Proposition of Higher Algebra identifies $Mod_{Sp}(Pr^L)$, the symmetric monoidal category of right modules over the monoidal category $Sp$ of spectra in $Pr^L$ the category of presentable categories, with the full subcategory $Pr^L_{St}$ of stable presentable infinity categories. In particular Lurie proves that the functor $-\otimes Sp: Pr^L \to Pr^L_{St}$ is a localization functor.

Question 1: Is there an analogue for ordinary (non $\infty$) categories? In other words, is $-\otimes Ab: Pr^L \to Mod_{Ab}(Pr^L)\cong Pr^L_{Ab}$ a localization functor from the (ordinary) category of presentable categories to the (ordinary) category of presentable abelian categories?

Proof strategy: I believe that Lurie's argument can be modified to prove that $-\otimes Ab$ is a localization functor if I could prove that every presentable category that is a right $Ab$-module is also abelian.

I can verify that $C$ is pre-abelian:

  • Since $C$ is presentable we know that that for any $a \in Ab$ the functor $- \otimes a$ has a right adjoint $[a, -]$. Thanks to the appendix of this paper (shared with me by Zhen Lin) we have that $C$ is enriched and tensored in $Ab$. Moreover, the $Ab$ action corresponds to the tensor.

  • Since $C$ is presentable, it is complete and cocomplete. In particular it has a $0$ object, biproducts, kernels, and cokernels.

I cannot verify that $C$ is actually abelian which requires that:

  • Every monomorphism is a kernel and every epimorphism is a cokernel.

Alternate strategy:

In Lurie's proof (summarized well in 5.3.1 of Groth's notes) he uses the description of $C \otimes Sp = Fun^R(C^{op}, Sp)$ and the fact that $Sp = lim(S \xrightarrow{\Omega} S \xrightarrow{\Omega} \ldots)$ to show that $C \otimes Sp$ is the stabilization $Sp(C)$. Then he shows stabilization of a stable category is itself.

Mimicking this, I can show that $C \otimes \mathbb{Z} \cong Fun^R(C^{op}, \mathbb{Z})$ which is obviously pre-abelian since it is presentable and $Ab$-enriched.

  • Is $C \otimes \mathbb{Z}$ abelian? I think it might be because it is a subcategory of a presheaf category.

  • If $C$ is already abelian does the tensor map $C \otimes \mathbb{Z} \to C$ give an equivalence?

  • $\begingroup$ For the first property, I believe that the result in the appendix of this paper is what you want. $\endgroup$
    – Zhen Lin
    Sep 4, 2013 at 8:09

1 Answer 1


A locally presentable category $\mathcal{C}$ has a (unique) structure of an $Ab$-module if and only if it is additive. Such a category need not be abelian.

This is one reason to prefer the setting of stable $\infty$-categories to the theory of abelian categories. The identification of presentable stable infty-categories with $Sp$-modules implies (among other things) that there is a robust theory of tensor products of presentable stable infty-categories. In the abelian setting, developing an analogous theory requires some fairly restrictive assumptions.

Edit: Actually, I think it's not as bad as I suggested; there's a well-behaved tensor product on Grothendieck abelian categories. But that's a smaller collection of categories than the $Ab$-modules.


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.