Classification of Erdős type subspaces of nonseparable \(\ell ^p\)-spaces (Q409699): Difference between revisions

For a topological space \(X\), \(w(X)\) denotes the weight of \(X\). The local weight of \(X\), denoted by \(lw(X)\), is defined by \(lw(X) = \min\{w(U): U\) is an open subset of \(X\}\). The symbol \(X_D\) denotes the set \(X\) equipped with the discrete topology. Let \(A\) be an arbitrary set and \(p \geq 1\). The \(\ell_A^p\) space is defined by \(\ell_A^p= \{x =(x_\alpha)_{\alpha \in A} \in {\mathbb R}^A: \sum_{\alpha \in A} |x_\alpha|^P < \infty\}\). Given a fixed family \(\{E_\alpha : \alpha \in A\}\) of subsets of \(\mathbb R\), the \textit{Erdős type space} corresponding to \(\{E_\alpha : \alpha \in A\}\), denoted by \({\mathcal E}_A\), is defined by \({\mathcal E}_A =\left(\prod_{\alpha \in A} E_\alpha \right) \cap \ell_A^p\). Since sets of cardinality equal to \(A\) generate a Banach space that is isomorphic to \(\ell_A^p\), \(A\) can be thought as being an infinite cardinal number \(\mu\). For \(\mu=\omega\), \(p = 2\) and \(E_n = \mathbb Q\) for every \(n\in \omega\), one defines Erdős space \(\mathfrak C = \{x \in \ell^2 : \forall n\in \omega, x_n \in \mathbb Q\}\). Note that it was shown in [\textit{J. J. Dijkstra, J. van Mill} and \textit{K. I. S. Valkenburg}, J. Math. Soc. Japan 60, No. 3, 793--818 (2008; Zbl 1154.54021)] that the three topological dimension functions ind, Ind and dim agree for the space \({\mathcal E}_\mu\). Generalizing a result on \({\mathcal E}_\omega\) in [\textit{J. J. Dijkstra} and \textit{J. van Mill}, Erdős space and homeomorphism groups of manifolds, Mem. Am. Math. Soc. 979, i-v, 62 p. (2010; Zbl 1204.57041)], the main result in this paper asserts the following: Suppose \({\mathcal E}_\mu\) is such that \(w({\mathcal E}_\mu) = \kappa\), \(lw({\mathcal E}_\mu) = \lambda\) and that there are infinitely many \(\alpha \in \mu\) for which \(E_\alpha\) is of the first category in itself. Then \({\mathcal E}_\mu \approx \mathfrak C \times (\lambda_D)^\omega \times \kappa_D\) if and only if every \(E_\alpha\) is a zero-dimensional \(F_{\sigma\delta}\)-subset of \(\mathbb R\) and ind \({\mathcal E}_\mu>0\).