Independent scholarship · Open accessThe monthly digest

Indexed text attachment

galois-theory-in-higher-categories.tex

Source for Galois Theory in Higher Categories

View paper · Download original file

Lines 3621–3720 of 4071. Source text is displayed literally.

3621 3622For an infinite discrete group $G$ and $B=BG$, the atlas3623coming from $*\to BG$ has $n$th term $\mathcal S_{/G^n}$.3624Equation~\eqref{eq:higher-quotient-actions} gives its category3625of actions as $\operatorname{Fun}(BG,\mathcal S)$. Its3626$0$-truncated objects are ordinary $G$-sets. Taking instead3627$B=K(\mathbb Z,2)$ yields the atlas3628$\mathcal S\to\mathcal S_{/B}$ whose first overlap is3629$\mathcal S_{/K(\mathbb Z,1)}$. That overlap is $1$-localic3630since it is a presheaf category on an ordinary groupoid,3631and it fails to be $0$-localic by Lemma~\ref{lem:geo-localic-space}.3632Moreover $\mathcal S_{/B}$ and $\mathcal S$ have the same3633ordinary topos of $0$-truncated objects. Straightening3634identifies the first ordinary topos with set-valued functors3635on $B$, which factor through the fundamental groupoid.3636For $K(\mathbb Z,2)$ that groupoid is contractible. The two3637higher topoi are inequivalent by Lemma~\ref{lem:geo-localic-space}.3638Theorem~\ref{thm:geo-localic-bound} concerns the specific3639class of surjective étale geometric morphisms and does not3640assume a theorem about all other notions of higher atlas.3641 3642\section{Hyperdescent and the role of truncation}3643\index{hypercover}3644\index{hypercompletion}3645 3646In any $\infty$-topos an augmented simplicial object3647$V_\bullet\to V$ is a hypercover if each map to its3648relative matching object is an effective epimorphism.3649In degree zero the matching object is $V$. In degree one3650it is $V_0\times_VV_0$. In higher degrees it is the3651object of compatible boundaries relative to $V$.3652Each matching condition asks for local existence of a lift.3653When the augmentation is a Čech nerve, there is a stronger3654identity. The levels are specified iterated pullbacks and3655the geometric realization of a cover is exactly $V$.3656 3657The realization of an arbitrary hypercover maps to $V$ by an3658$\infty$-connective morphism3659\cite[Lem.~6.5.3.11]{htt}. This means its relative3660$n$-truncation is terminal for every finite $n$. Applying3661\eqref{eq:higher-slice-vankampen} to $V_\bullet$ identifies3662$\lim_\Delta\mathcal X_{/V_n}$ with3663$\mathcal X_{/|V_\bullet|}$. Consequently the restriction3664comparison for all objects over $V$ is an equivalence3665precisely when $|V_\bullet|\to V$ is an equivalence.3666Indeed, if the restriction $g^*$ for a map $g$ is an3667equivalence, so is its left adjoint $g_!$. Applying $g_!$3668to the terminal object in the source slice gives $g$ as3669an object of the target slice, so $g$ is terminal there.3670Conversely an equivalence $g$ induces an equivalence of slices.3671 3672An $\infty$-topos is hypercomplete if every $\infty$-connective3673map is an equivalence. Hypercompletion is the left exact3674localization which forces these equivalences. Each3675$\infty$-connective morphism gives a constant simplicial3676hypercover in its target slice, according to the equivalent3677iterated-diagonal test for $\infty$-connectivity3678\cite[Lem.~6.5.3.5]{htt}. Thus effectiveness of3679all hypercovers is equivalent to hypercompleteness3680\cite[Thm.~6.5.3.12]{htt}. A bounded truncated object3681already satisfies the extra locality condition. Mapping3682into an $n$-truncated object cannot distinguish an3683$\infty$-connective map from an equivalence, by orthogonality3684to $n$-truncated maps. In particular,3685passing to ordinary sheaves of sets cannot test whether3686a given $\infty$-topos is hypercomplete.3687 3688A concrete instance of the distinction is the canonical3689$1$-localic lift of continuous $\mathbb Z_p$-sets. One can3690present it by the category of transitive quotients3691$\mathbb Z_p/p^n\mathbb Z_p$ with its atomic topology.3692This is an $\infty$-topos of ordinary sheaves of spaces3693in the sense of covering sieves, and it fails to be3694hypercomplete by Wieland's example3695\cite[Warning~7.2.2.31]{htt}. The site is ordinary and3696its underlying ordinary topos already has a classical3697group-theoretic presentation. Neither fact removes the3698hyperdescent distinction. The higher quotient3699\eqref{eq:higher-quotient-actions} uses an effective3700groupoid object, so it works in that topos before3701hypercompletion as well as in hypercomplete examples.3702 3703\begin{proposition}[Relative hypercompletion]3704\label{prop:relative-hypercompletion}3705Let $L\colon\mathcal X\to\mathcal X^\wedge$ be hypercompletion.3706For every $U\in\mathcal X$, the induced functor3707\[3708 L_{/U}\colon\mathcal X_{/U}\longrightarrow3709                  (\mathcal X^\wedge)_{/LU}3710\]3711is the hypercompletion of the slice. In particular, applying3712$L$ to a hypercover $V_\bullet\to V$ gives a slice3713comparison3714\[3715 (\mathcal X^\wedge)_{/LV}3716 \ \simeq\3717 \lim_{[n]\in\Delta}3718              (\mathcal X^\wedge)_{/LV_n}.3719\]3720\end{proposition}