The Arithmetic Seed

Chapter 10 gave us the helical attractor on a smooth contact 3-manifold — the dm³ ODE on cylindrical $(r,\theta,z)$ space. The attractor was the unit helix $r=1$. Now we make the same move with a different seed: instead of the smooth Hopf normal form, we start with the Riemann zeta function. Instead of a synthetic ODE, we have an analytic object built from all prime numbers simultaneously. The contact form that results is not smooth — it is woven from the entire prime counting machine.

The Tuning Fork Argument

Strike a tuning fork. It vibrates at a single frequency — not because you imposed that frequency, but because the metal has no choice. The length, mass, and elasticity of the tine determine the frequency the way a theorem follows from its axioms.

The Riemann Hypothesis says something similar about the prime numbers. The non-trivial zeros of the Riemann zeta function — points $s = \sigma + it$ in the complex plane where $\zeta(s) = 0$ and $0 < \sigma < 1$ — appear to lie on the single vertical line $\sigma = \tfrac{1}{2}$. Not because someone placed them there. Because the arithmetic structure of the primes, encoded in $\zeta(s)$, has no other choice. This has not been proved. But Chapter 11 asks a prior question: can we hear the structure that would force this?

The Prototype Recalled

Chapter 2 began with the harmonic oscillator $\dot{x} = y,\ \dot{y} = -x$. Its orbits in the phase plane are circles. When we promoted $t$ to a coordinate, those circles became helices in $\mathbb{R}^3_{(x,y,t)}$, governed by the contact form $\alpha = dy + x\,dx$. The key was non-integrability: $\alpha \wedge d\alpha \neq 0$. The planes twisted; the helix was forced to move forward.

Chapter 11 performs the same lift with $\zeta(s)$ as the starting object. The phase plane coordinates are the real and imaginary parts $U, V$ of $\zeta(\sigma+it)$. The parameter promoted to a coordinate is the imaginary part $t$ of $s$. The contact form that emerges carries the entire prime counting function as its coefficient.

The Zeta Phase Plane

Write $\zeta(\sigma+it) = U(\sigma,t) + i\,V(\sigma,t)$. Fix $\sigma$ and let $t$ increase: the pair $(U(\sigma,t), V(\sigma,t))$ traces a curve in the $UV$-plane — the zeta trajectory at height $\sigma$. A non-trivial zero at $s_0 = \sigma_0 + it_0$ is the moment this curve passes through the origin: $U = V = 0$ simultaneously.

The Riemann Hypothesis in this language: the trajectory only passes through the origin when $\sigma = \tfrac{1}{2}$. For $\sigma \neq \tfrac{1}{2}$ in the critical strip $0 < \sigma < 1$, the curve never touches the origin.

The interactive diagram below makes this visual. Adjust $\sigma$ to see how the trajectory changes. At $\sigma = 0.5$ the trajectory is shown in teal and its zeros are marked.

The Lift to 3D

Just as in Chapter 2, we promote $t$ to a coordinate. The lifted curve is: $$\gamma(t) = \bigl(U(\sigma,t),\; V(\sigma,t),\; t\bigr) \in \mathbb{R}^3_{(U,\,V,\,t)}.$$ A non-trivial zero at $s_0 = \sigma_0 + it_0$ is a piercing event: the curve $\gamma$ passes through the axis $\{U=V=0\}$ at height $t = t_0$. The Riemann Hypothesis says all piercing events occur only on the plane $\sigma = \tfrac{1}{2}$.

The Arithmetic Contact Form

The tangent velocity of the lifted curve satisfies $\partial_t V = g(\sigma,t) \cdot \partial_t U$ (from the Cauchy–Riemann equations applied to $\log\zeta$). This means the curve lies in the kernel of the 1-form:

Why von Mangoldt? Because it appears in the logarithmic derivative: $-\zeta'/\zeta(s) = \sum_{n=1}^\infty \Lambda(n)/n^s$. The coefficient $g(\sigma,t)$ is the imaginary part of this, encoding all primes simultaneously in a single real function.

Non-Integrability

The exterior derivative is $d\alpha_{\mathrm{arith}} = -(\partial_t g)\,dt \wedge dU$ where: $$\partial_t g = \sum_{n=1}^{\infty} \frac{\Lambda(n)\log n}{n^\sigma}\cos(t\log n).$$

What this means: The contact planes cannot fold flat. They twist continuously, driven by the prime frequencies. Trajectories in $\ker\alpha_{\mathrm{arith}}$ are forced to move forward in $t$. The primes are the engine of the twisting.

What this does not mean: It does not yet force zeros to $\sigma = \tfrac{1}{2}$. That requires a global statement about where piercing events occur — the subject of Ch 14.

Zeros as Poles of the Contact Structure

At a non-trivial zero $s_0 = \sigma_0 + it_0$, the logarithmic derivative $\zeta'/\zeta(s)$ has a simple pole. The coefficient $g(\sigma_0, t_0)$ therefore diverges — the contact planes twist infinitely rapidly at the exact moment $\gamma$ pierces the floor $\{U = V = 0\}$.

In the smooth dm³ system of Ch 10, convergence to the attractor was smooth throughout. In the arithmetic case, zeros are singularities of the contact structure itself — the tightest winding occurs precisely where the primes assert their constraint most forcefully.

The G-Chain in Arithmetic

The operator chain $G = U \circ F \circ K \circ C$ reappears in the arithmetic setting:

Lean 4 Connection

The AXLE repository contains a type stub for this chapter's contact form:

-- AXLE/lean/RH/ArithmeticContact.lean
import Mathlib.NumberTheory.VonMangoldt

namespace AXLE.Arithmetic

noncomputable def g_arith (σ t : ℝ) : ℝ :=
  -- Im(-ζ'/ζ(σ+it)) via von Mangoldt series
  sorry -- AXLE Issue #18: connect to Mathlib.vonMangoldt

-- Non-integrability · Ch11 main result
theorem arith_contact_nonintegrable (σ : ℝ) (hσ : 1 < σ) :
    ∀ t : ℝ, ∃ ε > 0, ∀ t' ∈ Set.Ioo (t - ε) (t + ε),
    (∑ n : ℕ, vonMangoldt n / (n : ℝ)^σ * Real.log n
             * Real.cos (t' * Real.log n)) ≠ 0 := by
  sorry -- AXLE Issue #19: Baker–Wüstholz linear independence

end AXLE.Arithmetic

Issue #19 is closeable using Mathlib's Baker–Wüstholz transcendence machinery (linear forms in logarithms) — non-trivial, approximately 80 lines.

Honest Inventory

Exercises