On orthogonal factorization systems and double categories

Branko Juran · 2025

Read On orthogonal factorization systems and double categories by Branko Juran (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-01363.txt.

Introduction

Branko Juran proves a link between two deep structures in category theory. Factorization systems split every map into two set kinds of parts. Double categories track maps in two directions at once. The note shows the world of factorization systems, at the infinity level, sits fully inside the world of double categories. A straightening result for the double side then narrows back to known fibered forms on the factorization side. The work is brief and purely structural, aimed at readers in higher category theory.

Catalog record

CREATORBranko Juran
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE108 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-01363

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

On orthogonal factorization systems and double categories Branko Juran Abstract We prove that the ∞ \infty -category of orthogonal factorization systems embeds fully faithfully into the ∞ \infty -category of double ∞ \infty -categories. Moreover, we prove an (un)straightening equivalence for double ∞ \infty -categories, which restricts to an (un)straightening equivalence for op-Gray fibrations and curved orthofibrations of orthogonal factorization systems. 1 Introduction Orthogonal factorization systems and double categories are very classical objects in category theory, their study goes back to work of MacLane [ Mac50 ] and Ehresmann [ Ehr63 ] , respectively. Their \infty -categorical analogs, introduced by Joyal [ Joy08 ] and by Haugseng [ Hau13 ] , play an equally important role in higher category theory. Both concepts deal with categorical structures equipped with two distingushed classes of morphisms. An orthogonal factorization system is a category together with the choice of two classes of morphisms, so that any morphism can uniquely be factored as a composite of one morphism from the first class, followed by one from the second. Similarly, we can think of a double category as a category with two different types of morphisms, vertical and horizontal morphisms. These two different types of morphisms cannot be composed with each other but part of the data are so-called squares which witness compatibilities between them. From this point of view, it seems like double categories are a generalization of orthogonal factorization systems. The goal of this paper is to make this precise in the context of \infty -categories. We will construct a functor \DCLR\colon\OFS\lhook\mathrel{\mkern-3.0mu}\mathrel{\mathchoice{{}\hbox{$\displaystyle{\meno}$}}{{}\hbox{$\textstyle{\meno}$}}{{}\hbox{$\scriptstyle{\meno}$}}{{}\hbox{$\scriptscriptstyle{\meno}$}}}\mathrel{\mkern-3.0mu}\rightarrow\operatorname{DCat} from the \infty -category of orthogonal factorization systems into the \infty -category of double \infty -categories ( 3.3 ). It sends an orthogonal factorization system to the double \infty -category which has the objects of the underlying \infty -category as its objects, horizontal morphisms are morphisms belonging to the first class, vertical morphisms are morphisms from the second class and squares are commutative squares. The essential image of this functor consists of the double \infty -categories where there is a unique square filling every choice of a “bottom left corner” : {\lx@inpgf@ignorespaces\bullet} {\lx@inpgf@ignorespaces\bullet} {\lx@inpgf@ignorespaces\bullet} {\lx@inpgf@ignorespaces\bullet} This precisely encodes that the “wrong order” composition of morphisms from each class can uniquely be rewritten as composition in the “correct order” , arguably the most important feature of an orthogonal factorization system. We will then prove the following: Theorem A ( Theorem 3.19 ) . The functor \DCLR\colon\OFS\lhook\mathrel{\mkern-3.0mu}\mathrel{\mathchoice{{}\hbox{$\displaystyle{\meno}$}}{{}\hbox{$\textstyle{\meno}$}}{{}\hbox{$\scriptstyle{\meno}$}}{{}\hbox{$\scriptscriptstyle{\meno}$}}}\mathrel{\mkern-3.0mu}\rightarrow\operatorname{DCat} from the \infty -category of orthogonal factorization systems into the \infty -category of double \infty -categories is fully faithful. The essential image consists of the double \infty -categories fulfilling the equivalent conditions from Proposition 3.1 . A similar result in the context of ordinary 1 -categories has recently been obtained by Štěpán [ Ště23 , Thm. 3.30] . We use A to deduce several other results about orthogonal factorization systems. Firstly, we verify that for {\mathcal{C}}^{{\dagger}} an orthogonal factorization system, the functor \DCLR from A induces an equivalence between curved orthofibrations (or op-Gray fibrations, respectively) over {\mathcal{C}}^{{\dagger}} , defined in [ HHLN23a ] , and (cocart,right)-fibrations (or (cart,right)-fibrations, respectively) over \DCLR({\mathcal{C}}^{{\dagger}}) , defined in [ Nui24 ] , ( Proposition 4.5 ): \displaystyle\Ortho\left({\mathcal{C}}^{{\dagger}}\right)\cong\mathrm{CoR}\left(\DCLR\left({\mathcal{C}}^{{\dagger}}\right)\right)\,, (1) showing that we can regard the former as a special case of the latter. Then we prove the following (un)straightening equivalence for these fibrations: Theorem B ( Theorem 4.6 ) . For \mathbb{C} a double \infty -category, let \mathrm{CoR}\left(\mathbb{C}\right)^{\simeq} denote the space of (cocart,right)-fibrations. There is a natural equivalence \mathrm{CoR}\left(\mathbb{C}\right)^{\simeq}\cong\map_{\operatorname{DCat}}\left({(\mathbb{C})}^{2\op},\operatorname{Sq}^{\mathrm{oplax}}\left(\Cat_{1}^{(2)}\right)\right) of functors from double \infty -categories to spaces. Here, \operatorname{Sq}^{\mathrm{oplax}}\left(\Cat_{1}^{(2)}\right) is the large double \infty -category having \infty -categories as its objects, functors of \infty -categories as horizontal and vertical morphisms and natural transformations as squares. This confirms an expectation outlined in [ Nui24 , Remark 2.14] . Combining ( 1 ) and B yields an (un)straightening equivalence for fibrations of orthogonal factorization systems. Among orthogonal factorization there are those which are called adequate (in the sense of Barwick [ Bar17 ] ), those admitting certain pullbacks that make it possible to define the span category. We check that under the embedding from A , an orthogonal factorization system {\mathcal{C}}^{{\dagger}} is adequate if and only if {\DCLR({\mathcal{C}}^{{\dagger}})}^{1\op} , the opposite in the horizontal direction of the associated double \infty -category, is also contained in the essential image of \DCLR . In this case, there is an equivalence {\DCLR\left({\mathcal{C}}^{{\dagger}}\right)}^{1\op}\cong\DCLR\left(\mathrm{Span}\left({\mathcal{C}}^{{\dagger}}\right)\right)\,. We hence recover the span category construction of adequate orthogonal factorization systems, which usually involves a certain amount of simplicial combinatorics. In fact we do so by computing the whole automorphism group of the \infty -category of adequate orthogonal factorization systems \OFS^{\bot} , refining [ HHLN23a , Cor. 5.7] : Theorem C ( Theorem 5.5 ) . There is an equivalence of groups \Aut\left(\OFS^{\bot}\right)\cong\mathbb{Z}/2\mathbb{Z} with generator given by the span category functor. Organization of the paper In Section 2 , we collect some background material on double \infty -categories and orthogonal factorization systems. Section 3 is dedicated to the proof of A . In Section 4 , we study fibrations, proving B . Finally, we discuss adequate orthogonal factorization systems in Section 5 and prove C . Conventions • We use the theory of \infty -categories as developed by Lurie in [ Lur09a ] and [ Lur17 ] . However, we make no essential use of the concrete model and the arguments should apply in most other models. Since most of the categories appearing in this paper are higher categories, we use the term category to refer to an \infty -category and say (1,1) -category if the mapping spaces are discrete. • We fix three nested Grothendieck universes U\in V\in W . We call a category very large if it is defined in W and large if is defined in V . If we just say category, we assume that it is defined in U . • We write \mathcal{S} for the large category of spaces (or animae, homotopy types, \infty -groupoids, …) and \Cat_{1} for the large category of categories. Moreover, we write (-)^{\simeq}\colon\Cat_{1}\rightarrow\mathcal{S} for the core functor. • The simplex category is denoted by \Delta . Its objects are finite posets \left[n\right]=\left\{0,1,\dots,n\right\} . We write d_{i}\colon\left[n-1\right]\rightarrow\left[n\right] for the injective map omitting i and \rho_{i}\colon\left[1\right]\rightarrow\left[n\right] for the map sending 0 to i-1 and 1 to i . We use the same notation to denote the category obtained from this poset via the inclusions \operatorname{Poset}\hookrightarrow\Cat_{1} . • We use \Lambda_{2}^{2}=(0\rightarrow 2\leftarrow 1) for the cospan (1,1) -category (i.e. the 2 -horn of the 2 -simplex). • For \mathcal{C} a category, we write \operatorname{Ar}\left(\mathcal{C}\right)=\Fun(\left[1\right],\mathcal{C}) for the arrow category of \mathcal{C} . • We write \PSh(\mathcal{C})=\Fun(\mathcal{C}^{\op},\mathcal{S}) for the large presheaf category of a category \mathcal{C} . • A full subcategory is called reflective if the inclusion functor admits a left adjoint. By [ Lur09a , Prop. 5.5.4.15] , a subcategory of a large presentable category is reflective if it can be characterized as the class of local objects with respect to a small set of morphisms. 2 Preliminaries Definition 2.1 . A simplicial space is a presheaf on \Delta , i.e. a functor X\colon\Delta^{\op}\mathrel{\mathchoice{{}\hbox{$\displaystyle{\meno}$}}{{}\hbox{$\textstyle{\meno}$}}{{}\hbox{$\scriptstyle{\meno}$}}{{}\hbox{$\scriptscriptstyle{\meno}$}}}\mathrel{\mkern-3.0mu}\rightarrow\mathcal{S}\,. We call it a Segal space if \prod_{i=1}^{n}\rho_{i}\colon X_{n}\mathrel{\mathchoice{{}\hbox{$\displaystyle{\meno}$}}{{}\hbox{$\textstyle{\meno}$}}{{}\hbox{$\scriptstyle{\meno}$}}{{}\hbox{$\scriptscriptstyle{\meno}$}}}\mathrel{\mkern-3.0mu}\rightarrow X_{1}\times_{X_{0}}\dots\times_{X_{0}}X_{1} is an equivalence. We call a Segal space complete if {\lx@inpgf@ignorespaces X_{0}} {\lx@inpgf@ignorespaces

Read the complete text