Profinite detection of free products and free factors

Andrei Jaikin-Zapirain, Henrique Souza, Pavel Zalesski · 2026

Read Profinite detection of free products and free factors by Andrei Jaikin-Zapirain, Henrique Souza, Pavel Zalesski (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-2603-16674.txt.

Introduction

Jaikin-Zapirain, Souza, and Zalesski detect free products from profinite completions. Their groups are fundamental groups of graphs of virtually free groups. Edge groups in the graph are virtually cyclic. If the group is LERF, it splits as a free product exactly when its completion does. The split there is a free profinite product. One-ended free factors can also be read off the completion. A rigid word in a free group is then universally rigid.

Catalog record

CREATORAndrei Jaikin-Zapirain, Henrique Souza, Pavel Zalesski
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE85761
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.16674

Browse the full library · About the Tower of Babel

Full text

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

Profinite detection of free products and free factors Andrei Jaikin-Zapirain, Henrique Souza, Pavel Zalesski Abstract . Let G G be the fundamental group of a graph of finitely generated virtually free groups with virtually cyclic edge groups. We show that G G is cohomologically good if G G is residually finite. If G G is LERF, we prove that G G splits non-trivially as a free product if and only if its profinite completion G ^ \widehat{G} splits non-trivially as a free profinite product. Moreover, we are able to detect one-ended free factors of G G from G ^ \widehat{G} . As an application, we deduce that any profinitely rigid word in a finitely generated free group is universally profinitely rigid. 1. Introduction A central theme in modern infinite group theory is the extend to which the finite quotients of a group determine its algebraic structure. Problems of this sort, colloquially known as profinite rigidity or profinite detection problems, asks which properties of a finitely generated residually finite group G can be recovered from its profinite completion \widehat{G} . In this setting, the strongest question one might ask is: given two finitely generated and residually finite groups G and H such that their profinite completions \widehat{G} and \widehat{H} are isomorphic, must G and H be themselves isomorphic? Groups for which this holds are called profinitely rigid in the absolute sense . Going past the realm of abelian groups, very few groups are known to be profinitely rigid in the absolute sense, with some recent known examples coming from three-dimensional hyperbolic geometry ( [ 5 , 6 ] ), some lamplighter groups ( [ 4 ] ) and free meta-abelian groups ( [ 40 ] ). A major open problem in the field, due to Remeslennikov, asks whether free groups are profinitely rigid in the absolute sense ( [ 20 , Problem 5.48] ). A weaker notion is that of relative profinite rigidity, where one asks whether the profinite completion \widehat{G} determines the isomorphism class of G among groups in a given class \mathcal{C} of finitely generated residually finite groups. A major achievement in this direction is due to Wilton ( [ 34 ] ), who showed that free groups are rigid in the class of hyperbolic fundamental groups of graphs of free groups with cyclic edge groups. Later, the hyperbolicity assumption was dropped by Morales in [ 26 ] . Recently, two broader questions have come into consideration. We say that profinite completions detect free products in \mathcal{C} if every finitely generated residually finite group G\in\mathcal{C} whose profinite completion \widehat{G} decomposes as a non-trivial free profinite product \widehat{G}=L\amalg K also decomposes non-trivially as a free product G=H*U . We say that profinite completions detect free factors in \mathcal{C} if for every finitely generated residually finite group G with a finitely generated subgroup H\leq G such that \widehat{G}\simeq\overline{H}\amalg K there exists a subgroup U\leq G such that G\simeq H*U . These two properties are related, but we have no reason to believe that one implies the other. It is known that, in the class of virtually free groups, profinite completions detect free products ( [ 12 ] ) and free factors ( [ 15 ] ). In this paper, we will consider this properties in the class of fundamental groups of graphs of virtually free groups with virtually cyclic edge groups, a class henceforth denoted by \mathcal{C} . A group is called LERF ( locally extended residually finite , or subgroup separable ) if every finitely generated subgroup is closed in the profinite topology. Equivalently, G is LERF if for every H\leq G finitely generated and every x\not\in H there is a finite index subgroup N\leq G containing H but not x . If G is LERF, then the profinite topology of G induces the profinite topology on each of its finitely generated subgroups ( [ 29 , Lem. 11.1.4] ). The main result of the paper is that, assuming the LERF condition, profinite completions detect free products in \mathcal{C} . Theorem 1.1 . Let G be a finitely generated LERF group that decomposes as the fundamental group of a graph of virtually free groups with virtually cyclic edge groups. If \widehat{G} decomposes as a non-trivial free profinite product, then G decomposes as a non-trivial free product. At this moment, we don’t know whether the LERF condition in Theorem 1.1 is necessary or if it can be replaced by the condition of residual finiteness. We also obtain the following corollary: Corollary 1.2 . Let G=F_{1}*_{c_{1}=c_{2}}F_{2} be the free product of two free groups amalgamated along non-trivial proper cyclic subgroups \langle c_{i}\rangle<F_{i} . If \widehat{G}=L\amalg K splits non-trivially as a free profinite product, then one of the F_{i} splits non-trivially as a free product F_{i}=H_{i}*K_{i} such that c_{i}\in H_{i} . One of the main ingredients of the proof of Theorem 1.1 is the proof of a Kurosh-style decomposition for virtually free groups. Theorem 1.3 . Let G be a finitely generated virtually free group and suppose that \widehat{G}\leq P=L\amalg K . Then G\simeq\left(\underset{g\in L\backslash P/G}{\ast}G\cap{{}^{g}L}\right)*\left(\underset{h\in K\backslash P/G}{\ast}G\cap{{}^{h}K}\right)*F where F is a free subgroup of G . With respect to the problem of detecting free factors, we can prove the following statement. Theorem 1.4 . Let G=\pi(Y,\mathcal{G}) be the fundamental group of a graph of virtually free groups with virtually cyclic edge groups. Assume that G is finitely generated and LERF. Let H be a finitely generated subgroup of G and moreover assume that either (1) H has at most one end or (2) H\leq{{}^{g}\mathcal{G}(v)} is contained in some G -conjugate of a vertex group. If \overline{H} is a free profinite factor of \widehat{G} , then H is a free factor of G . Another result we obtain, which is of independent interest, concerns the property of being cohomologically good (or cohomologically separable ); that is, if A is a finite G -module, then the restriction map on cohomology \mathrm{H}_{\mathrm{ct}}^{*}(\widehat{G},A)\to\mathrm{H}^{*}(G,A) is an isomorphism in every degree. The profinite completion of a finitely generated LERF group G from \mathcal{C} has a very transparent structure: it is the profinite fundamental group of the induced graph of profinite groups given by the profinite completions of the vertex and edge groups of G . This implies that they are cohomologically good (see, for instance, [ 16 , Thm. 3.9] ). We extend this result to finitely generated residually finite groups in the class \mathcal{C} . Theorem 1.5 . Let G be the fundamental group of a graph of virtually free groups with virtually cyclic edge groups. If G is finitely generated and residually finite, then G is cohomologically good. In particular, if G is not virtually free, then \widehat{G} has virtual cohomological dimension 2. This result was previously known in the case when G is the fundamental group of a finite graph of cyclic groups by a result of Cohen and Wykowski ( [ 10 ] ). We also present an application of Theorem 1.3 concerning profinite properties of words in free groups. Following Hanani, Meiri and Puder ( [ 18 , Def. 1.3] ), we say that a word w in a finitely generated free group F is profinitely rigid if any other word u\in F

Read the complete text