On the algebra of totally degenerate theta functions (Q2365898): Difference between revisions

Let \(C\) be a singular curve of genus \(g\) over the complex field, with \(g\) nodes as singularities and having a normalization \(\pi:\mathbb{P}^ 1\to C\). It is known [cf. \textit{D. Mumford}, ``Tata lectures on theta. II'', Prog. Math. 43 (1984; Zbl 0549.14014); chapter III b, \S5] that the (generalized) Jacobian \(\text{Jac} C\) is isomorphic to the product of \(g\) copies of the multiplicative group \(\mathbb{C}^*\). This space can be compactified, as analytic space, to \(\text{Jac} C \cong (\mathbb{P}^ 1)^ g/ \sim_ \nu\), where \(\sim_ \nu\) is a suitable equivalence relation determined by a symmetric matrix \(\nu\) such that, for \(1\leq i\), \(j\leq g\), \(\nu_{ii}=0\) and \(\nu_{ij}\) is invertible for \(i\neq j\). The author proves that, when \(k\) is a ring and \(\nu\) is a symmetric matrix with entries in \(k\), satisfying the above conditions, then the quotient \(A_ \nu\) of \((\mathbb{P}^ 1)^ g \times k\) with respect to the equivalence relation induced by \(\nu\), is a stable quasi-abelian scheme and he explicitly describes that scheme as \(\text{Proj} R\), where \(R\) is a graded ring of totally degenerate theta functions relative to \(\nu\). By means of this description the author proves that the quotient \(A_ \nu\) is a projective scheme and that the canonical map \((\mathbb{P}^ 1)^ g \times k \to A_ \nu\) is finite.