🔢 mathematics

Absolutely flat algebras in tensor categories

This paper investigates the relationship between semisimple, artinian simple, and artinian absolutely flat algebras within tensor categories, demonstrating that these notions coincide for commutative algebras in specific symmetric categories of moderate growth while diverging in general cases, and further provides novel examples of tensor categories with infinite-length objects and a generalization of Deligne's theorem on tannakian categories.

Original authors: Kevin Coulembier, Alexander Sherman

Published 2026-08-14

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Kevin Coulembier, Alexander Sherman

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are a master builder working in a universe where the laws of physics are slightly different. In our world, if you build a structure out of simple, solid bricks, you know exactly how it will behave: it's stable, predictable, and if it falls apart, it breaks into those same simple bricks. Mathematicians have a special toolkit called "algebra" to describe these structures. For centuries, they've known that in the standard world of numbers (like the ones you use to count apples), three specific ways of describing a "perfectly built" structure are actually the same thing. They call these structures "semisimple," "products of simple algebras," and "absolutely flat." Think of them as three different names for the same type of unbreakable, perfectly organized Lego tower.

But what happens if you move your construction site to a stranger, more exotic universe? In this paper, the authors explore a mathematical landscape called "tensor categories." You can think of these as alternative universes where the rules of combining objects are a bit more flexible and strange than in our normal world. The big question the authors ask is: If you build a tower in these exotic universes, do those three names still mean the same thing? Do the towers still break down into simple, perfect pieces? Or does the strange physics of these new worlds cause the rules to break, making a tower look perfect from one angle but actually crumbling from another? This matters because these exotic universes aren't just made-up puzzles; they are the mathematical language used to describe quantum physics, particle interactions, and the deep structure of reality. If the rules change, our understanding of the universe's building blocks might need a rewrite.

The authors, Kevin Coulembier and Alexander Sherman, dive deep into this mystery to see if the "perfect tower" rules hold up in these strange new lands. They find that the answer is a complicated "it depends." In the wild, chaotic corners of these mathematical universes—specifically in categories with "superexponential growth"—the rules definitely break. They construct examples of towers that look like they are made of simple, solid bricks (semisimple) but actually refuse to break apart into those simple pieces. They also find towers that are "simple" but lack the "absolutely flat" property that usually guarantees they are perfectly stable. In these chaotic zones, the three names are no longer synonyms; they describe very different, and sometimes broken, structures.

However, the story gets much more hopeful when they look at a specific, well-behaved neighborhood within these exotic universes: the "moderate growth" categories. These are the places where the mathematical objects don't grow too wildly fast. Here, the authors prove that the old rules actually do survive. If you restrict your building to these moderate zones, a tower that is semisimple is a product of simple algebras, and it is absolutely flat. They show that in these specific, well-ordered environments, the three concepts coincide perfectly, just like they do in our familiar world of numbers.

The paper also tackles a specific type of building block: commutative algebras (structures where the order of combining things doesn't matter). The authors demonstrate that in the chaotic, super-fast-growing universes, you can find a "simple" commutative tower that isn't "absolutely flat," meaning it has hidden weaknesses. But in the moderate growth universes, every simple commutative tower is guaranteed to be absolutely flat and artinian (a technical term meaning it has a finite, manageable size). They even provide a detailed map of what these perfect towers look like in the moderate zones, describing them as being built from specific "exact subgroups" of a larger symmetry group.

In short, the paper draws a line in the sand. It argues that while the beautiful, unified rules of classical algebra can shatter in the most extreme mathematical environments, they remain robust and reliable in the "moderate" regions that are most relevant to our current understanding of physics and geometry. The authors don't just say this might be true; they provide rigorous proofs for the moderate cases and concrete counterexamples for the chaotic ones, showing us exactly where the magic holds and where it breaks.

Technical Summary: Absolutely Flat Algebras in Tensor Categories

1. Problem Statement and Motivation

The paper investigates the structural relationship between three classes of algebras within the ind-completions of tensor categories (rigid abelian monoidal categories): semisimple algebras, artinian simple algebras, and artinian absolutely flat algebras (von Neumann regular algebras).

In classical ring theory over a field $k$ , for an artinian algebra, these three notions are equivalent: an algebra is semisimple if and only if it is a product of simple algebras, if and only if it is absolutely flat. The authors ask whether these equivalences persist in the broader context of tensor categories. Specifically, they address:

Is every semisimple algebra a product of simple algebras?
Is every artinian simple algebra absolutely flat?
Under what conditions do these properties coincide?

The motivation stems from several open questions in the field:

Whether the extension of scalars of a tensor category remains a tensor category (equivalent to asking if field extensions are absolutely flat).
Whether Deligne tensor products of pretannakian categories are always pretannakian.
The Etingof-Ostrik conjecture regarding the "exactness" of simple algebras in finite tensor categories.

2. Methodology and Framework

The authors employ a systematic approach combining categorical algebra, representation theory, and algebraic geometry within the framework of pretannakian categories (symmetric artinian tensor categories).

Definitions: They define the Jacobson radical $J(A)$ , semisimplicity ( $J(A)=0$ ), and absolute flatness ( $Tor^A_1 \equiv 0$ ) in the context of ind-algebras. They distinguish between left/right artinian and bi-artinian properties.
Well-Fibered Categories: A central methodological tool is the concept of well-fibered pretannakian categories. A category $\mathcal{C}$ $C$ is well-fibered if it admits a symmetric tensor functor to a "basic" pretannakian category $\mathcal{D}$ $D$ that is both Geometrically Reductive (GR) and Maximally Nilpotent (MN).
- GR: For every epimorphism $X \twoheadrightarrow 1$ , some symmetric power $S^n X \twoheadrightarrow 1$ splits.
- MN: Simple non-trivial objects have finite symmetric algebras, and non-split extensions of the unit eventually vanish under symmetric powers.
Reduction to Group Schemes: Using Tannakian formalism, the authors reduce problems in well-fibered categories to the study of algebras in $\text{Rep}_f G$ (finite-length representations of an affine group scheme $G$ ) over a base field $k$ .
Algebraic Geometry in Tensor Categories: They utilize the theory of algebraic groups and schemes developed in prior works (e.g., [C2], [CS]) to analyze commutative algebras as coordinate rings of homogeneous spaces $G/H$ .

3. Key Contributions and Results

3.1. General Relations and Counterexamples

The paper establishes that in complete generality, the classical equivalences fail.

Implications: The authors prove that for artinian algebras, absolute flatness implies the vanishing of $\text{Ext}^1$ , which implies the algebra is a product of simple algebras, which implies semisimplicity. However, the reverse implications do not hold universally.
Counterexamples:
- Semisimple but not a product of simples: In the Delannoy category (superexponential growth), there exists a commutative semisimple algebra that is not a product of simple algebras (Lemma 4.5).
- Simple but not absolutely flat: In categories of superexponential growth (e.g., $(\text{Rep } GL)_t$ ), there exist simple commutative ind-algebras that are not absolutely flat (Proposition 4.6).
- Non-artinian simple algebras: The paper provides the first known example of a tensor category with finite-dimensional morphism spaces containing objects of infinite length (Appendix A), highlighting subtleties in the definition of artinian algebras.

3.2. Main Theorems for Moderate Growth

The primary positive results are established for pretannakian categories of moderate growth, specifically those that are well-fibered.

Theorem A (Commutative Case): Let $\mathcal{C}$ be a well-fibered pretannakian category over an algebraically closed field $k$ .

Every simple commutative algebra in $\text{Ind } \mathcal{C}$ is artinian and absolutely flat.
Every semisimple commutative algebra in $\text{Ind } \mathcal{C}$ is a product of simple algebras.
If $\text{char}(k) > 0$ , every semisimple algebra (not necessarily commutative) in $\mathcal{C}$ is a product of simple algebras and is absolutely flat.

This theorem implies that for Frobenius exact pretannakian categories of moderate growth (in positive characteristic) and all such categories in characteristic 0, the classical equivalences hold.

Theorem B (Tannakian/Group Scheme Case): Let $G$ be an affine group scheme over an algebraically closed field $k$ .

For a simple commutative $G$ -equivariant algebra $A$ , let $L$ be the algebraic closure of the field of invariants $A^G$ . There exists an exact subgroup $H \subseteq GL$ such that $A \otimes_{A^G} L \cong L[GL/H]$ .
Semisimple commutative $G$ -equivariant algebras are products of simple ones.
Finite-dimensional semisimple $G$ -equivariant algebras are products of simple ones.

3.3. Structural Characterizations

Artinian Characterization: In well-fibered categories, a commutative ind-algebra is artinian if and only if it satisfies the descending chain condition (DCC) on ideals, or equivalently, if it is a finite direct product of artinian local algebras (Theorem 5.2).
Nakayama Property: The authors utilize the Nakayama property (valid in well-fibered categories) to prove that the Jacobson radical coincides with the intersection of maximal ideals and that simple modules correspond to maximal ideals.

3.4. Appendices

Appendix A: Constructs a symmetric tensor category with finite-dimensional morphism spaces but objects of infinite length, demonstrating that the ind-completion of a tensor category is not always a "nice" tensor category in the sense of having finite-length objects.
Appendix B: Generalizes Deligne's theorem on the neutrality of tannakian categories, extending it from $\text{Vec}$ to arbitrary GR+MN pretannakian categories.
Appendix C: Fills a technical gap regarding extensions of scalars in tensor categories.

4. Significance and Claims

The paper claims to resolve the question of the equivalence of semisimplicity, product-of-simples, and absolute flatness for commutative algebras in pretannakian categories of moderate growth, provided the category is well-fibered.

Conjectural Scope: The authors note that it is a major conjecture in the field (supported by [BEO, C2]) that every pretannakian category of moderate growth is well-fibered. If this conjecture holds, their results apply to all such categories.
Limitations: The paper explicitly states that in categories of superexponential growth (which are not of moderate growth), these equivalences fail.
Open Questions:
- Whether the equivalence holds for non-commutative algebras in super-tannakian categories of characteristic 0 (Question 5.16).
- Whether the counterexamples found in superexponential growth can be refined to distinguish between non-artinian simple algebras and artinian simple algebras that are not absolutely flat (Remark 4.7).

The work provides a rigorous structural description of simple commutative algebras in these categories, identifying them with coordinate rings of homogeneous spaces $G/H$ where $H$ is an exact subgroup, thereby linking abstract tensor category theory to concrete algebraic geometry.