🔢 mathematics

Weak Harnack inequality and Cartan property for nonlocal $W^{s,1}$ -minimizers

This paper establishes a weak Harnack inequality and Cartan-type properties for nonlocal $W^{s,1}$ -minimizers in doubling metric measure spaces, proving their semicontinuity and demonstrating that these results are novel even in the classical Euclidean setting.

Original authors: Panu Lahti, Yuxin Li, Khanh Nguyen

Published 2026-03-24

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Panu Lahti, Yuxin Li, Khanh Nguyen

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 standing in a vast, foggy landscape. This landscape isn't just flat ground; it's a complex, bumpy terrain where the rules of distance and volume are a bit different from our everyday world. In mathematics, this is called a metric measure space.

The paper you shared is like a set of new survival guides for people trying to navigate this foggy terrain. Specifically, it's about understanding the behavior of "minimizers"—think of them as hikers trying to find the path of least resistance (or the smoothest, most efficient route) across this strange landscape.

Here is the breakdown of the paper's discoveries, translated into everyday language:

1. The Hiker's Dilemma: The "Nonlocal" Problem

In the old days (classical math), if you wanted to find the smoothest path, you only looked at your immediate surroundings. If the ground was bumpy right next to your foot, you adjusted. This is called a "local" problem.

But this paper deals with "nonlocal" problems. Imagine that your hiking path doesn't just depend on the ground under your feet, but also on the ground miles away. Maybe a rock a mile to the left pulls your path slightly to the right. This is the "nonlocal" aspect. The paper focuses on a specific type of hiker called a $W^{s,1}$ -minimizer.

The Analogy: Think of these hikers as trying to minimize "friction." The "friction" isn't just about how rough the ground is locally, but how much the hiker's path clashes with the terrain everywhere else in the universe.

2. The Weak Harnack Inequality: The "Foggy Weather Report"

The first major discovery is something called the Weak Harnack Inequality.

The Metaphor: Imagine you are in a thick fog. You can't see the top of a mountain peak, but you know the average height of the ground around you. In classical math, if you know the average height, you can predict the peak with high precision.
The New Discovery: In this nonlocal world, the "fog" is thicker. You can't predict the exact peak, but the authors proved a safety rule: If the average height of the ground around you is low, the peak cannot be impossibly high. It's bounded.
Why it matters: This rule stops the hiker's path from going wild and shooting off to infinity. It guarantees that even in this weird, long-range world, the path stays somewhat "sensible" and doesn't explode.

3. Semicontinuity: The "Smoothness" Guarantee

Once you know the path can't explode, the next question is: Is it smooth? Can you walk on it without tripping?

The Reality: In this nonlocal world, the path might have tiny, invisible "jumps" or "glitches" (mathematicians call these discontinuities). You can't promise the path is perfectly smooth like a polished floor.
The Promise: However, the authors proved that the path is semicontinuous.
The Analogy: Imagine a staircase. It's not a ramp (smooth), but it's not a pile of random rocks either. You can step up, but you can't suddenly fall through the floor. The path has a "floor" you can trust. You can define a "best guess" for the height at any point, and that guess won't suddenly jump up or down without warning.

4. The Cartan Property: The "Thin Ice" Theory

This is the most abstract and fascinating part. The paper asks: "What happens if the hiker tries to walk over a patch of ground that is 'thin'?"

The Concept of "Thin": In this math world, a set of points can be "thin" (like a single hair on a vast beach) or "thick" (like a solid rock). If a set is "thin," it usually means you can walk right over it without it affecting your path.
The Discovery: The authors proved a Cartan Property.
The Analogy: Imagine you are walking on a frozen lake. Most of the lake is solid ice (thick). But there is a tiny, invisible crack (thin).
- The Old View: You might think, "It's just a crack, I'll step over it."
- The New View: The authors show that if you try to cross this "thin" crack, the path might actually spike up to infinity right at that crack, but only if you look at it from a very specific angle.
- The "Weak" vs. "Strong" Property:
  - Weak Property: To cross the thin crack, you might need to use a team of several different hikers (mathematical functions) working together. One hiker might stumble, but if you take the "best" height from the whole team, you can cross safely.
  - Strong Property: If the crack is "heavy" enough (has positive capacity), then a single hiker can cross it, but the path will shoot up to the sky (infinity) right at the crack, proving that the crack is actually a massive barrier, not a thin one.

5. Why This Matters (The "So What?")

You might ask, "Who cares about hikers in a math fog?"

New Territory: The authors note that these results are new, even for the flat, familiar world of Euclidean geometry (our normal 3D space). They found rules that no one knew existed before.
Real World Applications: While this sounds abstract, "nonlocal" math is used to model things where things affect each other over long distances, like:
- Finance: How a stock price in New York affects a market in Tokyo instantly.
- Biology: How a disease spreads not just to your neighbor, but to people you've met recently.
- Physics: How particles interact in quantum mechanics.

Summary

This paper is a rulebook for navigating a strange, long-range world.

Rule 1: Even if the world is weird and connected from far away, paths can't go crazy (Weak Harnack).
Rule 2: The paths are mostly stable and predictable, even if they have tiny jumps (Semicontinuity).
Rule 3: If you try to walk over a "thin" spot, the path might react violently, but we now have a precise map of exactly how it reacts (Cartan Property).

The authors, Panu Lahti, Yuxin Li, and Khanh Nguyen, have essentially drawn a new map for mathematicians, showing them that even in the most complex, nonlocal landscapes, there is still order, structure, and predictability.

1. Problem Statement and Context

The paper investigates the regularity and fine properties of nonlocal minimizers in the specific case where the integrability exponent is $p=1$ . While the theory of nonlocal operators (such as the fractional $p$ -Laplacian) is well-developed for $1 < p < \infty$ , the case $p=1$ presents unique challenges due to the lack of strict convexity and the connection to the space of functions of Bounded Variation ($BV$).

The authors study $s$ -subminimizers and $s$ -superminimizers in a general complete, connected, doubling metric measure space $(X, d, \mu)$ . These functions minimize a nonlocal functional $F_s$ defined by:
$F_s(u) = \iint_{\Omega \times \Omega} \frac{|u(x)-u(y)|}{\mu(B_{x,y})d(x,y)^s} d\mu(x)d\mu(y) + 2\iint_{\Omega \times \Omega^c} \frac{|u(x)-u(y)|}{\mu(B_{x,y})d(x,y)^s} d\mu(x)d\mu(y)$
where $B_{x,y} = B(x, d(x,y)) \cup B(y, d(y,x))$ and $0 < s < 1$ .

The primary goals are:

To establish a Weak Harnack Inequality for $s$ -subminimizers.
To derive semicontinuity results for these minimizers.
To prove Cartan-type properties (describing the behavior of superminimizers near "thin" sets) for $s$ -superminimizers.

2. Methodology

The authors employ a combination of nonlocal analysis, geometric measure theory, and variational methods adapted to metric measure spaces.

Functional Framework: The work is set in the homogeneous Sobolev space $\dot{W}^{s,1}(\Omega)$ , defined via the Gagliardo semi-norm. The authors utilize the Obstacle Problem formulation, where solutions are minimizers of the energy subject to a constraint $u \geq \psi$ .
Caccioppoli–De Giorgi Iteration: A core technique involves deriving a Caccioppoli-type inequality for the truncated function $(u-k)_+$ . This is achieved by testing the variational inequality with specific cut-off functions and utilizing the definition of sub/superminimizers.
Fractional Poincaré Inequalities: The proofs rely heavily on sharp fractional Poincaré inequalities.
- For general doubling spaces, they use inequalities from [DLV23].
- For spaces supporting a 1-Poincaré inequality, they utilize a sharp inequality from [KLLZ25] which provides better control over the constants as $s \to 1^-$ .
Capacity and Thinness: The authors define the $(s, 1)$ -capacity ( $C_{s,1}$ ) and the concept of $s$ -thinness. A set $A$ is $s$ -thin at $x$ if the ratio of the capacity of $A \cap B(x,r)$ to the capacity of the ball $B(x,r)$ vanishes as $r \to 0$ .
Coarea Formula: The relationship between the $W^{s,1}$ norm and the $s$ -perimeter of level sets is utilized via the coarea formula: $\|u\|_{\dot{W}^{s,1}} = \int_{-\infty}^{\infty} \|\mathbb{1}_{\{u>t\}}\|_{\dot{W}^{s,1}} dt$ .

3. Key Contributions and Results

A. Weak Harnack Inequality (Theorem 1.1)

The authors prove a weak Harnack inequality for $s$ -subminimizers. Unlike the case $p > 1$ , the inequality for $p=1$ does not involve a "tail" term (a nonlocal integral over the complement of the domain).

Result: For an $s$ -subminimizer $u$ in $\Omega$ , and for any $k_0 \in \mathbb{R}$ , $x_0 \in \Omega$ , and $0 < r < R$ with $B(x_0, R) \subset \Omega$ :
$\text{ess sup}_{y \in B(x_0, r)} (u(y) - k_0) \leq C \left( \frac{R}{R-r} \right)^Q \dashint_{B(x_0, R)} (u - k_0)_+ d\mu$
where $Q$ is the doubling dimension.
Significance: The absence of the tail term is a distinct feature of the $p=1$ case compared to $p>1$ . Furthermore, if the space supports a 1-Poincaré inequality, the constant $C$ is uniformly bounded as $s \to 1^-$ , ensuring stability.

B. Semicontinuity (Corollary 1.2)

Using the Weak Harnack Inequality, the authors establish the semicontinuity of minimizers.

Result: If $u$ is an $s$ -superminimizer, its lower approximate limit $u^\wedge$ is lower semicontinuous. If $u$ is an $s$ -subminimizer, its upper approximate limit $u^\vee$ is upper semicontinuous.
Context: This is the best possible regularity result. Full Hölder continuity (typical for $1 < p < \infty$ ) fails for $p=1$ due to the existence of nonlocal minimal surfaces with irregular boundaries.

C. Weak Cartan Property (Theorem 1.3)

This is a major contribution regarding the fine structure of superminimizers near sets of zero capacity.

Result: If a set $A$ $A$ is $s$ $s$ -thin at a point $x \in A \setminus A$ $x \in A ∖ A$ (the measure-theoretic boundary), there exist $k$ different $s$ $s$ -superminimizers (specifically characteristic functions of sets $E_j$ $E_{j}$ ) such that:
1. Their lower limits cover the set $A$ near $x$ (i.e., $\max_j \mathbb{1}_{E_j}^\wedge \equiv 1$ on $A$ ).
2. Their upper limits vanish at $x$ (i.e., $\mathbb{1}_{E_j}^\vee(x) = 0$ ).
3. The set where the maximum of their upper limits is positive is $s$ -thin at $x$ .
Interpretation: This implies that while a single superminimizer might not "see" the thin set $A$ at $x$ (limit is 0), a finite collection of them can reconstruct the set $A$ in the limit from the interior. This is termed a "weak" Cartan property because it requires a union of $k$ functions, whereas a "strong" property would require only one.

D. Strong Cartan Property (Theorem 1.4)

Under the additional assumption that the point $x$ has strictly positive capacity ( $C_{s,1}(\{x\}) > 0$ ), the authors prove a stronger result.

Result: There exists a single $s$ -superminimizer $u$ such that $\lim_{A \ni y \to x} u^\wedge(y) = \infty$ while $u^\vee(x) < \infty$ .
Significance: This demonstrates that if the point itself carries capacity, the singularity can be captured by a single function, contrasting with the weak property where multiple functions are needed for general thin sets.

4. Significance and Novelty

Extension to Metric Spaces: The results are established in the general setting of doubling metric measure spaces, not just Euclidean spaces.
The $p=1$ Case: The paper fills a gap in the literature. While $p$ -harmonic functions for $p>1$ are well-understood, the $p=1$ case (related to Total Variation and minimal surfaces) in the nonlocal setting was largely open.
New in Euclidean Setting: The authors explicitly state that these results appear to be new even in the classical Euclidean setting $\mathbb{R}^n$ .
Analogy with Local Theory: The theory developed is shown to be "mostly analogous" to the local case of $BV$ (Bounded Variation) super/subminimizers, bridging the gap between local and nonlocal variational problems in the $L^1$ regime.
Stability: The analysis of the constants as $s \to 1^-$ provides insight into the convergence of nonlocal problems to their local counterparts.

In summary, this paper provides a foundational regularity theory for nonlocal $W^{s,1}$ minimizers, establishing that despite the lack of smoothness typical of $BV$ functions, these minimizers possess robust semicontinuity properties and satisfy Cartan-type principles describing their behavior on sets of small capacity.