Two-generated verbally closed subgroups of a free solvable group G are retracts of G
V.A. Roman'kov, E.I. Timoshenko

TL;DR
This paper proves that in free solvable groups of finite rank, any two-generated verbally closed subgroup is necessarily a retract of the entire group, revealing a structural property of these subgroups.
Contribution
It establishes that all two-generated verbally closed subgroups in free solvable groups are retracts, a new insight into subgroup structure.
Findings
Every two-generated verbally closed subgroup is a retract.
The result applies to free solvable groups of finite rank.
Provides a structural characterization of such subgroups.
Abstract
We prove that every verbally closed two-generated subgroup of a free solvable group G of a finite rank is a retract of G.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsFinite Group Theory Research · semigroups and automata theory · Advanced Topology and Set Theory
Two-generated verbally closed subgroups of a free solvable group are retracts of
V.A. Roman’kov, E.I. Timoshenko
Abstract
We prove that every verbally closed two-generated subgroup of a free solvable group of a finite rank is a retract of
1 Introduction
Algebraically closed objects play an important part in modern algebra. In this paper we study verbally closed and algebraically closed subgroups of free solvable groups.
We first recall that a subgroup is called algebraically closed in a group if for every finite system of equations with constants from the following holds: if has a solution in then it has a solution in . Then we recall that a subgroup of a group is called a *retract * of , if there is a homomorphism (termed retraction) which is identical on . It is easy to show that every retract of is algebraically closed in . Furthermore, if is finitely presented and is finitely generated then the converse is also true (see [1]). This result still holds for any finitely generated group which is equationally Noetherian. Recall that a group is called equationally Noetherian if for any every system of equations in variables with coefficients from is equivalent (has the same solution set in ) to some finite subsystem of itself (see [2], [3].
Let be the free group of countably infinite rank with basis For and a group by we denote the set of all -elements in , i.e., . The verbal subgroup is the subgroup of generated by . The -width of an element is the minimal natural number such that is a product of -elements in or their inverses; the width of is the supremum of widths of its elements. The first question about -width goes back to the Ore’s paper [4] where he asked whether the -width (the commutator width) of every element in a non-abelian finite simple group is equal to (Ore Conjecture). The conjecture was established in [5]. Observing paper [6] and monographs [7], [8] present results about verbal width in groups.
Two important questions arise naturally for an extension and a given word :
- •
when it is true that ?
- •
when for a given ?
To approach these questions Mysnikov and the author introduced in [1] a new notion of verbally closed subgroups.
Definition 1.1**.**
A subgroup of is called verbally closed if for any word and element equation has a solution in if and only if it has a solution in , i.e., for every .
Theorem 1.2**.**
[1*]**.
Let be a free group of a finite rank. Then for a subgroup of the following conditions are equivalent:*
- a)
* is a retract of .* 2. b)
* is a verbally closed subgroup of .* 3. c)
* is an algebraically closed subgroup of .*
This result clarifies the nature of verbally or algebraically closed subgroups in . Surprisingly, the ”weak” verbal closure operator in this case is as strong as the standard one.
Similar result is true for any free nilpotent group of a finite rank.
Theorem 1.3**.**
[9*]**.
Let be a free nilpotent group of a finite rank and class . Then for a subgroup of the following conditions are equivalent:*
- a)
* is a retract of .* 2. b)
* is a verbally closed subgroup of .* 3. c)
* is an algebraically closed subgroup of .* 4. d)
* is a free factor of the free group in the variety of all nilpotent groups of the class *
2 Preliminaries
In this section we collect some known or simple facts on verbally or algebraically closed subgroups.
Proposition 2.1**.**
[1]. Let be a group extension. Then the following holds:
If is a retract of then is algebraically closed in
- 2)
Suppose is finitely presented and is finitely generated. Then is algebraically closed in if and only if is a retract.
- 3)
Suppose is finitely generated relative to and is equationally Noetherian. Then is algebraically closed in if and only if is a retract.
Further in the paper denotes a free metabelian group of rank and denotes a free metabelian nilpotent of class group of rank
The following statement was proved in [10] (see also [8]).
Lemma 2.2**.**
Let denotes or with basis . Let tuple is a solution of the equation
[TABLE]
Then
Lemma 2.3** ([9], Lemma 1.1).**
Let be a group and let be a verbal subgroup of If is a verbally closed subgroup of then its image is verbally closed in .
Proof.
Suppose that equation () has a solution in . Let be a preimage of this solution in Then is a solution of the equation in Here is a preimage of in There also is a preimage of in Then . Let be a corresponding to word that has value 1 in The equation is solvable in . Then it has a solution in The image of the first components of this solution will be a solution of in ∎
Lemma 2.4**.**
Let be a two-generated verbally closed noncyclic subgroup of a free solvable group ). Then the image in the abelianization (that is a free abelian group of rank ) is a direct factor of rank , and is a free solvable group of rank and class
Proof.
By Lemma 2.3 is verbally closed in . By Theorem 1.3 is a direct factor of Obviously, is non-trivial. We need to prove that is not cyclic. By induction on , we can assume that the image of in is generated by the image of (so it is cyclic) and that . Let be an expression of in the free generators of The equation is solvable in . Hence, it has a solution in Then we can write the components in the form Then where and is a value of -th partial Fox derivation in Details about Fox derivatives see in [8]. It is easy to compute directly without Fox derivatives. We can collect exponents of from the expression Note that belongs to the fundamental ideal of , because is a commutator word. Then and This is impossible because has no module torsion (see [8]).
By the G. Baumslag’s result [11] is free solvable group of rank and class with basis Recall that Baumslag proved in [11] the following statement: A subgroup H of a free solvable group is itself a free solvable group if and only if there exists a set of generators of which freely generates, modulo some term of the derived series of , a free abelian group. He also gave an obvious formula of the solvability class of See also [12]. ∎
3 Description of verbally and algebraically closed cyclic subgroups of free solvable groups
We start with the case of a cyclic subgroup.
Lemma 3.1**.**
Let be a free solvable group of a finite rank and let be a cyclic subgroup of generated by a non-trivial element . Then the following conditions are equivalent:
* is verbally closed in ;*
- 2)
* is a retract of ;*
- 3)
the image of in the abelianization (that is free abelian group of rank ) is primitive.
Proof.
Let be a basis of . The element can be expressed uniquely in the form
[TABLE]
where and is a product of commutators of words in (a commutator word). Then
To show that 1) 3) assume that has a non primitive image in , i.e., either or
We first suppose that , so . Replacing each by a variable in (2) one gets an equation , with as a constant from , which has a solution in . However, this equation does not have a solution in , since is abelian, so for any . This shows that is not verbally closed in - contradiction. So . Then in this case The equation
[TABLE]
still has a solution in , but for any one has
[TABLE]
for some . Hence, the equation does not have a solution in , so is not verbally closed - contradiction.
To show that 3) 2) assume that is primitive in the abelianization of . Then there are integers such that Now we define a homomorphism by putting for Since is abelian , so and is a retraction. Hence is a retract, as claimed.
- 1) follows from Proposition 2.1, statement 1). ∎
4 Description of verbally and algebraically closed two-generated subgroups of free metabelian groups
Further in the Section, is denoted a basis of group that is or
Theorem 4.1**.**
Let be a two-generated subgroup of that is or , Then the following conditions are equivalent:
* is a retract of .* 2. 2)
* is a algebraically closed subgroup of .* 3. 3)
* is an verbally closed subgroup of .*
Proof.
The implications 1) 2) 3) are obvious. We are to prove that 3) 1). By Lemma 2.3 the image of in the abelianization is verbally closed. By Lemma 2.4 this image is a direct factor of of rank 2. We can assume that and . Let and be expressions of and respectively. Then an equation
[TABLE]
has a solution in Thus it has a solution in It means that there are elements in such that
[TABLE]
Now we consider the case By the Baumslag’s theorem (see above) (4) is valid in the free metabelian group . By Lemma 2.2
[TABLE]
Recall that an element of a group is said to be test element if any endomorphism for which is an automorphism of Timoshenko proved in [13] that every non-trivial element of is a test element for Hence the following map
[TABLE]
defines an automorphism of Let Then this automorphism is identical By Bachmuth’s theorem [14] it is an inner automorphism (see also [17]). Then there is an element for which
[TABLE]
It follows that Then because the centralizer of any non-trivial element of the commutant of is equal to . This statement was proved by Mal’cev in [12].
Define an endomorphism of by the map
[TABLE]
The image lies in Moreover,
[TABLE]
[TABLE]
Thus is identical on and so is a retraction and is a retract.
Now let As above the following map
[TABLE]
defines an automorphism of A composition of this automorphism with itself is identical and we can finish our proof as above.
Let This group is free in a nilpotent variety . Then a pair of elements which induce a pair of free generators in , generate a free -factor in . This factor is obviously a retract of ∎
5 Description of verbally and algebraically closed two-generated subgroups of free solvable groups
To prove the main result of this section in the case of group we will use a concrete example of a test element of , that was constructed in [15]. The proof in the general case of group will be slightly different.
Example 5.1**.**
Let be a basis of For any pair of positive integers and we denote by the element , and by we denote the element Then
[TABLE]
is a test element of .
The following lemma was proved in [15].
Lemma 5.2**.**
Let be an endomorphism of for which where is the element in Example 5.1. Then is an automorphism identical modulo In this case is inner by theorem proved in [16] (see also [17]).
The following more general result was proved in [18]
Lemma 5.3**.**
Let be a free solvable group of rank and class with basis and let Then there is a positive number such that element is a test element. Moreover, for every automorphism if then is an inner automorphism.
Theorem 5.4**.**
Let be a two-generated subgroup of Then the following conditions are equivalent:
* is a retract of .* 2. 2)
* is a algebraically closed subgroup of .* 3. 3)
* is an verbally closed subgroup of .*
Proof.
The implications 1) 2) 3) are obvious. We are to prove that 3) 1).
Case of group Let be a basis of . By Theorem 4.1 and Lemmas 2.2 and 2.3 we can assume that and . Let and be two expressions of and respectively. Let be the element described by Example 5.1. Then equation
[TABLE]
has a solution in . Hence it has a solution in The map defines an endomorphism with the image in It maps to and to By Lemma 5.2 the map defines an inner automorphism of Then there is for which
[TABLE]
It follows that In any group the centralizer of any non-trivial element coincides with [12]. Hence Let be an endomorphism of defined by the map: The image lies in , and
[TABLE]
Hence is a retraction, and is a retract.
Case of group . Now is defined in Lemma 5.3. We repeat all the arguments as above until we get the map which defines an automorphism of Then By Lemma 5.3 is an inner automorphism of For some we have equalities (13). Then we finish proof as in the previous case.
∎
6 Open problems
Problem 6.1**.**
Is it true that for any subgroup of any free solvable group of rank and class the following conditions are equivalent:
is a retract of , 2. 2)
is a algebraically closed subgroup of , 3. 3)
is an verbally closed subgroup of ?
It is likely that the answer to this question is negative. By Timoshenko’s theorem (see [18] or [19]) the test rank of is Recall, that test rank of a group is the minimal number of elements of such that every endomorphism fixing any of this elements is automorphism. In the general case does not contain a test element. Hence, the methods of this paper do not work completely.
Problem 6.2**.**
Is it true that for any two-generated subgroup of any free polynilpotent group of rank and class the following conditions are equivalent:
is a retract of , 2. 2)
is a algebraically closed subgroup of , 3. 3)
is an verbally closed subgroup of ?
By Timoshenko’s theorem [18] (see also [19]) any free solvable group contains test elements, all belong to For the affimative solution to this problem sufficient to prove that any contains a test element with properties as in Lemma 5.2. Note that the reference to the statement in [16] (or [17]) in Lemma 5.2 works in the general case. Namely, any automorphism of identical is inner.
Note, that by Gupta and the second author’s theorem [20] (see also [19]) any free polynilpotent group of class for has a test element. We need in analog of Lemma 2.4 to solve this problem affirnatively.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] A. Myasnikov, V. Roman’kov. Verbally closed subgroups of free groups. J. Group Theory. 2014. 17, 29-40.
- 2[2] G. Baumslag, A. Myasnikov and V. Roman’kov. Two theorems about equationally Noetherian groups. J. Algebra. 1997. 194, 654-664.
- 3[3] G. Baumslag, A. Myasnikov and V. Remeslennikov. Algebraic geometry over groups I. Algebraic sets and ideal theory. J. Algebra. 1999. 219, 16-79.
- 4[4] O. Ore. Some remarks on commutators. Proc. Amer. Math. Soc. 1951. 2, 307-314.
- 5[5] M. Liebeck, E. O’Brian, A. Shalev A. and P. Tiep. The Ore conjecture. J. European Math. Soc. 2010. 12, 939-1008.
- 6[6] V. Roman’kov. Equations over groups. Groups, Complexity, Cryptology. 2012. 4, No. 2, 191-239.
- 7[7] D. Segal. Words: notes on verbal width in groups. London Math. Soc. Lect. Notes Ser. 361. Cambridge: Cambridge Univ. Press, 2009. 134 p.
- 8[8] V.A. Roman’kov. Essays in algebra and cryptology. Solvable groups. Omsk: Omsk State University Publishing House, 2017. 207 p.
