ScalingStacks

2. Definitions

In this brief section, we give precise definitions of the concepts we have introduced in §​​ 1.

0MWD

Definition 2.1. Let ℰ→𝑓ℬ\mathcal{E}\xrightarrow{f}\mathcal{B} be a functor of ∞\infty-categories.

  1. (1)

    We say that a morphism e1→𝜑e2e_{1}\xrightarrow{\varphi}e_{2} in ℰ\mathcal{E} is ff-cocartesian (or simply cocartesian, if the functor ff is clear from the context) if it induces a pullback diagram

    ℰe2/{\lx@inpgf@ignorespaces\mathcal{E}_{e_{2}/}}ℰe1/{\lx@inpgf@ignorespaces\mathcal{E}_{e_{1}/}}ℬf(e2)/{\lx@inpgf@ignorespaces\mathcal{B}_{f(e_{2})/}}ℬf(e1)/{\lx@inpgf@ignorespaces\mathcal{B}_{f(e_{1})/}}−∘φ\scriptstyle{\lx@inpgf@ignorespaces-\circ\varphi}f\scriptstyle{\lx@inpgf@ignorespaces f}f\scriptstyle{\lx@inpgf@ignorespaces f}−∘f(φ)\scriptstyle{\lx@inpgf@ignorespaces-\circ f(\varphi)}

    in 𝒞​at∞{{\mathcal{C}\textup{at}}_{\infty}}. In this case, we call φ\varphi an ff-cocartesian lift (or simply a cocartesian lift) of f⁡(φ)f(\varphi) relative to e1e_{1}. We then say that the functor ff is a cocartesian fibration if every object of

    Fun​([1],ℬ)​×s,ℬ,f​ℰ\textup{Fun}([1],\mathcal{B})\underset{s,\mathcal{B},f}{\times}\mathcal{E}

    admits an ff-cocartesian lift.

  2. (2)

    Dually, we say that a morphism e1→𝜑e2e_{1}\xrightarrow{\varphi}e_{2} in ℰ\mathcal{E} is ff-cartesian (or simply cartesian) if it induces a pullback diagram

    ℰ/e1{\lx@inpgf@ignorespaces\mathcal{E}_{/e_{1}}}ℰ/e2{\lx@inpgf@ignorespaces\mathcal{E}_{/e_{2}}}ℬ/f⁡(e1){\lx@inpgf@ignorespaces\mathcal{B}_{/f(e_{1})}}ℬ/f⁡(e2){\lx@inpgf@ignorespaces\mathcal{B}_{/f(e_{2})}}φ∘−\scriptstyle{\lx@inpgf@ignorespaces\varphi\circ-}f\scriptstyle{\lx@inpgf@ignorespaces f}f\scriptstyle{\lx@inpgf@ignorespaces f}f(φ)∘−\scriptstyle{\lx@inpgf@ignorespaces f(\varphi)\circ-}

    in 𝒞​at∞{{\mathcal{C}\textup{at}}_{\infty}}. In this case, we call φ\varphi an ff-cartesian lift (or simply a cartesian lift) of f⁡(φ)f(\varphi) relative to e2e_{2}. We then say that the functor ff is a cartesian fibration if every object of

    Fun​([1],ℬ)​×t,ℬ,f​ℰ\textup{Fun}([1],\mathcal{B})\underset{t,\mathcal{B},f}{\times}\mathcal{E}

    admits an ff-cartesian lift.

0MWE

Remark 2.2. Let ℰ→𝑓ℬ\mathcal{E}\xrightarrow{f}\mathcal{B} be a functor of ∞\infty-categories. The crucial consequence of a morphism e1→𝜑e2e_{1}\xrightarrow{\varphi}e_{2} in ℰ\mathcal{E} being ff-cocartesian is that for any object e∈ℰe\in\mathcal{E}, the induced diagram

homℰ⁡(e2,e){\lx@inpgf@ignorespaces\hom_{\mathcal{E}}(e_{2},e)}homℰ⁡(e1,e){\lx@inpgf@ignorespaces\hom_{\mathcal{E}}(e_{1},e)}homℬ⁡(f⁡(e2),f⁡(e)){\lx@inpgf@ignorespaces\hom_{\mathcal{B}}(f(e_{2}),f(e))}homℬ⁡(f⁡(e1),f⁡(e)){\lx@inpgf@ignorespaces\hom_{\mathcal{B}}(f(e_{1}),f(e))}

is a pullback square in 𝒮\mathcal{S}. Dually, the crucial consequence of a morphism e1→𝜑e2e_{1}\xrightarrow{\varphi}e_{2} in ℰ\mathcal{E} being ff-cartesian is that for any object e∈ℰe\in\mathcal{E}, the induced diagram

homℰ⁡(e,e1){\lx@inpgf@ignorespaces\hom_{\mathcal{E}}(e,e_{1})}homℰ⁡(e,e2){\lx@inpgf@ignorespaces\hom_{\mathcal{E}}(e,e_{2})}homℬ⁡(f⁡(e),f⁡(e1)){\lx@inpgf@ignorespaces\hom_{\mathcal{B}}(f(e),f(e_{1}))}homℬ⁡(f⁡(e),f⁡(e2)){\lx@inpgf@ignorespaces\hom_{\mathcal{B}}(f(e),f(e_{2}))}

is a pullback square in 𝒮\mathcal{S}. (Recall Example 1.6, and see [Lur09, Proposition 2.4.1.10].)

0MWF

Remark 2.3. A cocartesian fibration ℰ→𝑓ℬ\mathcal{E}\xrightarrow{f}\mathcal{B} whose fibers f−1​(b)f^{-1}(b) are all ∞\infty-groupoids is called a left fibration. These assemble into the full subcategory ℒ​ℱ​ib​(ℬ)⊂co​𝒞​ℱ​ib​(ℬ)\mathcal{LF}\textup{ib}(\mathcal{B})\subset\textup{co}\mathcal{C}\mathcal{F}\textup{ib}(\mathcal{B}). Correspondingly, the (covariant) Grothendieck construction restricts to an equivalence

Fun​(ℬ,𝒞​at∞){\lx@inpgf@ignorespaces\textup{Fun}(\mathcal{B},{{\mathcal{C}\textup{at}}_{\infty}})}co​𝒞​ℱ​ib​(ℬ){\lx@inpgf@ignorespaces\textup{co}\mathcal{C}\mathcal{F}\textup{ib}(\mathcal{B})}Fun​(ℬ,𝒮){\lx@inpgf@ignorespaces\textup{Fun}(\mathcal{B},\mathcal{S})}ℒ​ℱ​ib​(ℬ){\lx@inpgf@ignorespaces\mathcal{LF}\textup{ib}(\mathcal{B})}Gr∼\scriptstyle{\lx@inpgf@ignorespaces\sim}∼\scriptstyle{\lx@inpgf@ignorespaces\sim}Gr

from the ∞\infty-category of functors ℬ→𝒮\mathcal{B}\rightarrow\mathcal{S} valued in the ∞\infty-category of spaces to the ∞\infty-category ℒ​ℱ​ib​(ℬ)\mathcal{LF}\textup{ib}(\mathcal{B}) of left fibrations over ℬ\mathcal{B}. As a result of the fact that all morphisms in a space are equivalences, given a left fibration ℰ→𝑓ℬ\mathcal{E}\xrightarrow{f}\mathcal{B}, every morphism in ℰ\mathcal{E} is ff-cocartesian.

Dually, a cartesian fibration whose fibers are all ∞\infty-groupoids is called a right fibration, the contravariant Grothendieck construction restricts to an equivalence

Fun​(ℬo​p,𝒞​at∞){\lx@inpgf@ignorespaces\textup{Fun}(\mathcal{B}^{op},{{\mathcal{C}\textup{at}}_{\infty}})}𝒞​ℱ​ib​(ℬ){\lx@inpgf@ignorespaces\mathcal{C}\mathcal{F}\textup{ib}(\mathcal{B})}Fun​(ℬo​p,𝒮){\lx@inpgf@ignorespaces\textup{Fun}(\mathcal{B}^{op},\mathcal{S})}ℛ​ℱ​ib​(ℬ){\lx@inpgf@ignorespaces\mathcal{RF}\textup{ib}(\mathcal{B})}Gr−\scriptstyle{\lx@inpgf@ignorespaces\textup{Gr}^{-}}∼\scriptstyle{\lx@inpgf@ignorespaces\sim}∼\scriptstyle{\lx@inpgf@ignorespaces\sim}Gr−\scriptstyle{\lx@inpgf@ignorespaces\textup{Gr}^{-}}

from the ∞\infty-category of functors ℬo​p→𝒮\mathcal{B}^{op}\rightarrow\mathcal{S} to the full subcategory ℛ​ℱ​ib​(ℬ)⊂𝒞​ℱ​ib​(ℬ)\mathcal{RF}\textup{ib}(\mathcal{B})\subset\mathcal{C}\mathcal{F}\textup{ib}(\mathcal{B}) of right fibrations over ℬ\mathcal{B}, and given a right fibration ℰ→𝑓ℬ\mathcal{E}\xrightarrow{f}\mathcal{B}, every morphism in ℰ\mathcal{E} is ff-cartesian.

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

Aaron Mazel-Gee

Original source: arXiv:1510.02402v1