Definable Functions to Quotients in Ordered Abelian Groups

Harper Wells · 2026

Read Definable Functions to Quotients in Ordered Abelian Groups by Harper Wells (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-2604-00122.txt.

Introduction

Wells studies families of functions that can be defined from an ordered abelian group into quotients built from its convex subgroups. For maps into a quotient by a convex subgroup, every such family turns out uniformly piecewise linear. When the quotient also involves multiples of the group or a prime-power construction, the conclusion weakens. It becomes a uniform piecewise boolean combination of linear maps into quotients defined from the same subgroup data.

Catalog record

CREATORHarper Wells
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE73925
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.00122

Browse the full library · About the Tower of Babel

Full text

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

Definable Functions to Quotients in Ordered Abelian Groups Harper Wells Abstract In this paper we study definable families of functions from an ordered abelian group into various naturally arising definable quotients. We show that for an ordered abelian group G G and definable family of convex subgroups { D } D ∈ 𝒟 \{D\}_{D\in\mathcal{D}} , any definable family of functions { f D } D ∈ 𝒟 \{f_{D}\}_{D\in\mathcal{D}} with f D : G d → G D f_{D}:G^{d}\rightarrow\frac{G}{D} is uniformly piecewise linear; for a prime p p , integers s , r ≥ 1 s,r\geq 1 , and groups D [ p s ] D^{[p^{s}]} defined later, if f D : G d → G D + p r ​ G f_{D}:G^{d}\rightarrow\frac{G}{D+p^{r}G} or f D : G d → G D [ p s ] + p r ​ G f_{D}:G^{d}\rightarrow\frac{G}{D^{[p^{s}]}+p^{r}G} we instead obtain that the definable family of functions is uniformly piecewise a boolean combination of linear functions to quotients by subgroups which are uniformly definable from D D . 1 Introduction The study of the model theory of ordered abelian groups goes back to at least the sixties, starting with Robinson and Zakon in [ 8 ] classifying regular ordered abelian groups, and Gurevich expanding on this and studying decidability of ordered abelian groups in [ 5 ] . Quantifier elimination in certain cases was studied by Weispfenning in [ 11 ] , and a relative quantifier elimination for all ordered abelian groups was introduced by Schmitt in [ 9 ] . Cluckers and Halupczok more recently proved a relative quantifier elimination in [ 3 ] which uses the imaginary expansion of the group, rather than Schmitt’s result involving the spines as an external structure. The study of definable functions in ordered abelian groups has been studied in the case of groups of finite regular rank by Belegradek, Verbovskiy, and Wagner in [ 2 ] , and then further by Cluckers and Halupczok in their quantifier elimination paper. For Cluckers and Halupczok, the motivation arose from studying definable functions in Henselian valued fields. The main goal of this paper is to study definable functions not just from the group to itself, but from the group to various definable quotients; the motivation for this arose from questions regarding elimination of imaginaries in a general ordered abelian group. When it comes to elimination of imaginaries, work of Vicaría in [ 10 ] deals with the case of bounded regular rank; in [ 4 ] , Derakhsan and Hrushovski prove a result for elimination of imagninaries in general products of structures, while in [ 7 ] Liccardo proves a negative result regarding elimination of imaginaries in pure ordered abelian groups. While extensive, these results still don’t constitute a complete image of elimination of imaginaries in ordered abelian groups; namely, not all ordered abelian groups are elementarily equivalent to products of ordered abelian groups of bounded regular rank, and even in the case of such products there is potential to improve the collection of sorts used to obtain elimination of imaginaries. One method for proving weak elimination of imaginaries down to a specified collection of sorts involves proving density of definable types, and finding codes for definable types in those sorts. A proof of this can be found in [ 6 ] . This second condition can also be broken down further into coding definable 1 -types and coding germs of definable functions into (finite subsets of) the sorts to which imaginaries are to be eliminated; this was the motivation for this project. In the aim of showing that germs of definable functions have codes in appropriate sorts, we show that uniform families of definable functions from the group into natural quotient sorts are given piecewise by uniform intersections of linear functions into certain other quotients which are uniformly definable from the target quotient. This then allows us to code the uniform family of functions with a combination of a definable set in one of the spines and a tuple of elements coming from quotient groups of the form we study in this paper. Specifically, we classify definable functions f:G^{d}\rightarrow\frac{G}{H} with H=D , H=D+p^{r}G , and H=\alpha^{[p^{s}]}+p^{r}G where D is a convex subgroup, p is prime, and \alpha is a member of one of the spines; for the definition of spines and \alpha^{[p^{s}]} see Definitions 2.1 and 2.2 . For H=D or H=D+p^{r}G we have the following: Theorem 1.1 . Suppose either H=D or H=D+p^{r}G for some D\leq G convex and r\geq 1 . Also suppose that d,e\in\mathbb{N}_{>0} , and f:G^{d}\rightarrow(\frac{G}{H})^{(e)} is a B -definable function, where by X^{(e)} we mean subsets of X of size e . Then we can find some k\in\mathbb{N} , \textrm{acl}^{eq}(B) -definable sets \{X_{i}\subseteq G^{d}\}_{i\leq k} , \textrm{acl}^{eq}(B) -definable linear functions f_{i}:X_{i}\rightarrow\frac{G}{H} such that f=\bigcup\limits_{i\leq k}f_{i} . For H=\alpha^{[p^{s}]}+p^{r}G the result is a bit more technical; for a full breakdown of the individual function case see Theorem 3.2 . These results are also immediately uniformisable in the case where H=D and H=\alpha^{[p^{s}]}+p^{r}G , while some modifications are required for the uniform case for H=D+p^{r}G . In Section 2 we go over some required definitions and additional preliminaries, while in Section 3 we state the non-uniform versions of the main results in full. In Sections 4 , 5 and 6 we state and prove the uniform versions of the theorems, obtaining the non-uniform versions as corollaries. Finally, in Section 7 we give various examples showing why we cannot have an immediate uniform adaptation of Theorem 1.1 for all of the various cases. 2 Preliminaries and Notation Ordered Abelian Groups In this section, we cover various preliminaries regarding the model-theoretic treatment of ordered abelian groups (oags). First, we recall that when we consider (\mathbb{Z},+,0,<) as an oag, we obtain quantifier elimination in the language of Presburger arithmetic, L_{Pres}:=\{0,1,+,-,<,\{\equiv_{p}\}_{p\in\mathbb{P}}\} , where \mathbb{P} is the set of primes, and a\equiv_{p}b\Longleftrightarrow(\exists z)(a-b=pz) . From this, we see that divisibility essentially controls the model theory of (\mathbb{Z},+,0,<) , and the same turns out to hold true for a general ordered abelian group, although we have to take extra care when considering what is meant by divisibility. In particular, we have the following definitions: Definition 2.1 . Let (G,0,+,<) be an oag, and g\in G,n\in\mathbb{N}_{>0} . Then we have the following definitions: (1) Let H_{n,g} be the largest convex subgroup of G such that g\notin H_{n,g}+nG . We define \mathcal{S}_{n}:=G/\sim , where a\sim b\Longleftrightarrow H_{n,a}=H_{n,b} , and let s_{n}:G\rightarrow\mathcal{S}_{n} be the canonical projection map. If g\in nG then we set H_{n,g}=\{0\} . (2) Let H^{\prime}_{n,g}=\bigcup\limits_{\begin{subarray}{c}\alpha\in\mathcal{S}_{n}\\ g\notin\alpha\end{subarray}}\alpha , where the empty union is \{0\} . We define \mathcal{T}_{n}:=G/\sim^{\prime} , where a\sim^{\prime}b\Longleftrightarrow H^{\prime}_{n,a}=H^{\prime}_{n,b} , and t_{n}:G\rightarrow\mathcal{T}_{n} is the canonical projection map (3) Let H^{\prime+}_{n,g}=\bigcap\limits_{\begin{subarray}{c}\alpha\in\mathcal{S}_{n}\\ g\in\alpha\end{subarray}}\alpha . We define \mathcal{T}_{n}^{+}:=G/\sim^{+} , where a\sim^{+}b\Longleftrightarrow H^{\prime+}_{n,a}=H^{\prime+}_{n,b} , and t_{n}^{+}:G\rightarrow\mathcal{T}_{n}^{+} is the canonical projection map (4) The n -spine of G is Sp_{n}=\mathcal{S}_{n}\cup\mathcal{T}_{n}\cup\mathcal{T}_{n}^{+} , which forms a preorder via inclusion of convex subgroups; this collapses to a total order after equating elements which represent the same subgroup. Note . Each of H_{n,g},H^{\prime}_{n,g},H^{\prime+}_{n,g} are uniformly definable convex subgroups of G , hence each of \mathcal{S}_{n},\mathcal{T}_{n},\mathcal{T}_{n}^{+} is an imaginary sort of G . Throughout this paper (including in the above definitions), we

Read the complete text