Definability of mad families of vector spaces and two local Ramsey theories
Clement Yung · 2025
Read Definability of mad families of vector spaces and two local Ramsey theories by Clement Yung (2025) 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-2503-24207.txt.
Introduction
Clement Yung studies special families of subspaces in an endless vector space, in a 2025 math paper. Two subspaces count as almost disjoint when their overlap stays finite in size. A maximal family of such pairs, short for a mad family, cannot take any new member. How neatly such families can be defined was the open thread.
The author builds a full mad family of block subspaces from standard axioms alone. That settles a problem posed by Smythe in the yes. A tweak of the build gives a fully separable such family as well. A forcing tool then shows that in one classic model world, no such full families exist over a two-element field.
Catalog record
| CREATOR | Clement Yung |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 111 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-24207
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Definability of mad families of vector spaces and two local Ramsey theories Clement Yung Abstract . Let E E be a vector space over a countable field of dimension ℵ 0 \aleph_{0} . Two infinite-dimensional subspaces V , W ⊆ E V,W\subseteq E are almost disjoint if V ∩ W V\cap W is finite-dimensional. This paper provides some improvements on results about the definability of maximal almost disjoint families (mad families) of subspaces in [ 17 ] . We show that a full mad family of block subspaces exists assuming either 𝔭 = max { 𝔟 , 𝔰 } \mathfrak{p}=\max\{\mathfrak{b},\mathfrak{s}\} or a positive answer to a problem in [ 17 ] , improving Smythe’s construction assuming 𝔭 = 𝔠 \mathfrak{p}=\mathfrak{c} . We also discuss the abstract Mathias forcing introduced by Di Prisco-Mijares-Nieto in [ 11 ] , and apply it to show that in the Solovay’s model obtained by the collapse of a Mahlo cardinal, there are no full mad families of block subspaces over 𝔽 2 \mathbb{F}_{2} . 1. Introduction Let E be a vector space over a countable field \mathbb{F} of dimension \aleph_{0} . In [ 17 ] , Smythe studied almost disjoint families infinite-dimensional subspaces V,W\subseteq E , i.e. families of infinite-dimensional subspaces of E which pairwise intersections are finite-dimensional. Our paper focuses on the following problems raised by Smythe in [ 17 ] . Problem 1.1 . Can mad families of subspaces be definable? In particular: (1) Are there analytic mad families of subspaces? (2) Assuming the existence large cardinals, are there mad families of subspaces in \mathbf{L}(\mathbb{R}) ? Horowitz-Shelah proved in [ 5 ] that there are no analytic mad families of subspaces over \mathbb{F}_{2} , but the first problem remains largely open for other fields. Smythe remarked in [ 17 ] that there is a supercompact cardinal, then every set in Solovay’s model is \cal{H} -strategically Ramsey, so there are no full mad families. Strategically Ramsey subsets of E^{[\infty]} were introduced in [ 13 ] and [ 14 ] to study the combinatorics of Banach space theory and dichotomies, and Smythe introduced a localised version of the strategically Ramsey property in [ 16 ] . To the author’s knowledge, the precise large cardinal strength needed for the non-existence of mad families in \mathbf{L}(\mathbb{R}) is mostly unknown. Inspired by Mathias’ original proof of the non-existence of analytic mad families of [\omega]^{\omega} , Smythe approached this problem using a local Ramsey theoretic approach in [ 17 ] , replacing \cal{H} -Ramsey with \cal{H} -strategically Ramsey. Given an almost disjoint family \cal{A} , he extended Mathias’ definition of the “happy family” \mathfrak{F}{A} of subsets of \omega , to a “family” \cal{H}(\cal{A}) of subspaces of E . His approach requires \cal{H}(\cal{A}) to be full , a property which closely resembles the pigeonhole property of coideals of [\omega]^{\omega} . Calling a mad family \cal{A} full if \cal{H}(\cal{A}) is full, Smythe proved the following theorem: Theorem 1.2 (Smythe [ 17 ] ) . There exist no analytic full mad families of block subspaces . Block subspaces are subspaces with a basis of vectors with strictly increasing support (see § 2 for a precise definition). Block subspaces are more combinatorially tame, and since every infinite-dimensional subspace contains an infinite-dimensional block subspace, an almost disjoint family of subspaces is maximal with respect to block subspaces iff it is maximal. While it was clear from Smythe’s proof that the fullness property plays an important role in understanding the definability of mad families of subspaces, the extent which mad families of subspaces satisfy such property remains poorly understood. In particular, Smythe asked the following question in [ 17 ] : Problem 1.3 (Smythe, [ 17 ] ) . Does \mathsf{ZFC} prove the existence of a full mad family of subspaces? Even more, does \mathsf{ZFC} prove that every mad family of subspaces is full? In the same paper, Smythe proved that assuming \mathfrak{p}=\mathfrak{c} , there exists a full mad family. Another obstacle to understanding the definability of mad families of subspaces is the restriction to mad families of block subspaces. It is unknown if Theorem 1.2 holds for arbitrary mad families of subspaces (except for when |\mathbb{F}|=2 ). Furthermore, consider the following two cardinals: \displaystyle\mathfrak{a}_{\mathrm{vec},\mathbb{F}}^{*} \displaystyle:=\min\{|\cal{A}|:\text{$\cal{A}$ is a mad family of subspaces of $E$}\}, \displaystyle\mathfrak{a}_{\mathrm{vec},\mathbb{F}} \displaystyle:=\min\{|\cal{A}|:\text{$\cal{A}$ is a mad family of block subspaces of $E$}\}. It is unknown if they are equal in \mathsf{ZFC} . Problem 1.4 (Smythe, [ 17 ] ) . Do we lose any generality when restricting mad families to block subspaces? In particular, is \mathfrak{a}_{\mathrm{vec},\mathbb{F}}^{*}=\mathfrak{a}_{\mathrm{vec},\mathbb{F}} true in \mathsf{ZFC} ? The first of the two main results in the paper improves Smythe’s construction of a full mad family, and addresses Problems 1.1 and 1.4 as follows: Theorem 1.5 . If \mathfrak{a}_{\mathrm{vec},\mathbb{F}}^{*}=\mathfrak{a}_{\mathrm{vec},\mathbb{F}} or \mathfrak{p}=\max\{\mathfrak{b},\mathfrak{s}\} , then there exists a full mad family of block subspaces. Our method is inspired by a variant of the construction of a completely separable mad family of [\omega]^{\omega} in [ 9 ] assuming \mathfrak{s}\leq\mathfrak{a} , provided by Guzmán in [ 4 ] . Our construction utilises various results from the local Ramsey theory of block subspaces developed by Smythe in [ 16 ] and [ 17 ] , along with a few additional results that we shall prove. The second main result in this paper addresses the existence of full mad families of block subspaces over \mathbb{F}_{2} in Solovay’s model. We utilise the abstract Mathias forcing was developed in by Di Prisco-Mijares-Nieto [ 11 ] - the abstract Mathias forcing was developed to expand the local Ramsey theory of [\omega]^{\omega} to topological Ramsey spaces (i.e. closed triples (\cal{R},\leq,r) satisfying the axioms A1 - A4 introduced by Todorčević in [ 19 ] ). For example, if E is a countable vector space over \mathbb{F}_{2} , then the pigeonhole principle holds, so it is a topological Ramsey space. Let \operatorname{Coll}(\omega,<\!\kappa) be the Lévy collapse, which collapses the uncountable cardinal \kappa to \omega_{1} . A key application of the Mathias forcing of [\omega]^{\omega} was the following result: Theorem 1.6 (Mathias [ 8 ] ) . Let \kappa be a Mahlo cardinal, and let G be a \operatorname{Coll}(\omega,<\!\kappa) -generic filter. If \cal{H}\in\mathbf{V}[G] is a selective coideal of [\omega]^{\omega} , and \cal{X}\in\mathbf{L}(\mathbb{R})^{\mathbf{V}[G]} is a subset of [\omega]^{\omega} , then \cal{X} is \cal{H} -Ramsey. Using the abstract Mathias forcing, we obtain the following variant of a celebrated theorem of Mathias: Theorem 1.7 . Let \kappa be a Mahlo cardinal, and let G be a \operatorname{Coll}(\omega,<\!\kappa) -generic filter. Suppose that E is a countable vector space over \mathbb{F}_{2} . If \cal{H}\in\mathbf{V}[G] is a selective coideal E^{[\infty]} , and \cal{X}\in\mathbf{L}(\mathbb{R})^{\mathbf{V}[G]} is a subset of E^{[\infty]} , then \cal{X} is \cal{H} -Ramsey. As a consequence, we obtain the following corollary: Corollary 1.8 . Let \kappa be a Mahlo cardinal, and let G be a \operatorname{Coll}(\omega,<\!\kappa) -generic filter. Then there are no full mad families of block subspaces over \mathbb{F}_{2} in \mathbf{L}(\mathbb{R})^{\mathbf{V}[G]} . Our paper is organised as follows: (§ 2 ) We provide a concise introduction to various definitions and notations to block subspaces of E . (§ 3 ) We review the local Ramsey theory of block subspaces developed by Smythe, i.e. combinatorial properties of semicoideals of block subspaces (which Smythe calls family ). We