INVARIANT GRAPH AND RANDOM BONY ATTRACTORS

The paper proves Theorem A for an explicit open neighborhood U of a carefully constructed step skew product F, establishing: (1) A_max(G) is either a continuous invariant graph or a bony attractor, (2) existence of an invariant ergodic measure supported on the closure of the invariant graph, measurable isomorphism to the Bernoulli shift (hence Bernoulli and mixing), and continuity in the Hutchinson metric, and (3) negative fiber Lyapunov exponent; moreover, the bony subset of U is nonempty. These elements are stated and developed in Theorem A and ensuing propositions/lemmas (Theorem A: ; fiber Lyapunov exponent and its negativity: ; invariant graph and ergodic/Bernoulli measure: ; dichotomy continuous graph vs. bony attractor and closure-of-graph property: ; explicit nonempty bony subset: ; continuity in Hutchinson metric: ). By contrast, the model’s core mechanism assumes the graph transform T is a strict contraction on L^1 solely from E[L] < 1, concluding uniqueness via Banach’s fixed-point theorem. This is incorrect: from |Tφ − Tψ|(ω) ≤ L(σ^{-1}ω)|φ − ψ|(σ^{-1}ω) one only gets ∥Tφ − Tψ∥_1 ≤ ∫ L·|φ−ψ| dν, which does not imply ∥Tφ − Tψ∥_1 ≤ E[L]∥φ − ψ∥_1 in general. Without uniform-in-ω contraction (e.g., sup L < 1 or a different norm/weight), Banach’s contraction in L^1 does not follow. The paper instead uses the pullback (backward iteration) construction with Kingman’s theorem to produce an a.e. limit γ_G and attraction to the graph, avoiding this pitfall (). The model also invokes Egorov from L^1 convergence (not a.e. convergence), which is not justified, and sketches continuity of μ_G via a non-rigorous operator-norm estimate that again rests on the flawed L^1 contraction. While several high-level claims in the model match the paper’s conclusions (existence of invariant graph, negativity of the exponent, bony example), key steps rely on incorrect or unproven inequalities. Therefore, the paper’s argument is substantively correct, while the model’s proof outline contains critical errors.