On Zel’manov’s global nilpotence theorem for Engel Lie algebras

Michael Vaughan-Lee · 2025

Read On Zel’manov’s global nilpotence theorem for Engel Lie algebras by Michael Vaughan-Lee (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-2507-21677.txt.

Introduction

Michael Vaughan-Lee wrote a 2025 paper presenting a proof of Zel'manov's global nilpotence theorem for Engel Lie algebras. A Lie algebra is Engel when every element acts nilpotently by bracketing. Kostrikin proved such algebras are locally nilpotent in favorable characteristics. Zel'manov's stronger theorem says an n-Engel algebra over a field of characteristic zero is nilpotent outright.

The original proof contains striking, original ideas that are famously hard to follow. Vaughan-Lee wrote the note to understand the argument himself and to make it usable by others. He works through the proof in full, clarifying its moves. The theorem sits deep in the theory behind the restricted Burnside problem. An accessible account of it serves algebraists who need the result without years of decoding.

Catalog record

CREATORMichael Vaughan-Lee
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE34 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-21677

Browse the full library · About the Tower of Babel

Full text

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

On Zel’manov’s global nilpotence theorem for Engel Lie algebras Michael Vaughan-Lee Abstract I give a proof of Zel’manov’s theorem that if L L is an n n -Engel Lie algebra over a field F F of characteristic zero, then L L is (globally) nilpotent. This is a very important result which extends Kostrikin’s theorem that L L is locally nilpotent if the characteristic of F F is zero or some prime p > n p>n . Zel’manov’s proof contains some striking original ideas, and I wrote this note in an effort to understand his arguments. I hope that my efforts will be of use to other mathematicians in understanding this remarkable theorem. I am grateful to Christian d’Elbée for a number of helpful comments on earlier drafts of this note. 1 Introduction Efim Zel’manov [ 7 ] proved that if L is an n -Engel Lie algebra over a field F of characteristic zero, then L is (globally) nilpotent. This result extends Kostrikin’s theorem that if F has characteristic zero or prime characteristic p>n , then L is locally nilpotent (see [ 4 ] , [ 5 ] ). Kostrikin gives an expanded version of Zel’manov’s proof in his book Around Burnside [ 5 ] , but I feel there is room for another version of the proof of this remarkable result. Accordingly I present here a proof of Zel’manov’s theorem which actually closely follows his original proof, though it also draws on Kostrikin’s presentation of the proof in [ 5 ] . However my proof does differ from the proofs in [ 7 ] and [ 5 ] in one key point. The starting point for Zel’manov’s proof is Kostrikin’s theorem ( [ 4 ] , [ 5 ] ) that every (non-zero) n -Engel Lie algebra over a field F of characteristic zero or prime characteristic p>n contains a non-zero abelian ideal. Zel’manov uses this to show that every n -Engel Lie algebra of characteristic zero is the union of an ascending chain of ideals defined as follows. He sets I_{0}=\{0\} , and proceeds by transfinite induction. If \alpha is a limit ordinal he sets I_{\alpha}=\cup_{\nu<\alpha}I_{\nu} , and if \alpha is a successor ordinal and I_{\alpha-1} is defined he sets I_{\alpha} to be the inverse image in L of the sum of all abelian ideals in L/I_{\alpha-1} . However Kostrikin’s proof that every non-zero n -Engel Lie algebra over a field F of characteristic zero or prime characteristic p>n contains a non-zero abelian ideal actually shows that L has a finite chain of ideals L=I_{0}>I_{1}>\ldots>I_{k-1}>I_{k}=\{0\} with the property that I_{j}/I_{j+1} is the sum of all abelian ideals of L/I_{j+1} for j=0,1,\ldots,k-1 . Adjan and Razborov [ 1 ] show that the length k of this chain can be bounded by N(n,10).6^{n+12} where N(n,r) is defined by N(n,4)=6,\;N(n,r+1)=F(n,r+1,N^{2}(n,r).3^{(n+6)/2}), and where F is defined by F(n,r,0)=1,\;F(n,r,i+1)=n.r^{3F(n,r,i)}. So we let L be the free n -Engel Lie algebra of countably infinite rank over a field F of characteristic zero, and we let L=I_{0}>I_{1}>\ldots>I_{k-1}>I_{k}=\{0\} be a finite chain of ideals as described above. (The bound on k does not concern us.) We prove by reverse induction on j that the ideal I_{j} is fully invariant. Clearly the ideal I_{k} is fully invariant. So suppose that I_{j+1} is fully invariant, and consider the ideal I_{j}/I_{j+1} in L/I_{j+1} . Since I_{j+1} is fully invariant, L/I_{j+1} is relatively free, and any relation which holds in L/I_{j+1} is actually an identical relation. Let M=L/I_{j+1} and let I=I_{j}/I_{j+1} . So I is the sum of all abelian ideals of the relatively free Lie algebra M . We need to show that I is a fully invariant ideal of M . To this end it is sufficient to show that if \theta is an endomorphism of M and if a lies in an abelian ideal of M then the ideal of M generated by a\theta is abelian. We let the free generators of M be x_{1},x_{2},\ldots and we let a lie in the subring of M generated by x_{1},x_{2},\ldots,x_{r} . Since the ideal generated by a is abelian [a,x_{r+1},x_{r+2},\ldots,x_{r+m},a]=0 is an identical relation in M for all m>0 . Now let a_{1},a_{2},\ldots,a_{m} be arbitrary elements of M and let \varphi be an endomorphism of M such that x_{i}\varphi=x_{i}\theta for i=1,2,\ldots,r and such that x_{r+i}\varphi=a_{i} for i=1,2,\ldots,m . Then [a\theta,a_{1},a_{2},\ldots,a_{m},a\theta]=[a,x_{r+1},x_{r+2},\ldots,x_{r+m},a]\varphi=0, which shows that the ideal of M generated by a\theta is abelian, as claimed. We show by induction on the length of this series that L is nilpotent. Our base inductive step is to show that L/I_{1} is nilpotent using the fact that L/I_{1} is a relatively free n -Engel Lie algebra which is a sum of abelian ideals. For the general inductive step we suppose that L/I_{j} is nilpotent for some j with 1\leq j<k , and we prove that this implies that L/I_{j+1} is nilpotent. Our base inductive step is very easy. Let M=L/I_{1} . Then M is a relatively free n -Engel Lie algebra over F , and M is sum of abelian ideals. If x is a free generator of M then we can write x=a_{1}+a_{2}+\ldots+a_{k} for some k , where the elements a_{1},a_{2},\ldots,a_{k} all lie in abelian ideals of M . Since M is relatively free, this relation is an identical relation. Let \theta be the endomorphism of M which maps x to x , and maps all other free generators of M to zero. Then x=x\theta=a_{1}\theta+a_{2}\theta+\ldots+a_{k}\theta. The elements a_{i}\theta must all be scalar multiples of x , and as we showed above they must all lie in abelian ideals of M . So x lies in an abelian ideal of M , which implies that M satisfies the 2-Engel identity [y,x,x]=0 . It is well known that the 2-Engel identity in characteristic zero implies the identity [x,y,z]=0 . (See Example 2.7 from the introduction to [ 5 ] , or Theorem 3.1.1 of [ 6 ] .) So M is nilpotent of class 2. For the general inductive step we need some of the representation theory of the symmetric group. 2 Representation theory of the symmetric group We let N be a positive integer, and we consider the group ring \mathbb{Q} Sym (N) of the symmetric group on N letters, where \mathbb{Q} is the rational field. The identity element in \mathbb{Q} Sym (N) is a sum of primitive idempotents, and these are described in James and Kerber [ 3 ] : they correspond to Young tableaux . For each partition (m_{1},m_{2},\ldots,m_{s}) of N with m_{1}\geq m_{2}\geq\ldots\geq m_{s} we associate a Young diagram , which is an array of N boxes arranged in s rows, with m_{i} boxes in the i -th row. The boxes are arranged so that the j -th column of the array consists of the j -th boxes out of the rows which have length j or more. For example, if N=5 there are seven possible Young diagrams. We obtain a Young tableau from a Young diagram by filling in the N boxes with 1,2,\ldots,N in some order. We then let H be the subgroup of Sym (N) which permutes the entries within each row of the tableau, and we let V be the subgroup of Sym (N) which permutes

Read the complete text