A common extension of Arhangel'skii's Theorem and the Hajnal-Juhasz inequality
Angelo Bella, Santi Spadaro

TL;DR
This paper introduces a unified bound for the weak Lindelöf number of the $G_\delta$-modification of Hausdorff spaces, generalizing key cardinal inequalities in topology without requiring separation axioms beyond $T_2$.
Contribution
It provides a common extension of Arhangel'skii's Theorem and the Hajnal-Juhasz inequality, answering a longstanding open question in cardinal invariants.
Findings
Unified bound for the weak Lindelöf number of $G_\delta$-modification
Generalizes Arhangel'skii's and Hajnal-Juhasz inequalities
Does not require separation axioms beyond $T_2$
Abstract
We present a bound for the weak Lindel\"of number of the -modification of a Hausdorff space which implies various known cardinal inequalities, including the following two fundamental results in the theory of cardinal invariants in topology: (Arhangel'skii) and (Hajnal-Juhasz). This solves a question that goes back to Bell, Ginsburg and Woods and is mentioned in Hodel's survey on Arhangel'skii's Theorem. In contrast to previous attempts we do not need any separation axiom beyond .
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvanced Topology and Set Theory · Computability, Logic, AI Algorithms · Mathematical and Theoretical Analysis
A common extension of Arhangel’skii’s Theorem and the Hajnal-Juhász inequality
Angelo Bella
Dipartimento di Matematica e Informatica, viale A. Doria 6, 95125 Catania, Italy
and
Santi Spadaro
Dipartimento di Matematica e Informatica, viale A. Doria 6, 95125 Catania, Italy
Abstract.
We present a bound for the weak Lindelöf number of the -modification of a Hausdorff space which implies various known cardinal inequalities, including the following two fundamental results in the theory of cardinal invariants in topology: (Arhangel’skiĭ) and (Hajnal-Juhász). This solves a question that goes back to Bell, Ginsburg and Woods [6] and is mentioned in Hodel’s survey on Arhangel’skiĭ’s Theorem [15]. In contrast to previous attempts we do not need any separation axiom beyond .
Key words and phrases:
cardinality bounds, cardinal invariants, cellularity, Lindelöf, weakly Lindelöf, piecewise weakly Lindelöf.
2010 Mathematics Subject Classification:
54A25, 54D20, 54D10
The research that led to the present paper was partially supported by a grant of the group GNSAGA of INdAM
1. Introduction
Two of the milestones in the theory of cardinal invariants in topology are the following inequalities:
Theorem 1**.**
(Arhangel’skiĭ, 1969) [2] If is a space, then .
Theorem 2**.**
(Hajnal-Juhász, 1967) [13] If is a space, then .
Here denotes the character of , denotes the cellularity of , that is the supremum of the cardinalities of the pairwise disjoint collection of non-empty open subsets of and denotes the Lindelöf degree of , that is the smallest cardinal such that every open cover of has a subcover of size at most .
The intrinsic difference between the cellularity and the Lindelöf degree makes it non-trivial to find a common extension of the two previous inequalities. The first attempt was done in 1978 by Bell, Ginsburg and Woods [6], who used the notion of weak Lindelöf degree. The weak Lindelöf degree of () is defined as the least cardinal such that every open cover of has a -sized subcollection whose union is dense in . Clearly, and we also have , since every open cover without -sized dense subcollections can be refined to a -sized pairwise disjoint family of non-empty open sets by an easy transfinite induction. Unfortunately, the Bell-Ginsburg-Woods result needs a separation axiom which is much stronger than Hausdorff.
Theorem 3**.**
[6]** If is a normal space, then .
It is still unknown whether this inequality is true for regular spaces, but in [6] it was shown that it may fail for Hausdorff spaces. Indeed, the authors constructed Hausdorff non-regular first-countable weakly Lindelöf spaces of arbitrarily large cardinality.
Arhangel’skiĭ [3] got closer to obtaining a common generalization of these two fundamental results by introducing a relative version of the weak Lindelöf degree, namely the cardinal invariant , i.e. the least cardinal such that for any closed set and any family of open sets satisfying there is a subcollection such that .
Theorem 4**.**
[3]** If is a regular space, then .
O. Alas [1] showed that the previous inequality continues to hold for Urysohn spaces, but it is still open whether it’s true for Hausdorff spaces.
In [4] Arhangel’skii made another step ahead by introducing the notion of strict quasi-Lindelöf degree, which allowed him to give a common refinement of the countable case of his 1969 theorem and the Hajnal-Juhász inequality. He defined a space to be strict quasi-Lindelöf if for every closed subset of , for every open cover of and for every countable decomposition of there are countable subfamilies , for every such that . It is easy to see that every Lindelöf space is strict quasi-Lindelöf and every ccc space is strict-quasi Lindelöf. Arhangel’skii proved that every strict quasi-Lindelöf first-countable space has cardinality at most continuum.
However, Arhangel’skii’s approach cannot be extended to higher cardinals. Indeed, it’s not even clear whether is true for every strict quasi-Lindelöf space . This inspired us to introduce the following cardinal invariants:
Definition 5**.**
- •
The piecewise weak Lindelöf degree of () is defined as the minimum cardinal such that for every open cover of and every decomposition of , there are -sized families , for every such that .
- •
The piecewise weak Lindelöf degree for closed sets of () is defined as the minimum cardinal such that for every closed set , for every open family covering and for every decomposition of , there are -sized subfamilies such that .
As a corollary to our main result, we will obtain the following bound, which is the desired common extension of Arhangel’skii’s Theorem and the Hajnal-Juhász inequality.
Theorem 6**.**
For every Hausdorff space , .
For undefined notions we refer to [11]. Our notation regarding cardinal functions mostly follows [14]. To state our proofs in the most elegant and compact way we use the language of elementary submodels, which is well presented in [10].
2. A cardinal bound for the -modification
The following proposition collects a few simple general facts about the piecewise weak Lindelöf number which will be helpful in the proof of the main theorem.
Proposition 7**.**
For any space we have:
- (1)
. 2. (2)
. 3. (3)
. 4. (4)
If is then .
Proof.
The first two items are trivial. To prove the third item, let be a closed subset of and be an open collection satisfying . Suppose . For every let be a maximal collection of pairwise disjoint non-empty open subsets of such that for each there is some with . By letting , the maximality of implies that and so . Since , we have .
To prove the fourth item assume is a regular space and let be a cardinal such that . Let be a closed subset of and be an open cover of . If covers we’re done. Otherwise use regularity to choose, for every an open set such that and . Note that is an open cover of , so by , there is a -sized subfamily of such that . Hence and we are done. ∎
Corollary 8**.**
If is a regular space then .
Proof.
Combine Proposition 7, (4) and Arhangel’skii’s result that for every regular space . ∎
We state our main theorem in terms of the -modification of a space. Let be a cardinal number. By we denote the topology on generated by -sized intersections of open sets of . We call , the -modification of ; in case we speak of the -modification of and we often use the symbol instead. This construction has been extensively studied in the literature; various authors have tried to bound the cardinal functions of in terms of their values on (see, for example [8], [12], [16], [17], [18]) and results of this kind have found applications to other topics in topology, like the estimation of the cardinality of compact homogeneous spaces (see [5], [8], [9] and [18]).
By we denote the topology on generated by -sets, that is those subsets of such that there is a family of open sets with . In general, the topology of is coarser than the -modification of , but if is a regular space then .
Theorem 9**.**
Let be a Hausdorff space such that and has a dense set of points of character . Then .
Proof.
Let be a cover of by -sets. Let be a large enough regular cardinal and be a -closed elementary submodel of such that and contains everything we need (that is, , etc…).
For every choose open sets such that .
Claim 1. covers .
Proof of Claim 1.
Let . Since is a cover of we can find a set such that . Moreover, using , we can find a -sized subset of such that . Note that , for every . Moreover, by -closedness of , the set belongs to . Set . Note that and . Therefore and all the free variables in the previous formula belong to . Therefore, by elementarity we also have that and hence there exists a set such that , which is what we wanted to prove.
∎
Claim 2. has dense union in .
Proof of Claim 2.
Suppose by contradiction that . Then we can fix a point such that . Let be a local base at .
For every , let be a sequence of open sets such that . Note that . Let . Note that is an open cover of and .
For every , we can choose, using Claim 1, a set such that . Since , there is such that . Hence we can find an ordinal such that . This shows that is an open cover of . Let . Then is a decomposition of and hence we can find a -sized family for every such that . Note that by -closedness of the sequence belongs to and hence the previous formula implies that:
[TABLE]
So, by elementarity:
[TABLE]
But that is a contradiction, because , for every .
∎
Since , Claim 2 proves that , as we wanted.
∎
As a first consequence, we derive the desired common extension of Arhangel’skii’s Theorem and the Hajnal-Juhász inequality.
Recall that the closed pseudocharacter of the point in () is defined as the minimum cardinal such that there is a -sized family of open neighbourhoods of with . The closed pseudocharacter of () is then defined as .
Corollary 10**.**
Let be a Hausdorff space. Then .
Proof.
It suffices to note that in a Hausdorff space and hence if is a cardinal such that then is a discrete set. Thus if and only if . ∎
Remark. Corollary 10 is a strict improvement of both Arhangel’skii’s Theorem and the Hajnal-Juhász inequality. Indeed, if is the Sorgenfrey line and the Aleksandroff duplicate of the unit interval, then the space is first countable, and .
Recall that a space is initially -compact if every open cover of cardinality has a finite subcover (for we obtain the usual notion of countable compactness). The following Lemma essentially says that if is an initially -compact spaces such that , then it satisfies the definition of when restricted to decompositions of cardinality at most .
Lemma 11**.**
Let be an initially -compact space such that and be a closed subset of . If is an open cover of and is a -sized decomposition of , then there are -sized subfamilies such that
Proof.
Let . Then is an open cover of of cardinality , so by initial -compactness there is a finite subset of such that . Let now . We then have and hence by we can find a -sized subfamily of such that . Set now . Then and , as we wanted. ∎
Noticing that in the proof of Theorem 9 we only needed to apply the definition of to decompositions of cardinality , Theorem 9 and Lemma 11 imply the following corollaries.
Corollary 12**.**
[8]** Let be an initially -compact space containing a dense set of points of character and such that . Then .
Corollary 13**.**
(Alas, [1]) Let be an initially -compact space with a dense set of points of character , such that . Then .
3. Open Questions
Corollary 8 can be slightly improved by replacing regularity with the Urysohn separation property (that is, every pair of distinct points can be separated by disjoint closed neighbourhoods). Indeed, in a similar way as in the proof of Proposition 7 (4) it can be shown that if is Urysohn then , where is the weak Lindelöf number for -closed sets (see [7]). Moreover, for every Urysohn space . However it’s not clear whether regularity can be weakened to the Hausdorff separation property. That motivates the next question.
Question 3.1**.**
Is the inequality true for every Hausdorff space ?
Moreover, we were not able to find an example which distinguishes countable piecewise weak Lindelöf number for closed sets from the strict quasi-Lindelöf property.
Question 3.2**.**
Is there a strict quasi-Lindelöf space such that ?
Finally, Arhangel’skii’s notion of a strict quasi-Lindelöf space suggests a natural cardinal invariant. Define the strict quasi-Lindelöf number of () to be the least cardinal number , such that for every closed subset of , for every open cover of and for every -sized decomposition of there are -sized subfamilies such that . Obviously . It’s not at all clear from our argument whether the piecewise weak-Lindelöf number for closed sets can be replaced with the strict quasi-Lindelöf number in Corollary 10.
Question 3.3**.**
Let be a Hausdorff space. Is it true that ?
Even the following special case of the above question seems to be open.
Question 3.4**.**
Let be a strict quasi-Lindelöf space. Is it true that ?
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] O. T. Alas, More topological cardinal inequalities , Colloquium Matematicum 65 (1993), 165–168.
- 2[2] A.V. Arhangel’skiĭ, The power of bicompacta with first axiom of countability , Soviet Math. Dokl., 10 (1969), 951–955.
- 3[3] A.V. Arhangel’skiĭ, A theorem about cardinality , Russian Math. Surveys, 34 (1979), 153–154.
- 4[4] A.V. Arhangel’skiĭ, A generic theorem in the theory of cardinal invariants of topological spaces , Comment. Math. Univ. Carolin. 36 (1995), 303–325.
- 5[5] A.V. Arhangel’skiĭ, G δ subscript 𝐺 𝛿 G_{\delta} -modification of compacta and cardinal invariants , Comment. Math. Univ. Carolin. 47 (2006), 95–101.
- 6[6] M. Bell, J. Ginsburg, and G. Woods, Cardinal inequalities for topological spaces involving the weak Lindelöf number , Pacific J. Math., 79 (1978), 37–45.
- 7[7] A. Bella and F. Cammaroto, On the cardinality of Urysohn spaces , Canad. Math. Bull. 31 (1988), 153–158.
- 8[8] A. Bella and S. Spadaro, Cardinal invariants for the G δ subscript 𝐺 𝛿 G_{\delta} -topology , to appear in Colloq. Math.
