A remark on hypersurfaces with isolated singularities (Q858591)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | A remark on hypersurfaces with isolated singularities |
scientific article |
Statements
A remark on hypersurfaces with isolated singularities (English)
0 references
11 January 2007
0 references
Let \(F\) be a nodal hypersurface of degree \(m\) in \(\mathbb P^3\). A theorem due to \textit{F. Severi} [Ann. Mat. Pura. Appl. 25, 1--41 (1946; Zbl 0061.33901)] states that if \(d\) is at least \(2m-5\), then the subspace of \(| O_{\mathbb P^3}(d)| \) parametrising hypersurfaces passing through the nodes of \(F\), has codimension equal to the number of nodes. More generally one can ask the following (difficult) question: consider a given finite subscheme of length \(k\) of \(\mathbb P^n\). What is the dimension of the linear system of hypersurfaces of degree \(d\) vanishing on the subscheme? In particular one would like to ensure that the conditions on the linear system are linearly indepedent, i.e. the linear system has codimension \(k\) in the space of of hypersurfaces of degree \(d\). The authors generalise Severi's theorem in two directions: they consider hypersurfaces \(F\) in \(\mathbb P^n\) in arbitrary dimension and the singularities of \(F\) are not necessarily nodal, but only isolated. The main theorem states that if \(d\) is bigger than an explicit lower bound depending on the dimension and the degree of \(F\), and the multiplicities of the singularities (cf. the article for the exact formula), then the singular points impose linearly independent conditions on the form of degree \(d\). The idea of the proof is to tackle this very classical problem with the techniques of modern birational geometry. An application of the Nadel vanishing theorem [\textit{A. Nadel}, Ann. Math. (2) 132, 549--596 (1990; Zbl 0731.53063)] reduces the proof to constructing a \(\mathbb Q\)-divisor \(D\) in \(\mathbb P^n\) such that \((\mathbb P^n,D)\) is Kawamata log-terminal outside the singularities of \(F\) and has multiplicity at least \(n\) in the singular points of \(F\). The main part of the argument is then devoted to the (slightly technical) construction of such a divisor. The authors also discuss the optimality of their bound by comparing it with examples obtained via \texttt{Singular}.
0 references
hypersurface
0 references
isolated singularities
0 references
linear system
0 references
independent conditions
0 references