A counterexample to a problem on commuting matrices (Q1059135)

Let \({\mathfrak g}\) be a complex semisimple Lie algebra and \(C=\{(X,Y)\in {\mathfrak g}\times {\mathfrak g}:\) \([X,Y]=0\}\) be its commuting variety [see \textit{R. W. Richardson}, Compos. Math. 38, 311-327 (1979; Zbl 0409.17006)]. One has dim(\({\mathfrak g}_ X\cap {\mathfrak g}_ Y)\geq rank {\mathfrak g}\) for any (X,Y)\(\in C\). Let \(X\in {\mathfrak g}\). It was the question of M. Kashiwara if there exists \(Y\in {\mathfrak g}\) such that (X,Y)\(\in C\) and dim(\({\mathfrak g}_ X\cap {\mathfrak g}_ Y)=rank {\mathfrak g}\); this problem is connected with the study of the holonomic system of differential equations which governs an invariant eigendistribution on a real form of \({\mathfrak g}\). The author gives a counterexample to this problem when \({\mathfrak g}\) is simple of type \(F_ 4\) and \(X\in {\mathfrak g}\) is a certain nilpotent element.