Varieties of algebras with superinvolution of almost polynomial growth (Q295898): Difference between revisions

Assume \(A\) is an associative superalgebra over a fixed field of characteristic 0. Let \(A\) be equipped with a superinvolution \(*\), and denote by \(c_n^*(A)\) the sequence of the \(*\)-codimensions of \(A\). The authors study the growth of \(c_n^*(A)\). In the case of ordinary algebras (without involution as well), it is well known that the growth of \(c_n(A)\) is either polynomial or exponential (as long as \(A\) satisfies a polynomial identity). A well-known result of \textit{A. R. Kemer} [Sib. Math. J. 19, 37--48 (1978; Zbl 0411.16014)] describes the situation when the growth of \(c_n(A)\) is polynomial: The sequence \(c_n(A)\) is of polynomial growth if and only if neither the infinite dimensional Grassmann algebra \(G\) nor the algebra \(UT_2(F)\) of all upper triangular \(2\times 2\) matrices satisfy all identities of \(A\). The present paper establishes an analogous result for superalgebras with superinvolution. It is well known that \(c_n(A)\leq c_n^*(A)\leq 4^nc_n(A)\) if \(A\) is PI as an ordinary algebra. Therefore the growth of \(c_n^*(A)\) is at most exponential. The main result of the paper under review is the following. Let \(A\) be a superalgebra with a superinvolution \(*\), and assume \(\dim A<\infty\). Then \(c_n^*(A)\) is of polynomial growth if and only if neither of the following \(*\)-superalgebras satisfies all \(*\)-identities of \(A\): 1. The algebra \(F\oplus F\) with the trivial grading and with the exchange superinvolution. 2. The subalgebra \(M\) of \(UT_4(F)\) with diagonal entries \(u\), \(s\), \(s\), \(u\); entries of the second diagonal \(r\), 0, \(v\), and remaining entries zeros, with the trivial grading and the reflection superinvolution. 3. The above algebra \(M\) with the obvious nontrivial grading and with the reflection superinvolution. Moreover, the authors obtain that these three algebras are the only ones that generate varieties of almost polynomial growth (that is the corresponding variety is of exponential growth of its codimensions but every proper subvariety is of polynomial growth). Further, the authors show that in the case of finite dimensional superalegbras with superinvolution one has the dichotomy of the ordinary case, namely the codimensions either grow polynomially or exponentially.