How statement and proof provenance work
The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.
- Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
- AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
- AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.
These labels describe origin, not correctness: citations and verification chips remain separate evidence.
Smooth fundamental divisors from canonical double cuts and first jets
Statement
Assume AC. Let X be a smooth connected complex projective fourfold with ample Cartier A and . Suppose that for a general ordered pair of linearly independent sections of A, their scheme-theoretic intersection is a normal canonical surface of codimension two. Then a general member of |A| is smooth. No local factoriality or Weil-class assertion about an intermediate singular threefold is a hypothesis or conclusion.
Statement references
Facts & Assumptions
Given: Exactly the stated smooth fourfold, ample complete linear system, and actual canonical-doublecut premise. AC is retained from the suppliers.
Canonical codimension-two surface sections have a nonzero first jet proves directly by ambient point blowup, lci residue and finite component normalization/conductor that a normal canonical surface cut by two equations in a smooth fourfold cannot have both equations in the square of the maximal ideal. No ADE classification is imported.
Bertini smoothness away from the base locus supplies smoothness of a general member away from the base locus. Jacobian rank detects regularity at closed points identifies singular points of the actual codimension-two complete intersection by rank of its two defining differentials. A normal variety is regular in codimension one makes the singular locus of a normal surface zero-dimensional.
Image dimension and the generic fibre formula and Krull's height theorem supply the incidence dimension bound and the lower bound for the zero locus of at most r-1 functions. The finite rank stratification is constructed in the proof from minors and finite Noetherian induction. The Axiom of Choice.
Proof
Let Z be the reduced base locus of V. By [F2], a general member is smooth off Z. A section vanishes on Z, so its differential there is independent of the local A-trivialization up to the corresponding unit: the derivative of the transition unit is multiplied by the vanishing section. Thus there is an intrinsic map V tensor tensor A|Z.
Stratify Z into finitely many smooth irreducible locally closed strata on which this derivative map has constant rank and locally free image. To construct the strata, take its determinantal rank loci; on the dense smooth open of each irreducible component of each difference the image and kernel are bundles. Remove those opens and repeat on their smaller-dimensional closed complements. Finite Noetherian induction terminates. Fix one resulting stratum of dimension d and derivative image bundle F of rank r.
For , the sections having zero derivative at a point of this stratum form the projective kernel bundle I. Its dimension is d+dim P(V)-r, unless its fibres are empty. If its image in P(V) is not dense, a general section has no singular point on this stratum. If the image is dense, dimension gives .
In the latter case choose a general s1 that has a zero of its derivative at a point z of the stratum. Choose s2 general with ds2(z) nonzero and with the actual double cut T canonical as hypothesized. Such a choice is possible: derivative evaluation at z has positive rank, and the canonical-doublecut premise holds for general s2 after fixing a general s1. Shrink the stratum near z so ds2 is nowhere zero in F.
The locus where ds1 and ds2 are dependent is the zero locus of the image of ds1 in F/O ds2, of rank r-1. It contains z and is locally defined by at most r-1 functions on the smooth d-dimensional stratum. Its local dimension at z is at least by [F3].
Every point of that dependence locus lies in Z and hence in T; the two defining gradients have rank at most one. Because T is the actual codimension-two complete intersection in smooth X, [F2] makes all those points singular on T. This gives a positive-dimensional singular locus of the normal surface T, contradicting [F2]. Thus the dense-incidence alternative cannot occur on any positive-rank stratum.
On a rank-zero stratum every section and every first derivative vanish, so every double cut is singular there. The actual normal-doublecut premise excludes every positive-dimensional such stratum. There remain only finitely many fixed rank-zero points.
At each remaining point p, a local trivialization of A gives and with f1,f2 in for every pair of sections. A general pair is a normal canonical codimension-two surface by hypothesis. This is precisely forbidden by [F1]. Hence no fixed rank-zero point exists.
There are finitely many strata. For each positive-rank stratum the incidence image has proper closure, and the rank-zero strata have just been excluded. Choose a general section outside their finitely many proper closures and the off-base Bertini exceptional locus. Its divisor is smooth everywhere on X, proving the claim.
The geometric first-jet argument only needs a smooth projective fourfold and the stated canonical-doublecut premise; the Fano hypotheses retain the exact intended source/consumer scope. The canonical-doublecut construction is a separate supplier obligation. This research candidate does not assert production admission, proof acceptance or native gate coverage.
Remarks
The geometric first-jet argument only needs a smooth projective fourfold and the stated canonical-doublecut premise; the Fano hypotheses retain the exact intended source/consumer scope. The canonical-doublecut construction is a separate supplier obligation. This research candidate does not assert production admission, proof acceptance or native gate coverage.