Indexed text attachment
galois-theory-in-higher-categories.tex
Source for Galois Theory in Higher Categories
View paper · Download original file
Lines 231–330 of 4071. Source text is displayed literally.
231 232\begin{proof}233A point of the fiber consists of a map to \(E\) and a homotopy234identifying its composite with \(a\). These are the data in the235slice mapping space, with the same assertion for families236parametrized by a test space. The pullback universal property237rewrites that homotopy fiber as maps from the terminal238object of \(\X_{/A}\) to \(a^*E\).239\end{proof}240 241If \(F\) is a fixed fiber, its space of automorphisms \(\Aut(F)\)242consists of the equivalence components of \(\Map(F,F)\).243Composition makes it a group object in spaces. Its group of244components \(\pi_0\Aut(F)\) classifies automorphisms up to homotopy.245A family with fiber \(F\) can depend on homotopies within a single246component of \(\Aut(F)\), even when the group of components of247the automorphism object acts trivially.248 249\section{Cartesianness and realizations}\label{sec:grammar-van-kampen}250 251A cartesian family over a diagram \(U:I\to\X\) is a natural252transformation \(V\to U\) whose square over every arrow is a253pullback. Let \(B=\colim_I U\). Pullback defines a comparison254\begin{equation}\label{eq:grammar-slices}255\X_{/B}\longrightarrow\lim_{i\in I^{\op}}\X_{/U_i}.256\end{equation}257The target consists of cartesian families. The identifications258over arrows and their higher coherences are part of the data.259A colimit for which this comparison is an equivalence is called260van Kampen.261\index{colimit!van Kampen}262 263\begin{proposition}[A test for descent of a diagram]\label{prop:grammar-descent}264Suppose the ambient category admits pullbacks and the required265colimits, and these colimits are preserved by base change. The266comparison \eqref{eq:grammar-slices} is fully faithful. It is an267equivalence exactly when for each cartesian \(V\to U\), the squares268\[269\begin{tikzcd}[column sep=large]270V_i \arrow[r] \arrow[d] & \colim_j V_j \arrow[d] \\271U_i \arrow[r] & B272\end{tikzcd}273\]274are pullbacks.275\end{proposition}276 277\begin{proof}278Realization takes \(V\) to \(\colim_I V\to B\). For any \(W\to B\),279maps from the realization to \(W\) are the limit of maps from280\(V_i\to B\) to \(W\), by the universal property of the colimit in281the slice. Base change rewrites the latter mapping spaces as maps282from \(V_i\to U_i\) to \(W\times_B U_i\). Their limit is the space283of maps of cartesian families into the comparison image of \(W\).284Realization is therefore left adjoint to the comparison.285The counit \(\colim_I(W\times_B U_i)\to W\) is an equivalence by286the base-change assumption, so the right adjoint is fully faithful.287For each \(V\), the unit is an equivalence exactly when its components288\(V_i\to U_i\times_B\colim_j V_j\) are equivalences.289\end{proof}290 291A candidate over the base is determined by its restrictions as292soon as the comparison is fully faithful. Existence of a candidate293for every cartesian family requires the unit condition. Each294covering theorem below uses a specific argument for existence.295 296\begin{theorem}[Van Kampen property of higher topoi]\label{thm:grammar-vk}297Small colimits in an \(\infty\)-topos are van Kampen.298\end{theorem}299 300\begin{proof}301In spaces, straightening over a space \(K\) gives302\(\SSet_{/K}\simeq\Fun(K,\SSet)\), covariantly for a family303viewed by transport. The inclusion of spaces into304\(\CatInf\) preserves colimits because it is left adjoint to305the maximal subgroupoid functor. The internal-hom universal306property turns colimits of bases into limits of the functor307categories. In a presheaf \(\infty\)-category, the two criteria308in Proposition~\ref{prop:grammar-descent} then hold objectwise.309 310Write \(L:\mathcal P\rightleftarrows\X:i\) for a left exact311localization. A diagram of local objects and its cartesian312families can be viewed in \(\mathcal P\), since the inclusion313preserves limits. Realizing there gives pullback squares as in314Proposition~\ref{prop:grammar-descent}. Applying \(L\) preserves315those pullbacks and gives their realizations in \(\X\).316Universality of colimits in \(\X\) follows by the same localization317argument. If \(B_P\) is the colimit of the diagram in \(\mathcal P\),318pull back that colimit along \(iW\times_{iB}B_P\to B_P\) and apply319\(L\). This identifies \(W\) as the colimit of the diagram base320changed along \(W\to B\). The two criteria now apply in \(\X\).321The straightening input is developed in \cite[Ch.~2]{htt}.322\end{proof}323 324\section{The overlap diagram}325 326The \(\Cech\) nerve of \(p:U\to B\) has \(k\)-simplices327\[328U_k=\underbrace{U\times_B\cdots\times_B U}_{k+1}.329\]330Its operators omit or repeat factors. It is a groupoid object because