Dividual posterior

What the plate operates. Proposition 4' gives the classical parametric-rate Bayesian posterior contraction under three conditions (prior support, testability via U2, KL-neighborhood prior mass). The plate visualizes the contraction directly.

The two parameters.

• d (dimension). Higher-dimensional user-state spaces give larger posterior variance at the same sample size. The log term in $\sqrt{d\log t/t}$ grows slowly, so the parametric rate dominates; the curve shape is preserved across $d$.
• Effective-sample multiplier (axiom U1). (U1) gives a lower bound on the per-tick excitation probability $p_0>0$. The effective sample size is $\ge p_{0}t$. Increasing the multiplier accelerates contraction; decreasing it slows the rate.

What the proposition authorizes the prose to claim. The dividual is the user-state-as-jointly-observable. Identifiability is supplied by (U2); excitation is supplied by (U1); parametric stability is supplied by (R1)+(R2). Deleuze's dividual takes operational form here as the platform's filter at the converged regime — a formal object carrying the same weight as foreclosure. The user-state continues to be whatever the user is; the platform converges on the user's state in the standard Bayesian-posterior sense.