ScalingStacks

[05WW]

Proposition 2.1.5. Let L:π’žβ‡†π’Ÿ:RL\colon\mathcal{C}\leftrightarrows\mathcal{D}\penalty\mskip 6.0mu plus 1.0mu\mathpunct{}\nonscript\mkern-3.0mu{:}\mskip 2.0muR be an adjunction between ∞\infty-categories. Assume that π’ž\mathcal{C} admits and LL preserves pullbacks, that π’Ÿ\mathcal{D} admits an initial object, and that RR is fully faithful. For every object Yβˆˆπ’žY\in\mathcal{C} we consider the following pullback diagram

Yred\textstyle{Y^{\operatorname{\scriptsize{red}}}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}Y\textstyle{Y\ignorespaces\ignorespaces\ignorespaces\ignorespaces}R⁑(βˆ…π’Ÿ)\textstyle{R\left(\varnothing_{\mathcal{D}}\right)\ignorespaces\ignorespaces\ignorespaces\ignorespaces}R​L​(Y),\textstyle{RL\left(Y\right),}

where the right vertical map is the unit map of YY and the bottom horizontal map is the image under RR of the essentially unique map βˆ…π’Ÿβ†’L⁑(Y)\varnothing_{\mathcal{D}}\to L\left(Y\right). The top horizontal map ρ:Yredβ†’Y\rho\colon Y^{\operatorname{\scriptsize{red}}}\to Y exhibits YredY^{\operatorname{\scriptsize{red}}} as a co-localization of YY with respect to π’žred\mathcal{C}^{\operatorname{\scriptsize{red}}} (dual to T.5.2.7.6).

[05WX]

Proof. First, we show that YredY^{\operatorname{\scriptsize{red}}} is in fact reduced. Applying LL to the defining diagram of YredY^{\operatorname{\scriptsize{red}}} and using the fact that LL preserves pullbacks, we see that the map L⁑(Yred)β†’L​R​(βˆ…)L\left(Y^{\operatorname{\scriptsize{red}}}\right)\to LR\left(\varnothing\right) is the pullback of the map L⁑(Y)β†’L​R​L​(Y)L\left(Y\right)\to LRL\left(Y\right), which is an equivalence (from the fact that the counit L​R​(Y)β†’YLR\left(Y\right)\to Y is an equivalence, the zig-zag identities and the 2-out-of-3 property). It follows that the map L⁑(Yred)β†’L​R​(βˆ…)L\left(Y^{\operatorname{\scriptsize{red}}}\right)\to LR\left(\varnothing\right) is an equivalence, but L​R​(βˆ…)β†’βˆ…LR\left(\varnothing\right)\to\varnothing is an equivalence as well (since RR is fully faithful) and we are done.

Now, we show that ρ\rho is a co-localization. Let ZZ be a reduced object. We have a homotopy pullback diagram of spaces

Map⁑(Z,Yred)\textstyle{\operatorname{Map}\left(Z,Y^{\operatorname{\scriptsize{red}}}\right)\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}Map⁑(Z,Y)\textstyle{\operatorname{Map}\left(Z,Y\right)\ignorespaces\ignorespaces\ignorespaces\ignorespaces}Map⁑(Z,R⁑(βˆ…))\textstyle{\operatorname{Map}\left(Z,R\left(\varnothing\right)\right)\ignorespaces\ignorespaces\ignorespaces\ignorespaces}Map⁑(Z,R​L​(Y))\textstyle{\operatorname{Map}\left(Z,RL\left(Y\right)\right)}

and we note that the space of maps from a reduced object to any object in the essential image of RR is contractible. ∎

Original mathematics by the credited authors. Source collection and HTML conversion remain in progress.

Tomer Schlank, Lior Yanovski

Original source: arXiv:1808.06006v3

Original source page 9

Original source Β· 1808.06006v3