Skip to content

Hodge Conjecture

Jean-Paul Niko · 2026-03-19 · 7 pages

The Strategy

Induction on \(p\):

  • Base case \(p=1\): Lefschetz \((1,1)\) theorem. Done.
  • Inductive step: Given \(\alpha \in \mathrm{Hdg}^p(X)\), decompose \(\alpha = L\beta + \gamma^{\mathrm{prim}}\).
  • \(L\beta\): algebraic since \(\beta \in \mathrm{Hdg}^{p-1}\) is algebraic by induction, and \(L\) preserves algebraicity.
  • \(\gamma^{\mathrm{prim}}\): algebraic via Poincaré duality — reduce to the dual class in \(\mathrm{Hdg}^{n-p}\).

Key Lemma

Poincaré duality preserves algebraicity. If \([Z] \in \mathrm{Hdg}^p(X)\) is algebraic, then \([Z]^\vee \in \mathrm{Hdg}^{n-p}(X)\) is algebraic — it equals \([Z \cap H_1 \cap \cdots \cap H_{n-2p}]\), which is smooth and algebraic by Bertini.

The Wall

Algebraicity of \(\Lambda\) (the Lefschetz adjoint). The \(\mathfrak{sl}_2\) triple \((L, \Lambda, H)\) generates the Hard Lefschetz algebra. \(L\) preserves algebraicity (proved). Does \(\Lambda\)?

If yes: the \(\mathfrak{sl}_2\)-module generated by any algebraic class is entirely algebraic, the induction closes, the conjecture holds.

RTSG Statement

Every CS pattern (Hodge class) of pure type \((p,p)\) with rational coefficients has a PS realization (algebraic cycle). The Hodge conjecture is the statement that the instantiation map is surjective onto the intersection of QS-rational and PS-geometric classes.

Cross-references