Reason: S4.4 item 1: regularity, derivative formulas, restriction, and simplex bounds (B-tilde+K-tilde, sqrt(l)(B-tilde+K-tilde) Lipschitz, 3K-tilde second order, uniform continuity) for the extended aggregate observation drift, the observation-side counterpart of lem:extended-drift-regularity-2026a. Internally reviewed with every constant verified; validation clean.
Statement
Let l and l~ be \reftext{def:natural-numbers-2026a}{natural numbers} with l≥2 and l~≥1, let β~ be an \reftext{def:observation-rate-family-2026a}{observation-rate family} on l states with l~ observation channels and rate bound B~, let (U~,β~ˉ) be a \reftext{def:c2-observation-rate-extension-2026a}{twice continuously differentiable extension} of β~ with derivative bound K~, and let b~ˉ be the \reftext{def:extended-aggregate-observation-drift-2026a}{extended aggregate observation drift} of (U~,β~ˉ). Adopt the coordinate and partial-derivative notation ∂γ, ∂δ∂γ of the extension definition, write Δl for the \reftext{def:probability-simplex-2026a}{probability simplex}, and write d for the \reftext{def:euclidean-distance-rn-2026a}{Euclidean distance}. Then for every υ∈{1,…,l~}:
\textbf{(i) (Regularity, derivative formulas, and restriction.)} b~ˉυ is a \reftext{def:c1-map-euclidean-open-set-2026a}{C1 map} on U~, each ∂γb~ˉυ is again a C1 map on U~, and for all γ,δ∈{1,…,l} and Σ∈U~:
Moreover b~ˉ agrees on Δl with the \reftext{def:aggregate-observation-drift-2026a}{aggregate observation drift} of β~.
\textbf{(ii) (First-order bounds on the simplex.)} ∣∂γb~ˉυ(Σ)∣≤B~+K~ for all γ∈{1,…,l} and all Σ∈Δl, and
∣b~ˉυ(Σ)−b~ˉυ(Σ′)∣≤l(B~+K~)d(Σ,Σ′)for all Σ,Σ′∈Δl.
\textbf{(iii) (Second-order bounds and uniform continuity on the simplex.)} ∣∂δ∂γb~ˉυ(Σ)∣≤3K~ for all γ,δ∈{1,…,l} and Σ∈Δl; and for every real ε>0 there is a real δ∘>0, which may be chosen independently of γ, δ, and υ, such that ∣∂δ∂γb~ˉυ(Σ)−∂δ∂γb~ˉυ(Σ′)∣≤ε for all γ,δ∈{1,…,l}, all υ∈{1,…,l~}, and all Σ,Σ′∈Δl with d(Σ,Σ′)≤δ∘.
Curated associations between results. These are editable and subjective — they do not replace the dependency graph, which is derived from the references in the text.