🔢 mathematics

On positivity of the limit F-signature

This paper proves a conjecture by Carvajal-Rojas, Schwede, and Tucker regarding the positivity of the limit F-signature for complex KLT singularities, specifically establishing the result for three-dimensional non-weakly exceptional singularities via induction and for smooth hypersurfaces of very low degree through isotrivial normal toric degenerations inspired by K-stability theory.

Original authors: Yuchen Liu, Suchitra Pande

Published 2026-05-19

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Yuchen Liu, Suchitra Pande

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 an architect trying to understand the stability of a building. In mathematics, these "buildings" are shapes called singularities—places where a surface folds, pinches, or breaks in a way that isn't smooth.

Mathematicians have two different ways of looking at these shapes:

The Complex View: Looking at them in the "real world" of complex numbers (like the world we usually study in calculus).
The Finite View: Looking at them through a "pixelated" lens, where numbers are reduced to a finite set (like counting on your fingers, but with a specific number of fingers, $p$ ).

The paper by Yuchen Liu and Suchitra Pande tackles a big question: If a building is stable in the "real world," does it remain stable when we look at it through the "pixelated" lens, especially as we increase the number of pixels?

Here is a breakdown of their findings using simple analogies.

The Core Problem: The "F-Signature"

Think of the F-signature as a "stability score" for these shapes when viewed through the pixelated lens.

A score of 1 means the shape is perfectly smooth and stable (like a pristine sphere).
A score of 0 means the shape is completely broken.
A score greater than 0 means the shape is "strong" enough to hold together, even if it has some kinks.

The Conjecture:
Mathematicians Carvajal-Rojas, Schwede, and Tucker proposed a bold idea: If a shape is "KLT" (a specific type of stable, though not perfect, singularity) in the real world, then as we increase our pixel count (let $p$ go to infinity), the stability score should never drop to zero. It should always stay above a certain safe minimum.

Think of it like a bridge. The conjecture says: "If this bridge is safe in the real world, then no matter how many tiny cracks we simulate in our computer model, the bridge will never completely collapse; it will always retain some structural integrity."

The Authors' Solution: Two New Tools

The authors didn't just check the bridge; they built two new tools to prove the bridge holds up in many new cases.

Tool 1: The "Inductive Ladder" (Climbing Down)

Imagine you are trying to prove a bridge is safe, but it's too big to check all at once. The authors use a strategy called induction.

They realized that if a 3D shape (a complex building) is "not weakly exceptional" (a technical term meaning it doesn't have a unique, overly rigid structure), you can "peel off" a layer to reveal a 2D shape underneath (like a floor plan).
If you can prove the 2D floor plan is stable in the pixelated world, you can use that to prove the 3D building is stable.
The Result: They proved that for almost all 3D singularities (except the very rare, rigid ones), the stability score stays positive. They essentially climbed down the ladder from 3D to 2D, proved the 2D case, and used that to secure the 3D case.

Tool 2: The "Toric Transformation" (The Shape-Shifter)

Sometimes, a shape is too complicated to analyze directly. The authors used a technique called isotrivial degeneration.

Imagine you have a complex, twisted sculpture. You want to know if it's stable.
Instead of analyzing the twisted version, you slowly melt it down into a simpler, blocky shape (a "toric" variety, which is like a shape made of cubes and flat faces).
Crucially, this melting process is "isotrivial," meaning the twisted sculpture and the blocky shape are essentially the same thing, just viewed differently.
Because the blocky shape is much easier to analyze, and because stability properties don't suddenly vanish during this transformation, the authors could prove that if the blocky shape is stable, the original twisted sculpture is stable too.
The Result: They used this to prove the conjecture for smooth hypersurfaces (curved surfaces) of very low degree.

The "F-Adjunction" Trick

A key part of their logic involves a concept called F-adjunction.

Think of this as a rule that says: "If the foundation of a building is strong, the whole building is strong."
They found a special "foundation" (a divisor) inside the complex shape. They proved that if this foundation is stable in the pixelated world, the whole shape must be too.
They used a "birational" argument (a mathematical way of reshaping the building without tearing it) to move from the complex shape to a simpler one where the foundation is easier to inspect.

What They Actually Proved

The paper makes specific claims about where this "stability score" stays positive:

3D Shapes: For almost all 3D singularities (specifically those that are "non-weakly exceptional"), the stability score remains positive as the pixel count increases.
Low-Degree Surfaces: For smooth surfaces defined by simple equations (low-degree hypersurfaces), the conjecture holds true.
Spherical Shapes: They confirmed the conjecture for "spherical" varieties (shapes with a high degree of symmetry), which includes things like Grassmannians and flag varieties.

Summary

In everyday terms, Liu and Pande proved that for a vast class of mathematical shapes, structural integrity is preserved even when we zoom in to the finest possible digital resolution. They didn't just guess; they built a ladder to climb down from complex dimensions to simpler ones, and they used a shape-shifting trick to turn complicated curves into simple blocks, proving that the "stability score" never drops to zero.

Note: The paper is purely mathematical. It does not discuss applications to engineering, medicine, or climate science. It is a proof about the fundamental behavior of numbers and shapes.

Technical Summary: Positivity of the Limit F-Signature

Problem Statement
This paper investigates a conjecture proposed by Carvajal-Rojas, Schwede, and Tucker regarding the behavior of the F-signature for complex Kawamata log terminal (KLT) singularities under reduction to positive characteristic. Specifically, for a complex KLT singularity $(R, \mathfrak{m})$ , the conjecture posits that the F-signatures $s_p(R)$ of its reductions to characteristic $p$ remain bounded away from zero as $p \to \infty$ . In other words, $\liminf_{p \to \infty} s_p(R) > 0$ . While the F-signature is known to be positive if and only if a ring is strongly F-regular, and KLT singularities reduce to strongly F-regular rings for large $p$ , the uniformity of this positivity (i.e., the existence of a lower bound independent of $p$ ) had remained open for general cases.

Methodology
The authors develop a toolkit combining techniques from the Minimal Model Program (MMP), birational geometry, and the theory of F-singularities in positive characteristic. The core strategy involves reducing the local problem for KLT singularities to a global problem concerning log Fano pairs and their Frobenius-alpha ( $\alpha_F$ ) invariants.

Local-to-Global Reduction: The paper establishes that the conjecture for a KLT singularity $(R, \mathfrak{m})$ can be reduced to a conjecture about the $\alpha_F$ -invariant for log Fano pairs of one lower dimension. This is achieved by constructing a degeneration of the singularity to an orbifold cone over a Kollár component. The F-signature of the singularity is bounded below by the log discrepancy of the Kollár component multiplied by the F-signature of the component itself.
Inductive Approach via Special Divisors: For non-weakly exceptional singularities (those admitting more than one Kollár component), the authors utilize the existence of special divisors (log canonical places of Q-complements) derived from Zhuang's work. By running a $(K_Y + \Delta_Y + E)$ -MMP, they reduce the problem to a lower-dimensional log Fano pair $(E, \Delta_E)$ . They then employ "inversion of F-adjunction" to lift Frobenius splittings from the lower-dimensional pair back to the original variety, provided the characteristic $p$ is sufficiently large.
Isotrivial Degenerations: To handle specific classes of varieties, the authors construct isotrivial degenerations to normal toric varieties. Since the F-signature and $\alpha_F$ -invariant are semi-continuous in families, and toric varieties are known to satisfy the conjecture, this allows the transfer of positivity results to the generic fiber. This technique is applied to spherical varieties and smooth hypersurfaces of very low degree.

Key Contributions and Results

Equivalence of Conjectures: The paper proves that the positivity of the limit F-signature for KLT pairs of dimension $n$ is equivalent to the positivity of the limit $\alpha_F$ -invariant for log Fano pairs of dimension $n-1$ (Theorem 1.3).
Three-Dimensional Non-Weakly Exceptional Singularities: The authors prove the conjecture for all non-weakly exceptional three-dimensional KLT singularities over $\mathbb{C}$ (Theorem 1.5). This is achieved by reducing the problem to log del Pezzo surfaces with $\alpha < 1$ , which are further reduced to pairs on $\mathbb{P}^1$ with at most three components, a case previously established.
Global F-regularity in Mixed Characteristic: The techniques are extended to show that non-weakly exceptional three-dimensional KLT singularities over perfect fields of characteristic $p > 5$ (or complete DVRs with such residue fields) are $+$ -regular (and BCM-regular) (Theorem 1.6). This result is sharp, as counterexamples exist without the non-weakly exceptional assumption.
Hypersurfaces of Low Degree: Using isotrivial degenerations to normal toric hypersurfaces, the conjecture is proven for smooth hypersurfaces $X \subset \mathbb{P}^n$ of degree $d$ satisfying $n \geq \binom{2d-1}{d} + d - 1$ (Theorem 1.8). This includes a new proof for spherical Q-Fano varieties.
Uniform Bounds: For non-weakly exceptional threefold singularities, the paper establishes a uniform lower bound for the F-signature in terms of the normalized volume defined by Li (Proposition 7.1, Theorem 7.5).

Significance
The paper provides a partial inductive solution to the Carvajal-Rojas, Schwede, and Tucker conjecture, resolving it for a broad new class of three-dimensional singularities. By linking the asymptotic behavior of F-signatures to the geometry of K-stability (via $\alpha$ -invariants and test configurations) and utilizing the machinery of MMP in positive characteristic, the work bridges complex singularity theory and arithmetic geometry. The authors note that their methods are inspired by K-stability theory, particularly the use of degenerations and birational geometry to control invariants. The results confirm that for non-weakly exceptional singularities, the "positivity" of the F-signature is a robust property that persists uniformly as the characteristic tends to infinity.