Completely Syndetic Sets in Discrete Groups

Guy Salomon, Yotam Svoray, Ariel Yadin · 2025

Read Completely Syndetic Sets in Discrete Groups by Guy Salomon, Yotam Svoray, Ariel Yadin (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-2506-18784.txt.

Introduction

Salomon, Svoray, and Yadin study completely syndetic sets in discrete groups, sets so spread out that finitely many shifts can jointly cover any finite tuple of elements. Such sets matter for the structure theory of groups and their compactifications. Their 2025 paper asks which groups can be split into two such sets.

For groups built almost like the integers, no split into two completely syndetic sets exists, unlike wilder groups where splits abound. The team also fully characterizes these sets inside the integers, which settles which integer sets meet a key ideal in the Stone-Čech compactification. A final result may surprise: completely syndetic sets can still have arbitrarily small density, so wide coverage does not force bulk.

Catalog record

CREATORGuy Salomon, Yotam Svoray, Ariel Yadin
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE51 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-18784

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Completely Syndetic Sets in Discrete Groups Thanks: This research partially supported by the Israel Science Foundation, grant no. 954/21. YS thanks Ben-Gurion University of the Negev for hospitality while working on this project. Guy Salomon, Yotam Svoray, Ariel Yadin Abstract . We study completely syndetic ( 𝖢𝖲 \mathsf{CS} ) sets in discrete groups — subsets that for every n ∈ ℕ n\in\mathbb{N} admit finitely many left translates that jointly cover every n n -tuple of group elements. While for finitely-generated groups, the non-virtually nilpotent ones admit a partition into two 𝖢𝖲 \mathsf{CS} sets, we show that virtually abelian groups do not. We also characterize 𝖢𝖲 \mathsf{CS} subsets of ℤ \mathbb{Z} , and as a result characterize subsets of ℤ \mathbb{Z} whose closure in β ​ ℤ \beta\mathbb{Z} contains the smallest two sided ideal of β ​ ℤ \beta\mathbb{Z} . Finally, we show that 𝖢𝖲 \mathsf{CS} sets can have an arbitrarily small density. 1. Introduction Let G be a discrete countable group. A subset A\subset G is called (left) syndetic if there exists finitely many left translates of A whose union covers the entire group — that is, every g\in G lies in at least one of them. This notion generalizes the idea of having a finite index from subgroups to arbitrary subsets. Syndetic sets have appeared in various mathematical contexts over the past century [ 1 , 6 , 7 , 12 ] , particularly in connection with subsets of the integers. In \mathbb{Z} , the syndetic subsets are exactly those with bounded gaps (when viewed as increasing sequences). It is a fact that every discrete group can be partitioned into two disjoint syndetic sets; see Remark 2.2 . For instance, the even and odd integers form such a pair in \mathbb{Z} . Recently, Kennedy, Raum, and Salomon [ 11 ] introduced a stronger notion of syndeticity in their study of the universal minimal proximal and strongly proximal flows of discrete groups. A subset A\subset G is called (left) completely syndetic (abbreviated \mathsf{CS} ) if for every n\in\mathbb{N} , there exist finitely many left translates of A such that every n -elements of G all lie together in at least one of these translates; see Definition 2.1 below. \mathsf{CS} sets are “very large” sets in a certain sense. It is simple to verify that if A\subset B and A is \mathsf{CS} , then so is B . It stands to reason that “large” sets cannot fit together in a “small” group without overlapping. In other words, a natural question is: Which groups can be partitioned into two \mathsf{CS} sets? In [ 11 ] , it is shown that groups which cannot be decomposed this way must be strongly amenable — a property equivalent, in the finitely generated case, to being virtually nilpotent. Thus, “large” groups admit disjoint \mathsf{CS} sets. One of the main motivations for this work was to investigate what happens in the virtually nilpotent case (“small” groups). While we do not provide a complete answer, we do obtain several partial results and explore related phenomena. One such result is the following. Theorem \@upn 1.1 (Theorem 2.9 ) Virtually abelian groups cannot be partitioned into two \mathsf{CS} sets. Unfortunately, the case of virtually nilpotent groups that are not virtually abelian remains open. Question \@upn 1.2 (Question 2.10 ) Is there a virtually nilpotent group, necessarily not virtually abelian, that can be partitioned into two \mathsf{CS} sets? In fact, the virtually aspect is not a genuine obstacle: a group can be partitioned into two \mathsf{CS} sets if and only if some finite-index subgroup can. Therefore, in the finitely generated case, the question reduces to studying nilpotent groups; see Proposition 2.11 . Unlike syndetic sets in \mathbb{Z} , which are easy to describe, the structure of \mathsf{CS} subsets of \mathbb{Z} is more subtle. One of our aims is to give an explicit characterization of such sets. To this end, we introduce the notion of a uniform sub-Szemerédi set. A subset A\subset\mathbb{Z} is called uniform sub-Szemerédi if it contains arbitrarily long sub-arithmetic progressions of a uniformly bounded difference; see Definition 3.4 below. With this terminology, we obtain the following characterization: Theorem \@upn 1.3 (Theorem 3.6 ) A\subset\mathbb{Z} is \mathsf{CS} if and only if \mathbb{Z}\setminus A is not uniform sub-Szemerédi. Any \mathsf{CS} set is in particular thick. A set A\subset G is (left) thick if for any finite set F\subset G there exists some g\in G such that F\subset gA . One feature of \mathsf{CS} sets is that they are “so large” that they are simultaneously syndetic and thick. It is therefore natural to ask whether the converse is also true. Using the above characterization of \mathsf{CS} sets of the integers, we show this is false. Corollary \@upn 1.4 (Corollary 3.8 ) There are thick and syndetic sets of integers that are not \mathsf{CS} . The notion of complete syndeticity therefore expresses a very strong sense of “largeness” (at least stronger than that of thickness and syndeticity). Another common measure of the largeness of a set is its density. We show that these two notions of largeness — complete syndeticity and large density — are unrelated. Theorem \@upn 1.5 (Theorem 4.6 ) Let G be an infinite finitely generated group. For any \varepsilon>0 there exists a \mathsf{CS} set A\subset G with a density less than \varepsilon>0 . Let \beta G denote the Stone–Čech compactification of a countable discrete group G . This compactification carries a natural semigroup structure, turning it into a right topological semigroup (see the discussion preceding Proposition 2.3 ). A subset A\subset G is \mathsf{CS} if and only if its closure \overline{A}\subset\beta G contains a right ideal of \beta G (see Proposition 2.3 ). In this sense, \mathsf{CS} sets arise naturally in the context of the Stone–Čech compactification and are well worth studying from this perspective. Many of the results in this paper can be naturally formulated in this language. For instance, Theorem 1.1 translates to the statement that if G is virtually abelian, then \beta G contains a unique minimal closed right ideal, which then must necessarily be equal to K(\beta G) , the unique smallest two-sided ideal of \beta G . Combined with Theorem 1.5 , this leads to the following corollary: Corollary \@upn 1.6 (Corollary 3.7 ) Let A\subset\mathbb{Z} . Then \overline{A}\supset K(\beta\mathbb{Z}) if and only if \mathbb{Z}\setminus A is not uniform sub-Szemerédi. More broadly, many of our results admit natural reformulations in terms of the Stone–Čech compactification, though we refrain from translating each one explicitly. Beyond the introduction, the paper is organized into three sections. Section 2 explores the connection between \mathsf{CS} sets and closed right ideals in the Stone–Čech compactification \beta G of a group G . We establish a dichotomy: \beta G always contains either one or infinitely many minimal closed right ideals. In the case of virtually abelian groups, we show that \beta G has exactly one such ideal, which implies that these groups cannot be partitioned into two \mathsf{CS} sets. We also introduce the notion of intersection translates of a minimal closed right ideal and relate it to our central question. Section 3 focuses on \mathsf{CS} subsets of \mathbb{Z} , providing a complete characterization. Section 4 demonstrates that \mathsf{CS} sets can have

Read the complete text