\(\mathbb{Z}\)-graded polynomial identities of the Grassmann algebra (Q2656648)

The Grassmann algebra \(E\) of an infinite-dimensional vector space \(L\) plays a central role in the theory of PI-algebras. Its importance has been evidenced by many authors, in particular in a celebrated theorem of \textit{A. R. Kemer} [Math. USSR, Izv. 25, 359--374 (1985; Zbl 0586.16010); translation from Izv. Akad. Nauk SSSR, Ser. Mat. 48, No. 5, 1042--1059 (1984)], where he showed that over a field of characteristic zero, any associative algebra satisfies the same identities of the Grassmann envelope of some finite dimensional associative superalgebra. The above mentioned result is a key step in the positive solution of the Specht problem (also given by Kemer). He proved that over a field of characteristic zero any T-ideal is finitely generated as a T-ideal. It is worth mentioning that over infinite fields of positive characteristic it is no longer valid, and there are known examples of non finitely generated T-ideals. Despite the above result, it is usually a hard task to find a finite set of generators of a given T-ideal. In light of that, recently there has been intensive research on weaker types of identities, such as graded identities and identities with involution. The paper under review studies the graded polynomial identities of \(E\) when endowed with some gradings by the group \(\mathbb{Z}\) such that the elements of a given basis \(B=\{e_1, e_2, \dots, \}\) of \(L\) are homogeneous (the so called homogeneous gradings). The novelty in this case is that the grading group is infinity, while previous works dealing with gradings and graded identities of the Grassmann algebra considered only cases in which the grading group is finite. The canonical \(\mathbb{Z}\)-grading on \(E\) is the one induced by the length of the monomials. One can easily see that by passing to the quotient grading modulo \(2\mathbb{Z}\), it induces the canonical \(\mathbb{Z}_2\)-grading of \(E\). In the paper under review the authors introduce three families of homogeneous \(\mathbb{Z}\)-gradings of \(E\), denoted by \(E^k\), \(E^{k^*}\) and \(E^{\infty}\), defined in a similar way as the homogeneous \(\mathbb{Z}_2\)-gradings of \(E\) that appeared in the classification of homogeneous \(\mathbb{Z}_2\)-gradings (\(E_k\), \(E_{k^*}\) and \(E_{\infty}\)), given in [\textit{O. M. Di Vincenzo} and \textit{V. Ribeiro Tomaz da Silva}, Linear Algebra Appl. 431, No. 1--2, 56--72 (2009; Zbl 1225.16009)]. It turns out that the latter ones are obtained by the former ones by passing to the quotient grading modulo \(2\mathbb{Z}\). In fact, the authors present a general procedure for producing homogeneous \(G\)-gradings on \(E\) (where \(G\) is an arbitrary abelian group), and with these kind of examples, they exhibit non-isomorphic \(\mathbb{Z}\)-gradings on \(E\) satisfying the same \(\mathbb{Z}\)-graded identities. For the three families of gradings defined above, the authors describe a generating set for their \(T_{\mathbb{Z}}\)-ideal of \(\mathbb{Z}\)-graded identities. For the cases \(E^{k^{*}}\) and \(E^{\infty}\) their results hold for infinite fields of characteristic \(p>2\). For \(E^k\) a proof is presented only in the case of charateristic zero. As in the case of \(\mathbb{Z}_2\)-graded identities, the grading \(E^k\) is quite more complicated than the other ones.