Skip to the Main Content

Note:These pages make extensive use of the latest XHTML and CSS Standards. They ought to look great in any standards-compliant modern browser. Unfortunately, they will probably look horrible in older browsers, like Netscape 4.x and IE 4.x. Moreover, many posts use MathML, which is, currently only supported in Mozilla. My best suggestion (and you will thank me when surfing an ever-increasing number of sites on the web which have been crafted to use the new standards) is to upgrade to the latest version of your browser. If that's not possible, consider moving to the Standards-compliant and open-source Mozilla browser.

April 13, 2011

Enhanced 2-categories and Limits for Lax Morphisms

Posted by Mike Shulman

If you’re tired of all this type theory and are longing for some good old 2-category theory, this post is for you. Today on the arXiv we have the long-awaited paper:

  • Enhanced 2-categories and limits for lax morphisms, by Stephen Lack and Michael Shulman: arXiv.

The goal of this paper is to characterize the 2-categorical limits which exist in 2-categories of categories-with-structure and morphisms which preserve that structure laxly (up to a not-necessarily-invertible comparison map). However, it turns out that we can get a more useful theorem if instead of 2-categories, we work with richer structures called \mathcal{F}-categories (these are the “enhanced 2-categories” of the title).

Recall that one way to describe a notion of “category with structure” is by giving a 2-monad. And it’s easy to prove that for any ordinary monad TT on an ordinary category CC, the category TAlg=C TT Alg = C^T of TT-algebras inherits any limits that CC has (more precisely, the forgetful functor TAlgCT Alg\to C “creates limits”). We’d like to categorify this statement.

However, if TT is a 2-monad on a 2-category CC, there is more than one choice for what one might mean by a “2-category of TT-algebras”.

  • One thing to do is to “go completely pseudo” and consider the 2-category PsTAlgPs T Alg of pseudo TT-algebras and pseudo-TT-morphisms. We can then prove a “fully bicategorical” version of the usual theorem: PsTAlgPs T Alg inherits any bicategorical limits (bilimits) that CC has. These are limits whose cones commute up to isomorphism, and which satisfy their universal property up to equivalence.

  • On the other hand, we have the 2-category TAlg sT Alg_s of strict TT-algebras and strict TT-morphisms. This is the ordinary CatCat-enriched category of algebras, and as such it inherits all strict 2-categorical limits that CC has. Now bicategorical limits can be modeled by strict 2-limits; a 2-category theorist says that “pseudolimits are also bilimits”. Pseudolimits also have cones commuting up to isomorphism, but satisfy their universal property up to isomorphism rather than equivalence. In particular, if CC is complete as a strict 2-category, so is TAlg sT Alg_s, and both also have bicategorical limits.

    The advantage of strict algebras is that the types of “structure” on categories we consider are often not literally pseudoalgebra structures, but they are strict algebra structures. For instance, the usual notion of “monoidal category,” with a binary tensor product, unit object, and associator and unitor, describes the strict algebras for some strict 2-monad, but not the pseudo algebras for any 2-monad. (If TT is the 2-monad whose strict algebras are strict monoidal categories, then pseudo TT-algebras are “unbiased” monoidal categories; these are “equivalent” to ordinary ones, but only by a nontrivial theorem.)

  • On the other hand, strict monoidal functors between non-strict monoidal categories are rare, so instead of TAlg sT Alg_s it is often better to take a middle road: we consider the 2-category TAlg=TAlg pT Alg = T Alg_p of strict algebras and pseudo morphisms. However, now we no longer have an easy answer to the question “what limits does TAlgT Alg have?” But the answer is known: Blackwell, Kelly, and Power proved (in 2-dimensional monad theory) that TAlgT Alg inherits PIE-limits from CC. A PIE-limit is, roughly, a strict 2-limit which demands no equalities between 1-morphisms (a “non-evil strict 2-limit”). Since pseudolimits are PIE-limits, if CC is a complete strict 2-category (like CatCat) then TAlgT Alg has all bicategorical limits.

  • Sometimes, however, even pseudo morphisms are too strict, and we need to consider lax or colax morphisms, which preserve the structure only up to a not-necessarily-invertible comparison morphism. The resulting 2-categories TAlg lT Alg_l and TAlg cT Alg_c are not generally bicategorically complete, even when C=CatC=Cat. So what limits do they have?

In a previous paper (Limits for lax morphisms), Steve proved that the 2-category TAlg lT Alg_l has colax limits (limits whose cones commute up to not-necessarily-invertible 2-cells, in one of the possible directions), and dually TAlg cT Alg_c has lax limits. This is already a useful thing to know. For example, Eilenberg-Moore objects (objects-of-algebras for internal monads) are a lax limit, so this implies that (for instance) the category of algebras for a colax-monoidal monad on a monoidal category inherits a monoidal structure.

(Amusingly, it also implies the lifting of 1-limits to categories of algebras. For any small class 𝒳\mathcal{X} of 1-limits, there is a 2-monad TT on CatCat whose algebras are categories with 𝒳\mathcal{X}-limits. This 2-monad is colax-idempotent, which means that any functor between TT-algebras is a colax TT-morphism in a unique way; the colax structure is the canonical comparison map T(limF)lim(TF)T(\lim F) \to \lim (T F). Therefore, TAlg cT Alg_c is the 2-category of 𝒳\mathcal{X}-complete categories and all functors between them. It follows that for any monad on an 𝒳\mathcal{X}-complete category, the category of algebras is also 𝒳\mathcal{X}-complete.)

However, Steve also proved that TAlg lT Alg_l has a number of other limits besides colax ones. It has limits of all diagrams of strict morphisms (that is, the inclusion TAlg sTAlg lT Alg_s \to T Alg_l preserves limits). It has inserters of 2-cells fgf\to g for parallel 1-morphisms f,g:ABf,g\colon A\rightrightarrows B if ff is strict. It has equifiers of parallel 2-morphisms α,β:fg:AB\alpha,\beta\colon f \rightrightarrows g \colon A \rightrightarrows B if ff is strict. It has comma objects (fg)(f\downarrow g) if ff is strict. Moreover, all of these limits in TAlg lT Alg_l have a curious property: there is some specified collection of projections from the limit which are strict TT-morphisms, and which detect strictness in the sense that a map into the limit is strict if and only if its composites with these distinguished projections are all strict. (Blackwell-Kelly-Power also proved this latter fact for PIE-limits in TAlgT Alg.)

Note that none of these types of limits can be expressed as purely 2-categorical properties of TAlg lT Alg_l; they all require knowing which of the morphisms in TAlg lT Alg_l are strict. So the limit-structure of TAlg lT Alg_l becomes much richer if we enhance it by supplying the datum of the inclusion TAlg sTAlg lT Alg_s \to T Alg_l. The first main idea of the current paper is that instead of regarding this datum as a 2-functor TAlg sTAlg lT Alg_s \to T Alg_l, we can regard it as a single enriched category.

The category we are enriching over, denoted \mathcal{F}, has as objects categories equipped with a full subcategory, which is to say functors that are fully faithful and injective on objects. Thus an \mathcal{F}-category has a collection of objects, and (like a 2-category) between any two objects it has a hom-category — but also it has a specified full subcategory of that hom-category. In general, we call the morphisms in the specified subcategories tight, and the general morphisms loose; in the case of TAlg lT Alg_l they are of course the strict and lax TT-morphisms, respectively.

It turns out that all the odd-looking types of limits that Steve found TAlg lT Alg_l to admit, including both the requirements that some of the morphisms in the diagram be strict and the additional universal property that some of the projections are strict and detect strictness, can be expressed exactly as certain weighted limits in \mathcal{F}-enriched category theory. I find this absolutely amazing!

Once we get over our amazement, however, we realize that all we’ve done so far is introduce a language in which to talk about the limits which TAlg lT Alg_l has. We still need to prove a theorem about which ones those actually are; there are plenty of \mathcal{F}-limits that TAlg lT Alg_l generally doesn’t have.

To motivate the statement of this theorem, let’s go back to pseudo morphisms and PIE-limits. Recall that I said the PIE-limits were roughly “limits that don’t involve any equalities between morphisms”. Now a given type of 2-limit is described by a weight, which consists of a functor Φ:DCat\Phi\colon D \to Cat. There is a “projective” model structure on such weights, and one might guess that the PIE-weights would be the cofibrant objects therein—but that isn’t quite right. The cofibrant weights, which 2-categorists traditionally call flexible, are slightly more general than PIE-weights; splitting of idempotents is flexible, but not PIE. This makes some sense, since cofibrant objects are like projective things, which are usually retracts of free things, while PIE-weights are like free things.

(In general, TAlgT Alg does not admit all flexible limits: idempotent pseudo morphisms need not split. Interestingly, though, TAlgT Alg does have all flexible limits if TT itself is flexible as a 2-monad.)

Now free things are also like cell complexes (while projective things are like retracts of cell complexes), and we also know another way to characterize cell complexes. Namely, in an algebraic weak factorization system, the algebraically cofibrant objects are usually closely related to cell complexes. By definition, an object is “algebraically cofibrant” when it is a coalgebra for a cofibrant replacement comonad, usually denoted QQ. (By contrast, an object is cofibrant just when the counit QXXQ X \to X has a section, which is just the counit condition for a coalgebra; algebraicity adds the coassociativity condition.) It turns out that the PIE-weights are precisely the algebraically cofibrant objects for the model structure on weights. (Richard Garner and John Bourke also noticed this independently.)

Finally, one last important observation is that the cofibrant replacement comonad on weights has a universal property: it is the classifier for pseudo morphisms. In other words, for weights Φ:DCat\Phi\colon D \to Cat and Ψ:DCat\Psi\colon D \to Cat, to give a pseudo natural transformation ΦΨ\Phi \to \Psi is precisely to give a strict natural transformation QΦΨQ\Phi\to \Psi. A section of QΨΨQ \Psi \to \Psi, therefore, enables us to make any pseudonatural transformation out of Φ\Phi into a strict one, and this is essentially how we can show that TAlgT Alg admits PIE-limits. (This was not Blackwell-Kelly-Power’s original proof, however!)

By the way, this is in line with the general model-category philosophy that “weak maps” are maps from a cofibrant replacement to a fibrant one; in the projective model structure on weights, all objects are fibrant. There is also a dual “injective” model structure in which all objects are cofibrant and the fibrant replacement is a pseudo morphism coclassifier.

Now there is also a lax morphism classifier comonad Q lQ_l, with the corresponding property that lax natural transformations ΦΨ\Phi \to \Psi are precisely strict natural transformations Q lΦΨQ_l\Phi\to \Psi. And dually there is a colax morphism classifier Q cQ_c. So it is entirely reasonable to guess (with all the hindsight-inspired lead-up that I’ve given here) that the limits which TAlg lT Alg_l admits will have something to do with Q cQ_c-coalgebras. (You might have initially said Q lQ_l-coalgebras, but it turns out to be Q cQ_c; recall that TAlg lT Alg_l has colax limits, not lax ones.)

In fact, this is a theorem: a type of 2-categorical limit lifts to TAlg lT Alg_l, for any 2-monad TT, if and only if its weight is a Q cQ_c-coalgebra. That’s just a 2-categorical theorem, though, and as we saw above, the limit structure of TAlg lT Alg_l becomes much richer if we regard it as an \mathcal{F}-category instead. But we can mimic the above development for \mathcal{F}-weights as well, defining lax, colax, and pseudo \mathcal{F}-natural transformations and morphism classifier \mathcal{F}-comonads. (There are some tricky details here, but we’ll ignore them.) And we can prove that an \mathcal{F}-categorical limit lifts to TAlg lT Alg_l for all TT if and only if its weight is an \mathcal{F}-categorical Q cQ_c-coalgebra… plus an extra somewhat curious condition, which roughly says that “all the projections from the limit object are generated by the tight ones.”

We call a Q cQ_c-coalgebra satisfying this extra condition rigged (or more precisely “ll-rigged”). The intuition for this is that being rigged isn’t just about “being strict”—or, in \mathcal{F}-categorical language, “being tight”. Rather, it’s like the rigging on a ship: the parts that should be tight are tight, but the parts that should be loose are loose, and the two interact in just the right way. To explain the actual definition of rigging would take too much space (although it’s not really that complicated), so you’ll have to read the paper. It all works out quite nicely — although I should mention that some pretty weird-looking weights can still be rigged; there are some examples in section 6 of the paper.

There is one last thing I should say, since if I don’t, someone will probably ask about it. In the current paper, Steve and I restricted ourselves to enhanced 2-categories of strict and lax (or colax) morphisms. But clearly one could combine pseudo and lax instead; or strict, pseudo, and lax; or even all four types of morphisms! We fully expect analogous theorems to hold in these cases (although it may take some extra ideas to combine lax+colax). There should even be analogous theorems for pseudoalgebras over pseudomonads and \mathcal{F}-enriched bicategories. It’s just that the strict+lax case was the easiest one to start with, and made the paper long enough by itself.

Posted at April 13, 2011 8:26 AM UTC

TrackBack URL for this Entry:   http://golem.ph.utexas.edu/cgi-bin/MT-3.0/dxy-tb.fcgi/2385

4 Comments & 2 Trackbacks

Re: Enhanced 2-categories and Limits for Lax Morphisms

In a previous paper, Steve proved that the 2-category TAlg lT Alg_l has colax limits (limits whose cones commute up to not-necessarily-invertible 2-cells, in one of the possible directions), and dually TAlg cT Alg_c has lax limits. This is already a useful thing to know. For example, Eilenberg-Moore objects (objects-of-algebras for internal monads) are a lax limit, so this implies that (for instance) the category of algebras for a colax-monoidal monad on a monoidal category inherits a monoidal structure.

Let me see if I am following.

We have the 2-monad T:CatCatT : Cat \to Cat in 2CAT2 CAT such that TAlg cMonCat colaxT Alg_c \simeq MonCat_{colax} is the 2-category of monoidal categories and colax monoidal functors.

Now we have some colax monoidal 1-monad R:CCR : C \to C in CatCat, on a monoidal category CC. This is equivalently a 1-monad R:CCR : C \to C in TAlg cT Alg_c Ordinarily we would have said that its category of algebras is

RAlg barelaxlim(*TAlg cforgetCat), R Alg_{bare} \simeq laxlim(* \to T Alg_c \stackrel{forget}{\to} Cat) \,,

where on the right we have the functor that picks the object CTAlg cC \in T Alg_c and whose lax unit is RR after forgetting the monoidal structure everywhere. This gives us the bare category of algebras. But in fact since lax limits exist in TAlg c=MonCat colaxT Alg_c = MonCat_{colax} we can form

RAlglaxlim(*TAlg c), R Alg \simeq laxlim(* \to T Alg_c) \,,

and thus find RAlgTAlg cR Alg \in T Alg_c as a monoidal category. Since the forgetful functor is right 2-adjoint, I suppose, we have that indeed it maps RAlgR Alg to RAlg bareR Alg_{bare}, hence that indeed RAlgR Alg is RAlg bareR Alg_{bare} with its monoidal structure made explicit.

Is that the implication that you mean?

Posted by: Urs Schreiber on April 14, 2011 7:58 AM | Permalink | Reply to this

Re: Enhanced 2-categories and Limits for Lax Morphisms

Yes, exactly, except that the forgetful functor from TAlg cT Alg_c is not a right adjoint. (Not in the usual sense, anyway; I suppose it might be in some “lax” sense, but I don’t know whether that is true, whether anyone has made it precise, or whether it implies any preservation of (lax) limits). The fact that RAlgR Alg lies over RAlg bareR Alg_bare is not a formal consequence of the mere existence of EM-objects in TAlg cT Alg_c, but part of Steve’s theorem: not only does TAlg cT Alg_c have lax limits, but these limits are preserved by the forgetful functor. I should probably have mentioned that. Our theorem for \mathcal{F}-categories is the same.

Interestingly, it is not true that the forgetful 2-functor U:TAlg cCatU\colon T Alg_c \to Cat creates lax limits, in (what I believe to be) the usual sense. Specifically, although every diagram in TAlg cT Alg_c which has a lax limit in CatCat also has a lax limit in TAlg cT Alg_c which is preserved by UU, UU does not also reflect lax limits. This is in contrast to the case of 1-monads, where since the forgetful functor is conservative, it reflects any limits that it preserves. This is also something that gets better for \mathcal{F}-categories: the forgetful functor from the \mathcal{F}-category of strict+lax morphisms is conservative, at least on tight isomorphisms (these being strict algebra maps) and that is enough.

Posted by: Mike Shulman on April 14, 2011 6:24 PM | Permalink | PGP Sig | Reply to this

Re: Enhanced 2-categories and Limits for Lax Morphisms

So the limit-structure of TAlg lT Alg_l becomes much richer if we enhance it by supplying the datum of the inclusion TAlg sTAlg lT Alg_s \to T Alg_l.

This reminds me of the definition-as-a-2-functor of 2-category equipped with proarrows . Isn’t it quite similar? Can one grasp the similarity in a precise way?

The cofibrant weights, which 2-categorists traditionally call flexible, […] It turns out that the PIE-weights are precisely the algebraically cofibrant objects for the model structure on weights.

Hopefully this means that the flexible-weight-weighted limits are equivalent to PIE-limits, only that in PIE-limits more choices are made explicit?

How do you think of the model category theory entering the theory of 2-limits here? Are we seeing a piece of a shadow of genuine (,2)(\infty,2)-category theory? There ought to be an intrinsic analog of your theorem, one would hope, which does not mention model structures and resolutions, but does model \infty-weights on (,1)Cat(\infty,1)Cat-enriched category theory, or the like. Do you have a hunch of what such a statement might be?

Posted by: Urs Schreiber on April 14, 2011 8:21 AM | Permalink | Reply to this

Re: Enhanced 2-categories and Limits for Lax Morphisms

This reminds me of the definition-as-a-2-functor of 2-category equipped with proarrows.

Good eye! Yes, modulo strictness issues, a 2-category equipped with proarrows is precisely an \mathcal{F}-category in which every tight morphism has a loose right adjoint. (We remark on this in the introduction to the paper.) There are some interesting directions to go with this. For instance, the prototypical such \mathcal{F}-category ProfProf, whose tight morphisms are functors and whose loose ones are profunctors, has at least some rigged \mathcal{F}-weighted colimits.

Hopefully this means that the flexible-weight-weighted limits are equivalent to PIE-limits, only that in PIE-limits more choices are made explicit?

If I understand what you mean correctly, then I think the dual case of algebraically fibrant objects may be leading your intuition astray. In a cofibrantly generated AWFS, every fibrant object can be made algebraically fibrant by making choices. But this is not always true on the cofibrant side: not every projective is free, and not every retract of a cell complex is a cell complex.

How do you think of the model category theory entering the theory of 2-limits here?

There is actually no model category theory in the paper itself. I only brought it up to motivate the characterization of PIE-weights as coalgebras for the pseudo-transformation-classifier QQ. In the lax and colax cases, the corresponding colax and lax transformation classifier comonads are not cofibrant replacements in any model structure; they are purely 2-categorical (or \mathcal{F}-categorical) constructions.

I do think the results should have analogues for (,2)(\infty,2)-categories, once one has a particular model in mind which is tractable for carrying over the constructions. One should be able to define an (,2)(\infty,2)-category of algebras and lax or colax morphisms for any (,2)(\infty,2)-monad, construct classifier (,2)(\infty,2)-comonads for lax and colax (,2)(\infty,2)-transformations, and so on. Maybe one’s model for (,2)(\infty,2)-categories will involve model categories, or maybe not, but I think the fact that model categories can also be used to give a description of the pseudo transformation classifier is a bit of a red herring.

For instance, if we don’t try to go \infty but just make “everything pseudo” in the 2-categorical world, i.e. work with fully weak 2-monads on bicategories, pseudo algebras, and pseudo morphisms, then the notions of PIE-weight and flexible weight are no longer relevant. The bicategory of pseudoalgebras and pseudomorphisms for any pseudomonad should inherit all bilimits, by a simple 2-dimensional analogue of the standard theorem for 1-categories. However, studying the bicategory of lax or colax morphisms for such a pseudomonad will still require theory of this sort, and should be benefited by the introduction of \mathcal{F}-bicategories.

Posted by: Mike Shulman on April 14, 2011 6:42 PM | Permalink | PGP Sig | Reply to this
Read the post Two dimensional monadicity
Weblog: The n-Category Café
Excerpt: An extension of the monadicity theorem using F-categories to detect lax, pseudo, and colax morphisms provides new insight into the fundamental nature of such weak morphisms.
Tracked: January 11, 2013 8:00 PM
Read the post On Two-Dimensional Monad Theory
Weblog: The n-Category Café
Excerpt: Describes the approach to two-dimensional universal algebra taken in the paper of Blackwell, Kelly, and Power on two-dimensional monad theory.
Tracked: April 28, 2014 10:32 PM

Post a New Comment