The endomorphism tower of a finite symmetric group
Victoria Gould, Ambroise Grau, Marianne Johnson, Jamie Smith · 2026
Read The endomorphism tower of a finite symmetric group by Victoria Gould, Ambroise Grau, Marianne Johnson, Jamie Smith (2026) 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-2606-24274.txt.
Introduction
Gould, Grau, Johnson, and Smith climb a tower built layer by layer from all structure-preserving maps of a structure. Start with a monoid, a system with one associative operation. Its endomorphisms, maps keeping that operation, themselves form a monoid. Iterating gives the endomorphism tower. The team first shows that for any finite monoid this tower never stabilises in finitely many steps. It grows forever.
Attention then narrows to the symmetric group, all permutations of n points. Its automorphism tower famously settles immediately, except in special cases. Taking n at least seven avoids the exceptions. About endomorphisms above the first level, nothing was known. Gould's team determines the elements of the second level in full. Their list and products are explicit, verifying the tower's behaviour there. The 2026 result separates automorphism folklore from endomorphism reality. Allowing all self-maps, not just undoable ones, changes the tower's fate entirely.
Catalog record
| CREATOR | Victoria Gould, Ambroise Grau, Marianne Johnson, Jamie Smith |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 121 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2606.24274
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
The endomorphism tower of a finite symmetric group Thanks: Throughout this work, the second author was funded by Portuguese national funds through the FCT – Fundação para a Ciência e a Tecnologia, I.P., under the scope of the projects UID/00297/2025 ( https://doi.org/10.54499/UID/00297/2025 ) and UID/PRR/00297/2025 ( https://doi.org/10.54499/UID/PRR/00297/2025 ) (Center for Mathematics and Applications – NOVA Math). Victoria Gould, Ambroise Grau, Marianne Johnson, Jamie Smith Abstract . We consider the endomorphism tower of a monoid M M , that is, the sequence of monoids End i ( M ) \en_{i}(M) where End 0 ( M ) = M \en_{0}(M)=M and for all i ≥ 1 i\geq 1 , End i ( M ) \en_{i}(M) is the monoid of all endomorphisms of End i − 1 ( M ) \en_{i-1}(M) . We show that for a finite monoid M M this sequence does not stabilise in a finite number of steps. Our focus is then on the case where M = 𝒮 n M=\mathcal{S}_{n} , the symmetric group on a finite number n n of points. It is well known that other than in exceptional cases (which are avoided by taking n ≥ 7 n\geq 7 ), the corresponding automorphism tower of 𝒮 n \mathcal{S}_{n} stabilises at the first step. In spite of the natural nature of this question, nothing was known of the endomorphism tower above the level i = 1 i=1 . We determine (for each OPEN n ≥ 7 ) n\geq 7) the elements of End 2 ( 𝒮 n ) \en_{2}(\mathcal{S}_{n}) and their multiplication and thus verify that the monoids End i ( 𝒮 n ) \en_{i}(\mathcal{S}_{n}) for i = 0 , 1 , 2 i=0,1,2 all have group of units isomorphic to 𝒮 n \mathcal{S}_{n} . We show that the same is true of End 3 ( 𝒮 n ) \en_{3}(\mathcal{S}_{n}) . 2020 Mathematics Subject Classification 20B30, 20M15, 20M20 1. Introduction The most natural examples of groups and monoids come from taking the automorphism group or endomorphism monoid of a given mathematical object. Perhaps the best known examples are those where the object is a finite set of cardinality n or an n -dimensional vector space over a division ring D , yielding as groups of automorphisms the symmetric group \mathcal{S}_{n} and the general linear group GL(n,D) , respectively, and as monoids of endomorphisms the transformation monoid \mathcal{T}_{n} and the monoid of matrices M(n,D) , respectively. Exploring the structure of such naturally occurring monoids lends itself to new areas of study within algebra. There is a substantial literature supporting this direction of study (see, for example [ 2 , 3 , 4 , 5 , 9 ] ). A classic question posed in group theory concerns sequences of automorphism groups, as we now explain. Let G be a group and let \au(G) denote the group of automorphisms of G . We define a sequence \au_{i}(G) inductively by putting \au_{0}(G)=G and \au_{i}(G)=\au(\au_{i-1}(G)) for all i\geq 1 . This can be extended to transfinite ordinals by setting \au_{\alpha}(G) , for a limit ordinal \alpha , to be the direct limit of the groups \au_{\beta}(G) , for all \beta<\alpha , using the fact that there is a natural morphism from G to the subgroup \inn(G) of \au(G) . This sequence is known as the automorphism tower of G . One can then enquire as to the nature of the groups in this tower and, in particular, ask whether there exists an ordinal \mu such that \au_{\mu}(G) is isomorphic to \au_{\nu}(G) , for all \nu>\mu . If so, then choosing the least such \mu , the sequence is said to terminate, or stabilise, in \mu steps. A classical result of Weilandt [ 13 ] shows that if G is any finite centreless group then its automorphism tower terminates in a finite number of steps [ 13 ] . The corresponding result, showing that the automorphism tower of an infinite centreless group stabilises, possibly after an infinite number of steps, was proven by Thomas [ 12 ] (see also [ 10 ] ) and extended to non-centreless groups by Hamkins [ 8 ] . It appears to be still open as to whether the automorphism tower of an arbitrary finite group terminates after a finite number of steps. This article is intended as an opener to the analogous study of the endomorphism tower of a monoid. Let M be a monoid and let \en(M) denote the monoid of semigroup endomorphisms of M . We define a sequence \en_{i}(M) inductively by putting \en_{0}(M)=M and \en_{i}(M)=\en(\en_{i-1}(M)) for all i\geq 1 . This sequence is known as the endomorphism tower of M . (Note that the question of extending to an infinite sequence is not yet even well posed, since there are not necessarily canonical morphisms from M to \en(M) to allow us to define \en_{\alpha}(M) for a limit ordinal \alpha .) One then asks, given a monoid M , does there exist some finite i such that \en_{i}(M)\cong\en_{i+1}(M) , in which case we say the sequence terminates, or stabilises, in i steps. If M is finite, this question is easily resolved in the negative (see Proposition 3.2 ); the case for infinite M is still open. We are led therefore to investigate other patterns and similar stability conditions for these towers. For example, under what conditions is there a ‘nice’ pattern in the sequence of groups \au(\en_{i}(M)) for i\geq 0 ? We remark that if G is the group of units of M , that is, the invertible elements of M , then there is a natural morphism from G to the group of units \au(M) of \en(M) , again given by conjugation. The purpose of this article is to study the endomorphism tower of a finite symmetric group \mathcal{S}_{n} . In order to avoid exceptional behaviour, we focus on the case where n\geq 7 ; it is well known that this constraint guarantees that \au(\mathcal{S}_{n}) is isomorphic to \mathcal{S}_{n} , that is, the automorphism tower of \mathcal{S}_{n} terminates at i=0 . The elements and the structure of \en_{1}(\mathcal{S}_{n}) have been determined and used by several authors, for example in the study of digraphs constructed via endomorphisms of groups (see [ 1 ] ), and for studying the translational hull of this endomorphism semigroup (see [ 7 ] ). As noted above, the group of units of \en_{1}(\mathcal{S}_{n}) , that is, \au(\mathcal{S}_{n}) , is isomorphic to \mathcal{S}_{n} . We proceed to determine the elements and the structure of \en_{2}(\mathcal{S}_{n}) ; again, the group of units, that is, \au(\en_{1}(\mathcal{S}_{n})) , is isomorphic to \mathcal{S}_{n} . Using the structure of \en_{2}(\mathcal{S}_{n}) we then show that the group of units of \en_{3}(\mathcal{S}_{n}) , that is \au(\en_{2}(\mathcal{S}_{n})) , is once more isomorphic to \mathcal{S}_{n} . Our aim is to illustrate the techniques required for the study, and to note how the complexities arise. As mentioned earlier, this article is intended as an opener to the study of endomorphism towers. We are left with a plethora of interesting and non-trivial questions, some of which we articulate in the final section. The most immediate is that arising directly from our work: is \au(\en_{i}(\mathcal{S}_{n})) isomorphic to \mathcal{S}_{n} for all i\geq 0 and n\geq 7