A universal enveloping algebra of Malcev superalgebras (Q719800)

The classical Poincaré-Birkhoff-Witt Theorem ensures that any Lie algebra is a Lie subalgebra of \(A^{(-)}\) for some associative algebra \(A\), the product of \(A^{(-)}\) being the commutator \([x,y] = xy - yx\) of \(A\). Starting with an alternative algebra \(A\) this way we obtain a Malcev algebra \(A^{(-)}\). It remains an open problem whether any Malcev algebra is a subalgebra of \(A^{(-)}\) for some alternative algebra \(A\), see \textit{E. N.~Kuzmin} [Algebra Logika 10, 3--22 (1971; Zbl 0244.17019)]. An algebra having such a representation is called \textit{special}. \textit{J. M.~Pérez Izquierdo} and \textit{I. P.~Shestakov} [J. Algebra 272, No. 2, 379--393 (2004; Zbl 1077.17027)] have found a partial solution and have defined a universal enveloping algebra for Malcev algebras. Over commutative and associative rings of characteristic prime to \(6\) any free module with a Malcev algebra structure is a Malcev subalgebra of \[ N_{alt}(A) = \{ a \in A \mid (a,y,z) = - (y,a,z) = (y,z,a) \quad \forall_{y,z \in A}\} \] for some algebra \(A\), where \((x,y,z) = (xy)z - x(yz)\) denotes the associator. However, in general \(A \neq N_{alt}(A)\). So \(A\) is not alternative, and the problem about the speciality of Malcev algebras remains open. In the present paper the author extends the results of J. M.~Pérez Izquierdo and I. P.~Shestakov to superalgebras in the following way. She defines the generalized alternative nucleus \(N_{\text{alt}}(A)\) of a superalgebra \(A\) as \[ \text{span}\, \langle a \in A_{\bar{a}} \mid (a,y,z) = - (-1)^{\bar{a}\bar{y}} (y,a,z) = (-1)^{\bar{a}(\bar{y}+\bar{z})}(y,z,a) \quad \forall{y \in A_{\bar{y}}, z \in A_{\bar{z}}} \rangle. \] This space is proved to be closed under the super-commutator \([a,b] = ab - (-1)^{\bar{a}\bar{b}}ba\) and it satisfies the super-Malcev identity which one obtains as superization of Malcev's formula \[ [[x,y],[z,t]]=[[[x,y],z],t]+[[[y,z],t],x]+[[[z,t],x],y]+[[[t,x],y],z], \] (see \textit{A. A.~Sagle} [Trans. Am. Math. Soc. 101, 426--458 (1961; Zbl 0101.02302)]) and the super-anticommutativity \([x,y] = -(-1)^{\bar{x}\bar{y}}[y,x]\). So \(N_{\text{alt}}(A)\) is a Malcev superalgebra. For any Malcev super-algebra \(M\), which is a free module over a commutative and associative base ring of characteristic prime to \(6\), the author considers the superalgebra \(U(M)\) obtained as the quotient of the nonassociative \(\mathbb Z\)-tensor algebra \(\tilde{T}(M)\) by the submodule \(I\) spanned by the elements \[ ab - (-1)^{\bar{a}\bar{b}}ba - [a,b],\, (a,y,z) + (-1)^{\bar{a}\bar{x}}(y,a,z),\, (a,y,z) - (-1)^{\bar{a}(\bar{y}+\bar{z})}(y,z,a) \] for homogeneous \(a,b \in M\) and \(y,z \in \tilde{T}(M)\). The main result in the paper (Theorem 4.3) establishes the existence of a Poincaré-Birkhoff-Witt type basis for \(U(M)\). Namely, given a basis \(\{a_i \in M_{\bar{a}_{i}}\mid i\in \Lambda\}\) of \(M\) indexed by a totally ordered set \(\Lambda = \Lambda_{\bar{0}} \cup \Lambda_{\bar{1}}\), with the restriction that \(i_p < i_q\) whenever \(i_p\in \Lambda_{0}\) and \(i_q \in \Lambda_1\), the elements \( \iota(a_{i_1})(\cdots(\iota(a_{i_{n-1}})\iota(a_{i_n}))) \) with \( i_1\leq \cdots \leq i_n, n\geq 0 \) and \( i_p <i_{p+1} \) if \( a_{i_p} \in M_{\bar{1}} \) where \(\iota(a) := a + I \in U(M) = \tilde{T}(M)/I\) form a basis of \(U(M)\). In particular, \(M\) is isomorphic to the Malcev subalgebra \(\iota(M)\) of \(N_{\text{alt}}\,\iota(U(M))\).