🔢 mathematics

Descendability and descent in topological weaves

This paper establishes a criterion for descendability of finitely presented surjections of algebraic spaces within topological weaves, which is then applied to prove v-descent for rational motivic sheaves and étale motivic spectra under specific conditions, while also constructing a "forgetting supports" isomorphism for proper DM morphisms of Artin stacks.

Original authors: Adeel A. Khan

Published 2026-07-09

📖 4 min read🧠 Deep dive

CC0 1.0

Original authors: Adeel A. Khan

Original paper dedicated to the public domain under CC0 1.0 (http://creativecommons.org/publicdomain/zero/1.0/). ✨ 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 trying to understand a complex, multi-layered landscape (like a city or a mountain range). In mathematics, specifically in a field called algebraic geometry, these landscapes are called "spaces." To study them, mathematicians use tools called "sheaves," which are like bundles of information or data attached to every point in the space.

This paper is about a specific rule for how we can reconstruct the whole picture of a landscape by looking at smaller, overlapping pieces of it. The author, Adeel A. Khan, introduces a powerful new way to prove that this reconstruction works perfectly.

Here is a breakdown of the paper's ideas using simple analogies:

1. The Core Problem: The "Jigsaw Puzzle"

Imagine you have a giant jigsaw puzzle (the whole mathematical space). You want to know if you can rebuild the entire picture just by looking at a few specific, overlapping sections of it.

The Old Way: Usually, mathematicians check if the pieces fit together by looking at the edges. If the pieces overlap enough, they assume the picture is complete. This is called "descent."
The New Way (Descendability): The author proposes a stronger, more robust condition called "descendability." Think of this not just as checking if the edges match, but proving that the "glue" holding the pieces together is so strong that you can mathematically prove the whole picture must exist if you have the pieces. It's like proving that if you have a few specific Lego bricks, you can build any structure you want with them, not just a specific shape.

2. The Main Discovery: A Universal "Glue" Test

The paper proves a "criterion" (a test) to see if a specific type of map (a way of covering a space with smaller pieces) has this super-strong glue.

The Test: The author shows that if a map works well for two simple types of covers—finite étale covers (like a perfect, non-overlapping tiling of a floor) and finite radicial covers (like a map that looks the same but has hidden, invisible layers)—then it works for every complicated, finitely presented cover.
The Metaphor: Imagine you want to know if a new type of super-glue works on all materials. Instead of testing it on every possible material (wood, metal, glass, plastic), you prove that if it sticks perfectly to "wood" and "plastic," it will automatically stick to everything else. This saves a massive amount of work.

3. The Applications: Fixing "Motivic Sheaves"

The paper applies this "super-glue" test to two very important types of mathematical data called Motivic Sheaves. These are like high-definition, multi-dimensional maps used to study the deep structure of numbers and shapes.

Rational Motivic Sheaves: The paper proves that for these specific maps, you can always reconstruct the whole picture from any reasonable covering. It's like saying, "No matter how you slice this cake, you can always put it back together perfectly."
Étale Motivic Spectra: The paper also shows this works for a different type of map, but only under certain conditions (like when the "residue fields" have a specific, bounded complexity). It's like saying, "This glue works on all cakes, provided the cake isn't too tall or too complicated."

4. The "Forgetting Supports" Trick

The paper ends with a clever application involving "Artin stacks" (which are like spaces with some "fuzzy" or "stacked" points, similar to a pile of papers where some pages are stuck together).

The Problem: There are two ways to move data from a small space to a big space: one way keeps track of where the data came from (like keeping the receipt), and another way just moves the data and "forgets" the receipt. Usually, these two ways give different results.
The Result: The author proves that for certain "proper" (well-behaved) maps between these stacked spaces, the "receipt" doesn't matter. The two ways of moving the data are actually identical.
The Analogy: Imagine moving a box of books from a small room to a big library. Usually, you need a manifest (a list) to know which books came from where. The author proves that for certain types of rooms, you can move the books without the manifest, and the library will still know exactly which books are which. The "forgetting" of the list is safe because the glue (descendability) is so strong that the information is preserved automatically.

Summary

In short, this paper provides a universal key (the criterion for descendability) that unlocks the ability to reconstruct complex mathematical landscapes from their parts. It proves that for several major types of mathematical data, this reconstruction is not just possible, but mathematically guaranteed to be perfect. This allows mathematicians to simplify their proofs and solve problems about "stacked" spaces that were previously very difficult to handle.

Technical Summary: Descendability and Descent in Topological Weaves

Problem Statement
The paper addresses the problem of establishing descent properties for abstract "topological weaves"—a framework for six-functor formalisms on algebraic spaces and stacks introduced in [Kha]. Specifically, it investigates the conditions under which these formalisms satisfy descent for various topologies, most notably the $h$ -topology (generated by étale surjections and finitely presented proper surjections) and the $v$ -topology (universally subtrusive morphisms). While classical results exist for specific cohomology theories (e.g., étale cohomology with torsion coefficients), a general criterion for $h$ -descent in the context of motivic sheaves and spectra, particularly over non-excellent bases or with rational coefficients, remained to be unified and extended.

Methodology
The core methodological innovation is the shift from studying descent directly to studying descendability, a concept adapted from Mathew [Mat].

Descendability: A morphism $f: Y \to X$ is defined as descendable if the unit object $1_X$ in the stable $\infty$ -category $D(X)$ can be constructed from $f_*(1_Y)$ via a finite sequence of operations including cofibres, fibres, direct summands, and tensor products. This is quantified by an "index" $N$ , indicating the depth of the Čech nerve required to reconstruct the unit.
The Criterion: The paper establishes that if a weave satisfies descent for "simple" covers (finite étale and finite radicial surjections), it satisfies descent for all finitely presented surjections. This is achieved through a "bootstrapping" argument (Theorem 1.21) that utilizes stratifications of algebraic spaces and the localization property of the weave.
Application to Motivic Sheaves: The authors apply this criterion to specific motivic categories:
- Rational Motivic Sheaves ( $SH_{\mathbb{Q},+}$ ): By leveraging the fact that these are $\mathbb{Q}$ -linear and oriented, the Euler characteristic of any finite étale cover is invertible, ensuring descendability of index $\le 1$ .
- Étale Motivic Spectra ( $SH_{\text{ét}}$ ): The authors analyze the descent properties over noetherian finite-dimensional schemes with residue fields of uniformly bounded étale cohomological dimension. They utilize rigidity results and the vanishing of the "minus" part of the spectrum after inverting 2 to establish descendability.

Key Contributions and Results

General Descent Criterion (Theorem C): The paper proves that for a topological weave $D$ , the following are equivalent:
- Every finitely presented surjection of qcqs algebraic spaces is descendable.
- Every finite étale surjection and every finite radicial surjection is descendable.
  This reduces the complex problem of $h$ -descent to verifying descendability for two specific classes of morphisms.
Descent for Rational Motivic Sheaves (Theorem A): The presheaf $SH_{\mathbb{Q},+}(-)$ satisfies $v$ -descent (and thus $h$ -descent) on all algebraic spaces. This recovers and extends previous results by Cisinski–Déglise on quasi-excellent schemes to arbitrary algebraic spaces.
Descent for Étale Motivic Spectra (Theorem B): The presheaves $SH_{\text{ét}}(-)$ , $DA_{\text{ét}}(-;\Lambda)$ , and $DM_{\text{ét}}(-;\Lambda)$ satisfy $h$ -descent on noetherian finite-dimensional algebraic spaces whose residue fields have uniformly bounded étale cohomological dimension. For the latter two, this holds for any commutative ring $\Lambda$ . If 2 is not invertible in the coefficients, the result holds for $SH_{\text{ét}}[1/2]$ .
Forgetting Supports Isomorphism (Theorem D): The paper constructs a canonical isomorphism $f_! \simeq f_*$ for proper morphisms $f$ of Artin stacks in the context of rational motivic sheaves.
- Context: This applies when $f$ is representable by Deligne–Mumford (DM) stacks, or when $f$ has tame relative inertia and the base is equicharacteristic.
- Significance: This resolves a gap in previous arguments (e.g., in [Kha3]) where the commutation of $f_!$ with totalizations was assumed without justification. The authors show that because the covering morphisms are descendable, the totalization commutes with exact functors like $f_!$ , thereby validating the isomorphism.

Significance and Claims
The paper claims that the focus on descendability offers distinct advantages over standard descent:

Functoriality: It ensures that the Čech totalization $F \simeq \text{Tot}(f_{\bullet, *} f_\bullet^* F)$ is preserved by any exact functor, not just those preserving limits. This is crucial for proving the $f_! \simeq f_*$ isomorphism.
Stability: Descendability is preserved by symmetric monoidal functors, allowing results to transfer from the universal topological weave ($SH$) to specific oriented $\mathbb{Q}$ -linear weaves.
Generality: The results apply to algebraic spaces and stacks without requiring the base schemes to be quasi-excellent, provided the cohomological dimension of residue fields is bounded in the étale case.

The work unifies the descent theory for various motivic categories, providing a robust framework that handles both rational and integral coefficients (with specific constraints on the base) and extends naturally to the setting of Artin stacks.

1. The Core Problem: The "Jigsaw Puzzle"

2. The Main Discovery: A Universal "Glue" Test

3. The Applications: Fixing "Motivic Sheaves"

4. The "Forgetting Supports" Trick

Summary

More like this