Cohomology of the space of polynomial maps on A^1 with prescribed ramification

. It is, in a way, orthogonal to the many papers about polynomials with prescribed zeroes -- here we are prescribing, instead, the ramification data. Exploiting the topological properties of the poset that encodes the ramification behaviour, we use a sheaf-theoretic argument to compute

H^{*} (S i m p_{n}^{m} (m a t h b b C); m a t h b b Q)

as well as the 'etale cohomology

H_{e t}^{*} ({S i m p_{n}^{m}}_{/ K}; m a t h b b Q_{e l l})

for

c h a r K = 0

or

c h a r K > n + 1

. As a by-product we obtain that

H^{*} (S i m p_{n}^{m} (m a t h b b C); m a t h b b Q)

is independent of

n

, thus implying rational cohomological stability. When

c h a r K > 0

our methods compute

H_{e t}^{*} (S i m p_{n}^{m}; m a t h b b Q_{e l l})

provided

c h a r K > n + 1

and show that the 'etale cohomology groups in positive characteristics do not stabilize.

Review text

The paper under review starts with the space \({\mathrm{M}}_{\mathrm{n}}\) of all monic polynomials of degree \(n+1\) over \(K\) that vanish at \(0\), where \(K\) is an algebraically closed field. A lot of work has been made on subvarieties of \({\mathrm{M}}_{\mathrm{n}}\), and the author considers here the subvarieties \(Simp_n^m\), consisting of the morphisms with total ramification \(< m\). This subvariety is a Zariski open dense subset of \({\mathrm{M}}_{\mathrm{n}}\). The goal of the article is to compute the cohomology of \(Simp_n^m\), what is achieved in Theorem A. In order to state it, call \(\mathbf{p}(N)\) the number of partitions of a positive integer \(N\), and define \(\mathbf{c} : \mathbb{Z}^+ \rightarrow \mathbb{Z}^+ \) by \(\mathbf{c}(m) = \sum \mathbf{p}(n_1+1) \cdots \mathbf{p}(n_k+1)\), where the sum runs over \(k \geq 1\), \(n_1 + \dots + n_k=m\), \(n_1 \leq \dots \leq n_k\). Then, if \(m \geq 1, n \geq 3m\), Theorem A gives the singular cohomology of \(Simp_n^m\), \(H^i\), and its étale cohomology, \(H^i_{ \acute{e}t}\), in the following terms: (1) \(H^i(Simp_n^m(\mathbb{C});\mathbb{Q})\) is \(\mathbb{Q}\) for \(i=0\), \(\mathbb{Q}^{\oplus \mathbf{c}(m)}\) for \(i=2m-1\), and \(0\) otherwise. Besides, in the case \(i=2m-1\), \(H^{2m-1}(Simp_n^m(\mathbb{C});\mathbb{Q})\) is pure of weight \(-2m\) and Hedge type \((-m,-m)\). (2) If \(\kappa\) is a field of characteristic \(0\) or greater than \(n+1\), \(H^i_{ \acute{e}t}(Simp_{n\; / \overline{\kappa}} ^m(\mathbb{C});\mathbb{Q}_\ell)\) is \(\mathbb{Q}_\ell(0)\) for \(i = 0\), \(\mathbb{Q}_\ell(-m)^{\oplus \mathbf{c}(m)}\) for \(i = 2m-1\), and \(0\) otherwise, with \(\ell\) coprime with the characteristic of \(\kappa\). The proof of Theorem A is rather lengthy, and it runs throughout Sections 2 to 6 of the paper. However, the plan of the proof, and its development, are clearly explained. Its main tools are the topological properties of the poset that encodes the behavior of the ramification, and provides a stratification of \({\mathrm{M}}_{\mathrm{n}}\). As a noteworthy point of Theorem A we should note that the value of \(H^i\) is independent of \(n\).

Recommendations

Cites work

Cited in

(4)

This page was built for publication: Cohomology of the space of polynomial maps on \(\mathbb{A}^1\) with prescribed ramification

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2007744)