Homological stratification and descent

This paper introduces a homological stratification theory for rigidly-compactly generated tensor-triangulated categories, demonstrating its robust descent properties to provide a unified framework for answering when stratification descends and extending equivariant module spectrum results from finite groups to compact Lie groups.

The Gist

Imagine you are trying to understand a massive, complex city. This city isn't made of buildings and roads, but of mathematical objects called "categories." These objects interact, combine, and transform in intricate ways. Mathematicians want to map this city, to understand its neighborhoods, its hidden connections, and how to classify every single object within it.

This paper, written by Barthel, Heard, Sanders, and Zou, introduces a new, powerful way to map these mathematical cities. They call it "Homological Stratification."

Here is the breakdown of their ideas using simple analogies:

1. The Old Map vs. The New Map

For a long time, mathematicians had a standard map of this city called the Balmer Spectrum. It was like a street map that worked well in "nice" neighborhoods (where the rules were orderly and predictable). However, in messy, chaotic, or infinite neighborhoods, this old map would break down or miss important details.

The authors introduce a new map called the Homological Spectrum.

• The Old Map: Good for tidy, finite cities.
• The New Map: Works everywhere, even in the wildest, most chaotic parts of the mathematical universe. It doesn't care if the neighborhood is "noisy" or "infinite"; it just looks at the fundamental "residue fields" (the basic building blocks) to figure out where things belong.

2. What is "Stratification"? (The Layer Cake)

"Stratification" is just a fancy word for layering. Imagine the city is a giant layer cake.

• The Goal: To prove that if you know the ingredients of each layer, you know the whole cake.
• The Problem: Sometimes, layers are glued together so tightly that you can't tell where one ends and the next begins.
• The Solution: The authors show that their new map (Homological Stratification) can perfectly separate these layers. It proves that every object in the city belongs to a specific "layer" defined by its support (where it lives in the city).

3. The Superpower: "Descent" (The Traveling Agent)

This is the paper's biggest breakthrough. In math, "Descent" is like a traveling agent who visits different towns to gather information.

• The Scenario: Imagine you want to understand the whole country (the big category $T$), but it's too big to study all at once. So, you send agents to smaller towns ($S_i$) to study them.
• The Old Rule: If the agents found that the towns were "stratified" (well-organized), it didn't always guarantee the whole country was well-organized. The connection was weak.
• The New Rule: The authors prove that with their Homological Stratification, if the agents find the towns are well-organized, the whole country is automatically well-organized.
• The Metaphor: It's like saying, "If every single brick in a wall is perfectly square, then the whole wall is perfectly straight." Their method works so well that it unifies all previous attempts to prove this, making the "traveling agent" job much easier.

4. The "Nerves of Steel" Conjecture

There is a famous open question in this field called the "Nerves of Steel" conjecture.

• The Question: Is the old map (Balmer) and the new map (Homological) actually the same map? Do they show the exact same number of points?
• The Answer: The authors show that if the "Nerves of Steel" conjecture is true (meaning the maps are the same), then their new, powerful method (Homological Stratification) is exactly the same as the old, standard method.
• Why it matters: If the conjecture is true, you can use their new, easier-to-use method to solve problems that were previously very hard. If the conjecture is false, their new method is even better because it sees more details than the old map ever could.

5. Real-World Application: The Symmetry Groups

The paper isn't just theory; they apply it to Equivariant Spectra.

• The Analogy: Imagine studying a pattern that repeats itself under rotation or reflection (like a snowflake or a kaleidoscope). These are "symmetry groups."
• The Breakthrough: Previously, mathematicians could only easily map these patterns if the group was finite (like a triangle or a square). The authors extended this to Compact Lie Groups (which include continuous rotations, like a circle or a sphere).
• The Result: They successfully mapped the "city" of these continuous symmetries, proving that the layers can be classified just as neatly as the finite ones.

Summary

Think of this paper as the invention of a universal GPS for mathematical structures.

1. It works everywhere: It doesn't get confused by messy or infinite data.
2. It travels well: If you understand the small parts, you automatically understand the whole.
3. It unifies: It connects different theories that were previously separate.
4. It solves old problems: It finally answers the question, "When does the structure of a big system depend on the structure of its smaller parts?" with a resounding "Always, if you use our new map."

The authors have given mathematicians a robust, flexible tool to organize the chaotic universe of tensor-triangulated categories, ensuring that no matter how complex the system gets, it can be broken down into understandable, manageable layers.

Technical Summary

Here is a detailed technical summary of the paper "Homological Stratification and Descent" by Tobias Barthel, Drew Heard, Beren Sanders, and Changhan Zou.

1. Problem Statement

The paper addresses fundamental limitations in the existing theories of stratification for rigidly-compactly generated tensor-triangulated (tt) categories. Stratification is a classification theory that establishes a bijection between the localizing ideals of a category $\mathcal{T}$ and the subsets of a specific spectrum.

Two primary frameworks exist, both with significant drawbacks:

1. Cohomological Stratification: Relies on the action of a Noetherian ring $R$ (often $\text{End}^*(\mathbf{1})$). It fails when the parametrizing object is not affine or Noetherian.
2. Tensor-Triangular (tt) Stratification: Based on the Balmer spectrum $\text{Spc}(\mathcal{T}^c)$ and Balmer–Favi support. While more general, it requires the spectrum to be "weakly noetherian" (a topological condition) and lacks a unified, general descent theorem. Known descent results (e.g., Zariski, finite étale) are piecemeal.

The Core Question: Under what general conditions does stratification descend? Specifically, if a category $\mathcal{T}$ is covered by a family of categories $\mathcal{S}_i$ that are stratified, is $\mathcal{T}$ stratified?

2. Methodology

The authors introduce a new framework based on Homological Support ($\text{Supp}^h$), defined on the Homological Spectrum ($\text{Spch}(\mathcal{T}^c)$).

• Homological Spectrum: Points correspond to homological residue fields (homological primes). Unlike the Balmer spectrum, it does not require topological assumptions (like weak noetherianity).
• Homological Support: Defined via the non-vanishing of $\text{hom}(t, E_B)$, where $E_B$ is a pure-injective object associated with a homological prime $B$. This contrasts with the "naive" support defined by $E_B \otimes t \neq 0$.
• Homological Cosupport: A dual notion defined via $\text{hom}(E_B, t) \neq 0$.
• Weakly Descendable Functors: The authors generalize the concept of descent. A family of geometric functors $(f_i^*: \mathcal{T} \to \mathcal{S}_i)$ is weakly descendable if the unit $\mathbf{1}_{\mathcal{T}}$ lies in the localizing ideal generated by the images of the right adjoints $(f_i)_*$. This condition is weaker than standard "descendability" (which requires compact generation) but sufficient for stratification.

3. Key Contributions and Theoretical Framework

A. Definition of Homological Stratification (h-stratification)

A category $\mathcal{T}$ is h-stratified if the homological support induces a bijection:
$$\text{Supp}^h: \{ \text{localizing ideals of } \mathcal{T} \} \xrightarrow{\sim} \{ \text{subsets of } \text{Spch}(\mathcal{T}^c) \}$$
The authors prove that h-stratification is equivalent to:

1. The Homological Local-to-Global Principle (h-LGP): Every object is generated by its "local" pieces relative to the homological support.
2. Homological Minimality: The localizing ideal generated by each $E_B$ is minimal.

B. The Descent Theorem for h-stratification

The paper's most significant technical achievement is a general descent theorem.

• Theorem: If $(f_i^*: \mathcal{T} \to \mathcal{S}_i)$ is a weakly descendable family of geometric functors and each $\mathcal{S}_i$ is h-stratified, then $\mathcal{T}$ is h-stratified.
• Significance: This unifies and extends all previous descent results (Zariski, étale, nil, etc.) into a single, robust statement without requiring topological assumptions on the spectrum.

C. Comparison with tt-Stratification and the "Nerves of Steel" Conjecture

The authors relate h-stratification to the classical tt-stratification via the canonical map $\pi: \text{Spch}(\mathcal{T}^c) \to \text{Spc}(\mathcal{T}^c)$.

• Nerves of Steel Conjecture: Posits that $\pi$ is a bijection.
• Main Equivalence: If $\text{Spc}(\mathcal{T}^c)$ is weakly noetherian, then $\mathcal{T}$ is tt-stratified if and only if it is h-stratified and the Nerves of Steel conjecture holds.
• Implication: To prove tt-stratification in the weakly noetherian case, one only needs to prove h-stratification (which descends easily) and the Nerves of Steel conjecture.

4. Key Results

1. General Descent (Theorem D): Homological stratification descends along any weakly descendable family. This is a "universal" descent result.
2. Descent for tt-Stratification (Theorem F, G, H): By combining the h-descent theorem with the Nerves of Steel conjecture, the authors derive new descent theorems for tt-stratification:
   • Theorem F: Generalizes étale descent by removing separability conditions.
   • Theorem G: Applies to descendable dualizable commutative algebras in stable $\infty$-categories.
   • Theorem H: Generalizes nil-descent and quasi-finite descent to families of functors.
3. Equivariant Applications (Theorem I): The theory is applied to compact Lie groups. The authors extend previous results (which were limited to finite groups) to show that the stratification of equivariant module spectra $\text{Mod}_G(R_G)$ is determined by the stratification of the underlying non-equivariant ring $R$ and the conjugacy classes of closed subgroups of $G$.
4. Bousfield Lattice Classification: The paper shows that the h-detection property (if $\text{Supp}^h(t) = \emptyset \implies t=0$) is equivalent to the classification of Bousfield classes by subsets of the homological spectrum.

5. Significance

• Unification: The paper provides a uniform treatment of stratification that subsumes cohomological and tensor-triangular approaches.
• Removal of Topological Barriers: By utilizing the homological spectrum, the theory works for categories where the Balmer spectrum is not weakly noetherian, a common issue in non-Noetherian commutative algebra and stable homotopy theory.
• Descent Power: The introduction of "weakly descendable" functors provides the most general setting for descent currently known, resolving the question "When does stratification descend?" in the affirmative for h-stratification.
• New Applications: The extension of stratification results from finite groups to compact Lie groups opens new avenues in equivariant stable homotopy theory.
• Refinement: The theory acts as a refinement of tt-stratification. If the Nerves of Steel conjecture fails (i.e., the homological spectrum has "more points"), h-stratification provides strictly more refined information about the category's ideal structure.

In summary, this work establishes homological stratification as the robust, descent-friendly foundation for tensor-triangular geometry, capable of handling non-Noetherian and non-affine contexts while recovering classical results under standard conjectures.