A filter on a collection of finite sets and Eberlein compacta (Q2401740): Difference between revisions

From MaRDI portal
Importer (talk | contribs)
Changed an Item
ReferenceBot (talk | contribs)
Changed an Item
 
Property / cites work
 
Property / cites work: Q4449470 / rank
 
Normal rank
Property / cites work
 
Property / cites work: Continuous images of weakly compact subsets of Banach spaces / rank
 
Normal rank
Property / cites work
 
Property / cites work: Q3711330 / rank
 
Normal rank
Property / cites work
 
Property / cites work: A quintuple quotient quest / rank
 
Normal rank

Latest revision as of 09:13, 14 July 2024

scientific article
Language Label Description Also known as
English
A filter on a collection of finite sets and Eberlein compacta
scientific article

    Statements

    A filter on a collection of finite sets and Eberlein compacta (English)
    0 references
    0 references
    4 September 2017
    0 references
    The author proves an interesting combinatorial lemma about an ideal on \(\omega_1 \times \omega_1\) and uses it in order to find examples of Eberlein compacta which are not bisequential. Recall that an Eberlein compactum is a compact space which homeomorphically embeds into some Banach space with its weak topology. A topological space \(X\) is bisequential if, for every ultrafilter \(\mathcal U\) on \(X\) converging to \(x \in X\), there is a sequence of members of \( \mathcal U\) such that, for every neighborhood \(A\) of \(x\), the sequence is eventually contained in \(A\). Bisequential spaces can be characterized as biquotient images of metrizable spaces [\textit{E. A. Michael}, General Topology Appl. 2, 91--138 (1972; Zbl 0238.54009)]. The class of bisequential spaces is a subclass of the class of Fréchet spaces, but it behaves much better with respect to topological constructions. For example, countable products of bisequential spaces are bisequential, while there exist two compact Fréchet spaces whose product is not Fréchet. Let \(\mathcal I\) be the ideal on \(\omega_1 \times \omega_1\) consisting of all subsets with the property that all but countably many horizontal and vertical sections are countable. The main lemma in the paper asserts that, for every \(A \in \mathcal I\) and \( B_1,\ldots, B_ n \) subsets of \(\omega_1 \times \omega_1\) not in \(\mathcal I\), there is a strictly decreasing function from some subset of \( \omega_1\) to \( \omega_1\) whose graph omits \(A\) and intersects each \(B_i \). Let \(\mathcal X\) denote the topological space whose elements are those subsets of \(\omega_1 \times \omega_1\) which represent the graph of a strictly decreasing function with finite domain. Folklore results imply that \(\mathcal X\) is an Eberlein compactum. Peter Nyikos announced in 1992 that \(\mathcal X\) is not bisequential. The main theorem in the paper extends the result to the effect that, for every \(B \subseteq \omega_1 \times \omega_1\) such that \(B \not\in \mathcal I\), the space \(\mathcal X_B = \{A \in \mathcal X \mid A \subset B \} \) is not bisequential. Moreover, another proof is given of a result by \textit{A. G. Leiderman} and \textit{G. A. Sokolov} [Commentat. Math. Univ. Carol. 25, 233--246 (1984; Zbl 0586.54022)] showing that \(\mathcal X\) is not a uniform Eberlein compactum. The above constructions can be performed by considering any ordinal \(\alpha\) in place of \( \omega_1\). It follows from the case \( \alpha = \omega_1\) that, for every ordinal \(\alpha\), the resulting space is bisequential if and only if \(\alpha < \omega _1\). The author points out that the construction can be performed also by replacing \( \omega_1\) by some partially ordered set \(T\), and one obtains an Eberlein compactum \(\mathcal X_T\) in case \(T\) is a tree. The problem is left open to characterize those trees \(T\) of height \(\leq \omega _1\) such that \(\mathcal X_T\) is bisequential.
    0 references
    bisequentiality
    0 references
    Eberlein compactum
    0 references

    Identifiers