The first exact-likelihood, held-out test of cross-domain crisis entanglement

The polycrisis field's founding empirical claim is that crises are causally entangled across domains — that wars, disasters, outbreaks, defaults and coups make each other more likely. The claim has been argued conceptually and illustrated with co-occurrence plots, but never written down as a quantity that could be shown to be false. This paper supplies that object for the timing of crisis onset and then tests it: crisis streams are modelled as a multivariate point process, cross-domain coupling is estimated by exact likelihood against a shared-latent-regime rival, held out of sample, and referred to a circular-shift null that destroys cross-domain alignment while preserving each domain's own autocorrelation. Every verdict is gated on a decidability atlas computed on known-truth simulation before the data were read. The result is a negative that is informative because its power is mapped. On the field's own corpus the cross-domain excess is absent; on a finer country-month rebuild it survives the raw nulls and dies under a declared family of reporting baselines. Onset-time entanglement, as the field states it, does not survive an exact-likelihood test with a declared artifact channel. Correction carried in this version (2026-08-25 / 2026-08-30). The control this paper calls on to bound a shared media-attention confound is renamed to what it is: a variance-maximal shared-modulation absorption bound, an upper bound on how much of the cross-domain excess any single shared modulation could absorb. No attention data enters it — the covariate is the leading right singular vector of the standardised residual matrix, i.e. the single series that maximises shared cross-domain residual variance by construction. The earlier inference that a timing-destroyed placebo “confirms the specificity” of the absorption is withdrawn as circular: the direction is computed on the same residuals whose covariance is the excess under test, and the placebo compares the variance-maximal shared mode against a scrambled version of itself. The bound remains real, tight, pre-committed by name, and downside-only. Programme context. Fenocrisis is the crisis-entanglement repository of the Feno programme. Its discipline is stated once and applies to every record here: a verdict ships only where a decidability atlas computed on known-truth simulation says the design could have seen the effect; every number regenerates bit-for-bit from a pinned seed; the results ledger is append-only; and nothing forecasts, warns, or ranks — every figure is a retrospective model comparison on public, aggregate data. Part of the Feno program. The source repository is private; the reproduction materials for this record are contained in the record itself.