A Lefschetz (1,1) theorem for normal projective complex varieties (Q1975993)

Let \(X\) be a projective variety over \(\mathbb{C}\) and \(X_{an}\) be the analytic space associated to \(X\). The Lefschetz (1,1) theorem states that if \(X\) is smooth, then \[ \text{NS} (X)= \{\alpha\in H^2(X_{an}, \mathbb{Z}); \alpha_\mathbb{C} \in F^1H^2 (X_{an}, \mathbb{Z})\}, \] where \(F^\bullet\) denotes the Hodge filtration on the mixed Hodge structure \(H^2(X_{an},\mathbb{Z})\). \textit{L. Barbieri Viale} and \textit{V. Srinivas} [J. Reine Angew. Math. 450, 37-42 (1994; Zbl 0790.14011)] constructed, however, a singular surface \(X\) for which the equality does not hold. In the same paper they ask the following question. Let \(H^1(X, {\mathcal H}_X^1)\) be the subgroup of \(H^2(X_{an}, \mathbb{Z})\) consisting of Zariski-locally trivial cohomology classes. Is \(\text{NS} (X)=\{\alpha\in H^1 (X,{\mathcal H}^1_X)\); \(\alpha_C\in F^1H^2 (X_{an},\mathbb{Z})\}\)? Since \(H^1(X,{\mathcal H}^1_X) =\text{NS}(X)\) when \(X\) is smooth, the question has a positive answer for smooth projective varieties. The main theorem of the present article answers in the affirmative to the question under the assumption that \(X\) is normal.