Nondegenerate graded Lie algebras with degenerate transitive subalgebras (Q844421)

Let \(L=\bigoplus_{\in \mathbb{Z}}L_i \) be a graded Lie algebra over an algebraically closed field of characteristic \(p\). \(L\) is called irreducible if \(L_{-1}\) is an irreducible \(L_0\)-module, \(L\) is called transitive if for any \(j\geq 0\) and any \(0\neq x\in L_j\) we have that \([x, L_{-1}]\neq 0\). If also \(L_{-i}=[L_{-1}, L_{-i+1}]\) for all \(i>1\), \(L\) is called a transitive irreducible Lie algebra. Such an algebra is called degenerate if \([[L_{-1}, L_1], L_1]=0\) (otherwise \(L\) is called nondegenerate). A graded Lie subalgebra \(L'\) of \(L\) is called transitive if \(L_{-1}'=L_{-1}\). We say that a Lie algebra is classical and reductive if it is a finite sum of commuting ideals which are classical Lie algebras and at most a one-dimensional center. The Recognition Theorem is an essential ingredient in the classification of the finite-dimensional simple Lie algebras over algebraically closed fields of characteristic \(p\). It classifies finite-dimensional transitive irreducible graded Lie algebras with classical reductive \(L_0\)-component and it is extensively used in recognizing certain graded Lie algebras from their null components. Its first version was due to V. Kac in 1970 for \(p>5\) under the additional assumption of \(L_{-1}\) being a restricted \(L_0\)-module. A complete proof of such theorem for \(p>3\) can be found in the monograph [``The Recognition Theorem for graded Lie algebras in prime characteristic'', Mem. Am. Math. Soc. 920, 145 p. (2009; Zbl 1167.17004)] by \textit{G. Benkart, T. Gregory} and \textit{A. Premet}. In the paper under review the authors work in characteristic \(p>2\) and show that finite-dimensional nondegenerate transitive irreducible graded Lie algebras \(L=\oplus_{i\in \mathbb{Z}} L_i\) over algebraically closed fields of characteristic \(p > 2\) with classical reductive \(L_0\)-component cannot have degenerate transitive subalgebras \(L'\) with component \(L'_1\) of dimension greater than 1, since they show that under such hypothesis \(\text{dim}\;L_1'>1\) implies that \(L\) is infinite-dimensional. Their proof follows several steps, supposing that such finite-dimensional \(L\) can occur and, after closely studying the form of \(L\), getting a contradiction.