5
$\begingroup$

It is known that fully faithful functors are closed under pushouts in Cat (e.g. Lemma 4.9 of this paper). Are locally fully faithful 2-functors closed under (strict) 2-pushouts in the 2-category 2-Cat of 2-categories, (strict) 2-functors, and 2-natural transformations? I expect this to be true, but giving an explicit description of a 2-pushout is daunting. Is there a simpler way to reason to prove this by reasoning entirely locally (i.e. in the hom-categories)?

I expect the fully weak setting to be more difficult, but if it is known that locally fully faithful pseudofunctors are closed under pseudopushouts in a bicategory of bicategories, this would also answer my question.

Improve this question
$\endgroup$
8
  • $\begingroup$ To clarify: what precisely is your 2-category 2-Cat? Also, that ff functors are closed under pullback seems like an instance of limits commuting with limits (since a ff functor can be characterised by a limit diagram, IIRC). $\endgroup$
    – David Roberts ♦
    Nov 6, 2021 at 2:08
  • 2
    $\begingroup$ @DavidRoberts: I've clarified my question and corrected a typo: the relevant fact is that fully faithful functors are closed under pushout, not pullback. $\endgroup$
    – varkor
    Nov 6, 2021 at 2:40
  • $\begingroup$ OK, thanks! (I knew offhand that ff+**eso injective on objects** functors are closed under pushout, but I confess I didn't check your link) $\endgroup$
    – David Roberts ♦
    Nov 6, 2021 at 3:13
  • 1
    $\begingroup$ @varkor But I don't know what a 2-pushout is. Is it a cocomma object? Is it a lax pushout? Or is it simply a pushout? Oh... does it mean "Cat-enriched pushout"? $\endgroup$
    – Tim Campion ♦
    Dec 14, 2021 at 18:16
  • 3
    $\begingroup$ Ok -- so that's the same thing as a $Cat$-enriched pushout. Because $Cat$ has cotensors, this is the same thing as a pushout in the underlying 1-category $Cat$. $\endgroup$
    – Tim Campion ♦
    Dec 14, 2021 at 22:58

1 Answer 1

Reset to default
1
$\begingroup$

No. I'm not sure whether you're asking about pushout along an arbitrary 2-functor, or just about pushing out a locally fully faithful functor along another locally fully faithful functor. Either way the answer is no.

Consider the inclusion of 1-categories into 2-categories as the locally discrete ones. This functor is fully faithful. It has a right adjoint, so it preserves pushouts. And it identifies the faithful functors between 1-categories with the locally fully faithful functors between essentially discrete 2-categories. So it suffices to show that faithful functors of 1-categories are not closed under pushout in $Cat$.

So it suffices to show that injective homomorphisms of monoids are not closed under pushout in $Mon$.

For this, if you're asking about pushout along an arbitrary map, then it suffices to consider the pushout of the injection $\mathbb N \to \mathbb Z$ along the map $\mathbb N \to \mathbb N / (2=1)$, which is the non-injection $\mathbb N / (2=1) \to \ast$.

Otherwise, we're pushing out an injective homomorphism along another injective homomorphism. The result need not be an injective homomorphism -- see here (the first-linked paper at Benjamin Steinberg's answer there gives an example with $\leq 4$-element semigroups; adding disjoint unit elements these are $\leq 5$-element monoids).

Improve this answer
$\endgroup$
6
  • $\begingroup$ I'm probably missing something obvious, but why does it suffice to show that faithful functors are not closed under pushout? Why faithful and not fully faithful (which are closed under pushout)? $\endgroup$
    – varkor
    Feb 17, 2022 at 3:32
  • $\begingroup$ I'm including 1-categories into 2-categories as the locally discrete ones. This functor has a right adjoint, so it preserves pushouts. $\endgroup$
    – Tim Campion ♦
    Feb 17, 2022 at 4:12
  • $\begingroup$ It's still not clear to me why it suffices to consider only the faithful functors. I'm interested in locally fully faithful functors: why is it still sufficient if we drop fullness? $\endgroup$
    – varkor
    Feb 19, 2022 at 19:50
  • 1
    $\begingroup$ Let F: A —> B be a functor between 1-categories. Let i(F): i(A) —> i(B) be the corresponding 2-functor between locally discrete 2-categories. Then F is faithful iff i(F) is locally fully faithful. $\endgroup$
    – Tim Campion ♦
    Feb 19, 2022 at 20:00
  • 1
    $\begingroup$ If I were trying to prove your conjecture to be true, this should not suffice. But I’m proving it false, ie coming up with a counterexample. so it does suffice. $\endgroup$
    – Tim Campion ♦
    Feb 19, 2022 at 20:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.