Limit T-subspaces and the central polynomials in \(n\) variables of the Grassmann algebra. (Q1952134)

Let \(F\langle X\rangle\) be the free unitary associative algebra over a field \(F\) on the countable set of free generators \(X=\{x_1,x_2,\dots\}\). By analogy with T-ideals, a vector subspace \(V\) of \(F\langle X\rangle\) is called a T-space if it is closed under all endomorphisms of \(F\langle X\rangle\). T-spaces which are not finitely generated as T-spaces have played the key role in the construction of counterexamples to the Specht problem in positive characteristic. As in the case of T-ideals, the T-space \(V\) is `limit' if every T-space \(W\) which properly contains \(V\) is finitely generated as a T-space but \(V\) itself is not. Although the existence of limit T-ideals and T-spaces over a field of positive characteristic follows from the Zorn lemma, no examples of limit T-ideals are known. Recently \textit{A. P. Brandão} jun., \textit{P. Koshlukov, A. Krasilnikov} and \textit{É. Alves da Silva} [Isr. J. Math. 179, 127-144 (2010; Zbl 1207.16021)] have established that over an infinite field \(F\) of characteristic \(p>2\) the T-space \(C(G)\) of the central polynomials of the infinite dimensional Grassmann algebra \(G\) is a limit T-space. They conjectured that this is the only limit T-space in \(F\langle X\rangle\). In the paper under review the authors construct infinitely many limit T-spaces \(R_k\subset F\langle X\rangle\), \(k\geq 1\), when \(F\) is an infinite field of characteristic \(p>2\). The limit T-space \(R_k\) is generated by the central polynomials in \(2k\) variables of the Grassmann algebra \(G\) and the T-ideal generated by the commutator \([[x_1,x_2],x_3]\) and the product of \(k\) commutators \([x_1,x_2]\cdots[x_{2k-1},x_{2k}]\). As a by-product of their approach the authors establish that the T-subspace of \(F\langle x_1,\dots,x_n\rangle\) of the central polynomials in \(n\) variables of \(G\) is not finitely generated for even \(n\) and is finitely generated for odd \(n>3\). It is still an open problem what happens for \(n=3\).