10
$\begingroup$

Let $\mathcal A$ be a (small) non-zero abelian category equipped with a monoidal structure $\otimes$ which is right-exact in each variable. (Maybe feel free to assume more if that makes things easier -- for instance maybe assume more colimit-preservation properties, or even that $\otimes$ is symmetric monoidal.)

Question: Does there exist a functor $F: \mathcal A \to Ab$ which is nonzero, exact, and lax monoidal?

I think something like this comes up working over a field in the theory of Tannakian categories. But I'm interested in not working over a field. Moreover, I only want the functor to be lax monoidal -- it need not be strong monoidal. I also don't need the functor to be faithful, just nonzero.

This seems related to asking about algebra structures on injective modules. For instance, to answer my question affirmatively, it would suffice to show that $\operatorname{Ind}(\mathcal A^{\mathrm{op}})$ contains a nonzero injective object $I$ admitting an algebra structure (for then you can take $F = \operatorname{Hom}(-,I)$). But according to wikipedia, in general the injective hull of an algebra need not be an algebra, so perhaps the answer to my question is no?

Improve this question
$\endgroup$
2
  • $\begingroup$ @MartinBrandenburg If $\mathcal A = QCoh(X)$ or $\mathcal A = Sh(X)$, then then you could take $F$ to be the stalk functor at your favorite point. (The nonzeroness is maybe a rather weak desideratum! Although it implies anyway that the unit is carried to something nonzero -- if it didn't, I'd be asking for that property specifically.) $\endgroup$
    – Tim Campion
    Commented Dec 8, 2022 at 19:42
  • 2
    $\begingroup$ Sorry I deleted my first comment. My first impression was (and still is) that this is too good to be true. Also, it is a bit weird that $\otimes$ is not assumed to be exact, even though you want $F$ to be exact. But in case it turns out to be true, we will probably need that $\mathcal{A}$ is small. Here is an idea: Take any full monoidal subcategory $\mathcal{P} \subseteq \mathcal{A}$. Then $\mathcal{F} := \bigoplus_{Q \in \mathcal{P}} \hom(Q,-)$ can be endowed with a lax monoidal structure. It is non-zero iff $\mathcal{A}$ is non-zero. If $\mathcal{P}$ consists of projectives, $F$ is exact. $\endgroup$
    – Martin Brandenburg
    Commented Dec 8, 2022 at 20:45

0

Reset to default

You must log in to answer this question.

Browse other questions tagged .