Abstract.

The aim of this paper is to present new proofs of selected inequalities between cardinal invariants. The novelty relies on the application of some theorems on strong sequences. This new look for proofs of known theorems shows hope for proving new inequalities between invariants

keywords:
strong sequences; cardinal invariants; Lindelöf space; cellularity; character.
MSC:
54A25; 03E05; 03E10; 03E17; 03E20.

1. Introduction

In 1968 Archangel’skii solved a forty years old problem of Alexandroff and Urysohn by proving that if \(X\) is a first-countable Lindelöf space then \({|X|} \leqslant 2^{\aleph_{0}}\).

Very soon after publishing the above theorem Archangel’skii posed the following question: whether this theorem remains true if the condition of first-countability is relaxed to the requirement that all points are \(G_{\delta}\). This question was solved by Shelah [21] in a negative one.

Archangel’skii proved that the cardinality of every Lindelöf \(T_{1}\) space of countable pseudocharacter is less than the first measurable cardinal while Juhász provided examples of Lindelöf \(T_{1}\) spaces with countable pseudocharacter of arbitrarily large cardinality below the first measurable. However it is still open the following question whether the cardinality of every Lindelöf first-countable \(T_{1}\) space is bounded by the continuum.

Archangel’skii’s Theorem has become the inspiration for investigations concerning cardinal properties of topological spaces especially a number of cardinal invariants and soon as had a lot of generalizations, for example it was proved in [2] the following result: let \(X\) be a weakly Lindelöf first-countable normal space. Then \({|X|} \leqslant 2^{\aleph_{0}}\).

The question of whether normality in the above theorem can be relaxed to regurality is still open. The partial answer for this question was given in [5]. For this purpose some special topological games were used (see e.g. [5]).

Theorems concerning cardinal properties of topological spaces are usually proved with using straight combinatorial methods or forcing methods if special assumptions are needed. S. Spadaro et al. used topological games and model theory in proofs.

On the other hand some of the results concerning cardinal invariants were proved by M. Turzański, see [22] and J. Jureczko, see [13] with using the method of strong sequences. Strong sequences were introduced in 1965 by Efimov for proving several celebrated theorems in dyadic spaces. This method has been succesfully used in several papers by M. Turzański and by J. Jureczko.

The aim of this paper is to show that the strong sequences method will be used for selected results concerning cardinal invariants. This is the first paper on this topic. Further investigations in this topic are in [14, 15].

The paper is organized as follows. In Section 2 there are given definitions and previous results needed in Section 3. In Section 3 there are presented new proofs of selected theorems mentioned in Section 2. In Section 4 there are given some remarks concerning further investigations and there are cited theorems proved by applying Theorem 2.3 but their proofs were published in another papers.

2. Definitions and previous results

In this section we recall definitions and results used in Section 3 however, we refer the reader to [7, 8, 12, 20] for undefined notions.

2.1. Selected definitions

In the whole paper, unless we assume otherwise, \(X\) denote the Hausdorff space.

Let \(X\) be a topological space. A set \(A \subseteq X\) is called regularly open if \(A = {int\overline{A}}\) and regularly closed if \(A = \overline{i⁢n⁢t⁢A}\). The family of all regularly open sets in X is denoted by \(RO{(X)}\). A space \(X\) is called almost regular if for each regularly closed set \(A \subseteq X\) and each point \(x \notin A\) there are disjoint regularly open sets \({U,V} \subseteq X\) such that \(A \subset U\) and \(x \in V\).

For a topological space \(X\) we can consider a finer topology on \(X\) by declaring countable intersections of open subsets of \(X\) i.e. \(G_{\delta}\)-sets to be a base. The new space is called \(G_{\delta}\)-topology of \(X\) and is denoted by \(X_{\delta}\).

2.2. Selected cardinal invariants

In this subsection we recall some cardinal invariants of topological spaces used in the further part of this paper.

The Lindelöf number of \(X\):

\[{{L{(X)}} = {{\min{\{{\kappa:{{\text{~each open cover of~}X\text{~has a subcover of cardinality}} \leqslant \kappa}}\}}} + \omega}}.\]

If \({L{(X)}} = \aleph_{0}\) then we say that \(X\) is a Lindelöf space.
The weak Lindelöf number of \(X\):

\[wL{(X)} = \min{\{\kappa:\text{~each open cover of~}X\text{~has a subfamily of cardinality} \leqslant \kappa}\]
\[\text{~whose union is a dense subspace of~}X\} + \omega.\]

If \({wL{(X)}} = \aleph_{0}\) then we say that \(X\) is a weakly Lindelöf space. It is immediate that \({wL{(X)}} \leqslant {L{(X)}}\).
The piecewise weak Lindelöf degree for closed sets of \(X\):

\[pwL_{c}{(X)} = \min{\{\kappa:\text{~for each closed set~}F \subset X,\text{~for each open family~}\mathcal{U}}\]
\[\text{~covering~}F\text{~and for each decomposition~}{\{\mathcal{U}_{i}:i \in I\}}\text{~of~}\mathcal{U}\text{~there are~}{( \leqslant \kappa)}\text{-sized}\]
\[\text{~subfamilies~}\mathcal{V}_{i} \subset \mathcal{U}_{i}\text{~such that~}F \subset \bigcup{\{\overline{\bigcup\mathcal{V}_{i}}:i \in I\}}\} + \omega.\]

It is easy to see that \({pwL_{c}{(X)}} \leqslant {L{(X)}}\).
The cellularity of \(X\):

\[c{(X)} = \sup{\{|\mathcal{V}|:\mathcal{V}\text{~is a cellular family in~}X\}} + \omega.\]

If \({c{(X)}} = \aleph_{0}\) then we say that \(X\) satisfies the countable chain condition (abbr. X has c.c.c). Note that \({wL{(X)}} \leqslant {c{(X)}}\) and \({pwL_{c}{(X)}} \leqslant {c{(X)}}\).
The density of \(X\),

\[{{d{(X)}} = {{\min{\{{{|S|}:{{S \subset X},{\overline{S} = X}}}\}}} + \omega}}.\]

The character of \(X\) at \(p\), (\(p \in X\)):

\[{{\chi{(p,X)}} = {\min{\{{{|\mathcal{V}|}:{\mathcal{V}\text{~is a local base for~}p}}\}}}}.\]

The character of \(X\)

\[{{\chi{(X)}} = {{\sup{\{{\chi{(p,X)}}:{p \in X}\}}} + \omega}}.\]

The local \(\pi\)-character of \(X\) at \(p\), (\(p \in X\)):

\[{{\pi\chi{(p,X)}} = {\min{\{{{|\mathcal{V}|}:{\mathcal{V}{\text{~is a local~}\pi\text{-base for~}}p}}\}}}}.\]

The local \(\pi\)-character of \(X\)

\[{{\pi\chi{(X)}} = {{\sup{\{{\chi{(p,X)}}:{p \in X}\}}} + \omega}}.\]

The pseudocharacter of \(X\) at \(p\), (\(p \in X\)):

\[{{\psi{(p,X)}} = {\min{\{{{|\mathcal{V}|}:{\mathcal{V}\text{~is a pseudobase base for~}p}}\}}}}.\]

The pseudocharacter of \(X\)

\[{{\psi{(X)}} = {{\sup{\{{\chi{(p,X)}}:{p \in X}\}}} + \omega}}.\]

The Hausdorff pseudocharacter of \(X\),

\[H\psi{(X)} = \min{\{\kappa:\text{~for every~}x \in X\text{~there is a family~}\mathcal{V}_{x}\text{~of open}}\]
\[{\text{~neigbourhoods of~}x\text{~such that~}{|\mathcal{V}_{x}|}} \leqslant {\kappa\text{~and for every pair of distinct points~}}\]
\[x,y \in X\text{~there are~}V_{x} \in \mathcal{V}_{x}\text{~and~}V_{y} \in \mathcal{V}_{y}\text{~such that~}V_{x} \cap V_{y} = \varnothing\}.\]

Obviously, for each Hausdorff space \(X\) we have \({\psi{(X)}} \leqslant {H\psi{(X)}} \leqslant {\chi{(X)}}\).

2.3. Strong sequences

The method of strong sequences was introduced by Efimov in 1965 in [6] as a useful method for proving known theorems in dyadic spaces i.e. continuous images of the Cantor cube. Efimov proved, among other things, that there are no strong sequences in the subbase of the Cantor cube. So it became natural to ask, what would be the consequences of the existence of strong sequences in other spaces. This topic was taken up in the 90s’ of the last century by M. Turzański (Here we omit further historical data, which are included in paper [19]).

In author’s papers concerning strong sequences there was introduced the following definition of strong sequence. (Original and previous definitions of strong sequences are given in [6, 22]).

Let \((X,r)\) be a set with relation \(r\) and let \(\kappa\) be a regular cardinal number. A set \(A \subset X\) is called \(\kappa\)-directed if for each subset \(B \subset A\) of cardinality less than \(\kappa\) there is an element \(a \in A\) such that \({(b,a)} \in r\) for each element of \(b \in B\).

Definition 2.1.

A sequence \({(S_{\phi},H_{\phi})}_{\phi < \alpha}\), where \({S_{\phi},H_{\phi}} \subset X\) and \({|S_{\phi}|} < \kappa\) is called a \(\kappa\)-strong sequence if:

  • \(S_{\phi} \cup H_{\phi}\) is \(\kappa\)-directed,

  • \(S_{\psi} \cup H_{\phi}\) is not \(\kappa\)-directed whenever \(\psi > \phi\).

In papers [16, 17] we dealt with strong sequences in case of \(\kappa = \omega\). Then we simply say strong sequences instead of \(\omega\)-strong sequences. In paper [16] we proved the following theorem, which we called Strong Sequence Theorem.

Theorem 2.2 ([16]).

Let \(\beta\) and \(\lambda\) be cardinal numbers. Let \((X,r)\) be a set with relation \(r\) and \({|X|} \geqslant {(\beta^{\lambda})}^{+}\). If there is a strong sequence \({(S_{\alpha},H_{\alpha})}_{\alpha < {(\beta^{\lambda})}^{+}}\) such that \({|H_{\alpha}|} \leqslant \beta^{\lambda}\), for each \(\alpha < {(\beta^{\lambda})}^{+}\), then there is a strong sequence \({(T_{\alpha},H_{\alpha})}_{\alpha < {(\lambda)}^{+}}\) such that \({|T_{\alpha}|} < \omega\) and \(T_{\alpha} \subset S_{\alpha}\), for each \(\alpha < {(\lambda)}^{+}\).

In paper [18] we dealt with the following Generalized Strong Sequence Theorem and its consequences.

Let \(\beta\) and \(\eta\) be cardinal numbers. The symbol \(\beta \ll \eta\) means that the cardinal \(\eta\) is \(\beta\)-strongly inaccessible, i.e. \(\beta < \eta\) and \(\alpha^{\lambda} < \eta\) if only \(\alpha < \beta\) and \(\lambda < \eta\).

Theorem 2.3 ([18]).

Let \(\kappa,\beta,\mu,\eta\) be cardinal numbers such that \(\kappa < \eta\), \({\omega \leqslant \beta \ll \eta},\) \(\mu < \beta\) and \(\kappa,\beta,\eta\) be regular. Let \((X,r)\) be a set with relation \(r\) and \({|X|} \geqslant \eta\). If there is a \(\kappa\)-strong sequence \({(S_{\alpha},H_{\alpha})}_{\alpha < \eta}\) such that \({|H_{\alpha}|} \leqslant 2^{\mu}\) for each \(\alpha < \eta\), then there is a \(\kappa\)-strong sequence \({(T_{\alpha},H_{\alpha})}_{\alpha < \beta}\) such that \({|T_{\alpha}|} < \kappa\) and \(T_{\alpha} \subset S_{\alpha}\) for each \(\alpha < \beta\).

As a corollary of the above theorem we got the following result, sometimes called Ramsey-type theorem.

Corollary 2.4 ([18]).

Let \(\kappa,\beta,\mu,\eta\) be cardinal numbers such that \(\kappa < \eta\), \({\omega \leqslant \beta \ll \eta},{\mu < \beta}\) and \(\kappa,\beta,\eta\) be regular. Let \((X,r)\) be a set with relation \(r\) and \({|X|} \geqslant \eta\). If a set \(X\) has cardinality \(\eta\) then either \(X\) contains a \(\kappa\)-directed subset of cardinality \(\eta\), or \(X\) contains a set of cardinality \(\beta\) consisting of pairwise incomparable elements.

Strong Sequence Theorem formulated in Theorem 2.3 is true only for regular cardinals \(\beta\) and \(\eta\). In paper [13] there was proved the following theorem, where \(\eta\) is regular and \(\beta\) is singular.

Theorem 2.5 ([13]).

Let \(\kappa,\beta,\eta\) be cardinal numbers such that \(\kappa < \eta\) \(\omega \leqslant \beta \ll \eta\), with \(\kappa,\eta\) regular and \(\beta\) singular. Let \((X,r)\) be a set with relation \(r\) and \({|X|} \geqslant \eta\). If there is a \(\kappa\)-strong sequence \({(S_{\alpha},H_{\alpha})}_{\alpha < \eta}\) such that \({|H_{\alpha}|} \leqslant 2^{\beta}\) for each \(\alpha < \eta\), then there is a \(\kappa\)-strong sequence \({(T_{\alpha},H_{\alpha})}_{\alpha < {(\beta)}^{+}}\) such that \({|T_{\alpha}|} < \kappa\) and \(T_{\alpha} \subset S_{\alpha}\) for each \(\alpha < {(\beta)}^{+}\).

2.4. Selected previous results

In this subsection we recall selected inequalities among cardinal invariants which will be discussed in the next section. We start with two fundamental theorems in this topic saying about the cardinality of a Hausdorff space (proved in the 60s’ and 70s’ of the last century).

Theorem 2.6 (Archangel’skii, [1]).

If \(X\) is a Hausdorff space then \({|X|} \leqslant 2^{\chi{(X)}L{(X)}}\).

Theorem 2.7 (Juhász, [10]).

If \(X\) is a Hausdorff space then \({|X|} \leqslant 2^{\chi{(X)}c{(X)}}\).

In 1991 Hodel in [9] showed that \(\chi{(X)}\) in both above theorems can be replaced by \(H\psi{(X)}\). Later further generalizations of both Archangel’skii and Juhász’s Theorems were proven. The worth noting results belong to Bella and Spadaro. One of this result comes from 2020.

Theorem 2.8 (Bella, Spadaro, [4]).

If \(X\) is a Hausdorff space then \({|X|} \leqslant 2^{pwL_{c}{(X)}\chi{(X)}}\).

Theorem 2.8 was generalized in 2023 and is given in Theorem 2.9.

Theorem 2.9 (Bella, Carlson, Spadaro [3]).

Let \(X\) be a Hausdorff space. Then \({|X|} \leqslant 2^{pwL_{c}{(X)}H\psi{(X)}}\).

In parallel, research was also carried out for other spaces. In 1972 Juhász proved the following two theorems, where \(X_{\delta}\) denotes a topological space with a base consisting of \(G_{\delta}\)-sets of a space \(X\).

Theorem 2.10 (Juhász [11]).

If \(X\) is a compact Hausdorff space then \({c{(X_{\delta})}} \leqslant 2^{c{(X)}}\).

Theorem 2.11 (Juhász [11]).

If \(X\) is a compact Hausdorff space then \({wL{(X_{\delta})}} \leqslant 2^{c{(X)}}\).

The further theorems we wish to draw attention to here concern inequalities for normal spaces and almost regular spaces.

Theorem 2.12 (Bell, Ginsburg, Woods, [2]).

If \(X\) is a normal space then \({|X|} \leqslant 2^{\chi{(X)}wL{(X)}}\).

Theorem 2.13 (Shapirovski, [5]).

If \(X\) is an almost regular space then \({|{RO{(X)}}|} \leqslant {({\pi\chi{(X)}})}^{c{(X)}}\).

Theorem 2.14 (Bella, Spadaro, [5]).

If \(X\) is an almost regular space then \({\pi w{(X)}} \leqslant {({\pi\chi{(X)}})}^{c{(X)}}\).

It should be mentioned here that this list of the key theorems concerning inequalities between cardinal invariants is not exhausted. However, it is not our goal to list them all here. We only want to present proofs of selected theorems using the strong sequence method, which, although little known, is very useful in this topic.

3. Main results

In this section we will propose new proofs of selected theorems from Section 2. The novelty of proofs presented in this section lies in the usage of Theorem 2.3, i.e. Generalized Strong Sequences Theorem.

Proof 3.1 (Proof of Theorem 2.10).

(Some small parts of the proof from [11] are applied in the proof).

First note that \(G_{\delta}\)-sets of \(X\) form a base for open sets of \(X_{\delta}\). Thus, it is enough to prove that there are at most \(2^{c{(X)}}\) disjoint \(G_{\delta}\)-sets in \(X_{\delta}\).

Suppose in contrary that \({c{(X_{\delta})}} > 2^{c{(X)}}\). Observe that every non-empty \(G_{\delta}\)-set of \(X\) contains a non-empty closed \(G_{\delta}\)-set. Indeed. Choose \(x \in X\) arbitrarily and take the family \(\{ G_{n}:{n \in \omega}\}\) of open sets such that \(x \in {\bigcap_{n \in \omega}G_{n}}\). Next, for each \(n \in \omega\) consider open neigbourhoods \(\{ U_{n,m}:{m \in \omega}\}\) of \(x\) such that \(\overline{U_{n,1}} \in G_{n}\). It is possible because \(X\) is regular. Next, choose open sets \(U_{n,m} \subset X\) such that \({x \in U_{n,m} \subset \overline{U_{n,m}} \subset U_{n,{m - 1}}}.\) Define \(G = {\bigcap{\{ U_{n,m}:{{n,m} \in \omega}\}}}\). Since

\[{{\bigcap{\{ U_{n,m}:{{n,m} \in \omega}\}}} = {\bigcap{\{\overline{U_{n,m}}:{{n,m} \in \omega}\}}}}.\]

we conclude that \(G\) is closed and we have that \(G \subset {\bigcap_{n \in \omega}G_{n}}\). Hence, we will restrict our considerations to examing families of closed \(G_{\delta}\)-sets. We admit the fact as well-known that in a compact Hausdorff space every closed \(G_{\delta}\)-set has a countable neighbourhood base.

Now, we will construct an \(\omega^{+}\)-strong sequence \(\{{({\{ V_{\beta}\}},\mathcal{H}_{\beta})}:{\beta < {(2^{c{(X)}})}^{+}}\}\) of the following properties:

  • (1)

    \(V_{\beta}\) is a \(G_{\delta}\)-set in \(X\);

  • (2)

    \(\mathcal{H}_{\beta}\) is a (countable) neighbourhood base for \(V_{\beta}\) in \(X\);

  • (3)

    \(V_{\beta} \in {{P{(X)}} \smallsetminus {\{\mathcal{H}_{\alpha}:{\alpha < \beta}\}}}\);

  • (4)

    \(\exists_{{V,W} \subset X}\) \(V_{\beta} \subset V\) and \({\bigcup{\{ V_{\alpha}:{\alpha < \beta}\}}} \subset W\) and \({V \cap W} = \varnothing\);

  • (5)

    \({({\{ V_{\beta}\}},\mathcal{H}_{\beta})} = {({\{{\bigcup_{\alpha < \beta}V_{\alpha}}\}},{\bigcup_{\alpha < \beta}\mathcal{H}_{\alpha}})}\) for limit \(\beta\).

We start the construction with choosing an arbitrary \(G_{\delta}\)-set \(V_{0} \subsetneq X\). Let \(\mathcal{H}_{0}\) be a neighbourhood base for \(V_{0}\) in \(X\). By the well-known fact mentioned above \({|\mathcal{H}_{0}|} \leqslant \omega\). Let \(({\{ V_{0}\}},\mathcal{H}_{0})\) be the first pair of an \(\omega^{+}\)-strong sequence.

Assume that for some \(\alpha < \beta < {(2^{c{(X)}})}^{+}\) we have constructed the \(\omega^{+}\)-strong sequence

\[\{{({\{ V_{\alpha}\}},\mathcal{H}_{\alpha})}:{\alpha < \beta}\}\]

of properties \({(1)} - {(5)}\). Since the limit step is done, we will construct \(\beta = {({\alpha + 1})}\)-step.

Observe that for each \(\alpha < \beta < {(2^{c{(X)}})}^{+}\) we have that \({|\mathcal{H}_{\alpha}|} \leqslant \omega\). Hence

\[{{|{\bigcup{\{\mathcal{H}_{\alpha}:{\alpha < \beta}\}}}|} < {\omega \cdot \beta} < {(2^{c{(X)}})}^{+}}.\]

Hence \({{P{(X)}} \smallsetminus {\{\mathcal{H}_{\alpha}:{\alpha < \beta}\}}} \neq \varnothing\). Let \(V_{\beta} \in {{P{(X)}} \smallsetminus {\{\mathcal{H}_{\alpha}:{\alpha < \beta}\}}}\) be an arbitrary \(G_{\delta}\)-set in \(X\). Since \(X\) is normal there are open and disjoint sets \(U_{\alpha},U_{\beta}\) such that \(V_{\beta} \subset U_{\beta}\) and \({\bigcup{\{ V_{\alpha}:{\alpha < \beta}\}}} \subset U_{\alpha}\). Let \(\mathcal{H}_{\beta}\) be a neighbourhood base for \(V_{\beta}\) in \(X\). Let \(({\{ V_{\beta}\}},\mathcal{H}_{\beta})\) be the next pair of the \(\omega^{+}\)-strong sequence.

Thus, we have defined the \(\omega^{+}\)-strong sequence \(\{{({\{ V_{\beta}\}},\mathcal{H}_{\beta})}:{\beta < {(2^{c{(X)}})}^{+}}\}\) fulfilling properties \({(1)} - {(5)}\). Now, apply Theorem 2.3 for \(({P{(X)}}, \in )\), \(\kappa = \omega^{+}\), \(2^{\mu} = \omega\), \(\eta = {(2^{c{(X)}})}^{+}\) and \(\beta = {({c{(X)}})}^{+}\). By Theorem 2.3 we obtain the \(\omega^{+}\)-strong sequence \(\{{({\{ W_{\beta}\}},\mathcal{H}_{\beta})}:{\beta < {({c{(X)}})}^{+}}\}\) such that \(W_{\beta} \subseteq V_{\beta}\) for each \(\beta < {({c{(X)}})}^{+}\). By our construction \({W_{\alpha} \cap W_{\beta}} = \varnothing\) for distinct \({\alpha,\beta} < {({c{(X)}})}^{+}\). Hence, we obtain \({({c{(X)}})}^{+}\) of pairwise disjoint elements in \(X\). A contradiction.

Proof 3.2 (Proof of Theorem 2.14).

Suppose that \({\pi w{(X)}} > {({\pi\chi{(X)}})}^{c{(X)}}\). Let \(\mathcal{B}\) be a \(\pi\)-base of cardinality \(\pi w{(X)}\). We will construct a \({({\pi\chi{(X)}})}^{+}\)-strong sequence \(\{{({\{ U_{\beta}\}},\mathcal{W}_{\beta})}:{\beta < {({({\pi\chi{(X)}})}^{c{(X)}})}^{+}}\}\) of the following properties:

  • (1)

    \(U_{\beta} \in \mathcal{B}\);

  • (2)

    \(x_{\beta} \in U_{\beta}\);

  • (3)

    \(\mathcal{W}_{\beta}\) is a local \(\pi\)-base for \(x_{\beta} \in {\mathcal{B} \smallsetminus {\{\mathcal{W}_{\alpha}:{\alpha < \beta}\}}}\);

  • (4)

    \({({\{ U_{\beta}\}},\mathcal{W}_{\beta})} = {({\{{\bigcup_{\alpha < \beta}U_{\alpha}}\}},{\bigcup_{\alpha < \beta}\mathcal{W}_{\alpha}})}\) for limit \(\beta\).

Let \(U_{0} \in \mathcal{B}\) be an arbitrary set and let \(x_{0} \in U_{0}\) be an arbitrary element. Let \(\mathcal{W}_{0}\) be a local \(\pi\)-base for \(x_{0}\) in \(X\). Obviosuly \({|\mathcal{W}_{0}|} \leqslant {\pi\chi{(X)}} < {\pi w{(X)}}\) (by our assumption). Let \(({\{ U_{0}\}},\mathcal{W}_{0})\) be the first pair of a \({({\pi\chi{(X)}})}^{+}\)-strong sequence.

Assume that for some \(\alpha < \beta < {({({\pi\chi{(X)}})}^{c{(X)}})}^{+}\) we have constructed the \({({\pi\chi{(X)}})}^{+}\)-strong sequence \(\{{({\{ U_{\alpha}\}},\mathcal{W}_{\alpha})}:{\alpha < \beta}\}\) of properties \({(1)} - {(4)}\). Since the limit step is done, we will construct \(\beta = {({\alpha + 1})}\)-step.

Observe that for each \(\alpha < \beta < {({({\pi\chi{(X)}})}^{c{(X)}})}^{+}\) we have \({|\mathcal{W}_{\alpha}|} \leqslant {\pi\chi{(X)}} < {\pi w{(X)}}\) and hence

\[{{|{\bigcup{\{\mathcal{W}_{\alpha}:{\alpha < \beta}\}}}|} < {{\pi\chi{(X)}} \cdot \beta} < {({({\pi\chi{(X)}})}^{c{(X)}})}^{+}}.\]

Hence \({\mathcal{B} \smallsetminus {\{\mathcal{W}_{\alpha}:{\alpha < \beta}\}}} \neq \varnothing\). Let \(x_{\beta} \in {\mathcal{B} \smallsetminus {\{\mathcal{W}_{\alpha}:{\alpha < \beta}\}}}\). Such choice is possible because of our assumption (and \({\pi w{(X)}} = {{{d{(X)}} \cdot \pi}\chi{(X)}}\), (see [8, 3.8(b)])).

Since \(X\) is almost regular there are disjoint regularly open sets \({W_{\beta},W_{\alpha}} \subset X\) such that \(x_{\beta} \in W_{\beta}\) and \(A_{\alpha} \subset W_{\alpha}\), where \(A_{\alpha}\) is a regularly closed set such that \({\bigcup{\{ x_{\alpha}:{\alpha < \beta}\}}} \subset A_{\alpha}\). Let \(U_{\beta} \subset W_{\beta}\) be a set such that \(x_{\beta} \in U_{\beta} \in \mathcal{B}\). Let \(({\{ U_{\beta}\}},\mathcal{W}_{\beta})\) be the next pair of the \({({\pi\chi{(X)}})}^{+}\)-strong sequence.

Thus, we have defined the \({({\pi\chi{(X)}})}^{+}\)-strong sequence \(\{{({\{ U_{\beta}\}},\mathcal{W}_{\beta})}:{\beta < {({({\pi\chi{(X)}})}^{c{(X)}})}^{+}}\}\) of properties \({(1)} - {(4)}\). Now, apply Theorem 2.3 to \((\mathcal{B}, \in )\), \(\kappa = 2^{\mu} = {({\pi\chi{(X)}})}^{+}\), \(\beta = {({c{(X)}})}^{+}\) and \(\eta = {({({\pi\chi{(X)}})}^{c{(X)}})}^{+}\). By Theorem 2.3 we obtain a \({({\pi\chi{(X)}})}^{+}\)-strong sequence \({{\{{({\{ V_{\beta}\}},\mathcal{W}_{\beta})}:\beta < )}c{(X)})}^{+}\}\) such that \(V_{\beta} \subseteq U_{\beta}\) for each \(\beta < {({c{(X)}})}^{+}\). By our construction \({V_{\alpha} \cap V_{\beta}} = \varnothing\) for distinct \({\alpha,\beta} < {({c{(X)}})}^{+}\). Hence, we obtain \({({c{(X)}})}^{+}\) of pairwise disjoint elements in \(X\). A contradiction.

Proof 3.3 (Proof of Theorem 2.12).

Suppose that \({|X|} > 2^{\chi{(X)}wL{(X)}}\). We will construct a \({({wL{(X)}})}^{+}\)-strong sequence \(\{{({\{ U_{\beta}\}},\mathcal{C}_{\beta})}:{\beta < {(2^{\chi{(X)}wL{(X)}})}^{+}}\}\) of the following properties

  • (1)

    \(\mathcal{C}_{\beta} \subset {{P{(X)}} \smallsetminus {\{\mathcal{C}_{\alpha}:{\alpha < \beta}\}}}\) is a family such that \({|\mathcal{C}_{\beta}|} \leqslant {wL{(X)}}\) and \(\overline{\bigcup\mathcal{C}_{\beta}} = X\);

  • (2)

    \(U_{\beta} \in \mathcal{C}_{\beta}\);

  • (3)

    \({({\{ U_{\beta}\}},\mathcal{C}_{\beta})} = {({\{{\bigcup_{\alpha < \beta}U_{\alpha}}\}},{\bigcup_{\alpha < \beta}\mathcal{C}_{\alpha}})}\) for limit \(\beta\).

We start the construction with choosing an open cover \(\mathcal{D}_{0}\) of \(X\). Let \(\mathcal{C}_{0}\) be a subfamily of \(\mathcal{D}_{0}\) such that \({|\mathcal{C}_{0}|} \leqslant {wL{(X)}}\) and \(\overline{\bigcup\mathcal{C}_{0}} = X\). Let \(U_{0} \in \mathcal{C}_{0}\) be an arbitrary element. Let \(({\{ U_{0}\}},\mathcal{C}_{0})\) be the first pair of a \({({wL{(X)}})}^{+}\)-strong sequence.

Assume that for some \(\alpha < \beta < {(2^{\chi{(X)}wL{(X)}})}^{+}\) we have constructed the \({({wL{(X)}})}^{+}\)-strong sequence \(\{{({\{ U_{\alpha}\}},\mathcal{C}_{\alpha})}:{\alpha < \beta}\}\) of properties \({(1)} - {(3)}\). Since the limit step is done, we will construct \(\beta = {({\alpha + 1})}\)-step.

Observe that for each \(\alpha < \beta < {(2^{\chi{(X)}wL{(X)}})}^{+}\) we have \({|\mathcal{C}_{\alpha}|} \leqslant {wL{(X)}}\). Hence

\[{{|{\bigcup{\{\mathcal{C}_{\alpha}:{\alpha < \beta}\}}}|} < {{wL{(X)}} \cdot \beta} < {(2^{\chi{(X)}wL{(X)}})}^{+}}.\]

Hence \({{P{(X)}} \smallsetminus {\{\mathcal{C}_{\alpha}:{\alpha < \beta}\}}} \neq \varnothing\). Let \(\mathcal{D}_{\beta} \subset {{P{(X)}} \smallsetminus {\{\mathcal{C}_{\alpha}:{\alpha < \beta}\}}}\) be an open cover of \(X\) and let \(\mathcal{C}_{\beta}\) be a subfamily of \(\mathcal{D}_{\beta}\) such that \({|\mathcal{C}_{\beta}|} \leqslant {wL{(X)}}\) and \(\overline{\bigcup\mathcal{C}_{\beta}} = X\). Let \(U_{\beta} \in \mathcal{C}_{\beta}\) be an arbitrary element. Let \(({\{ U_{\beta}\}},\mathcal{C}_{\beta})\) be the next pair of the \({({wL{(X)}})}^{+}\)-strong sequence.

Thus, we have defined the \({({wL{(X)}})}^{+}\)-strong sequence \(\{{({\{ U_{\beta}\}},\mathcal{C}_{\beta})}:{\beta < {(2^{\chi{(X)}wL{(X)}})}^{+}}\}\) of properties \({(1)} - {(3)}\). Now, apply Theorem 2.3 to \(({P{(X)}}, \in )\), \(\kappa = {({wL{(X)}})}^{+}\), \(2^{\mu} = {wL{(X)}}\), \(\beta = {({\chi{(X)}})}^{+}\) and \(\eta = {(2^{\chi{(X)}wL{(X)}})}^{+}\). By Theorem 2.3 we obtain \({({wL{(X)}})}^{+}\)-strong sequence \(\{{({\{ V_{\beta}\}},\mathcal{C}_{\beta})}:{\beta < {({\chi{(X)}})}^{+}}\}\) such that \(V_{\beta} \subset U_{\beta}\) for each \(\beta < {({\chi{(X)}})}^{+}\).

Now, for each \(x \in {\bigcup_{\beta < {({\chi{(X)}})}^{+}}\mathcal{C}_{\beta}}\) we define \({\mathcal{B}{(x)}} = {\{{W_{\beta} \subset \mathcal{C}_{\beta}}:{{x \in W_{\beta}},{\beta < {({\chi{(X)}})}^{+}}}\}}\). By our construction \({\mathcal{C}_{\alpha} \cap \mathcal{C}_{\beta}} = \varnothing\) for distinct \({\alpha,\beta} < {({\chi{(X)}})}^{+}\). Hence \({|{B{(x_{0})}}|} = {({\chi{(X)}})}^{+}\) for some \(x_{0} \in {\bigcup_{\beta < {({\chi{(X)}})}^{+}}\mathcal{C}_{\beta}}\). A contradiction.

Proof 3.4 (Proof of Theorem 2.9).

We can prove this theorem by the slight modification of the proof of Theorem 2.12 presented in this section to obtain the contradiction with \({\psi{(X)}} \leqslant {H\psi{(X)}} \leqslant {\chi{(X)}}\).

Theorem 2.11 can be proved by the similar method but in the presence of Theorem 2.10 we can give the following proof.

Proof 3.5 (Proof of Theorem 2.11).

By the result from [11] we have \({wL{(X)}} \leqslant {c{(X)}}\) for each topological space. By Theorem 2.10 we have \({wL{(X_{\delta})}} \leqslant {c{(X_{\delta})}} \leqslant 2^{c{(X)}}\).

4. Further remarks and some generalizations

In papers [22] and [13] there are also other results in which proof it is used Theorem 2.3. We cite them below.

Theorem 4.1 and Theorem 4.2 come from [22] and say about the inequalities between cardinal invariants in regular spaces. Theorem 4.3 and Theorem 4.4 come from paper [13].

Theorem 4.1 ([22]).

If \(X\) is a regular space then \({w{(X)}} \leqslant {\chi{(X)}^{c{(X)}}}\).

Theorem 4.2 ([22]).

If \(X\) is a regular space then \({d{(X)}} \leqslant {\chi{(X)}^{c{(X)}}}\).

Theorem 4.3 ([13]).

If \(X\) is a Hausdorff space then \({|X|} \leqslant 2^{{\chi{(X)}} + {c{(X)}}}\).

Theorem 4.4 ([13]).

If \(X\) and \(Y\) are a topological spaces then \({c{({X \times Y})}} \leqslant 2^{{c{(X)}} + {c{(Y)}}}\).

Moreover, Theorem 2.3 can be succesfully used for proving other inequalities. Let \(\beta,\mu,\eta\) be cardinal numbers such that \(\omega \leqslant \beta \ll \eta\) and \(\eta\), \(\beta\) be regular. Let \(X\) be a set of cardinality of \(\eta\) (with some additional properties if they are needed). Let \(\beta,\gamma\) and \(\delta\), such that \({\gamma,\delta} \leqslant \eta\) denote selected cardinal invariants connected with \(X\). Then \({\delta \leqslant \gamma^{\beta}}.\) It is obviously possible, by Theorem 2.5, the case when \(\beta\) is singular.

Acknowledgements.
The author is very grateful to the anonymous reviewer for their careful reading of the manuscript, which allowed many errors and inaccuracies to be avoided in the final version and made the paper more readable.
Funding.
This research has not received external funding.
Author Contributions.
Conceptualization, methodology, formal analysis, investigation, writing - original draft, writing - review & editing, J. J. All authors have read and agreed to the published version of the manuscript

References

  • [1] A. Archangel’skii, The power of compacta with the first axiom of countability, Dokl. Akad. Nauk. S.S.S.R. 187 (1969), 967–968.
  • [2] M. Bell, J. Ginsburg, and G. Woods, Cardinal inequalities for topological spaces involving the weak Lindelöf number, Pacific J. Math. 79 (1978), no. 1, 37–45.
  • [3] A. Bella, N. Carlson, and S. Spadaro, Cardinal inequalities involving the Hausdorff pseudocharacter, Rev. R. Acad. Cienc. Exactas Fís. Nat. Ser. A Mat. RACSAM 117, no. 3 (2023), 129.
  • [4] A. Bella, and S. Spadaro, A common extension of Archangelskĭ’s theorem and the Hajnal-Juhász inequality, Canad. Math. Bull. 63, no. 1 (2020), 197–203.
  • [5] A. Bella, and S. Spadaro, Infinite games and cardinal properties of topological spaces, Houston J. Math. 41, no. 3 (2015), 1063–1077.
  • [6] B. A. Efimov, Dyadic bicompacta, Trudy Mosk. Mat. Obšč. 14 (1965), 211–247.
  • [7] R. Engeling, General Topology, Heldermann-Verlag, Berlin (1989).
  • [8] R. E. Hodel, Cardinal Functions I, in: Handbook of set-theoretical topology, K. Kunen, J. E. Vaughan (eds.), Elsevier Science Publishers B.V. 1984.
  • [9] R. E. Hodel, Combinatorial set theory and cardinal function inequalities, Proc. Amer. Math. Soc. 111, no. 2 (1991), 567–575.
  • [10] I. Juhász, Cardinal functions in topology, Mathematical Centre Tracts 34, Mathematisch Centrum, Amsterdam, 1975.
  • [11] I. Juhász, On two problems of A. V. Archangel’skii, General Topology and Appl. 2 (1972), 151–156.
  • [12] I. Juhász, Cardinal functions in topology - ten years later. 2nd edn. Mathematical Centre Tracts, vol. 123. Mathematisch Centrum, Amsterdam (1980).
  • [13] J. Jureczko, Strong sequences and partition relations, Ann. Univ. Pedagog Crac. Stud. Math. 16 (2017), 51–59
  • [14] J. Jureczko, Further remarks on cardinal properties of topological spaces, in preparation.
  • [15] J. Jureczko, Topological games and strong sequences, in preparation.
  • [16] J. Jureczko, On inequalities among some invariants, Mathematica Aeterna 6, no. 1 (2016), 87–98.
  • [17] J. Jureczko, Strong sequences and independent sets, Mathematica Aeterna 6, no. 2 (2016), 141–152.
  • [18] J. Jureczko, \(\kappa\)-strong sequences and the existence of generalized independent families, Open Math. 15, no. 1 (2017), 1277–1282.
  • [19] J. Jureczko, Some remarks on polarized partition relations, Bull. Iranian Math. Soc. 49, no. 3 (2023), 38.
  • [20] K. Kunen, Set Theory, Studies in Logic, vol. 34. College Publications, London (2011).
  • [21] S. Shelah, On some problems in general topology, Contemporary Math. 192 (1996), 91–101.
  • [22] M. Turzański, Strong sequences and the weight of regular spaces, Commentationes Mathematicae Universitatis Carolinae 31, no. 3 (1992), 557–561.