---
title: Zel'manov's Theorem on Global Nilpotence
url: https://www.emergentmind.com/topics/zel-manov-s-theorem
type: topic
---

# Zel'manov's Theorem on Global Nilpotence

Zel'manov's theorem, in the form treated in Vaughan-Lee's exposition, is the statement that if \(L\) is an \(n\)-Engel Lie algebra over a field \(F\) of characteristic \(0\), then \(L\) is globally nilpotent [2507.21677]. The theorem strengthens Kostrikin's earlier local nilpotence result by showing that the Engel identity does not merely force nilpotence on finitely generated subalgebras, but compels the entire Lie algebra to have finite lower central length. In the modern understanding presented there, the theorem is significant both for its structural conclusion and for the proof architecture, which combines ideal-theoretic reductions, multilinear commutator identities, symmetric-group representation theory, and a nilpotence criterion for \(\mathbb Z_2\)-graded Engel Lie algebras.

## 1. Statement and foundational definitions

Vaughan-Lee states the theorem in the following form [2507.21677]:

> **Theorem (Zel'manov).** If \(L\) is an \(n\)-Engel Lie algebra over a field \(F\) of characteristic \(0\), then \(L\) is **(globally) nilpotent**.

The Engel condition is the identity
\[
[x,\underbrace{y,\ldots,y}_{n\text{ times}}]=0
\]
for all \(x,y\in L\), with left-normed commutators defined by
\[
[x,y_1,\ldots,y_r]=[[\cdots[[x,y_1],y_2],\ldots],y_r].
\]
An \(n\)-Engel Lie algebra is therefore one in which repeated adjoint action by any element \(y\) annihilates every element \(x\) after \(n\) steps.

The theorem depends on a sharp distinction between two notions of nilpotence. A Lie algebra \(L\) is **locally nilpotent** if every finitely generated subalgebra is nilpotent. It is **globally nilpotent** if its lower central series terminates:
\[
\gamma_1(L)=L,\qquad \gamma_{i+1}(L)=[\gamma_i(L),L],\qquad \gamma_c(L)=0
\]
for some finite \(c\). Zel'manov's theorem is precisely the passage from the local Engel hypothesis to global nilpotence.

This distinction is not terminological only. Local nilpotence permits the nilpotence class to vary across finitely generated subalgebras, whereas global nilpotence requires a single finite class for the whole algebra. The theorem asserts that, in characteristic \(0\), the Engel identity is sufficiently rigid to force the stronger conclusion.

## 2. Kostrikin's theorem and the structural input

The immediate predecessor is Kostrikin's theorem: if \(L\) is an \(n\)-Engel Lie algebra over a field \(F\) of characteristic \(0\) or prime characteristic \(p>n\), then \(L\) is locally nilpotent [2507.21677]. Vaughan-Lee presents Zel'manov's theorem as a strengthening of this result in the characteristic-zero case.

The exposition emphasizes that Kostrikin's proof yields more than local nilpotence. Every nonzero \(n\)-Engel Lie algebra over such a field contains a nonzero abelian ideal, and one may construct a finite chain of ideals
\[
L=I_0>I_1>\cdots>I_k=0
\]
such that each quotient \(I_j/I_{j+1}\) is the sum of all abelian ideals of \(L/I_{j+1}\). Adjan-Razborov give an explicit bound on the length \(k\), although that bound is not required for the proof under discussion.

This chain is the principal structural input from Kostrikin's theorem. It organizes the Lie algebra into successive layers built from abelian ideal structure, and the proof of Zel'manov's theorem proceeds by forcing nilpotence up this chain one level at a time.

The field hypotheses matter. The theorem proved in the paper is stated only for characteristic \(0\), even though auxiliary results used in the proof also hold in characteristic \(p>n\), notably Higgins's theorem on solvable Engel Lie algebras. A common conflation is to identify Zel'manov's theorem with Kostrikin's broader characteristic-\(0\) or characteristic-\(p>n\) local nilpotence statement; Vaughan-Lee's formulation separates them sharply.

## 3. Fully invariant ideal chains and the base quotient

The first major reduction is to show, by reverse induction on the Kostrikin chain, that each ideal \(I_j\) is fully invariant [2507.21677]. The argument starts from the assumption that \(I_{j+1}\) is fully invariant and considers
\[
M=L/I_{j+1}, \qquad I=I_j/I_{j+1}.
\]
Here \(M\) is relatively free, and \(I\) is the sum of all abelian ideals of \(M\). To prove that \(I\) is fully invariant, the proof checks that if \(a\) lies in an abelian ideal and \(\theta\) is an endomorphism of \(M\), then the ideal generated by \(a\theta\) is again abelian.

The crucial identity used in this step is
\[
[a,x_{r+1},x_{r+2},\ldots,x_{r+m},a]=0
\]
as an identical relation in \(M\). After applying an endomorphism \(\varphi\), this becomes
\[
[a\theta,a_1,a_2,\ldots,a_m,a\theta]
=
[a,x_{r+1},x_{r+2},\ldots,x_{r+m},a]\varphi
=
0.
\]
This establishes the required stability under endomorphisms.

The base step is the quotient \(M=L/I_1\). Since \(M\) is relatively free, \(n\)-Engel, and a sum of abelian ideals, every free generator \(x\) lies in an abelian ideal; this is shown by using an endomorphism that fixes \(x\) and kills the other generators. The consequence is the \(2\)-Engel identity
\[
[y,x,x]=0.
\]
In characteristic \(0\), the paper invokes the classical implication
\[
[y,x,x]=0 \quad \Longrightarrow \quad [x,y,z]=0.
\]
Hence \(M\) is nilpotent of class \(2\). This establishes the first nontrivial nilpotent quotient in the chain.

## 4. The inductive step from quotient nilpotence to global nilpotence

The central inductive step starts with
\[
M=L/I_{j+1}, \qquad I=I_j/I_{j+1},
\]
and assumes that \(M/I\) is nilpotent [2507.21677]. The objective is to show that \(M\) itself is nilpotent. Because \(M/I\) is nilpotent, some commutator of bounded weight lies in \(I\):
\[
[x_1,x_2,\ldots,x_m]\in I.
\]
Since \(I\) is the sum of abelian ideals, that commutator can be written as
\[
[x_1,x_2,\ldots,x_m]=a_1+\cdots+a_{k-1},
\]
where each \(a_i\) lies in an abelian ideal.

From this decomposition the proof derives a family of identical relations with \(k\) copies of \([x_1,\ldots,x_m]\). After substituting sums of new generators and collecting multilinear terms, one obtains a symmetric multilinear identity of the form
\[
\sum [[x_{(1\sigma_1,1)},\ldots,x_{(1\sigma_m,m)}],\ldots, [x_{(k\sigma_1,1)},\ldots,x_{(k\sigma_m,m)}]]=0,
\]
summed over permutations \(\sigma_1,\ldots,\sigma_m\in \mathrm{Sym}(k)\).

The target is then a much larger identity,
\[
\bigl[\,[x_{(1,1)},\ldots,x_{(1,K)}],\,[x_{(2,1)},\ldots,x_{(2,K)}],\ldots, [x_{(N,1)},\ldots,x_{(N,K)}]\,\bigr]=0,
\]
for a suitable \(N\), where
\[
N=(Tk)^{2^K}.
\]
Once this identity is obtained, it implies that \(M\) is solvable. Higgins's theorem is then invoked to conclude that \(M\) is nilpotent.

This step is the logical core of the theorem. The argument does not merely propagate nilpotence by a simple extension principle; it manufactures a strong multilinear identity from the abelian-ideal decomposition and then turns that identity into solvability and finally nilpotence.

## 5. Symmetric-group representation theory in the proof

The passage from the smaller multilinear identity to the larger one is accomplished באמצעות the representation theory of \(\mathrm{Sym}(N)\) over characteristic \(0\) [2507.21677]. The proof decomposes the identity element into primitive idempotents associated with Young tableaux and studies the effect of the corresponding symmetrizations and antisymmetrizations on the commutator expressions.

A key combinatorial fact is that every Young tableau on \(N\) letters has either a first row or a first column of length at least \(N^{1/2}\). This row-or-column dichotomy is exploited iteratively. Applying a primitive idempotent reduces the problem to symmetrizing or antisymmetrizing over subsets of size \(N^{1/2}\), then \(N^{1/4}\), and so on. Repetition of this reduction eventually leads to a nested expression involving \(K\) permutations.

What is notable here is not only the presence of symmetric-group methods, but the role they play: they are used to transfer an identity with limited symmetry into one with sufficiently large combinatorial support to imply solvability. This is one of the striking features emphasized in Vaughan-Lee's account. The proof thereby places representation-theoretic combinatorics directly inside the structure theory of Engel Lie algebras.

## 6. The \(\mathbb Z_2\)-graded lemma and the final reduction

A central intermediate result is the lemma on \(\mathbb Z_2\)-graded Engel Lie algebras [2507.21677]:

> If \(L=L_0\oplus L_1\) is an \(n\)-Engel Lie algebra with a \(\mathbb Z_2\)-grading and \(L_0\) is nilpotent of class at most \(m-1\), then \(L\) is nilpotent of class at most
> \[
> \frac{n^{(n-1)(m-1)+1+m}-1}{n-1}.
> \]

The proof uses the derived series and the Engel identity to show that sufficiently long commutators with many entries from \(L_0\) vanish. The specific vanishing criterion highlighted in the paper is
\[
[L_1,\underbrace{L_0,\ldots,L_0}_{k}]=0
\quad \text{if } k>(n-1)(m-1).
\]
Proposition 4.6 of Kostrikin's *Around Burnside* is used to rewrite long commutators as linear combinations of shorter ones. If \(k>(n-1)(m-1)\), one of the resulting subcommutators has weight at least \(m\), and therefore vanishes because \(L_0\) has nilpotence class \(m-1\). This yields solvability, after which Higgins's theorem gives nilpotence with the explicit bound above.

The final reduction constructs a Lie subring \(L\) generated by \(x_1,\ldots,x_K\), equips it with various \(\mathbb Z_2\)-gradings, and defines the ideal \(J\) generated by \(m\)-fold commutators in the even part \(C_0\). Since \(L/J\) satisfies the graded lemma, the argument deduces
\[
[x_1,x_2,\ldots,x_K]\in J.
\]
Accordingly, \([x_1,\ldots,x_K]\) becomes a linear combination of terms of the form
\[
[[c_1,c_2,\ldots,c_m],a_1,\ldots,a_q],
\]
with \(c_i\in C_0\). Substituting these decompositions into the large symmetric identity and using the combinatorics of repeated indices reduces the problem to proving that a certain sum vanishes whenever one index repeats at least \(k\) times. A transposition argument then shows that the expression is symmetric in appropriate entries, so the alternating-symmetrized identity obtained from the Kostrikin relation forces the sum to be zero. This completes the derivation of the large identity and hence the proof of nilpotence.

## 7. Significance, scope, and interpretation

The significance of Zel'manov's theorem lies in the conversion of a local identity condition into a global structural conclusion [2507.21677]. The Engel identity is a polynomial identity involving repeated commutation by a single element, but the theorem shows that, over characteristic \(0\), this condition forces the entire Lie algebra to be nilpotent. Nilpotence is a highly rigid structural property, and its derivation from an Engel identity is accordingly nontrivial.

Historically and conceptually, the theorem is described as a major part of the Lie-algebraic route to the restricted Burnside problem. It strengthens Kostrikin's theorem, depends on Kostrikin's structural decomposition into abelian-ideal layers, and introduces deep use of symmetric-group representation theory into the analysis of Lie identities. The proof also exhibits a recurrent theme in modern structural algebra: a local or pointwise identity acquires global force only after one builds enough invariant structure to control extensions.

A common misunderstanding is to treat the theorem as merely another formulation of local nilpotence for Engel algebras. Vaughan-Lee's presentation shows that the difference between local nilpotence and global nilpotence is exactly where the hard work lies. A plausible implication is that the theorem's depth resides not in recognizing nilpotent behavior on finite pieces—Kostrikin had already achieved that—but in deriving a uniform global constraint from a chain of abelian-ideal quotients, multilinear commutator identities, and representation-theoretic symmetrization.

In this form, Zel'manov's theorem stands as a landmark result showing that the Engel condition is not a weak formal identity. In characteristic \(0\), it is a mechanism that forces finite-depth nilpotent structure throughout the Lie algebra.

Source: https://www.emergentmind.com/topics/zel-manov-s-theorem