Infinitesimally PI radical algebras (Q5939292)

One of the famous problems in ring theory is the Kurosh-Levitzki problem whether a finitely generated nil algebra is nilpotent. The examples of Golod and Shafarevich show that this is false in general. On the other hand, the Kaplansky theorem gives an affirmative solution to the Kurosh-Levitzki problem in the case when the algebra satisfies a polynomial identity. In the paper under review the author generalizes the Kaplansky theorem to a large class of algebras. His first main result is the following. Let \(A\) be an associative algebra generated by a finite set \(X\) and let \(L\) be the Lie subalgebra of \(A\) generated by \(X\). If every commutator in \(X\) is nilpotent in \(A\) and \(L\) satisfies a non-trivial Lie polynomial identity, then \(A\) is nilpotent. Then the author extends the notion of infinitesimally PI groups given by Zelmanov to associative algebras: If \(\text{gr}(A)=\bigoplus_{i\geq 1}A^i/A^{i+1}=\bigoplus_{i\geq 1}A_i\) is the associated graded algebra of \(A\), then \(A\) is called infinitesimally PI if the Lie subalgebra \({\mathcal L}(A)\) of \(\text{gr}(A)\) generated by \(A_1\) satisfies a polynomial identity. The second main result is: If \(A\) is a radical algebra which is finitely generated as an \(A\)-module, infinitesimally PI and every Lie commutator in a basis \(X\) of \(A_1\) is nilpotent in \(\text{gr}(A)\), then \(A\) is nilpotent. In particular, if a finitely generated nil algebra is infinitesimally PI, then it is nilpotent. The proofs of these results essentially depend on the results of Zelmanov on Engel Lie algebras. The first application of the above results concerns the Kaplansky problem and its reiteration by Passman: If \(G\) is a finitely generated group and \(F\) is a field of positive characteristic \(p\), does the coincidence of the augmentation ideal and the Jacobson radical of the group algebra \(F[G]\) imply that \(G\) is a finite \(p\)-group? Zelmanov's positive solution of the restricted Burnside problem and arguments due to Lichtman show that this problem has a positive solution for groups of finite exponent. In the paper under review the author answers affirmatively the above problem for infinitesimally PI groups. He also considers the Kaplansky problem for the restricted enveloping algebra \(u(L)\) of a finitely generated restricted Lie algebra \(L\) over a field of characteristic \(p\): If the augmentation ideal of \(u(L)\) coincides with the Jacobson radical, does it follow that \(L\) is finite dimensional with nilpotent \(p\)-map? Again, the author proves that this holds for infinitesimally PI Lie algebras.