Theorem 12 (crisis-boundary bifurcation classification)

The framework's regularity hypothesis (R1) requires the operator to contract: $\kappa :={\kappa }_{\Phi }<1$, where ${\kappa }_{\Phi }\le (\beta /\alpha )L_{\mathcal{D}}$ is the operator modulus of §0.3 — not the value-sensitivity product $L_{\hat{R}}{L}_{\mathcal{D}}$. As $\kappa \to 1^{-}$, the closure amplification factor $(I-D_{\theta }\Phi {)}^{-1}$ from Theorem 9 diverges. For $\kappa \ge 1$, the apparatus's stable-point analysis fails outright. This section classifies what dynamical regimes obtain at and beyond the crisis boundary.

Linearization at the stable point. Near ${\theta }^{\ast }$, the iteration ${\theta }_{n+1}=\Phi ({\theta }_n)$ has linearization ${\theta }_{n+1}-{\theta }^{\ast }\approx J^{\ast }({\theta }_n-{\theta }^{\ast })$ where $J^{\ast }:=D_{\theta }\Phi ({\theta }^{\ast })$. By (R1), $\rho (J^{\ast })\le \kappa$. The local-stability classification depends on the location of $J^{\ast }$'s eigenvalues relative to the unit circle. All inside: asymptotically stable (the apparatus's regime). On the unit circle: marginal; bifurcation analysis required. Outside: unstable.

The bifurcation regimes. At $\kappa =1$ a single eigenvalue of the linearization touches the unit circle (the generic codimension-1 case). Where it touches fixes the bifurcation type. Three types are generic in a one-parameter family:

• Saddle-node — eigenvalue at $+1$, quadratic non-degeneracy. Two fixed points collide and annihilate at threshold. Above threshold: iteration diverges or jumps to a distant attractor.
• Pitchfork (non-generic) — eigenvalue at $+1$, cubic non-degeneracy. It is not generic in a one-parameter family; it replaces the saddle-node only when $\Phi$ commutes with a ${\mathbb{Z}}_2$symmetry — which the framework's symmetric cohorts can supply. A single stable fixed point becomes unstable; two symmetric stable fixed points appear.
• Period-doubling (flip) — eigenvalue at $-1$. Stable fixed point becomes unstable; stable 2-cycle appears. With increasing parameter: period-2 → period-4 → … → chaos via the Feigenbaum cascade.
• Neimark–Sacker — complex eigenvalue pair $e^{\pm i\omega }$ on the unit circle. Stable fixed point becomes unstable; stable invariant circle appears with amplitude $\sqrt{\kappa -1}$ in the super-critical case.

All three — saddle-node, period-doubling, Neimark–Sacker — are generic codimension-1 events; which one obtains is a property of $\Phi$ at $({\theta }^{\ast },\kappa =1)$; the three are equally typical. The pitchfork is the lone exception — it requires the ${\mathbb{Z}}_2$-equivariance of a symmetric cohort and is non-generic without it.

Empirical signatures. Each regime has an observable platform-dynamical signature:

Real platforms likely exhibit a mixture: Neimark–Sacker oscillations at intermediate timescales, saddle-node crises at longer timescales, period-doubling where self-correction overshoots, and the pitchfork where a symmetric cohort supplies the ${\mathbb{Z}}_2$ symmetry it requires.

What the theorem authorizes the prose to claim. The platform itself is in crisis. Repeated algorithm-rebuild events, monetization pivots, rage-bait cycles, viral cascade rhythms — these are bifurcation phenomena at the contractivity boundary the framework names. The platform reads them as anomalies; the theorem reads them as structure. The platform-crisis chapter of Part II develops the substantive content; Theorem 12 supplies the formal vocabulary.

Platform-history reading.The framework's bifurcation classification offers a vocabulary for events the industry typically describes as engineering anomalies, “algorithmic shifts,” or content-policy upheavals. Twitter's recurrent algorithmic crises in 2014–2016 — the move from chronological to algorithmic timeline, the shadow-banning controversies, the platform's stated identity instability — plausibly trace a saddle-node trajectory. The viral content rhythms on TikTok's For You Page (content categories rising to platform dominance and decaying in characteristic cycles) exhibit the rotational dynamics Neimark–Sacker classifies. Facebook's repeated News Feed pivots between video-priority and friends-and-family-priority read as period-doubling responses to overshooting engagement targets. The framework does not claim that these specific events are these specific bifurcations — empirical work would have to demonstrate that — but supplies the formal vocabulary against which those events could be classified.

The reader can operate the crisis-boundary plate to view the bifurcation diagram for each regime — the three generic cases and the symmetry-protected pitchfork.