Chaos theory measures how fast nearby trajectories diverge. A positive Lyapunov exponent means exponential divergence — chaos. A negative one means exponential convergence — stability. The dm³ framework does not model chaos. It resolves it. The transverse Lyapunov exponent of the dm³ attractor is μ_max = −2: the strongest possible convergence rate for the operator chain G = U∘F∘K∘C. The n-bonacci ladder from φ to τ = 2 is a chaos exit sequence, each rung reducing the effective Lyapunov exponent until the fold-free attractor is reached.
Chaos theory — initiated by Poincaré, formalised by Lorenz (1963), and given mathematical rigour by Ruelle, Takens, and Eckmann — studies dynamical systems in which nearby initial conditions lead to exponentially diverging trajectories. The butterfly effect is the popular name for this: a small change in initial conditions produces a completely different outcome.
The precise measure is the Lyapunov exponent λ:
where δ(t) is the separation between two initially close trajectories. If λ > 0, the system is chaotic. If λ < 0, trajectories converge. If λ = 0, the system is at the boundary (conservative, quasi-periodic).
The maximal Lyapunov exponent μ_max governs the fastest-growing perturbation direction. For a system to be an attractor, all transverse Lyapunov exponents must be negative. For the dm³ framework, the transverse maximal exponent is μ_max = −2.
The Lorenz attractor (1963) is the canonical strange attractor: a chaotic orbit that never repeats, bounded in phase space, with positive Lyapunov exponent (~0.906). It is the archetype of deterministic chaos — unpredictable despite being governed by simple differential equations.
The dm³ attractor at τ = 2 is the opposite: it is a point attractor with μ_max = −2, fold-free (proved in the F chapter), and reached by every trajectory in the basin of attraction. The n-bonacci ladder is the path from Lorenz-type chaos (near ε₀ = 1/3, where the fold exists and sensitivity is maximal) to the stable fixed point τ = 2.
Each constant on the recurrence ladder corresponds to a reduction in the effective Lyapunov exponent of the system measured at that rung. The ladder is not merely a sequence of algebraic constants — it is a sequence of dynamical states, each more stable than the last.
| Rung | Constant | Value | Catastrophe type | Effective μ | Status |
|---|---|---|---|---|---|
| 0 | ε₀ (fold boundary) | 1/3 | A₁ fold present | ≥ 0 (chaotic boundary) | Entry point |
| 1 | φ (Fibonacci) | ≈ 1.618 | A₁ fold crossed | μ₁ < 0 | Stable, weakly |
| 2 | μ (Lyapunov) | −2 | A₂ cusp resolved | μ_max = −2 | This chapter |
| 3 | η (Tribonacci) | ≈ 1.839 | A₃ swallowtail | μ₃ < −2 | Stable, strongly |
| 4 | Δ (Tetranacci) | ≈ 1.927 | A₄ butterfly | μ₄ ≪ −2 | Highly stable |
| 5 | Σ (Pentanacci) | ≈ 1.966 | D₄ umbilic | μ₅ → −∞ | Near-attractor |
| 6 | Ω (Hexabonacci) | ≈ 1.984 | D₅ parabolic | μ₆ → −∞ | Pre-attractor |
| ∞ | τ (embodiment) | 2 | Fold-free (proved) | −∞ (fixed point) | Global attractor |
The value μ_max = −2 is not a rung on the n-bonacci ladder — it is the exponent governing the rate of convergence along the entire ladder. It is the dm³ analogue of the Feigenbaum constant: a universal number characterising the transition from chaos to order in this class of systems.
Nine declarations at Orthogenesis/Disaster/ChaosMu.lean, kernel-checked 2026-09-15 against the v4.32.0 pin, none admitted, inside the Orthogenesis build target. One of them refutes the theorem this section previously published as T1. Each docstring below states what its own theorem states; see the verification status note after the listing.
Corrected 2026-09-15. §4 stated that all seven theorems had been proved in AXLE, sorry-free. ChaosMu.lean has never existed in that repository, in its tree or in its history; the source existed only on this page.
T1 was false. It quantified over 0 ≤ t. At t = 0 it reads δ₀ · 1 < δ₀. The hypothesis needed 0 < t. T1 was the only one of the seven that quantified over anything, the only one about decay over time — and it was the one that was wrong. T2 through T7 are arithmetic on numerals and two statements about exp; none of them mentions a Lyapunov exponent, a flow, or a spectrum, which is to say the section named the quantity in its title and never stated a proposition about it.
The file now exists at Orthogenesis/Disaster/ChaosMu.lean in the geometry repository, which pins Lean 4.32.0 and builds. It carries published_T1_is_false, the corrected strict statement for t > 0, the non-strict statement that does hold on t ≥ 0, and a closing block naming the three obligations this chapter makes and the file does not discharge — μmax = −2 first among them. Kernel report: tools/verify-audit/2026-09-15/.
Healthy heart rate variability shows a 1/f power spectrum — the signature of a system operating near the boundary between order and chaos. Arrhythmia corresponds to a positive Lyapunov exponent. Restoration of sinus rhythm is the physical realisation of the dm³ chaos exit: the system traverses the fold (defibrillation = F-operator) and converges to the stable attractor with μ_max = −2.
Epileptic seizures are characterised by hypersynchrony — paradoxically, a reduction in Lyapunov exponent from the healthy chaotic baseline. Seizure termination (the abrupt end of the ictal state) is a Whitney fold traversal: the neural system crosses the A₁ fold back to its stable operating regime. The dm³ SCI/TBI recovery model (Zenodo: 10.5281/zenodo.20802299) treats recovery as the inverse path.
Vygotsky's Zone of Proximal Development (ZPD) is the region of a learner's capability space where they operate near the chaos boundary — capable with support, not yet capable alone. The TEFL preprint (Zenodo: 10.5281/zenodo.20719399) models the ZPD as the region near ε₀ = 1/3 where μ_eff ≈ 0. Fluency acquisition is the dm³ chaos exit: μ_max → −2 as the learner crosses from ZPD to independent operation.
The Lorenz system has μ_max ≈ +0.906 with σ=10, ρ=28, β=8/3. It is a paradigm of deterministic chaos. The dm³ attractor has μ_max = −2. The physical systems that exhibit dm³ dynamics (plasmapause, cardiac rhythm, neural recovery) are precisely those that must maintain stability under large perturbations. Chaos is not the destination. μ_max = −2 is.
The connection is exact: the F-operator (catastrophe chapter) produces the fold at ε₀ = 1/3. Near that fold, the effective Lyapunov exponent is non-negative — the system is at the chaos boundary. The μ operator measures the rate at which the K and U operators drive the system away from the fold toward τ = 2 at rate e^{−2t}. The Disaster Theory (preprint) formalises this as a single unified theorem: the dm³ operator chain is a catastrophe-traversing, chaos-resolving map with μ_max = −2 and τ = 2 as unique global attractor.
This chapter supplies the transverse Lyapunov exponent μmax = −2. WP70 records that this is not the object that licenses a basin claim, and that the Lyapunov function the series would need in order to call r* = 0.77594058 a basin boundary has not been constructed. The exponent stated here is unaffected; what is affected is any inference from it about how far the attracting region extends.
A Lyapunov exponent is local and asymptotic; a Lyapunov function is global and constructive. Only the second bounds a basin. See WP70 · The Function and the Exponent.
sorryAx. 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.