Abelian versus triangulated quotients

Henning Krause · 2025

Read Abelian versus triangulated quotients by Henning Krause (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-07194.txt.

Introduction

Henning Krause, writing alone in two thousand twenty-five, shows a bridge between two great worlds of categories: those built on sets with sums, and those built on triangles. The note sits in deep algebra. He proves that a place-based cut of a triangle world yields a like cut of a sum world, when one passes to their sum forms. From that, a growth of the base world gives a case where a cut exists just in the big world.

Catalog record

CREATORHenning Krause
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE16 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-07194

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Abelian versus triangulated quotients Henning Krause Abstract . It is shown that any localisation of triangulated categories induces (up to an equivalence) a localisation of abelian categories when one passes to their abelianisations. From this one obtains for any enlargement of Grothendieck universes an example of an abelian category and a Serre subcategory within the smaller universe such that the corresponding quotient does only exist within the bigger universe. The second part of this note provides an analogue for the abelian hull of an arbitrary category. 2020 Mathematics Subject Classification 18E10 (primary); 18E35, 18G80 (secondary) 1. Introduction Localisations of categories naturally arise in the study of abelian and triangulated categories [ 4 , 8 ] , but the construction of a localisation raises some set-theoretic issues. More specifically, given a category \mathcal{C} and a class of morphisms \Sigma in \mathcal{C} , there is a universal functor \mathcal{C}\to\mathcal{C}[\Sigma^{-1}] that inverts all morphisms in \Sigma ; see [ 5 ] . For the localised category \mathcal{C}[\Sigma^{-1}] it is a priori not clear that for any given pair of objects the morphisms between them form a set (and not a proper class). Fortunately there are explicit criteria in the literature which ensure that there is no reason to worry, but often this issue is ignored. For an abelian category \mathcal{A} and a Serre subcategory \mathcal{C} the quotient category \mathcal{A}/\mathcal{C} is by definition the localisation \mathcal{A}[\Sigma^{-1}] with respect to the class \Sigma of morphisms with kernel and cokernel both in \mathcal{C} ; see [ 4 , III.1] . In this note we present examples of such a quotient such that for some pair of objects the morphisms between them do not form a set. For a concise exposition it is convenient to use the setting of Grothendieck universes. 2. Abelian quotients We fix a Grothendieck universe \mathfrak{U} which is sufficiently big to contain the set of integers; see [ 4 , I.1] . If not mentioned otherwise, we assume throughout that all categories are \mathfrak{U} -categories , which means that for any pair of objects X,Y the set \operatorname{Hom}(X,Y) belongs to \mathfrak{U} . Given a pair of categories \mathcal{C},\mathcal{D} we write \operatorname{\mathcal{H}\!\!\;\mathit{om}}(\mathcal{C},\mathcal{D}) for the category of functors \mathcal{C}\to\mathcal{D} , with \operatorname{Hom}(F,G) given by the set of natural transformations F\to G for any pair of functors F,G . This need not be a \mathfrak{U} -category. * * * For a pair \mathcal{A},\mathcal{B} of abelian categories let \operatorname{\mathcal{H}\!\!\;\mathit{om}}_{\mathrm{ex}}(\mathcal{A},\mathcal{B}) denote the full subcategory of \operatorname{\mathcal{H}\!\!\;\mathit{om}}(\mathcal{A},\mathcal{B}) consisting of the exact functors \mathcal{A}\to\mathcal{B} . Given a Serre subcategory \mathcal{C}\subseteq\mathcal{A} , the quotient functor \mathcal{A}\to\mathcal{A}/\mathcal{C} can be characterised as follows. Lemma 1 . [ 4 , III.1] For an exact functor F\colon\mathcal{A}\to\mathcal{A}^{\prime} between abelian categories and a Serre subcategory \mathcal{C}\subseteq\mathcal{A} , the following are equivalent. (1) The functor F induces for any abelian category \mathcal{B} a fully faithful functor \operatorname{\mathcal{H}\!\!\;\mathit{om}}_{\mathrm{ex}}(\mathcal{A}^{\prime},\mathcal{B})\longrightarrow\operatorname{\mathcal{H}\!\!\;\mathit{om}}_{\mathrm{ex}}(\mathcal{A},\mathcal{B}) with essential image consisting of all functors \mathcal{A}\to\mathcal{B} that annihilate \mathcal{C} . (2) The functor F induces an equivalence \mathcal{A}/\mathcal{C}\xrightarrow{\raisebox{-1.20552pt}[0.0pt][0.0pt]{$\scriptstyle{\sim}$}}\mathcal{A}^{\prime} .∎ 3. The abelianisation of a triangulated category Let \mathcal{T} be a triangulated category. Given any abelian category \mathcal{A} we write \operatorname{\mathcal{H}\!\!\;\mathit{om}}_{\mathrm{coh}}(\mathcal{T},\mathcal{A}) for the full subcategory of \operatorname{\mathcal{H}\!\!\;\mathit{om}}(\mathcal{T},\mathcal{A}) consisting of the cohomological functors \mathcal{T}\to\mathcal{A} . Recall from [ 8 , II.3] that the abelianisation \mathcal{T}\to\operatorname{Ab}(\mathcal{T}) is a fully faithful cohomological functor that induces for any abelian category \mathcal{A} an equivalence \operatorname{\mathcal{H}\!\!\;\mathit{om}}_{\mathrm{ex}}(\operatorname{Ab}(\mathcal{T}),\mathcal{A})\xrightarrow{\ \raisebox{-1.20552pt}[0.0pt][0.0pt]{$\scriptstyle{\sim}$}\ }\operatorname{\mathcal{H}\!\!\;\mathit{om}}_{\mathrm{coh}}(\mathcal{T},\mathcal{A}). For a triangulated subcategory \mathcal{S}\subseteq\mathcal{T} we write \mathcal{T}/\mathcal{S} for the corresponding quotient category [ 8 , II.2] . The quotient functor has the following property. Lemma 2 . [ 8 , II, Théorème 2.2.6] Let \mathcal{T} be a triangulated category and \mathcal{S}\subseteq\mathcal{T} a triangulated subcategory. The quotient functor \mathcal{T}\to\mathcal{T}/\mathcal{S} induces for any abelian category \mathcal{A} a fully faithful functor \operatorname{\mathcal{H}\!\!\;\mathit{om}}_{\mathrm{coh}}(\mathcal{T}/\mathcal{S},\mathcal{A})\longrightarrow\operatorname{\mathcal{H}\!\!\;\mathit{om}}_{\mathrm{coh}}(\mathcal{T},\mathcal{A}) with image consisting of all functors \mathcal{T}\to\mathcal{A} that annihilate \mathcal{S} .∎ We obtain a commutative diagram {\lx@inpgf@ignorespaces\mathcal{T}} {\lx@inpgf@ignorespaces\mathcal{T}/\mathcal{S}} {\lx@inpgf@ignorespaces\operatorname{Ab}(\mathcal{T})} {\lx@inpgf@ignorespaces\operatorname{Ab}(\mathcal{T}/\mathcal{S})} and claim that the bottom horizontal functor is equivalent to a quotient functor. Let \operatorname{Ab}_{\mathcal{S}}(\mathcal{T}) denote the Serre subcategory of \operatorname{Ab}(\mathcal{T}) that is generated by the image of the embedding \mathcal{S}\to\mathcal{T}\to\operatorname{Ab}(\mathcal{T}) . Proposition 3 . Let \mathcal{T} be a triangulated category and \mathcal{S}\subseteq\mathcal{T} a triangulated subcategory. The canonical functor \operatorname{Ab}(\mathcal{T})\to\operatorname{Ab}(\mathcal{T}/\mathcal{S}) induces an equivalence \operatorname{Ab}(\mathcal{T})/\operatorname{Ab}_{\mathcal{S}}(\mathcal{T})\xrightarrow{\ \raisebox{-1.20552pt}[0.0pt][0.0pt]{$\scriptstyle{\sim}$}\ }\operatorname{Ab}(\mathcal{T}/\mathcal{S}). Proof. For any abelian category \mathcal{A} we obtain the following commutative diagram. {\lx@inpgf@ignorespaces\operatorname{\mathcal{H}\!\!\;\mathit{om}}_{\mathrm{ex}}(\operatorname{Ab}(\mathcal{T}/\mathcal{S}),\mathcal{A})} {\lx@inpgf@ignorespaces\operatorname{\mathcal{H}\!\!\;\mathit{om}}_{\mathrm{ex}}(\operatorname{Ab}(\mathcal{T}),\mathcal{A})} {\lx@inpgf@ignorespaces\operatorname{\mathcal{H}\!\!\;\mathit{om}}_{\mathrm{coh}}(\mathcal{T}/\mathcal{S},\mathcal{A})} {\lx@inpgf@ignorespaces\operatorname{\mathcal{H}\!\!\;\mathit{om}}_{\mathrm{coh}}(\mathcal{T},\mathcal{A})} \scriptstyle{\lx@inpgf@ignorespaces\wr} \scriptstyle{\lx@inpgf@ignorespaces\wr} The bottom horizontal functor is fully faithful by Lemma 2 . Thus the top horizontal functor is fully faithful, and its essential image is given by the functors that annihilate \operatorname{Ab}_{\mathcal{S}}(\mathcal{T}) . Then the assertion follows from Lemma 1 . ∎ * * * Now assume that there is a universe \mathfrak{V} strictly containing \mathfrak{U} and fix a set I in \mathfrak{V}\smallsetminus\mathfrak{U} . Let \mathbb{F} be a field in \mathfrak{U} and consider the polynomial algebra \Lambda=\mathbb{F}\langle I\rangle . Modules over \Lambda are nothing but \mathbb{F} -vector spaces with a distinguished set of endomorphisms indexed by the elements of I . Let \mathbb{F} denote the trivial \Lambda -module (so the endomorphisms given by the elements of I are zero) and let \mathcal{A} denote the category of \Lambda -modules M that admit a submodule M^{\prime}\subseteq M such that M^{\prime} and M/M^{\prime} are finite direct sums of copies of \mathbb{F} . The category \mathcal{A} is abelian and we write \mathbf{D}^{\mathrm{b}}(\mathcal{A})=\mathbf{K}^{\mathrm{b}}(\mathcal{A})/{\mathbf{Ac}}^{\mathrm{b}}(\mathcal{A}) for its bounded derived category , which is the triangulated quotient of the homotopy category with respect to the subcategory of acyclic complexes. Then \mathbf{K}^{\mathrm{b}}(\mathcal{A}) is a \mathfrak{U} -category, whereas \mathbf{D}^{\mathrm{b}}(\mathcal{A}) is not. For instance, there is an isomorphism \operatorname{Ext}^{1}_{\Lambda}(\mathbb{F},\mathbb{F})\cong\operatorname{Hom}_{\mathbb{F}}(\mathbb{F}[I],\mathbb{F})\cong\mathbb{F}^{I} where \mathbb{F}[I] denotes the free \mathbb{F} -module with basis I . This example is a variation of [ 2 , Example 6.A] due to Freyd. The calculation of \operatorname{Ext}^{1}_{\Lambda}(\mathbb{F},\mathbb{F}) shows that the set of objects of \mathcal{A} belongs to \mathfrak{V} , and therefore \mathbf{D}^{\mathrm{b}}(\mathcal{A}) is a \mathfrak{V} -category. Proposition 4 . The abelian category \operatorname{Ab}(\mathbf{K}^{\mathrm{b}}(\mathcal{A})) is a \mathfrak{U} -category, and there exists a Serre subcategory such that the corresponding quotient is not a \mathfrak{U} -category. Proof. Let \operatorname{Ab}_{\mathrm{ac}}(\mathbf{K}^{\mathrm{b}}(\mathcal{A})) denote the Serre subcategory of \operatorname{Ab}(\mathbf{K}^{\mathrm{b}}(\mathcal{A})) that is generated by all acyclic complexes in \mathbf{K}^{\mathrm{b}}(\mathcal{A}) . Then Proposition 3 yields an equivalence \operatorname{Ab}(\mathbf{K}^{\mathrm{b}}(\mathcal{A}))/\operatorname{Ab}_{\mathrm{ac}}(\mathbf{K}^{\mathrm{b}}(\mathcal{A}))\xrightarrow{\ \raisebox{-1.20552pt}[0.0pt][0.0pt]{$\scriptstyle{\sim}$}\ }\operatorname{Ab}(\mathbf{D}^{\mathrm{b}}(\mathcal{A})) of \mathfrak{V} -categories. We have seen that \mathbf{D}^{\mathrm{b}}(\mathcal{A}) is not a \mathfrak{U} -category, and therefore \operatorname{Ab}(\mathbf{D}^{\mathrm{b}}(\mathcal{A})) is not a \mathfrak{U} -category. ∎ 4. The abelian hull of an arbitrary category For a category \mathcal{C} let \operatorname{add}(\mathcal{C}) denote its additive hull , which is obtained by taking formal finite direct sums of the objects of the \mathbb{Z} -linearisation \mathbb{Z}\mathcal{C} . The canonical functor \mathcal{C}\to\operatorname{add}(\mathcal{C}) induces for any additive category \mathcal{A} an equivalence \operatorname{\mathcal{H}\!\!\;\mathit{om}}_{\operatorname{add}}(\operatorname{add}(\mathcal{C}),\mathcal{A})\xrightarrow{\ \raisebox{-1.20552pt}[0.0pt][0.0pt]{$\scriptstyle{\sim}$}\ }\operatorname{\mathcal{H}\!\!\;\mathit{om}}(\mathcal{C},\mathcal{A}), where \operatorname{\mathcal{H}\!\!\;\mathit{om}}_{\operatorname{add}}(-,-) denotes the full subcategory of additive functors. For an additive category \mathcal{C} let \operatorname{mod}(\mathcal{C}) denote its completion under finite colimits, which is obtained by adding formal cokernels of the morphisms in \mathcal{C} , for instance by using the category of coherent functors \mathcal{C}^{\mathrm{op}}\to\mathrm{Ab} into the category of abelian groups [ 1 ] . The canonical fully faithful functor \mathcal{C}\to\operatorname{mod}(\mathcal{C}) induces for any additive category \mathcal{A} with cokernels an equivalence \operatorname{\mathcal{H}\!\!\;\mathit{om}}_{\operatorname{mod}}(\operatorname{mod}(\mathcal{C}),\mathcal{A})\xrightarrow{\ \raisebox{-1.20552pt}[0.0pt][0.0pt]{$\scriptstyle{\sim}$}\ }\operatorname{\mathcal{H}\!\!\;\mathit{om}}_{\operatorname{add}}(\mathcal{C},\mathcal{A}), where \operatorname{\mathcal{H}\!\!\;\mathit{om}}_{\operatorname{mod}}(-,-) denotes the full subcategory of additive functors that are right exact. The abelian hull of an additive category \mathcal{C} is by definition \operatorname{Ab}(\mathcal{C}):=\operatorname{mod}((\operatorname{mod}(\mathcal{C})^{\mathrm{op}}))^{\mathrm{op}}. This construction goes back to Freyd and Gruson [ 3 , 6 ] . The canonical inclusion \mathcal{C}\to\operatorname{Ab}(\mathcal{C}) induces for

Read the complete text