoid closure operator has:

• Unique fixed point when a_cl < 4
• Geometric convergence rate a_cl/4
• Non-idempotent structure (Cl(Cl(u)) ≠ Cl(u)) is a feature, not a bug

Why important: Non-idempotence provides enhanced stability. The Hessian at a non-idempotent fixed point has more positive eigenvalues than any idempotent alternative.

Temporal Persistence#

Theorem T-Persist-Full (5-Component Chain):

1. Core inheritance under gentle transport
2. Basin containment (core doesn't dissolve in boundary noise)
3. Morphological persistence (interior-boundary structure preserved)
4. Multi-formation weak coexistence
5. Kinetic metastability via energy barriers

Why important: Explains how the same formation can persist across time without requiring identity tracking or explicit correspondence.

Multi-Formation Behavior#

Key Discovery (Phase 14 audit): Multi-formation is kinetic, not thermodynamic.

• Single formation (K=1) is energetically preferred on ALL connected graphs
• Multiple formations (K>1) coexist as metastable local minima
• Stability maintained by energy barriers ∝ β^0.89, not energy preference
• Dynamics: nucleation → coarsening → coalescence (governed by barrier heights, noise, and spatial structure)

T-Persist-K-Unified: Originally proposed as a unified parametrization of three regimes via coupling parameter Λ_coupling = λ_rep · ω_jk / min(μ_j, μ_k). The associated dynamical conjecture P-Unified-1 was FALSIFIED per exp49–exp50; Λ_coupling has been reclassified as a structural classifier (not a dynamical predictor). Its dynamical role is replaced by the kinetic predictions MK-1..MK-4 (nucleation, coarsening exponent, barrier scaling, closure-enhanced metastability).

5. THEOREM REGISTRY (current)#

Proof Registry#

Aligned with Perception_theory canonical CV-1.5.2 (2026-05-02; 2026-05-04 audit pass):

Key Theorem Groups#

Group 1: Existence & Uniqueness (5)

• T1: Energy minimizer existence
• T6a/T6b: Closure fixed point, uniqueness, convergence
• T20: Axiom consistency

Group 2: Phase Transitions (4)

• T8-Core: Formation birth on all graphs (spectral universality)
• T8-Full: Phase transition for full energy
• T-FORMATION-BIRTH: General graph (Phase 14 upgrade)

Group 3: Stability & Dynamics (8)

• T14: Gradient flow convergence (Łojasiewicz)
• T7-Enhanced: Non-idempotent metastability advantage
• T3/T6-Stability: Minimizer stability under perturbations
• A2/A3: Axiom monotonicity + contraction

Group 4: Predicates & Energy (5)

• Predicate-Energy Bridge: Sep = exact equality; Bind reverse inequalities
• T11: Γ-convergence to modified perimeter functional
• Deep Core Dominance: Spectral bounds on coexistence

Group 5: Temporal & Transport (6)

• T-Persist-Full: 5-component temporal inheritance chain
• T-Persist-K-Sep: Well-separated formation persistence
• T-Persist-K-Weak: Weakly-interacting persistence
• T-Persist-K-Unified: Parametric unification

Group 6: Multi-Formation Kinetics (4)

• MK-1: Nucleation via spectral eigenvectors
• MK-2: Coarsening exponent (α < 1/2)
• MK-3: Barrier scaling ∝ β^0.89
• MK-4: Closure-enhanced metastability (30% barrier reduction)

Group 7: W4 Pre-Objective Mechanism (3, all Cat A, 2026-04-24)

• T-PreObj-1: F=1 single-disk minimizer non-critical under full SCC; gradient flow attracts to multi-peak F≥2; IC-protocol dichotomy
• T-PreObj-1G: T-PreObj-1 graph-class independent (any finite connected graph under (G1)–(G4))
• Lemma 4: $M\in {\mathbb{R}}^{2\times 2}$ positive definite under linear independence; destabilization magnitude ${\Lambda }^{T}M\Lambda >0$
• F-1 Resolution Corollary: F-1 SPLIT-RESOLVED via T-Merge(b) + T-PreObj-1 (i)

Group 8: W4-extended Pre-Objective Goldstone (1 Cat A + 1 Cat C finding, 2026-04-26)

• T-V5b-T: Pre-Objective Goldstone on Translation-Invariant Graphs (sub/super-lattice spectral dichotomy on torus T^d / cycle C_n; 2D commensurability split; 1D 1-fold Goldstone; nodal count = 2 universal)
• V5b-F (Cat C): Partial Goldstone on Boundary-Modified Graphs (NQ-173 carry)

Group 9: W5 Day 1 G0 σ-framework supporting structures (Cat A canonical merge, CV-1.5, 2026-04-27)

• T-σ-Lemma-1: σ-framework supporting lemma 1 (Cat A)
• T-σ-Lemma-2: σ-framework supporting lemma 2 (Cat A)
• T-σ-Lemma-3: σ-framework supporting lemma 3 (Cat A)
• T-σ-Theorem-3: σ-framework supporting theorem 3 (Cat A)
• T-σ-Theorem-4: originally promoted Cat A at CV-1.5; retroactively reclassified to Cat B at CV-1.5.1, with the NQ-187 caveat (registered 2026-05-04)

Group 10: W5 Day 3 EOD D-6a Multi-Static (CV-1.5.1, 2026-04-29)

• T-Commitment-14-Multi-Static (Cat A def): Commitment 14 multi-static σ formulation
• T-σ-multi-A-Static (Cat A def): multi-formation σ static existence
• T-σ-multi-D-Static (Cat A def): multi-formation σ static derivation
• T-σ-Multi-1 (Cat B target): multi-formation σ unification target
• V5b-T-zero (Cat A def, sub): zero-mode Goldstone substatement (replaces V5b-T' WITHDRAWN)
• V5b-F-empirical (Cat B target, sub): empirical boundary-modified Goldstone substatement
• Commitment 16 (K-status two-tier decomposition): K_field (architectural cap, modeling-layer commitment set externally at instantiation) vs K_act (dynamic stratum index, kinetically determined per CN6 refined) — direct OAT-1 outcome resolving OP-0009-K
• OP-0008 registered HIGH (σ^A K-jump non-determinism); OP-0009 registered HIGH (Multi-Formation Foundations, 7 sub-items)
• MO-1 re-activation rider: MO-1 sidestep applies to single-formation σ on Σ_m only; multi-formation σ phase will re-engage MO-1 as conditional blocker

Group 11: W5 Day 6 T-L1-F first multi-formation Cat A conditional theorem (CV-1.5.2, 2026-05-02)

• T-L1-F (Hard-Bar / Active-Count Bridge under L1-J Regime (P0)–(P11)): the first multi-formation canonical Cat A conditional theorem, operating on shared-pool architecture I9' alternative (${\tilde{\Sigma }}_M^{K_{\mathrm{field}}}$). Establishes a hard-barrier / active-count bridge under the L1-J regime conditions (P0)–(P11)

Group 12: W5 Day 7 L-M Soft-Count Corollary working draft (Cat B sketched, 2026-05-03)

• L-M Soft-Count Corollary: Cat B sketched in working/MF/ksoft_kact_bridge_L1M_soft_count_corollary.md; CV-1.6 promotion target via L1-M-AUDIT (W6 G1)

Plus auxiliary results (symmetries, consistency conditions, numerical verifications).

6. WHAT REMAINS OPEN (Explicitly Marked)#

W4 close (2026-04-24): Critical 3 → 0. F-1 (K=2 vacuity) SPLIT-RESOLVED via T-Merge(b) + T-PreObj-1; M-1 (K=1 preference) LAYER-CLARIFIED (proved theorem misframed as problem); MO-1 (Morse inapplicability) SIDESTEPPED by σ-framework on Σ_m. v2.0 release path unblocked. CN15 Static/Dynamic Separation is the conceptual key.

Remaining open problems (post-W4, current as of CV-1.5.2):

1. OP-0005 K-Selection mechanism (HIGH) — partially addressed by σ-framework + CN15; full mechanism still open
2. OP-0006 Boundary precision (HIGH) — interface localization sharpness on graphs
3. OP-0008 σ^A K-jump non-determinism (HIGH) — registered CV-1.5.1 (2026-04-29); CV-1.7 Commitment 18 candidate
4. OP-0009 Multi-Formation Foundations (HIGH, 7 sub-items) — registered CV-1.5.1 (2026-04-29); OP-0009-K resolved via Commitment 16 (two-tier K_field/K_act decomposition); 6/7 sub-items PARTIALLY resolved
5. Near-bifurcation dynamics (μ → 0): Center manifold reduction for barrier-collapse behavior
6. Strong-regime merge dynamics (deeply overlapping formations): Kramers escape rates, noise-driven coalescence
7. Crisp recovery protocol: Explicit algorithm for object extraction from soft field
8. W5+ NQ carry: NQ-173 (V5b-F partial Goldstone), NQ-174 (ζ_*(graph) precise dependence), NQ-175 (3D extension), NQ-187 (T-σ-Theorem-4 Cat B reclassification caveat), σ supporting lemmas, multi-formation σ Phase 5 (would re-engage MO-1 per Group 10 rider)

7. DEVELOPMENT PROCESS: 12 ITERATIONS#

Current phase (I13-I15): Publication and dissemination

8. IMPLEMENTATION STATUS#

Python Package (scc/)#

scc/
├── graph.py           — GraphState (adjacency, Laplacian, Fiedler)
├── params.py          — ParameterRegistry (constraint validation)
├── operators.py       — Cl(), D(), C_t realizations
├── energy.py          — EnergyComputer (4-term energy + exact gradients)
├── optimizer.py       — find_formation() (projected gradient descent)
├── transport.py       — Temporal transport kernel + Sinkhorn OT
├── multi.py           — K-field architecture (multi-formation)
├── diagnostics.py     — DiagnosticVector (d = [Bind, Sep, Inside, Persist])
└── [+ 140+ tests, ~30 experiments]

Testing#

• 215 passing tests + 1 xfailed (216 collected) (all core components). Note: earlier CHANGELOG entries said "196/196" — corrected by the 2026-05-04 audit.
• Gradient verification to FD ~1e-9
• Axiom satisfaction proofs computational
• Phase transition predictions numerically validated
• Multi-formation coexistence confirmed (kinetic barrier measurements)

Experiments#

• 88+ experiments (exp1–exp88 + W4-extended E-0090–E-0097 + W5 numerical anchors + W6 OMS VP-1..VP-11) covering:
  • Phase transition verification
  • Parameter sweep (λ, β, α ranges)
  • Grid scaling (n = 25 to 400)
  • Graph diversity (lattices, random, real-world, synthetic)
  • Multi-formation dynamics (nucleation, coarsening, merge)
  • Transport concentration validation

9. PUBLICATION STATUS#

SCC Manuscripts (planned rewrite)#

Two SCC manuscripts — "Self-Referential Phase Fields on Graphs" (math focus, target: Journal of Mathematical Physics) and "Before Objects: Pre-Objective Perceptual Cohesion" (cognitive science focus, target: Cognitive Science) — were deleted from CODE/papers/ on 2026-05-04 as part of the W6 canonical audit, to be rewritten from scratch against the CV-1.11 (78-claim) baseline. They will appear on /papers/ when ready.

The closest published-side work is the Axiomatic Framework for Understanding Relations via Gauge-Invariant Cohomology, which establishes the relational substrate underpinning SCC.

10. DISTINCTIVENESS: What Makes This Novel?#

Bottom line: SCC is the first formal framework to:

1. Define coherence as a graded field property (not discrete)
2. Make that field self-defining (operators depend on field)
3. Achieve this through variational energy with four independent terms
4. Connect to temporal persistence via structured transport
5. Handle multi-formation coexistence through kinetic barriers

11. CRITICAL CONSTRAINTS & NON-NEGOTIABLES#

From Agent Instructions (binding protocol):

1. Ontological priority: The soft field u_t is primitive. Crisp objects are derivative.
2. Never collapse layers: Keep ontology, axiomatics, operator realizations, implementation strictly separate.
3. Never silently resolve open problems: Explicit open problem registry maintained.
4. Never reduce to familiar frameworks: Not fuzzy segmentation, clustering, tracking, or edge detection.
5. Four-term energy independence: Energy terms remain conceptually independent.
6. Idempotence deliberately omitted: Closure has stabilization tendency (A3), not primitive idempotence.

12. CURRENT CHALLENGES#

What Remains Hard#

1. Sharp interface limit: Can one rigorously recover crisp objects from soft fields in all regimes? (Γ-convergence gives partial answer)
2. Transition operator: Characterizing boundary morphology formally (T_t demoted; role now carried by diagnostics)
3. Strong-regime merge: Dynamics when two formations deeply overlap (kinetic escape rates)
4. Co-belonging role: C_t is conceptually necessary but dropped from energy/predicates (currently diagnostic only)

Why Hard#

• Self-referentiality creates fixed-point problems (Schauder's theorem helps, but domain structure is intricate)
• Temporal transport requires balancing self-consistency with computational tractability
• Multi-formation theory sits at boundary between thermodynamics (energy) and kinetics (barriers)

13. NEXT STEPS#

Immediate (CV-1.12 via H-SINK)#

• H-SINK (unblocked): Prove Sinkhorn-Lipschitz bound $L_g\le L_c$ in the SCC cost class (Bigot-Cazelles-Papadakis 2019) → unlocks T-Temporal-Identity (a,b,d) Cat A → CV-1.12 (+3A, 57A total). Dependency: P-F-A1 Package I (CV-1.9, complete).
• L1-M-AUDIT: Close R-1/R-2/R-3 audit items for T-L1-M canonical promotion.
• OMS formal rows: Register Appendix OMS §A–§M theorem items as formal rows in theorem_status.md (bookkeeping, non-blocking).
• Maintain: 215 + 1 xfailed test suite (clean throughout W6).

Medium Term (Publication)#

• Submit papers to target journals
• Respond to reviewer feedback (iterations expected)
• Publish/disseminate open-source scc/ package

Long Term (Extensions)#

• Center manifold reduction for near-bifurcation dynamics
• Noise-driven merge simulations (Kramers rates)
• Application to biological development, visual neuroscience
• Connections to predictive processing literature

14. HOW TO READ THIS CODEBASE#

For understanding the theory:

1. Start with the canonical spec (THEORY/canonical/canonical.md, currently CV-1.5.2; formal definitions)
2. Read the development overview documentation
3. Deep dive into the theorem proofs from Iteration 2

For understanding the implementation:

1. Read scc/operators.py (closure, distinction)
2. Read scc/energy.py (4-term energy + gradients)
3. Run the full test suite (215 passing + 1 xfailed, 216 collected)
4. Examine the experiments directory (parameter sweeps, phase transitions)

For understanding the context:

• This overview document
• The project instructions (CLAUDE.md)
• The session-level changelog

15. IN ONE SENTENCE#

SCC is a formal theory that answers the question: "What makes a coherent region coherent, before we call it an object?" — using a self-referential variational framework on graphs with four structural operators and 54 Cat A theorems (78 total claims, ~69% proved; CV-1.11, 2026-05-06).

Last updated: 2026-05-09 | CV-1.11 (2026-05-06) | 54A / 14B / 5C / 5R = 78 claims (~69% proved) | OP-0006 RESOLVED | OP-0005 3-way split (EQ partial, OBS partial, DYN open) | OMS-2.0 Accepted — Full | Next: CV-1.12 via H-SINK | Most-recent narrative: Week 6 weekly post