The Fréchet-Urysohn property of Pixley-Roy hyperspaces (Q649834): Difference between revisions

Let \(X\) be a regular space and \({\mathcal F}[X]\) the space of all nonempty finite subsets of \(X\) with the Pixley-Roy topology, that is, the topology generated by the sets of the form \([A, U]=\{B \in {\mathcal F}[X] : A \subset B \subset U\}\), where \(A\in {\mathcal F}[X]\) and \(U\) is open in \(X\). This hyperspace was introduced by \textit{C. Pixley} and \textit{P. Roy} [Proc.\ Auburn Topology Conf.\ 1969, 75--85 (1969; Zbl 0259.54022)] to construct a nonseparable Moore space with the countable chain condition. In the paper under review, the author proves that the following five conditions are equivalent: (1) \({\mathcal F} [X]\) is a \(k\)-space; (2) \({\mathcal F} [X]\) is sequential; (3) \({\mathcal F} [X]\) is Fréchet-Urysohn; (4) every finite power of \(X\) is Fréchet-Urysohn for finite sets; and (5) every finite power of \({\mathcal F} [X]\) is Fréchet-Urysohn for finite sets. Here, for a point \(x \in X\), a family \({\mathcal P}\) of nonempty subsets of \(X\) is said to be a \(\pi\)-network at \(x\) if every neighborhood of \(x\) contains some member of \({\mathcal P}\). According to \textit{G. Gruenhage} and \textit{P. J. Szeptycki} [Topology Appl.\ 151, 238--259 (2005; Zbl 1085.54016)], a space is said to be Fréchet-Urysohn for finite sets if for every \(x \in X\) and for every \(\pi\)-network \({\mathcal P}\) at \(x\) consisting of finite subsets of \(X\), there is a sequence \(\{P_n : n \in \omega\} \subset {\mathcal P}\) converging to \(x\) (that is, every neighborhood of \(x\) contains \(P_n\) for all but finitely many \(n \in \omega\)). This notion was first studied by \textit{E. A. Reznichenko} and \textit{O. V. Sipacheva} [Mosc.\ Univ.\ Math.\ Bull.\ 54, 33--38 (1999; Zbl 0949.54005)]. The author also discusses the metrizability of \({\mathcal F} [X]\). \textit{H. Tanaka} [Tsukuba J. Math.\ 7, 299--315 (1983; Zbl 0541.54015)] claimed that the following conditions are equivalent: (1) \({\mathcal F}[X]\) is metrizable; (2) \({\mathcal F}[X]\) is Lašnev; and (3) \({\mathcal F}[X]\) is a paracompact perfectly normal quasi-\(k\)-space. The author gives an example of a space \(X\) such that \({\mathcal F}[X]\) is a countable Fréchet-Urysohn space which is not metrizable. This shows that the implication \((3) \Rightarrow (1)\) in Tanaka's claim is incorrect. Also, the author improves \((2) \Rightarrow (1)\) in Tanaka's claim by proving that if \({\mathcal F}[X]\) is a paracompact \(k\)-space with a point-countable \(k\)-network, then it is metrizable.