Displacements

Abstract

Given a functor $p:E \rightarrow B$ and an object $e \in E$ , we define a \emph{displacement} of $e$ along a morphism $\varepsilon: p(e) \rightarrow b$, as a map $e \rightarrow \nabla_\varepsilon(e)$ satisfying a universal property analogue to that of a \emph{cocartesian lifting} (pushforward) \emph{\`a la} B\'enabou-Grothendieck-Street. There are many difficulties in geometry that come from the fact that forgetful functors such as $p: Var(\mathbf{C}) \rightarrow Top $ don't have displacements of objects along arbitrary maps. And this can be already seen abstractly, since the existence of a left adjoint to $p$, can be reduced to the existence of all displacements of the initial object. However some \emph{schematization functors} exist as approximations. In a broader context, if $B$ is a model category and $p$ is a right adjoint, then the right-induced model category on $E$ exists if and only if all displacements along any trivial cofibration $\varepsilon$, are weak $p$-equivalences. In these notes we provide some categorical lemmas that will be necessary for future applications. The idea is to have a \emph{homotopy descent process} for \emph{elementary displacements} when $p$ has a \emph{presentation} as a $2$-pullback of a family $\{p_i: E_i \rightarrow B\}_{i\in J}$. When suitably applied it should lead to techniques similar to Mumford's GIT through homotopy theory (simplicial presheaves).