Normal forms of nilpotent elements in semisimple Lie algebras (Q2243176): Difference between revisions

Let \(\mathfrak {g}\) be a finite dimensional semisimple Lie algebra over an algebraically closed field of characteristic 0. Let \(f \in \mathfrak {g}\) be a nonzero nilpotent element. It leads to an \(\mathfrak {s} \mathfrak {l}_2\)-triple \(\{h,e,f\}\), which generates a subalgebra \(\mathfrak {s}\). Let \(\mathfrak {g} = \oplus_{j=-d}^d \mathfrak {g}_j\) be the eigenspace decomposition of \(\mbox{ad}(h)\) with \(\mathfrak {g}_{\pm d} \neq 0\). Let \(\widetilde{d}\) denote \(d-1\) or \(d\), depending on whether an element of the form \(f+E\), where \(0 \neq E \in \mathfrak {g}_d\), is always nilpotent. Let \(Z(\mathfrak {s})\) be the centralizer of \(\mathfrak {s}\) in the adjoint group. A reducing subalgebra for \(f\) is a semisimple subalgebra \(\mathfrak {q}\), normalized by \(\mathfrak {s}\), such that \(Z(\mathfrak {s})(\mathfrak {q} \cap \mathfrak {g}_{\widetilde{d}})\) contains a non-empty Zariski open subset. The article proves that there is a unique minimal reducing subalgebra, up to conjugacy by \(Z(\mathfrak {s})\). Consequently, it provides a canonical decomposition of \(f\) as a sum of irreducible nilpotent elements.