On finite groups with exactly one noncommutator

Saveliy V. Skresanov · 2025

Read On finite groups with exactly one noncommutator by Saveliy V. Skresanov (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-2509-17587.txt.

Introduction

Saveliy V. Skresanov, writing alone, answers a group theory question posed in the Kourovka notebook in 2010. His 2025 arXiv paper asks about commutators, the elements of a group built from a standard four-factor product. MacHale's problem asks whether a finite group with more than two elements can have exactly one element that is not a commutator.

The answer turns out to be yes. Skresanov gives an infinite series of such groups, so the example is not a one-off curiosity. The smallest group in his construction is huge, with billions upon billions of elements, which hints at why the question stayed open. The paper settles the existence question in full. It also gives group theorists a new family to probe for how commutator sets can sit inside a group.

Catalog record

CREATORSaveliy V. Skresanov
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE8 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-17587

Browse the full library · About the Tower of Babel

Full text

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

On finite groups with exactly one noncommutator Saveliy V. Skresanov Abstract An element x x of a group G G is a commutator if it can be expressed in the form x = a − 1 ​ b − 1 ​ a ​ b x=a^{-1}b^{-1}ab for some a , b ∈ G a,b\in G . In 2010 MacHale posed the following problem in the Kourovka notebook: does there exist a finite group G G , with | G | > 2 |G|>2 , such that there is exactly one element of G G which is not a commutator? We answer this question in the affirmative and provide an infinite series of such groups, the smallest group in our construction having size 16609443840 16609443840 . 1 Introduction The study of commutators in finite groups has a long history, see, for instance, a survey [ 6 ] . It is well known that the derived subgroup of a finite group need not consist solely of commutators; the smallest such example having size 96 , see [ 2 ] . One can construct a perfect group with a noncommutator, see [ 4 ] and an improvement in [ 3 ] . In contrast, by the positive solution to the Ore conjecture [ 7 ] , every element of a nonabelian finite simple group is a commutator. In the 20th edition of the Kourovka notebook [ 9 ] MacHale asked the following question. Question 17.76. Does there exist a finite group G , with |G|>2 , such that there is exactly one element in G which is not a commutator? Let G be a finite group with exactly one noncommutator. Since every element outside the derived subgroup [G,G] is not a commutator, we must have |G\setminus[G,G]|\leq 1 . Hence either G is a cyclic group of order 2 , which is excluded from MacHale’s question, or G is perfect. If t is the noncommutator element of G , then t^{-1} and g^{-1}tg , g\in G , are also noncommutators. It follows that t^{2}=1 and t lies in the center Z(G) of G . Using GAP [ 1 ] , Staroletov [ 10 ] showed that there are no groups with one noncommutator among perfect groups of order at most 10^{6} , with the possible exception of some group orders which were not present in the GAP perfect groups library at the time. He also showed that there are no such examples among finite quasisimple groups, i.e. among groups G such that G/Z(G) is a nonabelian simple group (see also [ 8 , Table 1] ). We show that there is, in fact, a perfect group with exactly one noncommutator, so MacHale’s question can be answered in the affirmative. Our construction is inspired by a theorem of Isaacs [ 4 ] . He showed that if A is an abelian group and M is a nonabelian group, then the derived subgroup of the wreath product A\wr M , where M is assumed to act regularly, contains a noncommutator under some conditions (which were substantially relaxed by Guralnick in [ 3 ] ). In our construction, let M be the Mathieu group M_{11} acting transitively on 22 points; the point stabilizer will be isomorphic to the alternating group of degree 6 . We consider the wreath product H=C_{2}\wr M of the cyclic group of order 2 and M . Note that M no longer acts regularly and |H|=2^{22}\cdot|M|=2^{22}\cdot 7920 . Finally, we set G_{MacHale}=[H,H] . Theorem . The group G_{MacHale} has order 16609443840 and contains exactly one noncommutator. The proof of this theorem relies on a computation in GAP, but because of the size of the group a direct “bruteforce” computation is not feasible. In the next section we describe how one can check that G_{MacHale} has one noncommutator in reasonable time. The main result implies that there is an infinite series of groups with one noncommutator. Corollary . There are infinitely many finite groups with exactly one noncommutator. Proof. Let G be a finite nonabelian group with exactly one noncommutator t . Let T be a normal subgroup of G\times G generated by (t,t) , and set K=(G\times G)/T . Notice that |K|>|G| . We will prove that K also has exactly one noncommutator. Suppose that (t,1)T\in K is a commutator. Then there exist elements a,b\in G\times G such that (t,1)\in[a,b]T . If a=(a_{1},a_{2}) and b=(b_{1},b_{2}) , then (t,1)=([a_{1},b_{1}],[a_{2},b_{2}])T . Hence either t=[a_{1},b_{1}] , which is impossible since t is not a commutator in G , or 1=[a_{2},b_{2}]t . But that implies t=[a_{2},b_{2}] which is again a contradiction. Therefore (t,1) is not a commutator in K . Let (x,y)T\in K be an element distinct from (t,1)T . We may assume that x,y\neq t . Indeed, if x=t , then y\neq 1 , since (x,y)\not\in(t,1)T . So we may replace the representative (x,y) by (xt,yt)=(1,yt) , where 1,yt\neq t . Similar reasoning applies in the case when y=t . Now, if x,y\neq t , there exist elements a_{1},a_{2},b_{1},b_{2}\in G such that x=[a_{1},b_{1}] , y=[a_{2},b_{2}] , and hence (x,y)T=[a,b]T is a commutator in K , where a=(a_{1},a_{2}) , b=(b_{1},b_{2}) . Given a nonabelian group G with one noncommutator we constructed a larger group K with one noncommutator. By applying this procedure repeatedly to the original group G_{MacHale} , we obtain an infinite series of groups with one noncommutator. ∎ It follows immediately from the Corollary and the compactness principle that there is an infinite group of bounded exponent with exactly one noncommutator. We would like to end this section with an open question. It is not clear whether G_{MacHale} is the smallest nonabelian group with one noncommutator, so it would be interesting to determine the smallest group with such property. Question. What is the smallest nonabelian finite group with exactly one noncommutator? 2 How to check properties of G_{MacHale} We gave an explicit construction of G=G_{MacHale} in the previous section, but for GAP computations we can represent it as a permutation group of degree 44 generated by 4 elements: \displaystyle\tt G:=Group([ \displaystyle(1,2)(43,44), \displaystyle(1,2)(21,22), \displaystyle(1,39,13,43,25)(2,40,14,44,26)(3,37,15,41,27)(4,38,16,42,28) \displaystyle(5,35,23,17,31)(6,36,24,18,32)(7,33,21,19,29)(8,34,22,20,30), \displaystyle(1,23,27,11,41)(2,24,28,12,42)(3,22,26,10,44,4,21,25,9,43) \displaystyle(5,16,20,40,32,6,15,19,39,31)(7,13,17,37,29)(8,14,18,38,30)]); It can be easily checked in GAP that this group has order 16609443840 and its center has order 2 . Let t be the nonidentity element of the center. We need to verify that all elements of G besides t are commutators. Our group has 1280 conjugacy classes and it suffices to check whether a representative of each conjugacy class (besides t ) is a commutator. In the ancillary file published alongside this preprint we provide GAP code which defines a list commPairs containing 1279 pairs of permutations. Each pair a,b defines a commutator [a,b] and it can be checked that these commutators give representatives of 1279 conjugacy classes of G . The supplementary GAP file also contains a function CheckCommutators which can be used to check this property; it takes the author’s computer about 15 minutes to perform this check. The list commPairs was found by generating random commutators until all conjugacy classes (besides t ) were covered. Finally, in order to check that t is not a commutator, note that t acts on the set of conjugacy classes of G by left multiplication. Indeed, if g^{G} is a

Read the complete text