l-adic vs complex Perverse Sheaves - MathOverflow most recent 30 from http://mathoverflow.net 2013-06-20T06:56:38Z http://mathoverflow.net/feeds/question/78446 http://www.creativecommons.org/licenses/by-nc/2.5/rdf http://mathoverflow.net/questions/78446/l-adic-vs-complex-perverse-sheaves l-adic vs complex Perverse Sheaves Anonymous 2011-10-18T12:27:18Z 2011-10-18T13:09:29Z <p>Let $X$ be a scheme of finite type over $Spec(\mathbb{C})$. Let $X_{an}$ denote the associated complex analytic space. After fixing an isomorphism $\overline{\mathbb{Q}}_l\cong \mathbb{C}$, by $\S$6.1.2. of BBD (Asterisque 100) we know that there is fully faithfull functor</p> <p>$$\mathcal{F}: D_c^b(X, \overline{\mathbb{Q}}_l) \rightarrow D_c^b(X_{an}, \mathbb{C}),$$</p> <p>where the left hand side is the usual "derived category" of $l$-adic sheaves and the the right hand side is the usual derived category of complex sheaves with constructible (for an algebraic stratification) cohomology. As is well known this functor is not essentially surjective. However, on both sides we have two natural subcategories</p> <p>$$Perv_{l}(X) \subset D_c^b(X, \overline{\mathbb{Q}}_l)$$</p> <p>and</p> <p>$$Perv_{\mathbb{C}}(X) \subset D_c^b(X_{an}, \mathbb{C})$$ of l-adic Perverse sheaves and complex Perverse sheaves respectively. My question is this:</p> <p>Does $\mathcal{F}$ induce an equivalence of categories between $Perv_{l}(X)$ and $Perv_{\mathbb{C}}(X)$? </p> http://mathoverflow.net/questions/78446/l-adic-vs-complex-perverse-sheaves/78453#78453 Answer by ulrich for l-adic vs complex Perverse Sheaves ulrich 2011-10-18T13:09:29Z 2011-10-18T13:09:29Z <p>No. </p> <p>If $L$ is a local system on a smooth variety $X$ of dimension $d$ then $L[d]$ is perverse. As suggested by BBD, if we take $L$ to be a rank $n$ local system of <code>$\overline{\mathbb{Q}}_l$</code> vector spaces on $X_{an}$ then this is in the essential image of $\mathcal{F}$ iff the corresponding representation of <code>$\pi_1(X_{an})$</code> preserves a lattice (by which I mean a rank $n$ $\mathcal{O}$-submodule, where $\mathcal{O}$ is the ring of integers in some finite extension of <code>$\mathbb{Q}_p$</code>). The reason for this is that the algebraic fundamental group is compact and any compact subgroup of $GL_n(\overline{\mathbb{Q}}_l)$ is contained in $GL_n(\mathcal{O})$, for some $\mathcal{O}$ as above.</p> <p>This condition need not always hold: for example, one can consider representations of $\pi_1(X_{an})$ so that the eigenvalues of some element are algebraic but not $l$-adic integers. </p>