The monadic tower for \(\infty\)-categories (Q2069827)

It is shown that every right adjoint functor \(G:\mathcal{D}\rightarrow \mathcal{C}\) between presentable \(\infty\)-categories decomposes canonically as a \textit{coreflection} followed by possibly transfinitely many \textit{monadic functors}. Therefore we have the \textit{monadic tower} of \(G\) \[ \mathcal{D}\rightarrow\left( \dots\rightarrow\mathcal{C}_{\alpha} \rightarrow\dots\rightarrow\mathcal{C}_{2}\rightarrow\mathcal{C} _{1}\rightarrow\mathcal{C}_{0}=\mathcal{C}\right) \] where the functor \(\mathcal{C}_{\alpha+1}\rightarrow\mathcal{C}_{\alpha}\) is monadic for every ordinal \(\alpha\) and \[ \mathcal{C}_{\alpha} \xrightarrow{\sim} \varprojlim_{\beta<\alpha}\mathcal{C}_{\beta} \] for every limit ordinal \(\alpha\). The principal objective in this paper is to establish the following theorem as well as its two corollaries. The corresponding results for \textit{ordinary} categories are already familiar [\textit{H. Applegate} and \textit{M. Tierney}, Rep. Midwest Category Semin. 4, Lect. Notes Math. 137, 56--99 (1970; Zbl 0319.18004); \textit{E. J. Dubuc}, Kan extensions in enriched category theory. Berlin-Heidelberg-New York: Springer-Verlag (1970; Zbl 0228.18002); \textit{J. L. MacDonald} and \textit{A. Stone}, Cah. Topologie Géom. Différ. Catégoriques 23, 197--213 (1982; Zbl 0491.18002); \textit{J. Adámek} et al., J. Pure Appl. Algebra 59, No. 2, 111--123 (1989; Zbl 0675.18001)]. Theorem (Monadic Convergence). Let \(G:\mathcal{D}\rightarrow\mathcal{C}\) be a functor between presentable \(\infty\)-categories, which admits a left adjoint \(F\). Then we have \begin{itemize} \item[1.] The monadic tower of \(G\) stablizes for \(\alpha\gg0\) on \(\mathcal{C} _{\infty}:=\varprojlim_{\alpha}\mathcal{C}_{\alpha}\). \item[2.] The induced functor \(\mathcal{D}\rightarrow\mathcal{C}_{\infty}\) admits a fully faithful left adjoint \(\mathcal{C}_{\infty}\rightarrow\mathcal{D}\). \item[3.] The image of \(\mathcal{C}_{\infty}\rightarrow\mathcal{D}\) is the subcategory \(\mathcal{D}_{\infty}\subseteq\mathcal{D}\) generated under colimits by \(F(\mathcal{C}) \subseteq\mathcal{D}\). \end{itemize} Corollary (Transfinite Monadicity). A right adjoint functor between presentable \(\infty\)-categories is conservative iff it is a transfinite composition of monadic functors. For a general non-conservative functor \(G:\mathcal{D}\rightarrow\mathcal{C}\), the monadic convergence theorem provides a factorization of \(G\) as \[ \mathcal{D}\xrightarrow{G_{\infty}} \mathcal{D}_{\infty} \xrightarrow{\overline{G}} \mathcal{C} \] where \(G_{\infty}\) admits a fully faithful left adjoint \(F_{\infty}\) and \(\overline{G}\) is a transfinte composition of monadic functors. Corollary (Colocalization Sequence). Given \(G:\mathcal{D}\rightarrow\mathcal{C}\) in the \(\infty\)-category \(\mathrm{Pr}^{R}\) of presentable \(\infty\)-categories and right adjoints, the counits of the adjunctions \[ G:\mathcal{D}\leftrightarrows\mathcal{C}_{\alpha}:F_{\alpha} \] in the monadic tower of \(G\) assemble into a transfinite sequence in \(\mathcal{D}/Y\) \[ \begin{array} [c]{ccccccccccc} F_{0}G_{0}(Y) & \rightarrow & F_{1}G_{1}(Y) & \rightarrow & F_{2}G_{2}(Y) & \rightarrow & \dots & \rightarrow & F_{\alpha}G_{\alpha}(Y) & \rightarrow & \dots\\ \downarrow & & \downarrow & & \downarrow & & \dots & & \downarrow & & \dots\\ Y & = & Y & = & Y & = & \dots & = & Y & = & \dots \end{array} \] which stabilizes on \(F_{\infty}G_{\infty}(Y) \rightarrow Y\). Moreover, for each ordinal \(\alpha\), the term \(F_{\alpha}G_{\alpha}(Y)\) is canonically a colimt of objects in \(F_{\beta}(\mathcal{C})\) for \(\beta<\alpha\). This work was inspired by [\textit{E. Dror} and \textit{W. G. Dwyer}, Comment. Math. Helv. 52, 185--210 (1977; Zbl 0352.55007)] on homology localizations and completions of spaces in homotopy theory, where Dwyer and Farjoun (Dror has changed his name to this) constructed for each space \(X\) and a ring \(R\), which is either \(\mathbb{F}_{p}\) or a subring of \(\mathbb{Q}\), a functorial transfinite tower of spaces under \(X\) \[ \begin{array} [c]{ccccccccccc} \dots & = & X & = & \dots & = & X & = & X & = & X\\ \dots & & \downarrow & & \dots & & \downarrow & & \downarrow & & \downarrow\\ \dots & \rightarrow & T_{\alpha}X & \rightarrow & \dots & \rightarrow & T_{2}X & \rightarrow & T_{1}X & \rightarrow & T_{0}X \end{array} \] The author explains how the above long homology sequence can be derived within his setting.