Back to the Amitsur--Levitzki theorem: a super version for the orthosymplectic Lie superalgebra \(\mathfrak {osp}(1, 2n)\) (Q1431321)

The classical Amitsur-Levitzki theorem states that the linear Lie algebra \(\mathfrak{gl}\left( n\right) \) satisfies the polynomial identity \[ \sum_{\sigma\in S_{2n}}\text{sgn}\left( \sigma\right) X_{\sigma\left( 1\right) }...X_{\sigma\left( 2n\right) }=0,\qquad \forall X_{\sigma\left( i\right) } \in\mathfrak{gl}\left( n\right),\tag{1} \] where \(S_{2n}\) is the symmetric group [\textit{A. S. Amitsur} and \textit{J. Levitzki}, Proc. Am. Math. Soc. 1, 449--463 (1950; Zbl 0040.01101)]. In this work, the authors analyze the existence of analogous Amitsur-Levitzki type identities for the case of Lie superalgebras \(L_{0}\oplus L_{1}\). In order to be compatible with both blocks of the \(\mathbb{Z}_{2}\)-graduation, an identity of such type must be adapted to distinguish the sign of permutations when applied to the elements of the odd and even parts of \(L\), respectively. The proposed candidate for an identity is \[ \mathcal{A}_{k}\left( X_{1},..,X_{k}\right) :=\sum_{\sigma\in S_{k} }\text{sgn}\left( \sigma\right) \widetilde{\text{sgn}\left( \sigma\right) } X_{\sigma\left( 1\right) }...X_{\sigma\left( 2n\right) },\tag{2} \] where \(\widetilde{\text{sgn}\left( \sigma\right) }\) is the supersign, defined as follows: \[ \widetilde{\text{sgn}\left( \sigma\right) }:=\left( -1\right) ^{\text{card}\left\{ \left( i,j\right) | \;X_{\sigma\left( i\right) },X_{\sigma\left( j\right) }\in L_{1},\;i>j,\;\sigma\left( i\right) <\sigma\left( j\right) \right\} }.\tag{3} \] For the linear Lie superalgebra \(\mathfrak{gl}\left( p,q\right) \) it is easy to verify that no Amitsur-Levitzki identity holds as long as \(pq\neq0\). The reason for the failure lies in the non-finite generation of its ring of invariants, which has been shown to be one of the central arguments to the Amitsur-Levitski problem [cf. \textit{B. Kostant}, J. Math. Mech. 7, 237--264 (1958; Zbl 0087.25702)]. The authors conclude that only the orthosymplectic Lie superalgebra \(\mathfrak{osp}\left( 1,2n\right) \) could satisfy an identity like \(\left( 3\right) \). The main result is that for any \(X_{1},..,X_{4n+2}\in\mathfrak{osp}\left( 1,2n\right) \), we have\newline \[ \mathcal{A}_{4n+2}\left( X_{1},..,X_{4n+2}\right) =0. \] It is further conjectured that a similar result holds for the finite-dimensional representations of the orthosymplectic algebra (the previous identity corresponding to the standard representation).