On antichain numbers of algebras 𝒫 ( ω ) / 𝒥 \mathcal{P}(\omega)/\mathcal{J} and the splitting ideal
Aleksander Cieślak · 2026
Read On antichain numbers of algebras 𝒫 ( ω ) / 𝒥 \mathcal{P}(\omega)/\mathcal{J} and the splitting ideal by Aleksander Cieślak (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-19044.txt.
Introduction
Aleksander Cieślak studies the splitting ideal, a combinatorial ideal on the natural numbers, computing its cardinal invariants and its place in the Katětov order. He examines antichain numbers for power-set quotient algebras modulo Borel ideals. For a class including F-sigma and analytic P-ideals, he proves a lower bound for the antichain number from two cardinal invariants. He shows the bounding number bounds the convergent and boring ideals from below, while the reverse inequality is consistent for the Van der Waerden and linear-growth ideals.
Catalog record
| CREATOR | Aleksander Cieślak |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 95934 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.19044
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
On antichain numbers of algebras 𝒫 ( ω ) / 𝒥 \mathcal{P}(\omega)/\mathcal{J} and the splitting ideal Aleksander Cieślak Abstract . In this article, we study combinatorial properties of a certain ideal on ω \omega , called the Splitting ideal . We calculate its cardinal invariants and its position in the Katětov order among other definable ideals. We also study the antichain numbers 𝔞 ( 𝒥 ) \mathfrak{a}(\mathcal{J}) of algebras 𝒫 ( ω ) / 𝒥 \mathcal{P}(\omega)/\mathcal{J} for various Borel ideals. We show that min { 𝔟 , cov h + ( 𝒥 ) } ≤ 𝔞 ( 𝒥 ) \textrm{min}\{\mathfrak{b},\textrm{cov}^{+}_{h}(\mathcal{J})\}\leq\mathfrak{a}(\mathcal{J}) holds for a wide class of ideals, including all F σ F_{\sigma} -ideals, all analytic P P -ideals and many other examples. We also show that 𝔟 ≤ 𝔞 ( 𝒥 ) \mathfrak{b}\leq\mathfrak{a}(\mathcal{J}) holds for convergent ideal and for Boring ideal . Finally, we will show the consistency of 𝔞 ( 𝒥 ) < 𝔟 \mathfrak{a}(\mathcal{J})<\mathfrak{b} for the Van der Waerden’s ideal and the linear growth ideal. 2020 Mathematics Subject Classification Primary: 03E17, 03E05, 03E15 1. Introduction Ideals on countable sets play a fundamental role in infinite combinatorics and study of cardinal invariants of the continuum. This article is devoted to two subjects: study of a certain ideal, called the Splitting ideal and of antichain numbers of the algebras \mathcal{P}(\omega)/\mathcal{J} for various definable ideals. The splitting ideal is an interesting example of an ideal that was defined by Sabok and Zapletal in [ 29 ] in order to characterize adding splitting reals by certain class of idealized forcing notions. We will be interested in the cardinal invariants of the splitting ideal as well as its position in the Katětov order among other known definable ideals. In particular, we will show that, in the Katětov order, the Splitting ideal is above convergent ideal , but below the nowhere-dense ideal , the trace of measure zero ideal and the Fubini product of ideal of finite sets . We will also show that \mathrm{cov}^{*}(\mathcal{S}pl) is equal to the reaping number \mathfrak{r} . Here, the * -covering number is the smallest number of elements of the ideal that are needed to infinitely intersect any infinite subset of \omega . As a result, we will obtain a new upper bound for the covering number of the density zero ideal, i.e. \mathrm{cov}^{*}(\mathcal{Z})\leq\mathfrak{r} . This gives more conviction to the open problem whether \mathrm{cov}^{*}(\mathcal{Z})\leq\mathfrak{b} , where \mathcal{Z} is the asymptotic density zero ideal. Regarding the second subject, over the years, the Van Douwen’s diagram of the algebra \mathcal{P}(\omega)/\mathcal{J} for different ideals \mathcal{J} received attention of many researchers. In particular, several people were interested in the antichain numbers of such algebras. For an ideal \mathcal{J} on \omega , by \mathfrak{a}(\mathcal{J}) we mean the smallest size of an uncountable family of \mathcal{J} -positive sets, maximal with respect to the property that the intersection of any two of its elements is in \mathcal{J} . We will show that \min\{\mathfrak{b},\mathrm{cov}^{+}_{h}(\mathcal{J})\} is a lower bound for \mathfrak{a}(\mathcal{J}) for a wide class of definable ideals, called good ideals. This class includes all F_{\sigma} -ideals, all analytic P -ideals and many other known examples. The classical result saying that \mathfrak{b}\leq\mathfrak{a} can only be partially extended for algebras of the form \mathcal{P}(\omega)/\mathcal{J} . For example, \mathfrak{b}\leq\mathfrak{a}(\mathcal{J}) holds for all analytic ideals \mathcal{J} and for ideals like \mathcal{F}in\times\mathcal{F}in and \mathcal{ED} . However, for some ideals, the inequality \mathfrak{a}(\mathcal{J})<\mathfrak{b} is consistent. For example, Steprans in [ 32 ] showed that \mathfrak{a}(nwd)<\mathfrak{b} holds in the Laver’s model for Borel conjecture and Brendle in [ 8 ] showed that \mathfrak{a}(\mathcal{ED}_{fin})<\mathfrak{b} holds in the Hechler’s model. We will show that another examples of ( F_{\sigma} ) ideals for which \mathfrak{a}(\mathcal{J})<\mathfrak{b} is consistent is the Van der Waerden’s ideal and the linear growth ideal. Finally, we will show that \mathfrak{b}\leq\mathfrak{a}(\mathcal{J}) holds for convergent ideal conv and for Boring ideal \mathcal{BI} (studied in [ 23 ] and [ 4 ] ). By \mathcal{M} we will denote the \sigma -ideals of meager subsets of 2^{\omega} with the product topology. By \mathcal{N} we will denote the \sigma -ideals of the Lebesgue measure zero sets in 2^{\omega} with standard product measure. By \mathcal{E} we will denote the \sigma -ideal generated by closed sets of measure zero in 2^{\omega} . It is easy to see, that \mathcal{E}\subseteq\mathcal{N}\cap\mathcal{M} and the inclusion is proper. If \mathcal{I} is a \sigma -ideal on X , then the standard cardinal characteristics of \mathcal{I} are – \mathrm{add}(\mathcal{I})=\min\{|\mathcal{F}|:\mathcal{F}\subseteq\mathcal{I} and \bigcup\mathcal{F}\notin\mathcal{I}\} , – \mathrm{cov}(\mathcal{I})=\min\{|\mathcal{F}|:\mathcal{F}\subseteq\mathcal{I} and \bigcup\mathcal{F}=X\} , – \mathrm{non}(\mathcal{I})=\min\{|Y|:Y\subseteq X and X\notin\mathcal{I}\} , – \mathrm{cof}(\mathcal{I})=\min\{|\mathcal{F}|:\mathcal{F}\subseteq\mathcal{I} and \forall Y\in\mathcal{I} \exists Y^{\prime}\in\mathcal{F} Y\subseteq Y^{\prime}\} . These are known as the additivity , the covering , the uniformity and the cofinality of the ideal \mathcal{I} respectively. For f,g\in\omega^{\omega} we write f\leq^{*}g if g dominates f , i.e. f(n)\leq g(n) for almost all n\in\omega . We will call \mathcal{F}\subseteq\omega^{\omega} a dominating family if for every g\in\omega^{\omega} there is f\in\mathcal{F} which dominates g . We will call \mathcal{F}\subseteq\omega^{\omega} an unbounded family if there is no g\in\omega^{\omega} which dominates every f\in\mathcal{F} . The bounding and dominating numbers are defined as follows. – \mathfrak{b} is defined as a minimal cardinality of an unbounded family, – \mathfrak{d} as a minimal cardinality of an dominating family. Assume that A,B\in[\omega]^{\omega} . We will write A\subseteq^{*}B if A\setminus B is finite and A=^{*}B if A\subseteq^{*}B and B\subseteq^{*}A , i.e. A and B differ on finite set. We will say that A splits B if both A\cap B and B\setminus A are infinite. A family \mathcal{A}\subseteq[\omega]^{\omega} is a splitting family if for every B\in[\omega]^{\omega} there is A\in\mathcal{A} that splits B . A family \mathcal{A}\subseteq[\omega]^{\omega} is a reaping (unsplit) family if there is no single B\in[\omega]^{\omega} that splits all elements of \mathcal{A} . The splitting and reaping numbers are defined as follows. – \mathfrak{s} is defined as a minimal cardinality of a splitting family, – \mathfrak{r} as a minimal cardinality of a reaping family. It is well known that inequalities \mathfrak{s}\leq\mathfrak{d},\mathrm{non}(\mathcal{N}),\mathrm{non}(\mathcal{M}) and \mathrm{cov}(\mathcal{N}),\mathrm{cov}(\mathcal{M}),\mathfrak{b}\leq\mathfrak{r} hold and \mathfrak{s} together with \mathfrak{r} are mutually incomparable on the basis of ZFC. A family \mathcal{F}\subseteq[\omega]^{\omega} has the finite intersection property (fip. for short) if any intersection of finitely many elements of the family is infinite. A set A\in[\omega]^{\omega} is a pseudo-intersection of a family \mathcal{F}\subseteq[\omega]^{\omega} if A\subseteq^{*}F for any F\in\mathcal{F} . We define the pseudo-intersection number as – \mathfrak{p}=\min\{|\mathcal{F}|:\mathcal{F}\subseteq[\omega]^{\omega} has fip. but no pseudointersection \} It is well-known result of Bell, that the Martin’s number of \sigma\mathrm{-centered} partial orders \mathfrak{m}_{\sigma\mathrm{-centered}} is equal \mathfrak{p} . The following diagram summarizes the ZFC-provable inequalities between the cardinal invariants mentioned so far. {\lx@inpgf@ignorespaces\mathfrak{r}} {\lx@inpgf@ignorespaces\mathrm{cov}(\mathcal{N})} {\lx@inpgf@ignorespaces\mathrm{non}(\mathcal{M})} {\lx@inpgf@ignorespaces\mathrm{cof}(\mathcal{M})} {\lx@inpgf@ignorespaces\mathrm{cof}(\mathcal{N})} {\lx@inpgf@ignorespaces\mathfrak{b}} {\lx@inpgf@ignorespaces\mathfrak{d}} {\lx@inpgf@ignorespaces\mathrm{add}(\mathcal{N})} {\lx@inpgf@ignorespaces\mathrm{add}(\mathcal{M})} {\lx@inpgf@ignorespaces\mathrm{cov}(\mathcal{M})} {\lx@inpgf@ignorespaces\mathrm{non}(\mathcal{N})} {\lx@inpgf@ignorespaces\mathfrak{p}} {\lx@inpgf@ignorespaces\mathfrak{s}} An ideal \mathcal{J} on \omega is tall if for every A\in[\omega]^{\omega} there is B\in[A]^{\omega} such that B\in\mathcal{J} . An ideal \mathcal{J} is a P -ideal if for every \{A_{n}:n\in\omega\}\subseteq\mathcal{J} there is B\in\mathcal{J} such that A_{n}\subseteq B for every n\in\omega . An ideal \mathcal{J} is a P^{+} -ideal if for every decreasing sequence \{A_{n}:n\in\omega\}\subseteq\mathcal{J}^{+} there is B\in\mathcal{J}^{+} such that B\subseteq^{*}A_{n} for every n\in\omega . As usual, when we refer to the descriptive complexity of an ideal, we mean its complexity as a subset of Cantor space 2^{\omega} . Recall the standard cardinal invariants associated