Analytic measures and Bochner measurability (Q1383262)

The authors are concerned with abstract F. and M. Riesz theorems in a group action context, like that explored by \textit{K. de Leeuw} and \textit{I. Glicksberg} [Acta Math. 109, 179-205 (1963; Zbl 0126.12102)] and \textit{F. Forelli} [Acta Math. 118, 33-59 (1967; Zbl 0171.34201)]. Their setting is still more general and abstract, and when specialized to the de Leeuw-Glicksberg-Forelli context recovers the results of these authors under substantially weaker hypotheses. They give examples showing that their results cannot be obtained via the earlier approach. Here is the setup: \(\Sigma\) is a sigma algebra of subsets of a set \(\Omega\), \(M(\Sigma)\) the complex measures on \(\Sigma\) with the total variation norm. \(T=\{T_t\}_{t\in \mathbb{R}}\) is a set of isomorphisms of \(M(\Sigma)\) and \(\sup \{\| T_t^{\pm 1} \|: t\in\mathbb{R}\} <\infty\). \(\mu\) is called weakly analytic if for each \(S\in\Sigma\), \(t\mapsto T_t\mu (S)\) is Lebesgue measurable and lies in \(H^\infty (\mathbb{R})\). \(T\) is said to satisfy hypothesis \((A)\) if the only weakly analytic measure \(\mu\) such that for each \(S\in\Sigma\), \(T_t\mu (S)=0\) for Lebesgue-almost-all \(t\) is the measure \(\mu=0\). It is also required in much of the paper that for each \(t\) the adjoint \(T_t^*\) map the bounded measurable functions, viewed as a subspace of \(M(\Sigma)^*\), into itself. The authors prove that under these hypotheses, \(t\mapsto T_t\mu\) is Bochner measurable and uniformly continuous for every weakly analytic \(\mu\). They use the analytic Radon-Nikodým property studied by \textit{A. V. Bukhvalov} and \textit{A. A. Danilevich} [Math. Notes 31, 104-110 (1982); translation from Mat. Zametki 31, 203-214 (1982; Zbl 0496.30029)]. Suppose in addition to the above that \(\{T_t\}_{t\in\mathbb{R}}\) is a one-parameter group. A measure \(\nu\in M(\Sigma)\) is called quasi-invariant if \(\nu\) and \(T_t\nu\) are mutually absolutely continuous for all \(t\). When a weakly analytic measure \(\mu\) undergoes its Lebesgue decomposition \(\mu_a+\mu_s\) with respect to a quasi-invariant \(\nu\), then \(\mu_a\) and \(\mu_s\) are each weakly analytic (abstract F. and M. Riesz theorem). Let \(\Omega\) be a topological space, \(\Sigma\) its Baire sigma algebra, \(\{h_t\}_{t\in \mathbb{R}}\) a family of homeomorphisms of \(\Omega\) and define \(T_t\) by \(T_t\mu =\mu \circ h_t\). Then the preceding F. and M. Riesz type conclusion follows. If moreover \(\{h_t\}_{t \in\mathbb{R}}\) is a one-parameter group, then every weakly analytic \(\mu\) is quasi-invariant. The paper is well written, with a long informative introduction and frequent pauses to make contact with predecessor results and orient the reader to the goals being pursued. In a sequel paper the authors will investigate replacing \(\mathbb{R}\) throughout by a locally compact abelian group with ordered dual. Unfortunately, the authors' short bibliography overlooks some related work of \textit{H. Yamaguchi} [Hokkaido Math. J. 20, 109-121 (1991; Zbl 0742.43003); 21, 349-364 (1992; Zbl 0784.43001); 25, 567-589 (1996; Zbl 0874.43001)].