Pareschi–Popa generic vanishing conjecture

Let (A,Θ)(A,\Theta) be a principally polarized abelian variety of dimension gg, and let X⊂AX\subset A be a subvariety of dimension dd with 1≤d≤g−21\le d\le g-2. The conjecture asserts that XX represents the minimal cohomology class if and only if its ideal sheaf twisted by the principal polarization is a generic-vanishing sheaf:

[X]=[Θ]g−d(g−d)!⟺IX⊗OA(Θ) is GV[X]=\frac{[\Theta]^{g-d}}{(g-d)!}\quad\Longleftrightarrow\quad \mathcal I_X\otimes\mathcal O_A(\Theta)\text{ is }\mathrm{GV},

where the generic-vanishing condition means that, for every i>0i>0, codim⁡Pic⁡0(A){α∣Hi(A,IX⊗OA(Θ)⊗α)≠0}≥i\operatorname{codim}_{\operatorname{Pic}^0(A)}\{\alpha\mid H^i(A,\mathcal I_X\otimes\mathcal O_A(\Theta)\otimes\alpha)\ne0\}\ge i. The full Pareschi–Popa formulation further relates these conditions to the Jacobian and Fano-surface classifications in the relevant dimensions.

References

Primary source

arXiv

Additional references

Progress summary

Refreshed
Claimed progress

A new paper claims progress for divisors, but the original conjecture in higher codimension remains open.

Pareschi and Popa formulated the conjecture in 2006, asserting equivalences between minimal cohomology classes, generic-vanishing conditions, and Jacobian or Fano-surface classifications for subschemes of principally polarized abelian varieties.

Known results

  • The original paper proves (2)⟺(4)⇒(3)(2)\Longleftrightarrow(4)\Rightarrow(3) and (2)⇒(1)(2)\Rightarrow(1).
  • For d=1d=1 or d=g−2d=g-2, condition (2)(2) gives the Jacobian and Abel–Jacobi classifications.
  • In dimension 44, (1)(1), (2)(2), (4)(4), and (5)(5) are equivalent, and imply (3)(3).
  • The full higher-codimension equivalence was explicitly left conjectural.

September 2026 codimension-one result

Fanjun Meng’s preprint claims a codimension-one analogue for divisors, extending the framework but not proving the original conjecture. This is claimed progress, not a verified resolution.

Current status (as of September 2026): The classical partial implications are known, and a new divisor analogue is claimed, but the original conjecture in codimension greater than one remains open.

Sources

Solutions 0

No solutions have been posted yet.