---
title: Kemer-type Theorems in PI-Theory
url: https://www.emergentmind.com/topics/kemer-type-theorems
type: topic
---

# Kemer-type Theorems in PI-Theory

Searching arXiv for the cited paper and closely related Kemer/PI-algebra work to ground the article.
Kemer-type theorems are deep structural results in PI-theory that describe affine PI-algebras through the interaction of polynomial identities, radicals, and representability phenomena. In the framework developed around the Braun–Kemer–Razmyslov theorem, the characteristic conclusion is that nil behavior in affine PI-algebras is forced into nilpotence: for an affine PI-algebra over a commutative Noetherian ring, every nil ideal is nilpotent, and over a Jacobson Noetherian base ring the Jacobson radical is nilpotent [1405.0730]. In this setting, the phrase “Kemer-type theorem” refers to results proved by combining representation theory of symmetric groups, combinatorics of alternating identities, and reductions to finite or finite-dimensional models, with the Braun–Kemer–Razmyslov theorem serving as a canonical example [1405.0730].

## 1. Canonical statements and the Braun–Kemer–Razmyslov synthesis

Let \(C\) be a commutative ring, and let \(A\) be affine over \(C\), meaning that it is generated as a \(C\)-algebra by finitely many elements \(a_1,\dots,a_\ell\), written \(A=C\{a_1,\dots,a_\ell\}\). An algebra is finite over \(C\) if it is a finitely generated \(C\)-module. Over a field \(F\), \(A\) is a PI-algebra if it satisfies some nontrivial polynomial identity in the free algebra \(F\{x_1,x_2,\dots\}\). The Capelli polynomial of degree \(2k\) is
\[
\mathrm{Cap}_k(x_1,\dots,x_k;y_1,\dots,y_k)
=
\sum_{\pi\in S_k}\mathrm{sgn}(\pi)\,x_{\pi(1)}y_1x_{\pi(2)}y_2\cdots x_{\pi(k)}y_k,
\]
and a Capelli identity of degree \(2k\) means that \(\mathrm{Cap}_k\) vanishes under all evaluations in the algebra [1405.0730].

The classical field-theoretic form of the Braun–Kemer–Razmyslov theorem states that the Jacobson radical \(\mathrm{Jac}(A)\) of any affine PI-algebra \(A\) over a field is nilpotent [1405.0730]. Within the same narrative, several precursor statements are distinguished. Razmyslov’s theorem asserts that if an affine algebra over a field satisfies a Capelli identity, then its Jacobson radical is nilpotent [1405.0730]. Kemer’s theorem supplies the missing hypothesis by proving that affine PI-algebras satisfy some Capelli identity, first in characteristic zero and then in positive characteristic [1405.0730]. Braun’s theorem extends the nilpotence phenomenon to arbitrary commutative Noetherian base rings by showing that any nil ideal of an affine PI-algebra is nilpotent [1405.0730].

Belov–Rowen organize these results into a unified statement: if \(A\) is an affine PI-algebra over a commutative Noetherian ring \(C\), then any nil ideal of \(A\) is nilpotent; in particular, if \(C\) is Jacobson, then \(\mathrm{Jac}(A)\) is nilpotent [1405.0730]. This formulation is the most general version presented there and is the sense in which the theorem is treated as a central Kemer-type result.

## 2. Meaning of “Kemer-type theorem” in PI-theory

Kemer’s theory concerns the structure and classification of PI-algebras and their \(T\)-ideals of identities. The relevant themes include \(T\)-ideal representability, reduction to finite-dimensional and basic algebras, and finite basis and codimension growth [1405.0730]. In this context, a Kemer-type theorem is not merely a statement about existence of identities; it is a structural theorem asserting that affine PI-algebras are governed by rigid constraints on radicals and semisimple quotients, together with strong consequences for \(T\)-ideals [1405.0730].

The Braun–Kemer–Razmyslov theorem is archetypal in this sense. It identifies nilpotence of the Jacobson radical, or more generally of nil ideals, as a universal structural constraint in affine PI-algebras [1405.0730]. Its proof is Kemer-type because it proceeds from an arbitrary polynomial identity to a Capelli identity, then exploits the highly structured form of Capelli identities to control radicals by combinatorial and integrality arguments [1405.0730]. The passage from “some PI” to “a Capelli identity” is mediated by sparse identities and the representation theory of \(S_n\), while the passage from Capelli identities to nilpotence uses Zubrilin’s doubly alternating module, generic integrality, and Shirshov’s height theorem [1405.0730].

A plausible implication is that the phrase “Kemer-type” functions less as a narrow theorem label than as a methodological designation. In the present setting it denotes a synthesis of representation-theoretic, combinatorial, and structural techniques that convert polynomial identities into strong algebraic consequences about radicals and representability [1405.0730].

## 3. Fundamental constructions: \(T\)-ideals, Capelli identities, and doubly alternating modules

A polynomial identity for a \(C\)-algebra \(A\) is a nonzero polynomial \(f\in C\{x_1,x_2,\dots\}\) such that \(f(a_{i_1},\dots,a_{i_m})=0\) for all evaluations in \(A\), with at least one coefficient equal to \(1\). The identities of \(A\) form a \(T\)-ideal \(\mathrm{Id}(A)\), meaning an ideal stable under all endomorphisms of the free algebra. The corresponding relatively free algebra is \(C\{x_1,x_2,\dots\}/I\) for a \(T\)-ideal \(I\) [1405.0730].

Capelli polynomials are alternating in the \(x\)-variables. More generally, a polynomial \(f(x_1,\dots,x_n;y_1,\dots,y_n;t)\) is alternating in \(x_1,\dots,x_n\) if
\[
f(\dots,x_i,\dots,x_j,\dots)+f(\dots,x_j,\dots,x_i,\dots)=0
\]
for all \(i<j\), equivalently if substituting \(x_i=x_j\) gives zero. It is doubly alternating if it is alternating separately in the \(x\)’s and in the \(y\)’s [1405.0730].

The central object in Zubrilin’s approach is the module \(M\) generated by all doubly alternating polynomials in the free algebra \(C\{x,y,t\}\), together with the double Capelli polynomial
\[
\mathrm{DCap}_n=t_1\mathrm{Cap}_n(x_1,\dots,x_n;t)\,t_2\mathrm{Cap}_n(y_1,\dots,y_n;t_3).
\]
Modulo the \(T\)-ideal generated by \(\mathrm{Cap}_{n+1}\), the image of \(M\) is denoted \(\hat M\subset C\{x,y,t\}/\mathrm{CAP}_{n+1}\) [1405.0730]. The importance of \(\hat M\) lies in the fact that it carries a module structure encoding generalized integrality relations, and these relations become the mechanism by which one annihilates products involving the obstruction ideal and Capelli-generated ideals [1405.0730].

This construction is characteristic of Kemer-type arguments: alternating and doubly alternating polynomials are not treated as isolated identities but as carriers of module-theoretic and representation-theoretic structure. That structure is then transferred, via specialization and relatively free products, to the algebra under study [1405.0730].

## 4. \(\delta\)-operators, integrality obstruction, and the nilpotence mechanism

Fix \(n\) and a polynomial \(f\) multilinear in \(x_1,\dots,x_n\). For a new noncommuting indeterminate \(z\), the operators \(\delta_{k,z}^{(x,n)}\) are defined by
\[
f((z+1)x_1,\dots,(z+1)x_n,y,t)=\sum_{k=0}^n \delta_{k,z}^{(x,n)}(f)(x_1,\dots,x_n,y,t),
\]
where \(\delta_{k,z}^{(x,n)}(f)\) is the homogeneous component of degree \(k\) in \(z\), equivalently
\[
\delta_{k,z}^{(x,n)}(f)=
\sum_{1\le i_1<\cdots<i_k\le n}
f(x_1,\dots,zx_{i_1},\dots,zx_{i_k},\dots,x_n,y,t).
\]
Analogous operators \(\delta_{k,z}^{(y,n)}\) act on the \(y\)-variables. If \(f\) is alternating in \(x_1,\dots,x_n\), then each \(\delta_{k,z}^{(x,n)}(f)\) is again alternating [1405.0730].

The crucial symmetry is that if \(f\) is doubly alternating in \(x_1,\dots,x_n\) and \(y_1,\dots,y_n\), then for any polynomial \(h\in C\{t\}\),
\[
\delta_{k,h}^{(x,n)}(f)\equiv \delta_{k,h}^{(y,n)}(f)\mod \mathrm{CAP}_{n+1}.
\]
This allows one to define commuting operators \(\xi_{k,h}\) on \(\hat M\) by
\[
\xi_{k,h}\cdot f:=\delta_{k,h}^{(x,n)}(f)=\delta_{k,h}^{(y,n)}(f).
\]
The resulting module action is the algebraic core of Zubrilin’s method [1405.0730].

For a \(C\)-algebra \(A\), Belov–Rowen introduce commuting variables \(\xi_{1,a},\dots,\xi_{n,a}\) for each \(a\in A\), form
\[
A[\xi_{n,A}]=A[\{\xi_{k,a}\mid 1\le k\le n,\ a\in A\}],
\]
and define the integrality ideal
\[
I_{n,A}=
\left\langle
a+\xi_{1,a}a^{n-1}+\cdots+\xi_{n,a}\ \middle|\ a\in A
\right\rangle
\subset A[\xi_{n,A}].
\]
The obstruction to integrality is
\[
\mathrm{Obst}_n(A)=A\cap I_{n,A}\subset A.
\]
If every element of \(A\) is \(n\)-integral over the base ring, or some central subring, then \(\mathrm{Obst}_n(A)=0\), and \(\mathrm{Obst}_{n-1}(A)\supseteq \mathrm{Obst}_n(A)\) [1405.0730].

The link between the combinatorics and radical theory is the assertion that
\[
I_{n,C\{t\}}\cdot M=0,
\]
so the generic integrality relations annihilate the module of doubly alternating polynomials modulo Capelli [1405.0730]. For a PI-algebra satisfying \(\mathrm{Cap}_{n+1}\), this leads to the central vanishing theorem
\[
\mathrm{Obst}_n(A)\cdot \mathrm{CAP}_n(A)=0,
\]
where \(\mathrm{CAP}_n(A)\) is the ideal generated by evaluations of \(\mathrm{Cap}_n\) in \(A\) [1405.0730]. Combined with two additional inclusions,
\[
\mathrm{Jac}(A)^{2q}\subseteq \mathrm{CAP}_n(A)^q
\quad\text{and}\quad
\mathrm{Jac}(A)^m\subseteq \mathrm{Obst}_n(A)
\]
for suitable integers \(q,m\), one obtains nilpotence: for a nil ideal \(N\),
\[
N^{2q+m}\subseteq \mathrm{Obst}_n(A)\cdot \mathrm{CAP}_n(A)=0.
\]
This proves that nil ideals are nilpotent in the Capelli case over a commutative Noetherian base ring [1405.0730].

## 5. Sparse identities, symmetric groups, and the derivation of Capelli identities

The second major ingredient is the derivation of Capelli identities from arbitrary polynomial identities. Let \(V_n\) denote the multilinear polynomials in \(x_1,\dots,x_n\); it is identified with the group algebra \(F[S_n]\) by
\[
\sigma \leftrightarrow x_{\sigma(1)}\cdots x_{\sigma(n)}.
\]
A polynomial
\[
g=\sum_{\sigma\in S_d}\alpha_\sigma x_{\sigma(1)}\cdots x_{\sigma(d)}
\]
is a sparse identity for \(A\) if for any monomial \(f(x_1,\dots,x_d;t)\),
\[
\sum_\sigma \alpha_\sigma f(x_{\sigma(1)},\dots,x_{\sigma(d)};t)\in \mathrm{Id}(A).
\]
Capelli polynomials are sparse identities [1405.0730].

Sparse identities have a decisive combinatorial property: they yield a lexicographic reduction on words, allowing monomials involving many long words to be rewritten as linear combinations with fewer long words. Coupled with Shirshov’s height theorem, this gives spanning and representability consequences [1405.0730]. To construct sparse identities, Belov–Rowen use Young symmetrizers, dimensions \(s_\lambda=\dim J_\lambda\) from the hook formula, codimension estimates, and Regev’s bound
\[
c_m(A)\le (d-1)^{2m}
\]
for algebras satisfying a PI of degree \(d\) [1405.0730].

In characteristic zero, rectangular partitions are used to obtain strong identities and hence sparse identities of degree bounded by
\[
\deg g\le e^2(d-1)^2
\]
up to constants, leading to an explicit Capelli bound in Theorem 3.17 [1405.0730]. In characteristic \(p>0\), wide staircases and Fayers’ criterion yield sparse identities with degree roughly
\[
d'\approx (p-1)p\frac{(u+1)u}{2}\sim p^2u^2,
\]
with \(u\) chosen in terms of \(d\), producing the positive-characteristic Capelli theorem of Theorem 4.5 [1405.0730].

These sparse identities feed into the statement that every affine algebra satisfying a sparse identity of degree \(\le d\) and generated by \(r\) elements satisfies a Capelli identity \(\mathrm{Cap}_n\) for
\[
n\ge r+d.
\]
This is the mechanism by which Kemer’s theorem is realized in the paper: arbitrary polynomial identities are first strengthened to sparse identities and then converted into Capelli identities [1405.0730].

## 6. Extension from fields to arbitrary commutative Noetherian rings

The extension from fields to arbitrary commutative Noetherian rings requires both field-theoretic reduction and Noetherian induction. In the Capelli case, the combinatorial proof of Razmyslov’s theorem works over arbitrary commutative Noetherian \(C\), with Noetherianity entering through finiteness arguments and applications of Shirshov’s height theorem [1405.0730]. This yields the theorem that if \(A\) is affine over a commutative Noetherian ring \(C\) and satisfies a Capelli identity, then every nil ideal of \(A\) is nilpotent [1405.0730].

For a general affine PI-algebra over Noetherian \(C\), Belov–Rowen prove that \(A\) satisfies a product of Capelli identities. When \(C\) is an integral domain, one tensors with its field of fractions \(F\) to obtain \(A_F\), applies the field-based Capelli theorem to \(A_F\), and deduces that a multiple \(s\mathrm{Cap}_n\) vanishes in \(A\) for some nonzero \(s\in C\) [1405.0730]. Passing to \(A/sA\), one invokes Noetherian induction and a lemma showing that appropriate products of Capelli identities produce a single Capelli identity on \(A\) [1405.0730]. For a general Noetherian ring \(C\), the nilradical is expressed as an intersection of finitely many prime ideals, and the same argument is applied to each quotient \(A/P_jA\), after which a Capelli identity on \(A\) is recovered [1405.0730].

Once the existence of a Capelli identity is established over arbitrary Noetherian \(C\), the Capelli-case nilpotence theorem implies the full Braun–Kemer–Razmyslov conclusion [1405.0730]. If \(C\) is Jacobson Noetherian, Amitsur–Procesi imply that affine algebras over \(C\) are Jacobson and that their Jacobson radicals are nil, so Braun’s theorem upgrades nil to nilpotent [1405.0730]. This is the precise point at which the stronger statement about \(\mathrm{Jac}(A)\) is obtained.

## 7. Structural consequences, examples, and mathematical significance

The central structural consequence is that for any affine PI-algebra \(A\) over a Noetherian Jacobson ring \(C\), the Jacobson radical \(\mathrm{Jac}(A)\) is nilpotent [1405.0730]. Consequently,
\[
A/\mathrm{Jac}(A)
\]
is a semiprime PI-algebra, and under further standard hypotheses it decomposes into a subdirect product of prime PI-algebras finite over their centers [1405.0730]. Even without the Jacobson hypothesis, the fact that every nil ideal is nilpotent imposes strong restrictions on the radical structure of affine PI-algebras [1405.0730].

Several special cases clarify the theorem’s range. If \(A\) is a commutative affine algebra over a field, then \(\mathrm{Jac}(A)\) is nilpotent; this is the base case \(n=1\), since \(\mathrm{Cap}_2\) forces commutativity [1405.0730]. If \(A\subseteq M_n(K)\) for a commutative \(C\)-algebra \(K\), then any nil subalgebra \(N\subseteq A\) is nilpotent of index \(\le mn\), where \(m\) is the nilpotence index of the nilradical of \(K\); if \(K\) is reduced, then any nil subalgebra must be zero [1405.0730]. If \(A\) is finite over an affine commutative subalgebra \(C\) over a field, then \(\mathrm{Jac}(A)\) is nilpotent [1405.0730].

The theorem also interacts with \(T\)-ideal theory. The paper does not re-derive finite basis or representability of \(T\)-ideals, but its use of sparse identities, symmetric group representations, Young diagrams, hook formula, and Specht modules is the same kind of machinery used in Kemer’s classification of \(T\)-ideals and in the proof of the Specht problem for PI-algebras [1405.0730]. This suggests that radical nilpotence is one manifestation of a broader principle: affine PI-algebras are accessible because their identities admit a highly rigid combinatorial organization.

Relative to the earlier literature, Razmyslov’s theorem addressed affine algebras over fields satisfying Capelli identities, while Braun’s theorem treated affine PI-algebras over arbitrary Noetherian rings by structural methods involving Azumaya algebras and central closure [1405.0730]. Belov–Rowen’s contribution is to provide a complete combinatorial proof of Razmyslov’s theorem and its Noetherian generalization, to integrate Kemer’s Capelli theorem in all characteristics, and to give a characteristic-free, self-contained exposition that unifies Razmyslov’s combinatorics, Kemer’s sparse identities, and Braun’s nilpotence theorem into a single framework [1405.0730].

In this sense, Kemer-type theorems are best understood as structural PI-theorems obtained by a characteristic blend of methods: combinatorics of words and alternating identities, representation theory of \(S_n\), reduction to finite and basic algebras, and integrality arguments over commutative Noetherian rings [1405.0730]. The Braun–Kemer–Razmyslov theorem remains a flagship instance because it converts the existence of polynomial identities into a definitive statement about radicals: in affine PI-algebras, nil ideals cannot remain merely nil; they are forced to be nilpotent [1405.0730].

Source: https://www.emergentmind.com/topics/kemer-type-theorems