Surfaces generating the even primal cohomology of an abelian fivefold (Q2225618)

Let \(A\) be a principally polarized abelian variety of dimension \(g \geq 4\) with smooth symmetric theta divisor \(\Theta\). The primal cohomology of \(\Theta\) is defined in [\textit{E. Izadi} et al., J. Algebr. Geom. 26, No. 1, 107--175 (2017; Zbl 1355.14009)] as \(\mathbb{K} := \mathrm{Ker}(H^{g-1}(\Theta, \mathbb{Z}) \xrightarrow{i_*} H^{g+1}(A, \mathbb{Z}))\) where \(i: \Theta \hookrightarrow A\) is the inclusion. The action of \(-1\) splits \(\mathbb{K}_{\mathbb{Q}}:= \mathbb{K} \otimes \mathbb{Q}\) into the direct sum of its invariant piece \(\mathbb{K}_{\mathbb{Q}}^{+}\) and its anti-invariant piece \(\mathbb{K}_{\mathbb{Q}}^{-}\). The general Hodge conjecture predicts that \(\mathbb{K}_{\mathbb{Q}}\) is contained in the image, via Gysin pushforward, of the cohomology of a smooth (possibly reducible) variety of pure dimension \(g-3\). The conjecture was previously proved for \(g=5\) in [\textit{E. Izadi} et al., J. Algebr. Geom. 26, No. 1, 107--175 (2017; Zbl 1355.14009)]. By recent results in [\textit{E. Izadi} and \textit{J. Wang}, Am. J. Math. 142, No. 5, 1409--1438 (2020; Zbl 1466.14010)], the space \(\mathbb{K}_{\mathbb{Q}}^{+}\) consists of Hodge classes while \(\mathbb{K}_{\mathbb{Q}}^{-}\) is simple. By [\textit{E. Izadi} et al., J. Algebr. Geom. 26, No. 1, 107--175 (2017; Zbl 1355.14009)] and the Lefschetz \((1,1)\) theorem, the classes belonging to \(\mathbb{K}_{\mathbb{Q}}^{+}\) are algebraic. In the main result of the article the authors describe explicit surfaces in \(\Theta\) representing these classes. Hence giving a second proof of the Hodge conjecture in this case. They also determine the intersection pairing between these surfaces using Brill-Noether theory.