No label defined (Q4397330)

The main result of the paper is a generalization of a divergence result proved by \textit{A. del Junco} and \textit{J. Rosenblatt} [Math. Ann. 245, 185-197 (1979; Zbl 0408.28015)] for the case of Banach spaces to Baire topological spaces. Recall that a subset \(X_0\) of a topological space \(X\) is called to be superdense if it is residual (i.e. its component is of first Baire category) uncountable and dense. Theorem: Let \(X\) be a linear topological space which is a Baire space, let \((T,{\mathcal A},\mu)\) be a probability space and let \(S(T)\) be the space of all measurable step functions on \(T\), topologized by the convergence in measure. Let \((A_n)\) be a sequence of linear continuous operators \(A_n:X\to S(T)\) satisfying the condition \[ \forall\varepsilon> 0\quad\forall \text{ zero-neighborhood }U\subseteq X\quad \exists x_0\in U: \mu\Biggl\{\sup_n | A_nx_0|= \infty\Biggr\}\geq 1-\varepsilon. \] Then there is a superdense subset \(X_0\) of \(X\) such that \(\mu\left\{\sup_n | A_nx|= \infty\right\}= 1\) holds true for all \(x\in X_0\).