Special effect varieties and \((-1)\)-curves (Q976546)

Given a linear system of plane curves of degree \(d\) in \({\mathbb P}_{\mathbb C}^2\), having singularities of given multiplicities \(m_1,\dots,m_r\) at \(r\) generic points \(P_1,\dots,P_r\in {\mathbb P}^2\) (we will denote such a linear system by \({\mathcal L}_{2,d}(-\sum _i^rm_iP_i)\)), it is not known in which cases \({\mathcal L}_{2,d}(-\sum _i^rm_iP_i)\) is \textit{special}, i.e. its affine dimension is bigger than \(\max\{{d+2 \choose 2} - \sum _i^r {m_i+1\choose 2}\;,\;0 \}\), the expected value. A well-known conjecture states that \({\mathcal L}_{2,d}(-\sum _i^rm_iP_i)\) is special if and only if it possesses a fixed multiple component which is a (-1)-curve counted with multiplicity at least 2 (a (-1)-curve is a curve such that its strict transform in the blow up of \({\mathbb P}^2\) at the \(P_i\)'s is an exceptional divisor). In this paper, two new ways to express that conjecture are given: the first one is done by introducing the concept of ``\(\alpha\)-special effect curve'' and the second one via the idea of ``\(h^1\)-special effect curve''. In both cases it is shown that the existence of such curves for \({\mathcal L}_{2,d}(-\sum _i^rm_iP_i)\) (the first is defined by numerical conditions, the second by cohomological ones) implies that the linear system is special. Then in both cases it is conjectured that every special linear system appears this way. It is shown that the two conjectures are equivalent and also equivalent to the well-known one described above. The relevance of this approach is in that it seems quite apt to be generalized to other settings than divisors in the plane; examples in the cases of \({\mathbb P}^3\), Hirzebruch surfaces and \(K3\) surfaces are considered in the last section.