The growth of residually soluble groups

Sean Eberhard, Elena Maini · 2025

Read The growth of residually soluble groups by Sean Eberhard, Elena Maini (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-2511-07018.txt.

Introduction

Sean Eberhard and Elena Maini prove a growth theorem for residually soluble groups, classes of symmetries whose structure can be probed through soluble quotients. The question is how fast the number of elements within a given word-length grows, and what that rate reveals. Building on work of Wilson, they show that if the growth function grows slower than a certain subexponential bound, the group must be virtually nilpotent, a highly restricted class.

Concretely, the condition is that the logarithm of the growth function, divided by the fourth root of the length, tends to zero. The result proves Grigorchuk's Gap Conjecture for all exponents below one quarter within residually soluble groups, improving Wilson's earlier exponent of one sixth. Stronger versions of the conjecture are also discussed.

Catalog record

CREATORSean Eberhard, Elena Maini
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE37 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-07018

Browse the full library · About the Tower of Babel

Full text

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

The growth of residually soluble groups Thanks: SE is supported by The Royal Society (University Research Fellowship, URF\R1\221185). EM is supported by the Warwick Mathematics Institute Centre for Doctoral Training, and gratefully acknowledges funding by the Swinnerton-Dyer scholarship. Sean Eberhard, Elena Maini Abstract . Building on work of Wilson, we show that if G G is a finitely generated residually soluble group whose growth function γ \gamma satisfies ( log ⁡ γ ⁡ ( n ) ) / n 1 / 4 → 0 (\log\gamma(n))/n^{1/4}\to 0 as n → ∞ n\to\infty then G G is virtually nilpotent. This shows that Grigorchuk’s Gap Conjecture holds for all exponents β < 1 / 4 \beta<1/4 within the class of residually soluble groups (improving Wilson’s exponent 1 / 6 1/6 ). We also discuss stronger versions of the Gap Conjecture. 1. Introduction If G is a group generated by a finite set X and g\in G , we denote by \ell_{X}(g) the length of g with respect to X , i.e., the smallest integer n such that g can be written as a product of n elements in X\cup X^{-1} . For a finite subset S\subseteq G , we write \ell_{X}(S)=\max\{\ell_{X}(g):g\in S\} . The growth function of X is \gamma_{X}(n)=|\{g\in G:\ell_{X}(g)\leq n\}|. Usually, growth functions are only coarsely compared. Given two functions f,g:\mathbf{N}\to[0,\infty) , we write f\preceq g if there is a constant C>0 such that f(n)\leq Cg(Cn) for all n>0 . This relation defines a preorder on the set of functions. The equivalence class of the growth function \gamma_{X}(n) is a group property (i.e., independent of the choice of generating set X ), and in fact a quasi-isometry invariant. In 1968, Milnor [ 14 ] asked some highly influential questions about the relation between the growth function \gamma_{X}(n) and the group structure of G . One of his conjectures was that \gamma_{X}(n)\preceq n^{d} for some d (if and) only if G is virtually nilpotent; this is now the celebrated theorem of Gromov [ 9 ] . Milnor also asked whether the growth of a group is always either bounded by a polynomial or equivalent to e^{n} . This was proved for soluble groups by Milnor and Wolf [ 13 , 22 ] , for linear groups by Tits [ 19 ] , for elementary amenable groups by Chou [ 2 ] , but Grigorchuk [ 5 ] famously discovered a counterexample, i.e., a group of intermediate growth . In fact, Grigorchuk [ 6 ] showed that there are a continuum of groups with pairwise incomparable growth functions, which shows that the landscape of group growth functions is in truth completely dominated by the groups of intermediate growth. Notwithstanding the existence of groups of intermediate growth, there is undoubtedly a definite gap between polynomial growth and superpolynomial growth. This follows already from Gromov’s theorem and some logical nonsense (see [ 9 ] *p. 71), but without an explicit function. Pursuing this idea, Shalom and Tao [ 16 ] proved that if G is not virtually nilpotent then \gamma_{X}(n)\succeq n^{(\log\log n)^{c}} for some c>0 . This is the greatest lower bound currently known for the growth of non-virtually-nilpotent groups in general. On the other hand, Grigorchuk’s groups all have growth \succeq\exp(\sqrt{n}) . In particular, the growth of Grigorchuk’s first group itself is of the form \exp(n^{\alpha+o(1)}) , where \alpha\approx 0.767 (Erschler–Zheng [ 4 ] ), and this is still the slowest known superpolynomial growth function. For these reasons among others, Grigorchuk [ 8 ] has advocated the following conjecture, known as the Gap Conjecture , which we formalize with a free parameter \beta . Conjecture ( \operatorname{Gap}(\beta) ) . If G is a group generated by a finite set X and \gamma_{X}(n)\prec\exp(n^{\beta}) then G is virtually nilpotent. We have seen that Milnor’s conjecture \operatorname{Gap}(1) is false, but it does hold within certain large natural classes of groups including soluble groups and linear groups. “The” Gap Conjecture is \operatorname{Gap}(1/2) , and it was proved in [ 7 , 10 ] that \operatorname{Gap}(1/2) holds for residually nilpotent groups (such as all of Grigorchuk’s groups). In [ 20 , 21 ] , J. S. Wilson considered the wider class of residually soluble groups, and in the latter paper he proved that \operatorname{Gap}(1/6) holds within this class. Our main contribution in this paper is to show that Wilson’s exponent can be improved to 1/4 . Theorem 1.1 ( \operatorname{Gap}(1/4-\epsilon) for residually soluble groups) . If G=\left\langle X\right\rangle is a finitely generated residually soluble group such that \log\gamma_{X}(n)/n^{1/4}\to 0\qquad\text{as}~n\to\infty, then G is virtually nilpotent. In particular, \operatorname{Gap}(\beta) holds for residually soluble groups for all \beta<1/4 . Our proof of Theorem 1.1 builds on Wilson’s ideas, particularly the consideration of self-centralizing chief factors. We will review everything we need. The proof splits into two cases. If the self-centralizing chief factors of G have bounded rank then \operatorname{Gap}(1/2) holds for G . Otherwise there are self-centralizing chief factors of arbitrarily large rank, and then we can use a generalization of a lemma of Milnor to produce many distinct elements of short length in the chief factor. To optimize the approach, we use an ad hoc notion of “modified derived length”, which is a variant of derived length that allows class-2 nilpotent factors at a discount, and we bound the modified derived length of soluble linear groups. 2. Self-centralizing chief rank A key insight of Wilson [ 20 ] was to consider self-centralizing chief factors. A chief factor M/N of a finite group G is a minimal normal subgroup of a quotient G/N . If M/N is abelian (e.g, if G is soluble) then M/N\cong\mathbf{F}_{p}^{n} for some prime p and integer n\geq 1 ; the rank of M/N is n . A chief factor M/N of G is called self-centralizing if M/N=C_{G/N}(M/N) . If G/M is soluble, a Frattini argument shows that M/N is complemented in G/N , and it follows that G has a self-centralizing chief factor of order p^{n} if and only if G has a maximal subgroup of index p^{n} . Definition 2.1 . Let G be a finite soluble group. The self-centralizing chief rank (a.k.a. non-Frattini chief rank ) of G is the maximum rank of its self-centralizing chief factors. We write \mathfrak{S}(n) for the class of finite soluble groups of self-centralizing chief rank at most n . Remarks . (1) A chief factor M/N is called Frattini if M/N\leq\Phi(G/N) . Every self-centralizing chief factor is non-Frattini, and conversely every non-Frattini chief factor is equivalent to a self-centralizing chief factor. Refer to [ 3 ] *A.15. (2) Every nontrivial soluble group has positive self-centralizing chief rank, and the class \mathfrak{S}(1) coincides with the class of finite supersoluble groups: see [ 3 ] *VII 2.2(a, c). If G is a soluble group we denote by \delta(G) its derived length : the least integer k such that G^{(k)}=1 . We need the following well-known result of Zassenhaus and Newman (see [ 15 ] ). The weak form, due to Zassenhaus, is sufficient in this section, but in the next section we will use the precise bound found by Newman. Theorem 2.2 . Let \rho(n) be the supremum of the derived lengths \delta(G) of the soluble linear groups of degree n

Read the complete text