Alphabeta Math
LemmaStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedprecheck passjudge pass (gpt-6.1-sol)
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.

Strictness of mixed Hodge morphisms

Statement

Every morphism f ⁣:H→K of rational mixed Hodge structures is strict for W and for F: f(WnH)=f(H)∩WnK,f(FpHC)=f(HC)∩FpKC for all n,p. Its kernel, image and cokernel, with the induced or quotient filtrations, are mixed Hodge structures; the functor Gr⁡nW is exact; and a morphism from a structure of weights at least n (that is Wn−1H=0) to a structure of weights at most n (that is WnK=K) has pure image of weight n.

Facts & Assumptions

Given: A morphism f ⁣:H→K of rational mixed Hodge structures (Rational mixed Hodge structures); the complexified map is written fC=f⊗id and commutes with the conjugations because f is rational.

[F1]

The Deligne splitting of The Deligne splitting of a mixed Hodge structure: HC=⨁p,qIp,q(H) with WmHC=⨁p+q≤mIp,q(H), FaHC=⨁p≥aIp,q(H), the projection of Ip,q(H) onto Gr⁡p+qWHC being an isomorphism onto Hp+qp,q, and fC(Ip,q(H))⊆Ip,q(K). The conjugation formula is only Ip,q+Wp+q−2=σ(Iq,p)+Wp+q−2; no exact Fˉ partial-sum formula by the same pieces is used.

[F2]

On Gr⁡nWHC one has Gr⁡nWHC=⨁p+q=nHnp,q with Hnp,q=FpGr⁡nW∩FˉqGr⁡nW (Rational mixed Hodge structures).

Proof

technique · direct
1.1F1algebra

(Weight strictness.) Let y=f(x)∈f(H)∩WnK and decompose x=∑p,qxp,q along the splitting of HC; then y=∑p,qfC(xp,q) with fC(xp,q)∈Ip,q(K) by [F1]. Since y∈WnK=⨁p+q≤nIp,q(K) and the sum is direct, all components with p+q>n vanish: fC(xp,q)=0 whenever p+q>n. Hence x′:=∑p+q≤nxp,q lies in WnHC and satisfies fC(x′)=y, so y∈fC(WnHC). The reverse inclusion f(WnH)⊆f(H)∩WnK is the preservation of W by f.

1.2F1algebra

(Hodge strictness.) The same computation with the condition p<a instead of p+q>n: if y=fC(x)∈f(HC)∩FaKC, decompose x and use FaKC=⨁p≥aIp,q(K) to conclude that fC(xp,q)=0 for p<a, so x′=∑p≥axp,q lies in FaHC and maps to y.

2.1step 1.1step 1.2algebra

(Rational descent.) Step 1.1 gives equality of the complexifications of the rational subspaces A=f(WnH) and B=f(H)∩WnK. Images and intersections commute with extension from Q to C: in finite rational coordinates they are kernels or images of matrices, whose rank is unchanged on extension since the same nonzero rational minors remain nonzero. To descend AC=BC, map A into the rational quotient K/B. Its complexification is zero, hence the rational map is zero (a nonzero rational coordinate remains nonzero over C), so A⊆B; interchanging them gives equality. Together with step 1.2 this proves the stated rational W strictness and complex F strictness. Complex conjugation fixes the real scalar extension of a rational space and is not used to recover rational vectors.

3.1F1step 1.1step 2.1algebra

(Exactness of Gr⁡nW.) Let 0→A→iB→pC→0 be a short exact sequence of mixed Hodge structures. By [F1] the splitting identifies Gr⁡nWA, Gr⁡nWB, Gr⁡nWC with ⨁p+q=nIp,q(A) and its analogues, and the maps Gr⁡nWi, Gr⁡nWp with the direct sums of the maps induced by i and p between the pieces Ip,q. Each piecewise sequence 0→Ip,q(A)→Ip,q(B)→Ip,q(C)→0 is exact: injectivity follows by restriction of i; to lift c∈Ip,q(C), choose a preimage in BC and decompose it, whose (p,q) component maps to c by functoriality and uniqueness of the target decomposition. An element b∈Ip,q(B) lies in ker⁡p precisely when b=i(a) for some a∈A, and decomposing a into its I-pieces shows that all other pieces map to zero; since i is injective, they vanish and a∈Ip,q(A). A direct sum of short exact sequences is short exact, so 0→Gr⁡nWA→Gr⁡nWB→Gr⁡nWC→0 is exact after complexification. The rational graded sequence is exact as well, since kernels and images commute with scalar extension and rational ranks descend as in step 2.1.

3.2F1F2step 1.1step 1.2step 2.1algebra

(Kernel, image and cokernel.) The functorial splitting [F1] makes fC a direct sum of maps Ip,q(H)→Ip,q(K). Thus its kernel, image and cokernel decompose into the respective kernels, images and cokernels of these maps, and their induced or quotient W and F are the partial sums by p+q≤n and p≥a. The projection isomorphisms in [F1] identify their weight-n pieces, with these induced F filtrations, with the kernel, image and cokernel of Gr⁡nWf, computed componentwise in the pure decompositions Hnp,q,Knp,q. These identifications are the canonical rational graded maps: for a kernel this follows from W strictness, since if f(x)∈Wn−1K with x∈WnH, choose y∈Wn−1H with f(y)=f(x) and replace x by x−y; the image and cokernel follow by the same induced/quotient description. The graded rational map commutes with conjugation and preserves F, hence also Fˉ; between equal-weight pure structures it preserves each type Fp∩Fˉq with p+q=n. Consequently its pure kernels, images and cokernels have conjugation exchanging their (p,q) and (q,p) pieces. Their F partial sums and their conjugates are therefore n-opposed. Every weight-graded piece of the three rational filtered spaces satisfies the mixed Hodge axiom, proving the claim. This argument requires no exact conjugate decomposition on the mixed vector space.

4.1step 1.1step 3.2algebra∎

(Pure image.) Now suppose Wn−1H=0 and WnK=K. By step 3.2 the image is a mixed Hodge structure with the induced filtration, and step 1.1 gives Wn−1(im⁡f)=f(Wn−1H)=0 and Wn(im⁡f)=f(H)∩WnK=f(H)=im⁡f; hence the image is pure of weight n.

Source notes

Deligne II (1.2.10)-(1.2.12) and Theorem (2.3.5) are the source; the proof above follows Voisin's printed deduction of Theorem 2.7 from Theorem 2.8 and uses the splitting lemma of this page in place of Deligne's three-opposed-filtrations category theorem. The statement claims no canonical splitting of the induced structures on kernel, image or cokernel, and the argument is choice-free: all decompositions used are the canonical ones supplied by the splitting lemma.

Depends on

Used by

Dependency tree · two levels

9 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.

Sources