This volume is the second in the Principia Orthogona series. Volume I developed the singularity-theoretic and variational foundations of generative transitions: the operator sequence \(C \to K \to F \to U\), the curvature threshold \(\kappa^*\), the Whitney \(A_1\)–\(A_3\) singularity classification, and a symplectic preservation theorem for the fold map. The present volume constructs the explicit contact-geometric realization of those foundations.
Three main results: (A) a precise correspondence between the geometric fold operator \(F\) and the dm³ contact Hamiltonian dissipation \(H_{\mathrm{diss}}\); (B) equivalence of the curvature threshold \(\kappa^*\) and the embodiment threshold \(\tau\), with explicit values \(\tau=2\), \(\varepsilon_0=1/3\) verified in the dm³ toy model; (C) a bifurcation analysis showing that the four dm³ bifurcations correspond surjectively to the Whitney \(A_1\)–\(A_3\) singularity types, two-to-one on \(A_1\). Version V5 regenerates Appendix A from the Lean file (six of V4’s twelve entries named declarations that have never existed), restates result (C) as a surjection, corrects the Lean file path, and adds §5.3 on the \(D_6\)-equivariant DNLS results. Version V4 corrected the Contact Hopf bifurcation point in §5 (\(\gamma^* = e^{z_0} \to 2e^{z_0}\), propagated from the dm³ toy model V3 erratum). Version V3 added the certified inner-basin boundary \(r^* = 0.77594058\) (§4.7). Version 2a added: Lean 4 formal proof skeleton (VolumeTwo.lean), fully-reproducible figures (figures.py, exact §4.3 equations), interactive HTML dashboard, and the Mini-Beast companion document.
Volume I established that generative transitions are localised geometric events classified by the Whitney \(A_1\)–\(A_3\) hierarchy. The fold operator \(F\) was shown to act as a symplectic canonical transformation on \(T^*X\), and the full transition \(G = U \circ F \circ K \circ C\) was shown to be a piecewise-smooth symplectic map. What Volume I deliberately left open was the question of post-fold stability: once the fold has occurred and the unfolding \(U\) has selected a stable branch, what governs the long-term dissipative dynamics near that branch?
The Hamiltonian framework of Volume I cannot answer this question: Liouville's theorem forbids attractors in symplectic systems on compact manifolds. The answer is contact geometry. The contact manifold \(M = X \times \mathbb{R}\) with contact form \(\alpha = dz - \lambda\) provides the correct geometric setting for dissipative dynamics with limit cycle attractors, stochastic stability, and variational structure.
The central results are: a precise correspondence between the geometric fold operator \(F\) and the dm³ contact Hamiltonian dissipation (§2); a theorem establishing the equivalence of the curvature threshold \(\kappa^*\) and the embodiment threshold \(\tau\) (§3); explicit verification in the dm³ toy model (§4); and a bifurcation analysis (§5). Throughout, results of Volume I, Generative Contact Mechanics, and the dm³ Toy Model paper are taken as established.
Volume I established that generative transitions are classified by the Whitney \(A_1\)–\(A_3\) hierarchy. The fold operator \(F\) acts as a symplectic canonical transformation on \(T^*X\); the full transition \(G = U \circ F \circ K \circ C\) is a piecewise-smooth symplectic map. What Volume I left open: once the fold has occurred and \(U\) has selected a stable branch, what governs the long-term dissipative dynamics?
The Hamiltonian framework of Volume I cannot answer this: Liouville's theorem forbids attractors in symplectic systems on compact manifolds. The answer is contact geometry. The contact manifold \(M = X \times \mathbb{R}\) with contact form \(\alpha = dz - \lambda\), \(d\lambda = \omega\) provides the correct geometric setting.
The dm³ Toy Model [3] serves as a complete, explicit realization of the abstract contact-geometric framework and as the bridge certifying the realizability of the geometric fold theory of Volume I. The dm³ Toy Model proves that the generative transition framework is not merely definable, but fully instantiable by an explicit, smooth, globally analyzable dynamical system. Its role is logical, not empirical. See also the companion SIAM paper → and interactive dashboard →.
The fold operator \(F\) of Volume I is the piecewise-smooth, pre-contact limit of the dm³ operator \(A_{\mathrm{dm^3}} = \varphi^{T^*/4}\) of [2]. Under the contact extension \(M = X \times \mathbb{R}\), the impulsive momentum jump \(p^+ - p^- = \mu\mathbf{n}\) at the fold corresponds to the contact Hamiltonian correction \(H_{\mathrm{diss}} = -\gamma V e^{-\beta z}\) in the regularized limit \(\beta \to \infty\).
\[|\kappa|\uparrow\kappa^* \;\Longleftrightarrow\; \mu_{\max} < 0 \;\Longleftrightarrow\; \tau = \sqrt{c/\kappa_{\mathrm{noise}}} \in (0, \infty).\]
The curvature threshold \(\kappa^*\) and the embodiment threshold \(\tau\) are two parameterizations of the same event: the onset of transverse stability in the post-fold dissipative system.
The four bifurcations of the dm³ toy model [3] correspond to the Whitney singularity types of Volume I under the projection \(M = S \times \mathbb{R} \to S\). The correspondence is surjective, and injective away from \(A_1\): contact Hopf and saddle-node both arise from rank-1 loss in the radial direction and share the \(A_1\) type, so four bifurcations map onto three Whitney types, two-to-one at \(A_1\). It is not a bijection, and cannot be.
The operator sequence \(C \to K \to F \to U\), the threshold \(\kappa^*\), and all results of [1] hold.
The dm³ system, contact manifold \(M = S \times \mathbb{R}\), and all results of [2, 3] hold.
In Volume I, the fold is realized in \(T^*X\) as a symplectic discontinuity. At fold point \(s_0\):
generated by \(S(\gamma) = \mu\,\Theta(|\kappa(\gamma)| - \kappa^*)\) ([1], §12). The fold map \(\mathcal{F}: (\gamma,p) \mapsto (\gamma, p+\mu\mathbf{n})\) preserves \(\omega = d\gamma \wedge dp\) ([1], Theorem 12.1). To capture post-fold stabilization, we pass to the contact extension \(M = X \times \mathbb{R}\) with \(\alpha = dz - \lambda\), \(d\lambda = \omega\) ([2], Definition 6.1).
The contact dissipation \(H_{\mathrm{diss}}(x,z) = -\gamma V(x)e^{-\beta z}\) is a smooth regularization of \(S(\gamma) = \mu\,\Theta(|\kappa(\gamma)| - \kappa^*)\): as \(\beta \to \infty\) and \(z \to 0^+\), the correction \(-\gamma \nabla V e^{-\beta z}\) concentrates near \(\Gamma = \{V=0\}\) and mimics the threshold activation of \(\Theta\) at \(\kappa = \kappa^*\).
Structural properties: (i) \(H_{\mathrm{diss}}|_\Gamma = 0\); (ii) off \(\Gamma\): \(H_{\mathrm{diss}} < 0\); (iii) \(e^{-\beta z}\) weakens dissipation as action accumulates — the orbit earns its stability.
By [2] (Theorem C), every dm³ system near \(\Gamma\) is locally contact-diffeomorphic to the normal form:
| Volume I (geometric) | GCM / dm³ (contact) |
|---|---|
| Curvature threshold \(\kappa^*\) | Onset of transverse contraction |
| Fold impulse \(p^+ - p^- = \mu\mathbf{n}\) | Contact correction \(H_{\mathrm{diss}} = -\gamma V e^{-\beta z}\) |
| Rank-1 Jacobian loss | Dissipative contact normal form |
| Unfolding \(U\) | Gradient flow to \(\Gamma\) |
| Distributional generator \(S\) | Regularized \(H_{\mathrm{diss}}\) (Prop. 2.1) |
Theorem A is structural: the fold and the contact correction are two descriptions of the same event at different levels of regularization. The formal proof is open. thm_A_contact_realization_fold in VolumeTwo.lean states the hypotheses and carries the conclusion True as an explicit placeholder; the distributional convergence of \(e^{-\beta z}\) to \(\Theta\) as \(\beta \to \infty\) is not formalised.
If \(K\) drives curvature to \(\kappa^*\), \(F\) is rank-1, and \(U\) selects a nondegenerate branch, then \(\Gamma\) is hyperbolic with \(\mu_{\max} < 0\) and \(\dot V \leq -cV\) for some \(c > 0\).
Rank-1 loss at \(F\) and Morse nondegeneracy of \(\Phi\) yield transverse contraction. Floquet theory gives \(\mu_{\max} < 0\).
Under Lemma 3.1, \(\mathcal{L}V \leq -cV + \kappa_{\mathrm{noise}}\|\sigma\|^2\) and \(\tau = \sqrt{c/\kappa_{\mathrm{noise}}} \in (0,\infty)\).
\(\dot V \leq -cV\) plus the Itô correction \(\frac{1}{2}\|\sigma\|^2 \|\mathrm{Hess}\,V\|\) gives \(\mathcal{L}V \leq -cV + \kappa_{\mathrm{noise}}\|\sigma\|^2\) with \(\kappa_{\mathrm{noise}} = \frac{1}{2}\sup\|\mathrm{Hess}\,V\|\). Then \(\tau = \sqrt{c/\kappa_{\mathrm{noise}}} > 0\). (Proof requires Floquet theory + SDE stability — the 5-star sorry.)
If \(\mathcal{L}V \leq -cV + \kappa_{\mathrm{noise}}\|\sigma\|^2\) with \(c > 0\), then \(\mu_{\max} < 0\) and \(\dot V \leq -cV\) near \(\Gamma\).
If \(\tau \in (0,\infty)\), then the trajectory must have crossed \(\kappa^*\) and undergone a rank-1 fold.
By Lemma 3.3, \(\mu_{\max} < 0\), so a hyperbolic attracting cycle exists. Suppose \(|\kappa| < \kappa^*\) everywhere. Then [1] (Invariant 7.5, unconditional folding corollary) gives: below threshold no rank-1 fold occurs, hence no fold-generated attracting cycle. Contradiction.
V6 dependency note: this backward direction uses only the unconditional folding part of Invariant 7.5, so it needs no displacement condition (D). The unqualified global-injectivity form (previously cited here as "Invariant I5") is false without (D) — witnessed by the Gerono lemniscate — but Theorem 3.4, and therefore Theorem B (3.5), do not depend on it. Statements unchanged; citation narrowed.
\[ |\kappa|\uparrow\kappa^* \;\Longleftrightarrow\; \mu_{\max} < 0 \;\Longleftrightarrow\; \tau \in (0,\infty). \]
The curvature threshold \(\kappa^*\) is the geometric precursor of \(\tau\): curvature accumulation creates the conditions under which stochastic stability becomes meaningful.
Forward: Theorem 3.2. Backward: Theorem 3.4. Middle chain: \(\tau > 0 \Leftrightarrow c > 0 \Leftrightarrow \mu_{\max} < 0\) (Lemma 3.3). Scope: local to the fold neighbourhood and post-fold tubular neighbourhood of \(\Gamma\). Full proof requires Floquet theory + SDE regularity, and is not formalised: thm_B_threshold_equivalence in VolumeTwo.lean establishes only the middle⇔right link, and does so from assumptions on both sides.
There exists an explicit smooth dynamical system on a contact manifold — the dm³ toy model — in which all of the following hold by direct computation: (1) all eight dm³ axioms satisfied simultaneously; (2) explicit contact structure \(\alpha = dz - r^2 d\theta\); (3) contact normal form with \((\mu_{\max}, \omega, \beta) = (-2, 1, 1)\); (4) operator algebra closure; (5) stochastic stability with \(\tau = 2\); (6) global dynamics: attractor \(\Gamma_{12}\), four predicted bifurcations.
On \(M = \mathbb{R}^2_{>0} \times \mathbb{R}\) with polar coordinates \((r, \theta, z)\) and contact form \(\alpha = dz - r^2 d\theta\):
Limit cycle: \(\Gamma = \{r = 1\}\), period \(T^* = 2\pi\). The contact structure is non-degenerate: \(\alpha \wedge d\alpha = -2r\,dr \wedge d\theta \wedge dz \neq 0\) for \(r > 0\).
\(\mu_{\max} = -2\), \(\kappa_{\mathrm{noise}} = 1\), \(\tau = 2\). Verified by norm_num in Lean 4.
The transverse eigenvalue \(\lambda(z) = -2(1 - e^{-z})\) satisfies: \(\lambda(0) = 0\) (neutral, pre-embodiment); \(\lambda(z) < 0\) for \(z > 0\) (attracting, post-embodiment); \(\lambda(z) \to -2\) as \(z \to \infty\) (full dm³ rate).
Linearize at \(r = 1\): \(\dot\rho = -2(1-e^{-z})\rho + O(\rho^2)\). Generator: \(\mathcal{L}V = -4V(1-e^{-z}) + \sigma^2\), giving \(c \to 4\), \(\kappa_{\mathrm{noise}} = 1\), \(\tau = 2\).
The neutral stability at \(z = 0\) is the mathematical content of the embodiment threshold: the orbit earns its stability by accumulating action. The crossing \(z = 0 \to z > 0\) corresponds exactly to the curvature crossing \(\kappa \to \kappa^*\) in Volume I.
In coordinates \((\rho, \theta, z)\) with \(\rho = r - 1\), the system takes the contact normal form of [2] (Theorem C) with \((\mu_{\max}, \omega, \beta) = (-2, 1, 1)\):
\[ \varepsilon_0 = \frac{|\mu_{\max}|}{2(1 + \sup_\Gamma\|\mathrm{Hess}\,V\|)} = \frac{2}{2(1+2)} = \frac{1}{3}. \]
The basin asymmetry — \(\varepsilon_0 = 1/3\) is established for the outer basin \(\{r > r_{\mathrm{att}}\}\) only. The inner basin is asymmetric: trajectories from \(r_0 \in (2/3,\, r^*)\) lie inside the Gronwall ball and still escape to \(r \to 0\). The boundary \(r^* = 0.77594058\) is certified numerically by certify_rstar.py (DOP853, rtol 1e-12, atol 1e-14, bisection 1e-7), giving \(\varepsilon_0 = 1/3 < 2/3 < r^* \approx 0.776 < \kappa^* = \sqrt{7/9} \approx 0.882 < 1\). A numerical certificate is not a proof: no Lean declaration for this obligation exists.
From [3] (Theorem C), the dm³ toy model exhibits four bifurcations: (i) contact Hopf at \(\gamma = 2e^{z_0}\): limit cycle loses stability, new cycle bifurcates (corrected in V4; the earlier \(\gamma^* = e^{z_0}\) carried a spurious factor of 2 in the linearized transverse eigenvalue, and the canonical case \(\gamma = 2\) used throughout matches the corrected value at \(z_0 = 0\)); (ii) saddle-node at \(\eta \approx 0.15\): two cycles collide; (iii) Neimark–Sacker at detuning \(|\Delta| = \Delta^*\): resonant orbit loses stability, 2-torus bifurcates; (iv) slow-fast crossover at \(\beta = \beta^*\): smooth transition between contact and classical regimes.
Under projection \(M = S \times \mathbb{R} \to S\): \(A_1\) (codim 0) \(\leftrightarrow\) Contact Hopf + Saddle-node; \(A_2\) (codim 1) \(\leftrightarrow\) Neimark–Sacker [MODEL — see correction below]; \(A_3\) (codim 2) \(\leftrightarrow\) Slow-fast crossover.
Higher singularities excluded: in Volume I by the Morse condition; in [2] by contact normal form rigidity (Theorem C). Proves Theorem C for the A₁ and A₃ rows.
Correction notice (2026-08-07): The A₂ ↔ NS correspondence is [MODEL], not proved. The bare dm³ ODE linearises to a triangular 2×2 Jacobian J(z) = [[−2(1−e−z), 0], [2, 0]] with only real eigenvalues at every z — verified by direct computation. A Neimark–Sacker bifurcation requires a complex-conjugate eigenvalue pair, which is absent from the 3-equation system. The NS mechanism is the modulational instability of the DNLS amplitude equation (J hopping term, threshold λ_c = −2J/|A|²), correctly identified by the "2nd angular direction / detuning Δ" language in the table but not derivable from (r,θ,z) alone. [OPEN]
| Whitney Type | Codim | dm³ Bifurcation | Mechanism |
|---|---|---|---|
| \(A_1\) fold | 0 | Contact Hopf (H) | Rank-1 loss, radial direction |
| \(A_1\) fold | 0 | Saddle-node (SN) | Rank-1 loss, radial collision |
| \(A_2\) cusp | 1 | Neimark–Sacker (NS) | Rank-1, 2nd angular direction |
| \(A_3\) swallowtail | 2 | Slow-fast crossover (SF) | Rank-1, 3rd-order \(z\)-change |
\(A_1\) has two dm³ preimages (both from rank-1 loss in one transverse direction). Table 2 proves Theorem C.
(1) The fold operator \(F\) is the geometric precursor of the dm³ contact Hamiltonian \(H_{\mathrm{diss}}\); the distributional generator \(S\) is regularized by \(H_{\mathrm{diss}}\) in the limiting sense of Proposition 2.1 (Theorem A).
(2) \(\kappa^*\) and \(\tau\) are equivalent: each is finite and positive if and only if the other is, both detecting the onset of transverse stability (Theorem B).
(3) The four dm³ bifurcations correspond to Whitney \(A_1\)–\(A_3\) types, completing the singularity-theoretic classification of the toy model dynamics (Theorem C).
The Principia Orthogona series is a spiral, not a ladder. Each volume is a complete orbit — a full turn around the fixed point. \(G\) applied five times is \(G^5\) = Complete Completeness. \(G^6\) is still open (2026).
| G-Level / Volume | Core Concept | Fixed Point |
|---|---|---|
| G¹ / Vol. I | Abstract Operator Algebra | Orthogonal operator |
| G² / Vol. II (this) | Contact Geometry, g₃₃ = 33 | Contact fixed point |
| G³ / Vol. III | Biological Instantiations | Living form |
| G⁴ / Vol. IV | GTCT T1, IMPA | Temporal contact |
| G⁵ / Vol. V + AXLE | Banach FPT, formal proof | Complete Completeness |
| G⁶ / Issue 6 | χ(H*(X⁶)) = 33 ∀n | Open — 2026 |
Global Equivalence. Theorem B is local. A global version requires showing every \(\tau\)-stable dm³ system arises from a fold globally on \(X\).
Higher Resonances. Systematic treatment of \(k:m\) correspondence between higher \(A_k\) singularities and higher resonances.
Basin Asymmetry (AXLE Issue #13). \(\varepsilon_0 = 1/3\) is established for the outer basin only. The inner boundary r* = 0.77594058 is certified numerically (V3); the formal proof is open, and no Lean declaration for it exists in VolumeTwo.lean.
The Lean lives at AXLE/PrincipiaOrthogona_v2/VolumeTwo.lean. This table is generated from that file, not written alongside it. Reproduce with:
git clone https://github.com/TOTOGT/AXLE && cd AXLE
bash tools/verify-vol2/run.shThat builds the PrincipiaVol2 target at Lean v4.14.0, asks the kernel for the axiom dependencies of all nineteen declarations, and refuses on sorryAx or on any axiom outside {propext, Classical.choice, Quot.sound}. All nineteen pass, and two depend on no axioms at all.
The state described here is AXLE commit e44e8d1 — pinned to a commit rather than a branch, so it resolves to what was checked when this was written.
Passing that gate is necessary and not sufficient. A theorem whose conclusion is True, or whose hypotheses already contain its conclusion, is a true theorem and reports clean axioms. Three rows below are marked for exactly that reason, and the marks are the point of the table.
| Theorem / Lemma | Lean Name | Status | Notes |
|---|---|---|---|
| τ > 0 (Thm 3.2) | embodimentThreshold_pos | ✓ proved | sqrt_pos_of_pos |
| τ = 2 (Prop. 4.2) | toyModel_tau | ✓ proved | Real.sqrt_sq |
| λ(0) = 0 (Prop. 4.2) | eigenvalue_at_zero | ✓ proved | simp |
| λ(z) < 0 for z > 0 | eigenvalue_neg_pos_z | ✓ proved | Real.exp_lt_exp |
| λ(z) → μmax | eigenvalue_limit | ✓ proved | const_mul_atTop_of_neg, tendsto_exp_atBot |
| Thm 3.3 stability certificate | vol2_contact_Theorem_3_3 | ✓ proved | conjunction of the two preceding rows |
| ε0 = 1/3 (§4.6) | toyModel_epsilon0 | ✓ proved | norm_num |
| NJ|Γ = 0 (Thm 15.2, Level 1) | Theorem_15_2_integrability | ✓ proved | dimension count on a 1-dimensional tangent space |
| alternating m-form vanishes when m > dim | alternating_vanishes_beyond_dim | ✓ proved | map_linearDependent |
| Thm C — unique preimages of A₂, A₃ | thm_C_singularity_bijection | ✓ proved | the name overstates it: no surjectivity conjunct, and the map is not a bijection |
| A₁ surjectivity | thm_C_A1_surjective | ✓ proved (V5) | closes a row V4 listed as proved against a name that existed nowhere. No axioms. |
| the map is not a bijection | thm_C_not_bijective | ✓ proved (V5) | contact Hopf and saddle-node share the A₁ preimage, so injectivity fails. No axioms. |
| Thm A — correction vanishes off the fold | thm_A_regularization_pointwise | ✓ proved (V5) | the pointwise skeleton of the distributional claim |
| Thm A — correction constant on the fold | thm_A_regularization_at_fold | ✓ proved (V5) | the other half: it does not move with β at z = 0 |
| τ = |μ| iff |μ| = 2 | tau_eq_abs_mu_iff | ✓ proved (V5) | the toyModel_tau docstring caveat, now a theorem: τ = |μmax| is a property of the canonical value, not an identity |
| ε0 = 1/3, Waddington reading | epsilon_zero_waddington | duplicate | statement identical to toyModel_epsilon0; proof is := toyModel_epsilon0 |
| μmax + τ = 0 | entropy_lyapunov_duality | ⚠ arithmetic | reduces to −2 + 2 = 0; c = 4 is supplied by hand, not derived from μmax |
| Thm B — μmax < 0 ⇔ τ > 0 | thm_B_threshold_equivalence | ⚠ trivial | both branches discard the hypothesis; μmax < 0 is a field of DM3System, so each side is independently true |
| Thm A (full) | thm_A_contact_realization_fold | ⚠ placeholder ★★★★ | conclusion is True. Not a sorry — so no axiom check can flag it. Distribution theory unformalised. |
| Thm B (full chain |κ| ↑ κ*) | — | ⚠ open ★★★★★ | no declaration; proved at the prose level in §3.3 only. Floquet + SDE. |
| Gronwall asymmetry, r* = 0.77594058 | — | ⚠ open ★★★ Issue #13 | no declaration; numerical certificate only |
| NJ|ξ = 0 (Level 2d) | — (OP4) | ⚠ open ★★★★ | not yet statable: NJ needs Lie brackets of vector fields |
| NJ|M = 0 (Level 2d+t) | — (OP5) | ⚠ open ★★★★ | depends on OP4 |
-- lambda(z) < 0 for z > 0 (proved, no sorry, kernel-checked)
theorem eigenvalue_neg_pos_z (sys : DM3System) (z : ℝ) (hz : 0 < z) :
transverseEigenvalue sys z < 0 := by
unfold transverseEigenvalue
apply mul_neg_of_neg_of_pos sys.mu_neg
have hbz : -sys.beta * z < 0 := by nlinarith [sys.beta_pos]
have hexp : Real.exp (-sys.beta * z) < Real.exp 0 := Real.exp_lt_exp.mpr hbz
rw [Real.exp_zero] at hexp
linarith
Full source: github.com/TOTOGT/AXLE · PrincipiaOrthogona_v2/VolumeTwo.lean · verifier tools/verify-vol2/run.sh
Acknowledgments. The author acknowledges with gratitude the foundational influence of the teachings of Paramahamsa Nithyananda and the yogic scriptural tradition from which this work derives.
The toy model SIAM paper proves four theorems (A–D) with full global dynamical analysis. The interactive dashboard provides live exploration of all dm³ figures.
Toy Model Paper → Interactive Dashboard → ← Volume IsorryAx. A clean axiom report is not a reading of the statement: per R20, a theorem can assume its conclusion and still report clean. Follow the link before citing one as evidence.