Home An axiomatizable profinite group with infinitely many open subgroups of index 2
Article Publicly Available

An axiomatizable profinite group with infinitely many open subgroups of index 2

  • Or Ben Porath EMAIL logo and Mark Shusterman
Published/Copyright: September 14, 2016

Abstract

We show that a profinite group with the same first-order theory as the direct product over all odd primes p of the dihedral group of order 2p, is necessarily isomorphic to this direct product.

1 Introduction

We say that a profinite group Γ is axiomatizable if for every profinite group Λ with the same first-order theory as that of Γ (in the language of groups), we have ΛΓ. The study of axiomatizable profinite groups began in [4], where it is shown that finitely generated profinite groups are axiomatizable, and an example of a profinite group that is not axiomatizable is given (for instance, 20). More generally, it is shown in [3] that a strongly complete profinite group (that is, a profinite group all of whose finite index subgroups are open) is axiomatizable. Are there more axiomatizable profinite groups?

By [3, Corollary 3.8] a strongly complete profinite group is small (that is, it has only finitely many open subgroups of index n, for every n). Thus, a possible precise formulation given by [3, Question 3.15 (i)] for the question we have just raised is whether every axiomatizable profinite group is small. Here, a negative answer to this question is given.

Theorem 1.1.

The profinite group G given by the direct product of the dihedral groups Dp over all odd primes p, is axiomatizable.

It follows from [2, Chapter 5] that a negative answer to [3, Question 3.15 (i)] cannot come from abelian profinite groups, so our example is, in a sense, the simplest possible.

2 The group G and its properties

Definition 2.1.

For a profinite group Γ we set

Inv(Γ):={τAut(Γ)τ2=IdΓ}

and note that this defines a group if Aut(Γ) is abelian.

Definition 2.2.

For a prime number p we denote by Cp the cyclic group of order p. If p is odd, then Aut(Cp)Cp-1 so Inv(Cp) is a group isomorphic to C2. The semidirect product CpInv(Cp) is the dihedral group Dp. Let ρp be a generator of Cp, and let ϵp be a generator of Inv(Cp) so that they generate Dp and we have ϵpρpϵp=ρp-1.

Lemma 2.3.

For an odd prime p we have Cp={[a,b]a,bDp}.

Proof.

For one inclusion note that Dp/Cp is abelian, and for the other one take some ρpnCp. As p is odd, there exists a k such that 2kn(p). We find that [ϵp,ϵpρpk]=ϵpϵpρpkϵpρp-kϵp=ρpkϵpρp-kϵp=ρpkρpk=ρp2k=ρpn.

Remark 2.4.

A similar argument shows that CpDp is its own centralizer.

Definition 2.5.

We set G:=Dp, C:=Cp and E:=ϵp, where the products (here and in the sequel) are always taken over all odd primes p.

Since CpDp and Dp/CpC2 for all odd primes p, we conclude that C is a closed normal procyclic subgroup of G with G/C(/2)0. Hence,

(2.1)g2Cfor all gG

and G/C is not small. Therefore, G is not small as well. Furthermore, we have G=CE. Since the Sylow subgroups of C are normal, we see that

Aut(C)=pAut(Cp)pCp-1

is an abelian group, so

Inv(C)=pInv(Cp)=E.

Thus,

(2.2)GCInv(C).

Definition 2.6.

For a profinite group Γ we denote by Γ its profinite commutator, which is the closed subgroup of Γ generated by {[a,b]a,bΓ}.

It follows from Lemma 2.3 that

(2.3)G=C={[a,b]a,bG}

so the following first-order sentence holds in G:

(2.4)a,b,c,dr,s:[a,b][c,d]=[r,s].

Remark 2.7.

Let us present some more first-order sentences that hold in G.

  1. By (2.1), (2.3) the following first-order sentence is valid in G

    (2.5)gh,k:g2=[h,k].
  2. It follows from Remark 2.4 that C is its own centralizer in G. In view of (2.3), this is tantamount to the following first-order sentence

    (2.6)x((y,z:x[y,z]x-1=[y,z])a,b:x=[a,b]).
  3. Fix an odd prime p. We can think of ϵp,ρp as elements of E,C respectively, and thus also as elements of G. The first-order sentence

    (2.7)xy,z:x[y,z]x-1=[y,z]-1[y,z]p=1

    holds in G since we can take x=ϵp.

Definition 2.8.

We say that profinite groups Γ,Λ are elementarily equivalent if they have the same first-order theory, and denote this by ΓΛ.

3 The proof of Theorem 1.1

Let G~ be a profinite group for which GG~. Set

(3.1)C~:={[g,h]g,hG~}

and note that it is the image in G~ of the compact G~2 under the continuous map sending (g,h)G~2 to [g,h]G~. Hence, C~ is compact, and thus closed in G~. By (2.4), C~ is a subgroup of G~, and by (2.3), (3.1) we have CC~.

By [4, Theorem A], we have CC~ and by (2.5) every element of G~/C~ is of order dividing 2, so |G~/C~| is prime to |C~|=|C|. Since C~C is abelian and normal in G~, the action by conjugation of G~ on C~ gives rise to a continuous homomorphism τ:G~/C~Inv(C~). By the Schur–Zassenhaus theorem (see [1, Lemma 22.10.1]) we get that

(3.2)G~C~τG~/C~.

By (2.6) τ is injective, and in order to see that it is also surjective, first identify Inv(C~) with Inv(C). A generating set for Inv(C~) is thus given by {ϵp}p. By (2.7) the image of τ contains ϵp for each odd prime p, so τ is a surjection, and thus an isomorphism. We conclude that

G(2.2)CInv(C)C~Inv(C~)C~τG~/C~(3.2)G~.

Remark 3.1.

From our proof of Theorem 1.1 one can extract an explicit (infinite) set of axioms characterizing G up to an isomorphism. For that, one just has to find a set of first-order sentences characterizing C, and such a set can be obtained from the proof of [4, Theorem A]. Let us however write down such axioms for C explicitly.

First, we take a sentence saying that C is abelian. Since an abelian profinite group is isomorphic to the direct product of its Sylow subgroups, it suffices to include axioms assuring that the Sylow subgroups are elementary abelian, and cyclic. For that matter, include a sentence saying that C=C2, and fix an odd prime p. Finally, take a sentence saying that C/Cp2/p, so that the p-Sylow subgroup is isomorphic to /p.


Communicated by Pavel A. Zalesskii


Award Identifier / Grant number: UGC no. 40/14

Funding statement: The second author was partially supported by a grant of the Israel Science Foundation with cooperation of UGC no. 40/14.

Acknowledgements

We would like to sincerely thank Arno Fehm for telling us about the question that motivated this work, and for many helpful discussions. Special thanks go to the referee for the remarks given.

References

[1] M. D. Fried and M. Jarden, Field Arithmetic, 3rd ed., revised by M. Jarden, Ergeb. Math. Grenzgeb. (3) 11, Springer, Heidelberg, 2008. Search in Google Scholar

[2] N. Frohn, Model theory of absolute Galois groups, Dissertation, University of Freiburg, 2011. Search in Google Scholar

[3] P. Helbig, On small profinite groups, preprint (2015), http://arxiv.org/abs/1511.08760. 10.1515/jgth-2017-0014Search in Google Scholar

[4] M. Jarden and A. Lubotzky, Elementary equivalence of profinite groups, Bull. Lond. Math. Soc. 40 (2008), 887–896. 10.1112/blms/bdn069Search in Google Scholar

Received: 2016-8-9
Published Online: 2016-9-14
Published in Print: 2017-7-1

© 2017 Walter de Gruyter GmbH, Berlin/Boston

Downloaded on 19.9.2025 from https://www.degruyterbrill.com/document/doi/10.1515/jgth-2016-0044/html
Scroll to top button