Polarizability of internal homs #
The integral internal hom of two polarizable pure Hodge structures is polarizable.
Its lattice Hom_ℤ(V, W) is identified with V* ⊗ W by Mathlib's dualTensorHomEquiv;
the complexification of this identification respects the Hodge filtrations, by
TauCeti.Hodge.HodgeStructureOn.internalHom_F_eq_comap.
A polarizing form on the dual of V and a polarizing form on W therefore give a polarizing
form on the internal hom: pull their tensor product back along the inverse contraction.
The dual polarization exists without unimodularity of the original integral form, by
TauCeti.Hodge.IsPolarizable.dual.
Main results #
TauCeti.Hodge.IsPolarization.internalHom: the pullback of the tensor product of polarizing forms onV*andWpolarizes their internal hom.TauCeti.Hodge.IsPolarizable.internalHom: polarizable pure Hodge structures are closed under internal homs.
References #
Deligne, Théorie de Hodge II, §2.1; Peters--Steenbrink, Mixed Hodge Structures, §2.1.
Polarizing forms on the dual of the source and on the target induce a polarization of the integral internal hom, by pulling their tensor product back along the inverse contraction.
Polarizable pure Hodge structures are closed under internal homs. The internal hom of
weights n₁ and n₂ is polarizable of weight n₂ - n₁.