The Exceptional Bound

The algebraic exceptional set of a variety of general type consists of all subvarieties that are not of general type themselves — the curves, points, and other loci where the variety's complexity drops. For surfaces of log general type (surfaces with a boundary divisor that together make the canonical class ample), the exceptional set should be small. How small depends on the boundary.

When the boundary divisor B has at least three irreducible components, the exceptional set is finite or empty in most cases. The three-component condition is sharp: with fewer components, the exceptional set can be larger. Three components provide enough intersection structure on the surface to constrain rational and elliptic curves to a finite collection. Each component of B contributes positional constraints; three components overconstrain the possible non-general-type loci.

The finiteness is proved by bounding the degree of any exceptional curve against the intersection numbers of the boundary components. The arithmetic of the intersection form — the matrix of how the components of B meet each other and meet the canonical class — determines whether a rational or elliptic curve can exist and how many there can be. The three-component hypothesis ensures this matrix is sufficiently constrained.

The through-claim: the exceptional set is controlled not by the surface's global geometry but by the combinatorics of the boundary. Three boundary components create a lattice of intersection constraints that leaves no room for infinite families of exceptional curves. The boundary's local intersection data determines the surface's global non-general-type locus. Combinatorial richness at the boundary forces geometric simplicity in the interior.