Total computable two-variable laws ap,na_{p,n} detecting abelian Sylow pp-subgroups for n≥∣G∣n\ge|G|

11.15

Problem

For a prime p>2p>2, give an explicit sequence of two-variable words that eventually characterizes finite groups with abelian Sylow pp-subgroups.

Syl⁡p(G) abelian⟺∃N ∀n≥N, G⊨wp,n(x,y)=1.\operatorname{Syl}_p(G)\text{ abelian} \quad\Longleftrightarrow\quad \exists N\ \forall n\ge N,\ G\models w_{p,n}(x,y)=1.

Here explicit means a total computable word construction. The exhaustive search below does not claim a practical running-time bound or a short power-commutator formula.

The sequence and its scope

Fix a prime pp. The existence of a sequence of two-variable laws eventually characterizing finite groups with abelian Sylow pp-subgroups is already known. Brandl (2) gives an explicit power-commutator sequence for p=2p=2, and for arbitrary pp within the class of finite pp-soluble groups. Problem 11.15 of the Kourovka Notebook (4) asks for an explicit sequence when p>2p>2. Related commutator characterizations are studied in (3).

We give a total algorithm, with a uniform threshold equal to the order of the target group. The algorithm enumerates finite words and tests them in a finite symmetric group and a finite direct product. This is a search construction. Explicit recursive formulas, such as those considered for finite solubility in (1), impose an additional form on the sequence which is not part of our assertion.

We use [x,y]=x−1y−1xy[x,y]=x^{-1}y^{-1}xy. A word is represented by a finite string in x,x−1,y,y−1x,x^{-1},y,y^{-1}. Its value at (g,h)∈G2(g,h)\in G^2 is denoted by w(g,h)w(g,h). Thus G⊨w=1G\models w=1 means that w(g,h)=1w(g,h)=1 for every (g,h)∈G2(g,h)\in G^2.

Theorem 1. There is a total computable function (p,n)↦ap,n(p,n)\mapsto a_{p,n} from N2\mathbb N^2 to the set of two-variable words with the following properties. If pp is prime and GG is finite, then

n≥∣G∣⟹(G⊨ap,n=1 ⟺ every Sylow p-subgroup of G is abelian).(1) \tag{1} n\geq |G| \quad\Longrightarrow\quad \bigl(G\models a_{p,n}=1 \ \Longleftrightarrow\ \text{every Sylow }p\text{-subgroup of }G\text{ is abelian}\bigr).

For every pp-subgroup H≤GH\leq G and every g,h∈Hg,h\in H,

ap,n(g,h)=[g,h](n≥∣G∣).(2) \tag{2} a_{p,n}(g,h)=[g,h]\qquad(n\geq |G|).

In particular, if some Sylow pp-subgroup is nonabelian, there are fixed g,h∈Gg,h\in G such that ap,n(g,h)≠1a_{p,n}(g,h)\ne1 for every n≥∣G∣n\geq |G|.

It follows that the Sylow pp-subgroups of GG are abelian if and only if, for every pair (g,h)∈G2(g,h)\in G^2, one has ap,n(g,h)=1a_{p,n}(g,h)=1 for all sufficiently large nn. In this formulation the threshold may initially depend on the pair; Theorem 1 supplies the uniform threshold ∣G∣|G| whenever the condition holds.

Finite subgroup tests

Let SS be a finite group. For a subset T⊆ST\subseteq S, write

C(T)⟺1∈T,TT⊆T,T−1⊆T.\mathcal C(T)\quad\Longleftrightarrow\quad 1\in T,\quad TT\subseteq T,\quad T^{-1}\subseteq T.

These are precisely the subgroup conditions. Define

Pp(a,b)⟺∃T⊆S (C(T), a,b∈T, ∣T∣=pj for some 0≤j≤∣T∣).(4) \tag{4} \mathcal P_p(a,b) \quad\Longleftrightarrow\quad \exists T\subseteq S\ \bigl(\mathcal C(T),\ a,b\in T,\ |T|=p^j\text{ for some }0\leq j\leq |T|\bigr).

Lemma 2. The predicate Pp\mathcal P_p is decidable from the finite multiplication table of SS. If pp is prime, then Pp(a,b)\mathcal P_p(a,b) holds if and only if a,ba,b lie in a common pp-subgroup of SS.

Proof. Enumerate the subsets of SS, test the three subgroup conditions, and test the finitely many integers 0≤j≤∣T∣0\leq j\leq |T|. A successful subset is a subgroup of prime-power order. Conversely, a finite pp-subgroup TT has order pjp^j for some j≥0j\geq0. Since p≥2p\geq2, induction gives j≤pj=∣T∣j\leq p^j=|T|, so this subgroup occurs in the finite test. ◻

Universal evaluations and Sylow lifts

For a finite group SS and F=F(x,y)F=F(x,y), put

DS=∏(a,b)∈S2S,η:F⟶DS,η(w)(a,b)=w(a,b).D_S=\prod_{(a,b)\in S^2}S,\qquad \eta:F\longrightarrow D_S,\qquad \eta(w)_{(a,b)}=w(a,b).

Write Q=η(F)Q=\eta(F). The group QQ is finite and generated by η(x)\eta(x) and η(y)\eta(y).

A pair of words (u,v)(u,v) is called suitable for (p,S)(p,S) when

Pp(η(u),η(v))in DS,Pp(a,b)⟹u(a,b)=a,v(a,b)=bfor every (a,b)∈S2.\begin{align*} &\mathcal P_p\bigl(\eta(u),\eta(v)\bigr) &&\text{in }D_S,\tag{6a}\\ &\mathcal P_p(a,b)\Longrightarrow u(a,b)=a,\quad v(a,b)=b &&\text{for every }(a,b)\in S^2.\tag{6b} \end{align*}

Here the first occurrence of Pp\mathcal P_p uses the finite group DSD_S. By Lemma 2, suitability is a decidable predicate on finite word pairs.

Lemma 3. If pp is prime, then every finite group SS admits a suitable pair of words.

Proof. Let II be the set of pairs (a,b)∈S2(a,b)\in S^2 satisfying Pp(a,b)\mathcal P_p(a,b). For each i=(a,b)∈Ii=(a,b)\in I, choose a pp-subgroup Hi≤SH_i\leq S containing a,ba,b. Every word evaluated at (a,b)(a,b) lies in HiH_i, so restriction to these coordinates defines a homomorphism

ψ:Q⟶∏i∈IHi.\psi:Q\longrightarrow \prod_{i\in I}H_i.

The product is a pp-group. Indeed, for an element (hi)(h_i), choose integers eie_i such that hipei=1h_i^{p^{e_i}}=1; its p∑ieip^{\sum_i e_i}-th power is then 11. Thus L=ψ(Q)L=\psi(Q) is a pp-group.

Choose a Sylow pp-subgroup PP of QQ. Surjective homomorphisms of finite groups map Sylow subgroups to Sylow subgroups. Applying this to ψ:Q→L\psi:Q\to L gives ψ(P)=L\psi(P)=L, since a pp-group is its own Sylow pp-subgroup. There are therefore s,t∈Ps,t\in P with

ψ(s)=ψ(η(x)),ψ(t)=ψ(η(y)).\psi(s)=\psi(\eta(x)),\qquad \psi(t)=\psi(\eta(y)).

Choose word representatives u,v∈Fu,v\in F such that η(u)=s\eta(u)=s and η(v)=t\eta(v)=t. The subgroup PP, viewed inside DSD_S, verifies (6a). For i=(a,b)∈Ii=(a,b)\in I, equality of the ii-th coordinates gives u(a,b)=au(a,b)=a and v(a,b)=bv(a,b)=b, which is (6b). ◻

The word-producing algorithm

We specify a coding so that the search determines a single sequence. Define the pairing function

π(r,s)={s2+r,r<s,r2+r+s,r≥s.\pi(r,s)= \begin{cases} s^2+r,&r<s,\\ r^2+r+s,&r\geq s. \end{cases}

This is a bijection N2→N\mathbb N^2\to\mathbb N. Give the letters x−1,x,y−1,yx^{-1},x,y^{-1},y the codes ϵ(ℓ)=0,1,2,3\epsilon(\ell)=0,1,2,3, respectively. For a finite word ℓw\ell w, with first letter ℓ\ell, define

E(∅)=0,E(ℓw)=1+π(ϵ(ℓ),E(w)),E(u,v)=π(E(u),E(v)).(10) \tag{10} E(\varnothing)=0,\qquad E(\ell w)=1+\pi(\epsilon(\ell),E(w)),\qquad E(u,v)=\pi(E(u),E(v)).

Pairing inversion and recursive list decoding give a computable partial inverse. Let δ:N→W2\delta:\mathbb N\to\mathcal W^2 be its total extension, assigning (∅,∅)(\varnothing,\varnothing) to invalid codes, where W\mathcal W is the set of finite words. Then

δ(E(u,v))=(u,v).(11) \tag{11} \delta(E(u,v))=(u,v).

For prime pp and n≥0n\geq0, put Sn=Sym⁡({0,…,n−1})S_n=\operatorname{Sym}(\{0,\ldots,n-1\}). Starting at m=0m=0, test whether δ(m)\delta(m) is suitable for (p,Sn)(p,S_n), and increase mm by one until the test succeeds. Equivalently, set

m(p,n)=min⁡{m∈N:δ(m) is suitable for (p,Sn)},(up,n,vp,n)=δ(m(p,n)).(12) \tag{12} m(p,n)=\min\{m\in\mathbb N:\delta(m)\text{ is suitable for }(p,S_n)\}, \qquad (u_{p,n},v_{p,n})=\delta(m(p,n)).

Define

ap,n=up,n−1vp,n−1up,nvp,n.(13) \tag{13} a_{p,n}=u_{p,n}^{-1}v_{p,n}^{-1}u_{p,n}v_{p,n}.

For nonprime pp, set ap,n=∅a_{p,n}=\varnothing.

Each suitability test terminates by finite enumeration. Lemma 3 and (11) show that a successful code exists for every prime pp and every nn. Thus (12) is a terminating sequential search, and (13) defines a total computable function. The subgroups selected in Lemma 3 are used to prove termination; the search itself uses only the finite predicates (6a) and (6b).

The uniform tail

Proof of Theorem 1. Let GG be finite and n≥∣G∣n\geq |G|. The regular action of GG, extended by fixed points, gives an injective homomorphism ι:G→Sn\iota:G\to S_n. Write u=up,nu=u_{p,n} and v=vp,nv=v_{p,n}.

Suppose first that every Sylow pp-subgroup of GG is abelian. For g,h∈Gg,h\in G, evaluate the coordinates of DSnD_{S_n} at (ι(g),ι(h))(\iota(g),\iota(h)). By (6a), there is a pp-subgroup T≤DSnT\leq D_{S_n} containing η(u)\eta(u) and η(v)\eta(v). Its image under this coordinate projection is a pp-subgroup M≤SnM\leq S_n containing u(ι(g),ι(h))u(\iota(g),\iota(h)) and v(ι(g),ι(h))v(\iota(g),\iota(h)). The subgroup K=ι−1(M)K=\iota^{-1}(M) embeds in MM, and is therefore a pp-group. Word evaluation commutes with ι\iota, so u(g,h),v(g,h)∈Ku(g,h),v(g,h)\in K. Every pp-subgroup of GG is contained in a Sylow pp-subgroup; hence KK is abelian. Consequently

ap,n(g,h)=[u(g,h),v(g,h)]=1.a_{p,n}(g,h)=[u(g,h),v(g,h)]=1.

Next let H≤GH\leq G be a pp-subgroup and let g,h∈Hg,h\in H. The pair (ι(g),ι(h))(\iota(g),\iota(h)) lies in the pp-subgroup ι(H)\iota(H) of SnS_n. Equation (6b) gives

ι(u(g,h))=ι(g),ι(v(g,h))=ι(h).\iota(u(g,h))=\iota(g),\qquad \iota(v(g,h))=\iota(h).

Injectivity of ι\iota implies u(g,h)=gu(g,h)=g and v(g,h)=hv(g,h)=h, proving (2).

If a Sylow pp-subgroup is nonabelian, choose a noncommuting pair g,hg,h in it. This pair is independent of nn. Equation (2) gives

ap,n(g,h)=[g,h]≠1(n≥∣G∣),a_{p,n}(g,h)=[g,h]\ne1 \qquad(n\geq |G|),

which proves the converse in (1) and the persistent failure assertion. ◻

The pointwise eventual formulation follows as well. In the nonabelian case, the fixed pair just obtained fails at every stage beyond ∣G∣|G|. In the abelian case, the first part of the proof gives a law at every such stage.

References

Preprint · Lean (GitHub)

  1. T. Bandman, G.-M. Greuel, F. Grunewald, B. Kunyavskiı̆, G. Pfister and E. Plotkin, Two-variable identities for finite solvable groups, C. R. Math. Acad. Sci. Paris 337 (2003), no. 9, 581–586, doi:10.1016/j.crma.2003.09.003.
  1. R. Brandl, Finite varieties and groups with Sylow pp-subgroups of low class, J. Austral. Math. Soc. Ser. A 31 (1981), no. 4, 464–469, doi:10.1017/S1446788700024265.
  1. R. Brandl, Groups with abelian Sylow subgroups, J. Aust. Math. Soc. 88 (2010), no. 1, 43–47, doi:10.1017/S1446788709000184.
  1. E. I. Khukhro and V. D. Mazurov (eds.), The Kourovka Notebook: Unsolved problems in group theory, 21st ed., September 2026 update, Problem 11.15, arXiv:1401.0300v46.