Norms and Cayley-Hamilton algebras (Q2039137)

Cayley-Hamilton algebras (CH-algebras) were introduced by \textit{C. Procesi} [J. Algebra 107, 63--74 (1987; Zbl 0618.16014)] in order to settle an important embedding problem in the theory of PI algebras. This was developed in characteristic 0; one defines a CH-algebra as an algebra with trace that satisfies the \(n\)-th Cayley-Hamilton polynomial, and moreover \(tr(1)=n\). In the paper under review another approach is chosen, and it is used to define CH-algebras in positive characteristic (in fact in characteristic-free way). Instead of the traces one considers determinants. Namely, assume \(R\) is an algebra over a commutative ring \(A\), and \(n\) is a fixed positive integer. An \(n\)-norm \(N\) is a multiplicative polynomial function \(N\colon R\to A\) which is homogeneous of degree \(n\). Then the notion of a CH-algebra becomes as follows. Let \(R\) be equipped with a norm as above, such that for every commutative \(A\)-algebra \(B\), the elements \(a\in B\otimes_AR\) satisfy their characteristic polynomial \(N(t-a)=0\). \par The author relates the theory of CH-algebras to \(n\)-dimensional representations of the algebra \(R\), and to the invariant theory of matrices. The free CH-algebra is constructed in Section 2, and also the first and second fundamental theorem for matrix invariants are given in characteristic-free way. Then the author proceeds to describing the prime and the simple CH-algebras. Furthermore, the prime and the semiprime T-ideals in a free CH-algebra are classified (see the last section of the paper). The paper is expertly written by one of the founders of the theory. It will be very useful for people working with invariants, polynomial identities, combinatorial ring theory, and related topics.