5
$\begingroup$

I have a complex smooth projective scheme $X$ with the sheaf of Kähler differentials $\Omega_{X/\mathbb{C}}$ (or only $\Omega$). Denote its analytification $X^{an}$ with analytification morphism $h:X^{an} \rightarrow X$. I already know that $h^*\Omega_{X/\mathbb{C}} \cong \Omega_{X^{an}}$ the sheaf of holomorphic $1$-forms on $X^{an}$. My goal is to define a wedge product:

\begin{align*} H^{q_1}(X, \Omega^{p_1}) \times H^{q_2}(X,\Omega^{p_2}) &\rightarrow H^{q_1+q_2}(X, \Omega^{p_1+p_2}) \\ (\alpha, \beta) \rightarrow \alpha \wedge \beta \end{align*}

My first attempt is to use Čech cohomology and mimic the cup product of singular cohomology. In our setting Čech cohomology is the same as sheaf cohomology, moreover there is a finite standard affine covering of $X$, on which we can compute the cohomology. Let $\check{\alpha} \in \check{C}^{p_1}(X, \Omega^{q_1})$ and $\check{\beta} \in \check{C}^{p_2}(X, \Omega^{q_2})$ be the representative of $\alpha, \beta$, respectively, then we can define a represetative of $\alpha \wedge \beta$ on the intersection $i_0,i_1, \ldots, i_{p_1+p_2}$ as $\check{\alpha}_{i_0,i_1, \cdots, i_{q_1}} \wedge \check{\beta}_{i_{q_1}, i_{q_1 + 1}, \cdots, i_{q_1 + q_2}}$ where the wedge here means:

\begin{align*} \Omega^{p_1}(U_{ i_0,i_1, \cdots, i_{q_1}}) \times \Omega^{p_2}(U_{i_{q_1}, i_{q_1+1}, \cdots, i_{q_1 + q_2}}) &\rightarrow \Omega^{p_1 + p_2}(U_{i_0,i_1, \cdots, i_{q_1+q_2}}) \\ \gamma_1 \times \gamma_2 &\mapsto \gamma_1|_{U_{i_0,i_1, \cdots, i_{q_1+q_2}}} \wedge \gamma_2|_{U_{i_0,i_1, \cdots, i_{q_1+q_2}}}. \end{align*}

I can prove that this is well-defined. The same thing can be done to get a wedge product:

$$ H^{q_1}(X^{an}, \Omega^{p_1}_{X^{an}}) \times H^{q_2}(X^{an},\Omega^{p_2}_{X^{an}}) \rightarrow H^{q_1+q_2}(X^{an}, \Omega^{p_1+p_2}_{X^{an}}) $$

But we also know that $H^{q}(X^{an}, \Omega^{p}_{X^{an}}) = H^{p,q}_{\bar\partial}(X^{an},\mathbb{C})$ the Dolbeault cohomology and here we have the usual wedge product between differential forms. So my questions are:

1) Is there a better way to define the wedge product between the cohomology groups categorically without using Čech cohomology

2) With either definition how can we prove the compatibility between the second and third row:

$\require{AMScd}$ \begin{CD} H^{q_1}(X, \Omega^{p_1}) \times H^{q_2}(X,\Omega^{p_2})@>>> H^{q_1+q_2}(X, \Omega^{p_1+p_2})\\ @VVV @VVV\\ H^{q_1}(X^{an}, \Omega^{p_1}_{X^{an}}) \times H^{q_2}(X^{an},\Omega^{p_2}_{X^{an}}) @>>> H^{q_1+q_2}(X^{an}, \Omega^{p_1+p_2}_{X^{an}})\\ @VVV @VVV\\ H^{p_1,q_1}_{\bar{\partial}}(X^{an}, \mathbb{C}) \times H^{p_2,q_2}_{\bar{\partial}}(X^{an},\mathbb{C}) @>>> H^{p_1+p_2,q_1+q_2}_{\bar{\partial}}(X^{an},\mathbb{C}) \end{CD}

Thank you in advance.

$\endgroup$
3
  • $\begingroup$ Typo : the sheaves $h^* \Omega^{X/ \Bbb C}$ and $\Omega_{X_{an}}$ are not isomorphic (but the global sections are). Also, if you write $H^k(X, F) = Ext^k(\mathcal O_X, F) = Hom_{D^b(X)}(\mathcal O_X, F[k])$ then you get a product using composition of morphisms in $D^b(X)$. However I don't know a reference and I don't know how to prove that this product is the same as the one obtained from Cech cohomology. $\endgroup$ Nov 1, 2019 at 22:57
  • $\begingroup$ Are you sure the sheaves are not isomorphic, because according to this post they are math.stackexchange.com/questions/3412415/… $\endgroup$ Nov 1, 2019 at 23:08
  • $\begingroup$ Ah sorry your are right, they are isomorphic indeed. $\endgroup$ Nov 2, 2019 at 8:52

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.