Characterization and finite descent of local cohomological invariants

This paper establishes simple left-inverse characterizations of the singularity invariants $c(Z)$, $w(Z)$, and ${\rm HRH}(Z)$ for equidimensional varieties and utilizes a trace morphism to prove their descent under finite surjective morphisms.

The Gist

Imagine you are an architect trying to understand the structural integrity of a building. Some buildings are perfectly smooth and solid (like a modern glass skyscraper), while others have cracks, jagged edges, or weird corners (like an ancient ruin). In mathematics, these "buildings" are geometric shapes called varieties, and the "cracks" are called singularities.

For a long time, mathematicians had a few ways to measure how "bad" a crack was. But recently, new, more sophisticated tools were invented to measure these cracks with incredible precision. This paper is about three specific new tools (let's call them The Depth Gauge, The Width Ruler, and The Harmony Meter) and how the authors figured out exactly how to use them.

Here is the breakdown of their discovery, explained simply:

1. The Problem: How do we know a building is "good"?

In the past, if you wanted to know if a building had a "Du Bois" crack (a specific type of manageable flaw), you had to check a very complicated list of rules. It was like trying to diagnose a car engine by listening to the noise it makes from a mile away.

The authors wanted a simpler way. They asked: "Is there a simple test, like checking if a key fits in a lock, that tells us immediately if the building is good?"

2. The Solution: The "Left-Handed Key" (Left-Inverse)

The authors discovered a clever trick. Imagine you have a door (the building) and a key (a mathematical map).

• Usually, you try to push the key into the door.
• The authors found that if the building is "good" (has a certain level of smoothness), you can always find a magic reverse key (a "left-inverse") that fits perfectly and unlocks the door from the other side.

If you can find this reverse key, you instantly know the building is structurally sound. If you can't, the building has a deep flaw.

• The Depth Gauge ($c$): Checks if the cracks go deep.
• The Width Ruler ($w$): Checks how wide the cracks are.
• The Harmony Meter ($HRH$): Checks if the cracks are "harmonious" (a combination of depth and width).

The paper proves that for all three of these new tools, this "magic reverse key" test works perfectly. It turns a complex, multi-step math problem into a simple "Yes/No" question: Does the reverse key fit?

3. The Second Discovery: The "Shadow" Effect (Finite Descent)

Now, imagine you have a large, complex sculpture (Building Y) and you shine a light on it to cast a shadow on the wall (Building X). The shadow is a simpler, smaller version of the sculpture.

The authors asked a fascinating question: "If the shadow (X) is perfect, does that mean the original sculpture (Y) must also be perfect? Or, conversely, if the original sculpture has a flaw, does that flaw show up in the shadow?"

In math, this is called descent. They proved that if the shadow (X) is smooth and perfect, the original sculpture (Y) must also be smooth.

• The Analogy: Think of a mirror. If you look in a mirror and see a perfect reflection, you know the object in front of the mirror is also perfect. If the reflection is cracked, the object is cracked.
• The Result: They showed that these new "Depth," "Width," and "Harmony" measurements behave exactly like that mirror. If the "shadow" building is good, the "source" building is guaranteed to be good too.

4. Why is this important?

Before this paper, mathematicians had to use very heavy, complicated machinery to prove that a building was good.

• Old way: "Let's run a simulation, check 50 different equations, and hope for the best."
• New way (this paper): "Does the magic reverse key fit? Yes? Great, it's good. No? It's broken."

This makes it much easier to study complex shapes. It also allows mathematicians to take a known "good" shape and create new "good" shapes by projecting them (like casting shadows), knowing that the quality will be preserved.

Summary in a Nutshell

• The Goal: To find simple ways to tell if a geometric shape is "smooth" or "cracked."
• The Method: They found that for three new types of "crack detectors," you can simply check if a mathematical "reverse key" fits. If it does, the shape is healthy.
• The Bonus: They proved that if a "shadow" of a shape is healthy, the original shape must be healthy too. This allows mathematicians to build new healthy shapes from old ones with confidence.

It's like giving architects a simple "tap test" to check for structural integrity, rather than needing to dismantle the whole building to check the beams.

Technical Summary

Here is a detailed technical summary of the paper "Characterization and Finite Descent of Local Cohomological Invariants" by Bradley Dirks, Sebastián Olano, and Debaditya Raychaudhury.

1. Problem Statement

The paper addresses the study of singularity invariants for complex algebraic varieties, specifically focusing on three recently introduced invariants: $c(X)$, $w(X)$, and $HRH(X)$. These invariants, defined in prior works ([CDO25], [CDOR26], [DOR25]), extend classical notions of Du Bois and rational singularities using Saito's theory of mixed Hodge modules.

The central problems tackled are:

1. Characterization: Finding simple, structural characterizations of these invariants (specifically $c(X)$, $w(X)$, and $HRH(X)$) in terms of the existence of left-inverses for specific morphisms between complexes of differential forms.
2. Finite Descent: Determining whether these invariants satisfy descent properties under finite surjective morphisms. Specifically, if $f: Y \to X$ is a finite surjective morphism (or a quotient by a finite group), does the "singularity level" of $X$ relate to that of $Y$? (e.g., if $Y$ has high-level singularities, does $X$ inherit them?)

2. Methodology

The authors employ a combination of Mixed Hodge Module (MHM) theory, derived category techniques, and trace morphisms.

• Complexes of Interest: The authors introduce and utilize three specific complexes associated with a variety $X$ of dimension $n$:

  • $\Omega^p_X$: The graded pieces of the filtered Du Bois complex.
  • $I\Omega^p_X$: The intersection Du Bois complex (derived from the intersection complex Hodge module $IC^H_X$).
  • $P\Omega^p_X$: A newly introduced perverse Du Bois complex, derived from $H^0(Q^H_X[\dim X])$.
  • These fit into a factorization chain: $\Omega^p_X \to P\Omega^p_X \to I\Omega^p_X$.

• Injectivity Theorems: The core technical strategy involves proving injectivity theorems for Ext groups involving these complexes and the dualizing complex $\omega^\bullet_X$. The logic follows the pattern established by Kovács for Du Bois singularities: if a map $A \to B$ admits a left inverse, then the induced map on duals $D(B) \to D(A)$ is surjective. By proving injectivity of the dual maps under certain cohomological conditions, the authors establish the existence of left inverses.

• Trace Morphisms: To handle the descent problem, the authors construct a trace morphism ($Tr_f$) in the derived category of mixed Hodge modules ($D^b(MHM(X))$).

  • They generalize the construction of trace maps for finite group quotients to arbitrary finite surjective morphisms (assuming $X$ is normal).
  • This relies on the formal properties of mixed sheaves and the splitting of idempotents in the category of mixed Hodge modules.

• Hodge-Lyubeznik Numbers: As an alternative approach, the authors utilize Hodge-Lyubeznik numbers ($\lambda^{p,q}_{r,s}$) and their intersection variants ($I\lambda^p_r$) to detect the invariants, providing a second proof for the descent results.

3. Key Contributions and Results

A. Left-Inverse Characterizations (Theorems A & B)

The paper provides the first "left-inverse" characterizations for the invariants $c(X)$, $w(X)$, and $HRH(X)$.

• Theorem A: Establishes injectivity results for natural morphisms between Ext groups of the complexes defined above.
  • If $c(X) \ge k-1$, the map $\text{Ext}^i(P\Omega^k_X, \omega^\bullet_X) \to \text{Ext}^i(\Omega^k_X, \omega^\bullet_X)$ is injective.
  • If $w(X) \ge k-1$, the map $\text{Ext}^i(I\Omega^k_X, \omega^\bullet_X) \to \text{Ext}^i(P\Omega^k_X, \omega^\bullet_X)$ behaves like an isomorphism/injection in specific degrees.
• Corollary B: Translates these injectivity results into characterizations:
  • $c(X) \ge k \iff \Omega^p_X \to P\Omega^p_X$ has a left inverse for all $0 \le p \le k$.
  • $w(X) \ge k \iff P\Omega^p_X \to I\Omega^p_X$ has a left inverse for all $0 \le p \le k$.
  • $HRH(X) \ge k \iff \Omega^p_X \to I\Omega^p_X$ has a left inverse for all $0 \le p \le k$.

B. Finite Descent Results (Theorems C & D)

The authors establish that these invariants descend under finite surjective morphisms.

• Theorem C: Constructs a trace morphism $Tr_f: f_* Q^H_Y \to Q^H_X$ in $D^b(MHM(X))$ for finite surjective $f$ (with $X$ normal) or finite group quotients. This recovers and generalizes previous trace maps.
• Corollary D: Using the trace morphism and the left-inverse characterizations, they prove the descent inequalities:
  • $c(Y) \le c(X)$
  • $w(Y) \le w(X)$
  • $HRH(Y) \le HRH(X)$
  • Implication: If $Y$ has "better" singularities (higher invariant values), $X$ must also have at least that level of singularity control. This confirms that pre-$k$-rational singularities descend.

C. Hodge-Lyubeznik Inequalities (Corollary E)

The paper proves inequalities for Hodge-Lyubeznik numbers under finite surjective maps. For $f: Y \to X$ and $x \in X$ with fiber $F_x = \{y_1, \dots, y_j\}$:

• $\lambda^{p,q}_{r,s}(O_{X,x}) \le \sum \lambda^{p,q}_{r,s}(O_{Y,y_i})$
• $I\lambda^p_r(O_{X,x}) \le \sum I\lambda^p_r(O_{Y,y_i})$
  These inequalities provide an alternative route to proving the descent of the main invariants.

D. Descent of Q-Factoriality Defect (Corollary F)

The authors apply their methods to the $\mathbb{Q}$-factoriality defect $\sigma(X)$, defined as the dimension of the cokernel of the map from Cartier divisors to Weil divisors (tensored with $\mathbb{Q}$).

• They prove that $\sigma(X) \le \sigma(Y)$ for finite surjective maps between normal projective varieties.
• They extend this to the local analytic setting, showing $\sigma_{an}(X, x) \le \sum \sigma_{an}(Y, y_i)$.

4. Significance

• Unification of Singularity Theory: The work unifies various higher singularity invariants ($c, w, HRH$) under a single framework of left-inverse characterizations, similar to the classical characterization of Du Bois singularities.
• Generalization of Descent: While descent for Du Bois and rational singularities was known, this paper extends these results to the more subtle invariants $c, w,$ and $HRH$, which are crucial for understanding higher Du Bois and higher rational singularities.
• New Tools: The introduction of the perverse Du Bois complex ($P\Omega^k_X$) and the rigorous construction of trace morphisms in the context of mixed Hodge modules provide new technical tools for future research in birational geometry and singularity theory.
• Connection to Topology: The results bridge the gap between Hodge-theoretic invariants and topological properties (like $\mathbb{Q}$-factoriality and homology manifolds), showing how local cohomological properties behave under finite covers.

In summary, this paper provides a robust theoretical framework for analyzing and comparing singularity invariants of complex varieties, proving that these invariants are well-behaved under finite surjective morphisms, and offering concrete algebraic criteria (left-inverses) to detect them.