The Canonical Model Structure on Cat/Gpd

James Leslie - The Canonical Model Structure on Cat/Gpd
Posted on March 6, 2021 by James Leslie

When one first starts studying category theory, they learn that the correct notion of β€œthe same” for categories is that of equivalence, rather than the stronger notion isomorphism. For those that have studied some algebraic topology, the situation is very similar to identifying spaces if they are homotopic, rather than homeomorphic. This is often given as an analogy, but the two notions can be reconciled by giving model structures where the weaker form of identification forms the subcategory of weak equivalences. Here, we define and present the so called canonical model structure on the category of categories. The proof of each step given can also be applied to the category of groupoids.

We denote the category of (small) categories by π‚πšπ­\mathbf{Cat} and the category of (small) groupoids 𝐆𝐩𝐝\mathbf{Gpd}.

This proof can be found on the nLab’s β€œcanonical model structure on Cat” page, but here we go into a lot more detail of the constructions.

The model structure

Let CC be a category and (A,B)(A, B) a pair of subcategories. The pair (A,B)(A, B) is a weak factorisation system if:

  1. Both AA and BB are stable under retracts;
  2. AA is in the left lifting class of BB (AβŠ‚LBA \subset LB);
  3. Every morphism f:Xβ†’Yf : X \to Y in CC factors as f=pif = pi with i∈Ai \in A and p∈Bp \in B.

A model structure on a category π’ž\mathcal{C} consists of a triple of wide subcategories W,F,CW, F, C whose morphisms are called weak equivalences (anotated with a ∼\sim), fibrations (annotated with a double head β† \twoheadrightarrow) and cofibrations (annotated with a tail ↣\rightarrowtail), respectively. This triple must satisfy the following three axioms:

  1. WW satisfies 2-out-of-3;
  2. (F,W∩C)(F, W \cap C) is a weak factorisation system;
  3. (F∩W,C)(F \cap W, C) is a weak factorisation system.

A model category is a finitely bicomplete category, along with a model structure on it.

Before giving the main result, we need one more definition.

Let F:π’œβ†’β„¬F: \mathcal{A} \to \mathcal{B} be a functor. We say FF is an isofibration if for every aβˆˆπ’œa \in \mathcal{A} and isomorphism f:Faβ†’bβˆˆβ„¬f : Fa \to b \in \mathcal{B}, there is an isomorphism g:aβ†’aβ€²βˆˆπ’œg: a \to a' \in \mathcal{A} such that Ff=gFf = g.

Our main result is the following:

The category π‚πšπ­\mathbf{Cat} has a model structure on it with:

  1. WW being the class of categorical equivalences;
  2. FF being the class of isofibrations;
  3. CC being the class of functors injective on objects.

The rest of this section will be proving the above theorem. That π‚πšπ­\mathbf{Cat} and 𝐆𝐩𝐝\mathbf{Gpd} has all finite limits follows from it having a terminal object, binary products and equalisers. The theory of categories and groupoids are generalised algebraic theories, so the category of models (π‚πšπ­\mathbf{Cat} and 𝐆𝐩𝐝\mathbf{Gpd}, respectively) has all (finite) colimits.

We now break down each of the defining features of a model structure and show that they hold here.

2-out-of-3

Categorical equivalences are closed under 2-out-of-3.

This follows by basic diagram chasing. If we have the following commuting triangle, with any two maps being equivalences, we can take pseudo-inverses to construct a pseudo-inverse to the third map.

Retracts

The class of categorical equivalences is stable under retracts.

Suppose we have the following commuting diagram where gg is an equivalence. We show ff is also an equivalence:

Since gg is an equivalence, there exists a functor g′g' such that gg′≅1Vgg' \cong 1_V and g′g≅1Ug'g \cong 1_U. The morphism pg′jpg'j is then easily seen to be a pseudo-inverse to ff, showing that it is also an equivalence.

The class of isofibrations is stable under retracts.

Suppose the following diagram commutes and gg is an isofibration. We need to show that ff is an isofibration also:

Let x∈Xx \in X and Ξ±:fxβ†’y\alpha : fx \to y be an isomorphism in YY. We apply jj to get an isomorphism jΞ±:jfx=gixβ†’jyj\alpha : jfx=gix \to jy. As gg is an isofibration, there is an isomorphism Ξ²:ixβ†’u∈U\beta : ix \to u \in U such that gΞ²=jΞ±g\beta = j\alpha. Applying pp, we have an isomorphism pΞ²:pix=xβ†’pu∈Xp\beta : pix=x \to pu \in X. If we apply ff to this map, we have:

Hence, given an isomorphism Ξ±\alpha, there exists an isomorphism pΞ²p\beta such that fpΞ²=Ξ±fp\beta = \alpha, so ff is also an isofibration.

The class of functors injective on objects is stable under retracts.

Suppose we have the following diagram in π‚πšπ­\mathbf{Cat}, where gg is a functor injective on objects. We need to show that ff is injective on objects also:

Suppose xx and yy are objects in XX such that fx=fyfx = fy. We have the following chain of equalities:

Hence ff is also injective on objects.

Lifting properties

We first give a lemma that will be useful in solving the lifting problems in this section.

Isofibrations that are equivalences are exactly the equivalences that are surjective on objects.

Suppose that F:π’œβ†’β„¬F: \mathcal{A} \to \mathcal{B} is an isofibration and an equivalence. To see that it is surjective on objects, let b∈Bb \in B. As FF is an equivalence, it is fully faithful, so there is some aβˆˆπ’œa \in \mathcal{A} and isomorphism Ξ²:Faβ†’bβˆˆβ„¬\beta : Fa \to b \in \mathcal{B}. As FF is an isofibration, there is an isomorphism Ξ±:aβ†’aβ€²βˆˆπ’œ\alpha : a \to a' \in \mathcal{A} such that FΞ±=Ξ²F\alpha = \beta. Then, we must have that Faβ€²=bFa' = b, so FF is surjective on objects.

If FF is an equivalence that is surjective on objects, it is also an isofibration. Let aβˆˆπ’œa \in \mathcal{A} and Ξ²:Faβ†’b\beta : Fa \to b be an isomorphism in ℬ\mathcal{B}. As FF is surjective on objects, b=Faβ€²b = Fa' for some aβ€²βˆˆπ’œa' \in \mathcal{A}, so Ξ²\beta is a map Faβ†’Faβ€²Fa \to Fa'. By FF being full, there is a morphism Ξ±:aβ†’aβ€²\alpha : a \to a' such that FΞ±=Ξ²F\alpha = \beta. By FF being faithful, this Ξ±\alpha must be an isomorphism, so FF is an isofibration.

The class of functors injective on objects has the left lifting property with respect to functors that are isofibrations and equivalences.

Suppose we have the following commuting diagram, where ff is injective on objects and gg is an equivalence and isofibration:

As this commutes, we have the following ``object square’’ commuting in π’πžπ­\mathbf{Set}:

In particular, f0f_0 is injective and by Lemma 10, g0g_0 is surjective. As (inj,surj)(\text{inj}, \text{surj}) form a weak factorisation system on π’πžπ­\mathbf{Set}, there is a lift h0h_0. We now aim to turn h0h_0 into a functor. Let Ξ±:yβ†’yβ€²\alpha: y \to y' be a morphism in YY. We then have a morphism jΞ±:jy=gh0yβ†’gh0yβ€²=jyβ€²βˆˆV0j\alpha : jy = gh_0y \to gh_0y'=jy' \in V_0. As gg is fully faithful, there is a unique morphism Ξ²:h0yβ†’h0yβ€²βˆˆU\beta : h_0 y \to h_0 y' \in U such that gΞ²=jΞ±g\beta = j\alpha. We define hΞ±:=Ξ²h\alpha := \beta. As jj is a functor, it must preserve identity morphisms and as gg is fully faithful, the lift of an identity morphism from vv to uu must be the identity, so hh preserves identity morphisms. Similarly, by functorality of jj and fully faithfulness of gg, we have that composites are mapped to composites, so hh is a functor and makes the bottom triangle commute:

The top triangle commutes on objects, so we need to check it commutes on maps. Given Ξ±:xβ†’xβ€²βˆˆX\alpha : x \to x' \in X,

So iΞ±=hfΞ±i \alpha = h f \alpha by fully faithfulness of gg, meaning that the top triangle commutes. This means that hh is indeed a lift, which shows that functors injective on objects are in the left lifting class of fucntors that are equivalences and isofibrations.

The class of functors injective on objects which are also equivalences has the left lifting property with respect to functors that are isofibrations.

Suppose we have the following diagram commuting, with ff an equivalence that is injective on objects and gg an isofibration:

To construct a functor h:Yβ†’Uh:Y \to U, we use the axiom of choice to get some extra structure. As ff is essentially surjective, for every object y∈Yy \in Y, we choose an isomorphism Ξ±y:fxyβ†’y\alpha_y : fx_y \to y, picking the identity morphism wherever possible, i.e Ξ±fx=1fx\alpha_{fx} = 1_{fx} (we call the argument of ff xyx_y, which is well defined at ff is injective on objects). As gg is an isofibration, we choose for every y∈Yy \in Y a morphism Ξ²y:ixyβ†’uy∈U\beta_y : ix_y \to u_y \in U such that gΞ²y=jΞ±yg\beta_y = j\alpha_y, again, picking the identity morphism whenever possible, i.e Ξ²fx=1ix\beta_{fx} = 1_{ix}. On objects, we define h(y)=uyh(y) = u_y. For morphisms, the output is a little more complicated. Starting with a map Ξ³:yβ†’yβ€²\gamma : y \to y', we form the composite Ξ±yβ€²βˆ’1Ξ³Ξ±y:fxyβ†’fxyβ€²\alpha_{y'}^{-1} \gamma \alpha_y : fx_y \to fx_{y'}. We let Ξ±yβ€²βˆ’1Ξ³Ξ±yΒ―:xyβ†’xyβ€²\overline{\alpha_{y'}^{-1} \gamma \alpha_y} : x_y \to x_{y'} be the unique map that ff maps to Ξ±yβ€²βˆ’1Ξ³Ξ±y\alpha_{y'}^{-1} \gamma \alpha_y. Then, applying ii, we have a map iΞ±yβ€²βˆ’1Ξ³Ξ±yΒ―:ixyβ†’ixyβ€²i \overline{\alpha_{y'}^{-1} \gamma \alpha_y} : ix_y \to ix_{y'}. We can then compose with maps Ξ²yβˆ’1\beta_y^{-1} and Ξ²yβ€²\beta_{y'} to get Ξ²yβ€²i(Ξ±yβ€²βˆ’1Ξ³Ξ±yΒ―)Ξ²yβˆ’1:uyβ†’uyβ€²\beta_{y'} i \left(\overline{\alpha_{y'}^{-1} \gamma \alpha_y}\right) \beta_{y}^{-1} : u_y \to u_{y'}. This is what we define hΞ³h\gamma to be.

To see that hh is a functor, we see that it sends the identity to the identity:

We also see that hh preserves composites: let γ:y→y′\gamma : y \to y' and δ:y′→y″\delta : y' \to y''. Then:

We now show that it makes both the triangles commute:

Given x∈Xx \in X, we need ix=ufxix = u_{fx}, however this follows from our choice of maps being the identity wherever possible, which means Ξ²fx=1ix\beta_{fx} = 1_{ix}. Likewise, for maps Ξ³:xβ†’xβ€²\gamma: x \to x', our chosen isomorphisms are the identity, giving hfΞ³=iΞ³hf\gamma = i\gamma. Now, if y∈Yy \in Y, we see by definition that guy=jygu_y = jy, so the bottom triangle commutes on objects. Given a map Ξ³:yβ†’yβ€²βˆˆY\gamma : y \to y' \in Y, by the properties of Ξ²y\beta_y we have:

Hence hh is a lift, as required.

Factorisations

Every functor can be factored as a functor injective on objects, followed by a functor that is an isofibration and equivalence.

Given f:Xβ†’Yf:X \to Y, we form a new category ZZ with objects given by X0∐Y0X_0 \coprod Y_0. We define the hom-sets as follows:

We then have a functor X→ZX \to Z which is the identity on objects, and ff on maps, which, in particular, is injective on objects. We can also construct another functor Z→YZ \to Y which is formed by applying ff to objects from XX and the identity to objects from yy. It is then the identity morphism on maps, which means it is fully faithful. It is also surjective on objects as every object in YY is mapped to itself by this functor. The composite of these two morphisms is ff on objects and ff on maps, hence it is a factorisation of ff into a functor injective on objects, followed by a surjective equivalence (which by Lemma 10) is an equivalence and isofibration), as required.

Every functor can be factored as a functor injective on objects and an equivalence, followed by an isofibration.

Given f:Xβ†’Yf:X \to Y we form the comma category F↓1YF \downarrow 1_Y, then take the full subcategory spanned by objects (x,y,Ο•:fxβ†’y)(x, y, \phi : fx \to y) where Ο•\phi is an isomorphism in YY. Denote this category as F↓≅1YF \downarrow_{\cong} 1_Y Note that we have a functor Xβ†’F↓≅1YX \to F \downarrow_{\cong} 1_Y defined by sending x↦(x,fx,1fx)x \mapsto (x , fx, 1_{fx}) and Ξ±:xβ†’x′↦(Ξ±,fΞ±)\alpha : x \to x' \mapsto (\alpha, f\alpha). This is clearly injective on objects and fully faithful. We also have that every (x,y,Ο•)β‰…(x,fx,1fx)(x, y, \phi) \cong (x, fx, 1_{fx}) by the following commuting square in YY (with horizontal maps being isomorphisms):

We have a functor f↓≅1Yβ†’Yf \downarrow_{\cong} 1_Y \to Y given by projecting out the second component. This functor is also seen to be an isofibration: given (x,y,Ο•)(x, y, \phi) and an isomorphism Ξ±:yβ†’yβ€²βˆˆY\alpha : y \to y' \in Y, the following square commutes, with horizontal maps isomorphisms, so is an isomorphism in f↓≅1Yf \downarrow_{\cong} 1_Y:

Our projection functor sends this isomorphism to Ξ±\alpha, so it is an isofibration. The composite of the two functors defined is easily seen to be equal to ff, completing the factorisation of ff as an equivalence that is injective on objects, followed by an isofibration.

Putting it all together

Functors injective on objects and equivalences that are isofibrations, form a weak factorisation system on π‚πšπ­\mathbf{Cat}.

Lemma 9 shows functors injective on objects are stable under retracts. Lemmas 9 and 9 show that equivalences that are isofibrations are also stable under retracts. Lemma 13 gives the desired lifting data and Lemma 14 gives the desired factorisation.

Equivalences that are injective on objects and isofibrations form a weak factorisation system on π‚πšπ­\mathbf{Cat}.

Lemmas 7 and 9 show that the class of equivalences that are injective on objects is stable under retracts. Lemma 8 shows the class of isofibrations are stable under retracts. Lemma 12 gives the desired lifting data and Lemma 14 gives the desired factorisation.

The category π‚πšπ­\mathbf{Cat} has a model structure on it with:

  1. WW being the class of equivalences;
  2. FF being the class of isofibrations;
  3. CC being the class of functors injective on objects.

Lemma shows that WW satisfies 2-out-of-3. Corollaries 15 and 16 show that (C∩W,F)(C \cap W, F) and (C,F∩W)(C, F \cap W) are weak factorisation systems.

This proves Theorem 5. To see that this restricts to 𝐆𝐩𝐝\mathbf{Gpd} we only need to realise that if XX and YY are groupoids, the categories constructed in the Factorisations section are also groupoids. This means we have the following:

The category 𝐆𝐩𝐝\mathbf{Gpd} has a model structure on it with:

  1. WW being the class of categorical equivalences;
  2. FF being the class of isofibrations;
  3. CC being the class of functors injective on objects.