Singular loci of Bruhat-Hibi toric varieties (Q934075)

Let \(\mathbb{K}\) be an algebraically closed field of arbitrary characteristic. With any distributive lattice \(\mathcal{L}\) one associates the affine variety \(X(\mathcal{L})\) whose defining ideal in \(\mathbb{K}[x_{\alpha} : \alpha\in\mathcal{L}]\) is generated by \(x_{\alpha}x_{\beta}-x_{\alpha\vee\beta}x_{\alpha\wedge\beta}\). In fact, the variety \(X(\mathcal{L})\) is toric, and it is called a Hibi toric variety. For \(\mathcal{L}\) being the Bruhat poset of Schubert varieties in a minuscule flag variety \(G/P\), it is proved in [\textit{N. Gonciulea} and \textit{V. Lakshmibai}, Transform. Groups 1, No. 3, 215--248 (1996; Zbl 0909.14028)] that \(X(\mathcal{L})\) flatly deforms to the cone over \(G/P\). Later the same authors conjectured that the set of singular points of \(X(\mathcal{L})\) for a Bruhat poset \(\mathcal{L}\) is a union of subvarieties \(Z_{\alpha,\beta}\), where \((\alpha,\beta)\) is an unordered incomparable pair of join-meet irreducibles in \(\mathcal{L}\), and \(Z_{\alpha,\beta}\) is defined by vanishing of the coordinates corresponding to the elements of \([\alpha\wedge\beta, \alpha\vee\beta]\). It turns out that the conjecture does not extend to a general Hibi toric variety. The aim of the paper under review is to prove the conjecture for \(\mathcal{L}\) being the Bruhat poset of Schubert varieties in a minuscule \(G/P\). Moreover, it is shown that the singular locus is pure of codimension 3 in \(X(\mathcal{L})\), and the generic singularities are of cone type.