🔢 mathematics

A characterisation of $\infty$ -harmonic maps in terms of $1$-currents

This paper characterizes critical points of the $L^\infty$ -norm functional for maps between Riemannian manifolds by establishing a geometric condition equivalent to criticality, formulated via a vector-valued 1-current that is itself a critical point of a generalized mass functional.

Original authors: Roger Moser

Published 2026-06-10

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Roger Moser

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 trying to stretch a piece of fabric (a map) from one shape (Manifold M) onto another shape (Manifold N). Your goal is to do this in a way that minimizes the "worst-case" stretch. In math terms, you aren't looking at the average stretch (like a standard rubber band), but rather the single point where the fabric is stretched the tightest. This is called an $\infty$ -harmonic map.

The problem is that finding the "tightest point" is tricky. Standard math tools (calculus) work great for smooth curves and averages, but they break down when you are dealing with a single, sharp maximum. It's like trying to find the slope of a cliff edge; the math gets messy because the edge is too sharp.

Roger Moser's paper provides a new way to describe these maps without getting stuck on that sharp edge. Here is the breakdown of his ideas using simple analogies:

1. The Problem: The "Unsmooth" Cliff

Usually, to find the best way to stretch a map, mathematicians look for a "critical point"—a spot where if you wiggle the map slightly, the energy doesn't go down. But because we are looking at the maximum stretch (the $L^\infty$ norm), the energy function has a sharp corner. You can't take a derivative there.

Moser defines a new way to say a map is "optimal." Instead of asking "Is the slope zero?" (which doesn't exist), he asks: "If I wiggle the map, does the energy go up?" If the energy stays the same or goes up, the map is considered $\infty$ -harmonic. It's like balancing a ball on a flat plateau; you can't roll it down, so it stays put.

2. The Solution: The "Ghost River" (1-Currents)

Since we can't use standard calculus, Moser introduces a clever geometric tool called a 1-current.

Think of a 1-current not as a number, but as a collection of invisible, ghostly rivers flowing through your fabric.

These rivers don't just flow anywhere; they only flow through the parts of the fabric that are stretched to their absolute limit.
These rivers carry a "charge" or a vector that tells us the direction and intensity of the stretch at that specific spot.

Moser proves that if a map is $\infty$ -harmonic, there must exist this special system of rivers. The map and the rivers are locked together:

The River Defines the Map: The direction the river flows tells you exactly how the fabric is stretched at that point.
The Map Defines the River: The way the fabric is stretched tells you where the rivers must flow.

3. The "Flow" Rule (The Boundary Condition)

The paper introduces a rule for these rivers called the $u$ -boundary.
Imagine the rivers are flowing through a maze. The rule says: "The total amount of river flowing out of any closed loop must be zero, unless the loop is interacting with the fabric in a specific way."

In simpler terms, these rivers are in a state of perfect equilibrium. They aren't piling up or disappearing; they are flowing in a way that perfectly balances the tension of the fabric. If you tried to push the rivers to flow differently, the fabric would snap or the tension would increase.

4. The "Heavy Load" (Mass Functional)

The paper also shows that these rivers aren't just random; they are the most efficient way to carry the tension.

Imagine the rivers are carrying a heavy load (the "mass").
The paper proves that these specific rivers are critical points of a mass functional.
Analogy: Think of a river system that has carved the most efficient path through a landscape to carry the maximum amount of water with the least amount of energy. The rivers in this paper are the "most efficient" path for the tension of the fabric.

5. Two Types of Stretching

The paper looks at two main ways to measure "stretch":

The "Sum of Squares" ( $p=2$ ): This is like measuring the total energy of the stretch. Here, the math is very clean. The rivers perfectly determine the stretch, and the map perfectly determines the rivers. It's a two-way street.
The "Maximum Stretch" ( $p=\infty$ ): This is the strict "worst-case" scenario. The math is messier here because the "maximum" function is very sharp. The rivers still exist and still flow, but they don't tell us everything about the stretch in the same precise way. It's like knowing the river flows through the deepest canyon, but not knowing the exact depth of every rock in the canyon.

Summary

Roger Moser's paper is like giving a new pair of glasses to a mathematician.

Before: They tried to look at the "sharpest point" of a stretched fabric and got blind because the math was too jagged.
After: They stop looking at the jagged point directly. Instead, they look at the invisible rivers (1-currents) that flow through the fabric.
The Discovery: If the fabric is stretched in the most efficient way possible (an $\infty$ -harmonic map), these rivers must exist, they must flow in a balanced way, and they must carry the tension perfectly.

This allows mathematicians to study these difficult maps using the smooth, flowing language of rivers and currents, rather than getting stuck on the sharp, jagged edges of the original problem.

Technical Summary: A Characterisation of $\infty$ -Harmonic Maps in Terms of 1-Currents

Problem Statement
The paper investigates variational problems involving the $L^\infty$ -norm of the derivative for maps $u: M \to N$ between two compact Riemannian manifolds. The primary functional under study is:
$E_\infty(u) = \text{ess sup}_{x \in M} \sqrt{F(du(x))}$
where $F$ is a convex, quadratically homogeneous function derived from an orthogonally invariant norm (e.g., the Hilbert-Schmidt norm or the operator norm). Unlike standard $p$ -harmonic map problems ( $1 < p < \infty$ ), the functional $E_\infty$ is not differentiable, even in the Gâteaux sense, rendering classical Euler-Lagrange equations inapplicable. The paper seeks to define a rigorous notion of "critical points" for this functional and derive geometric characterizations of such points, extending previous work on scalar-valued $\infty$ -harmonic functions and vector-valued maps in Euclidean domains to the setting of Riemannian manifolds.

Methodology
The author employs a combination of geometric measure theory, convex analysis, and approximation techniques:

Definition of Criticality: Instead of relying on differentiability, the paper adopts a Fréchet subdifferential approach. A map $u \in C^{0,1}(M; N)$ is defined as $\infty$ -harmonic if the lower limit of the energy difference quotient is non-negative with respect to the uniform distance metric. This definition is shown to imply a quadratic lower bound on the energy discrepancy relative to the distance from $u$ .
Vector-Valued 1-Currents: The core technical tool is the introduction of a vector-valued 1-current $T$ on the domain manifold $M$ . Unlike standard currents, this object is adapted to the map $u$ via a "u-boundary" operator $\partial_u T$ . This operator is constructed using the embedding of $N$ into a Euclidean space and a parallelization of $N$ , allowing the definition of a covariant derivative-like structure on the trivial bundle $M \times \mathbb{R}^n$ .
Approximation by $p$ -Harmonic Maps: To handle the lack of differentiability and the singularity of the $L^\infty$ limit, the author approximates the $\infty$ -harmonic map $u$ by a sequence of $p$ -harmonic maps ( $p \to \infty$ ) associated with penalized functionals $E_p$ . This allows the use of classical Euler-Lagrange equations for the approximating sequence.
Measure-Section Pairs and Convergence: The analysis utilizes the concept of measure-section pairs (extending Hutchinson's work) to handle the convergence of gradients and associated measures. The author proves strong $L^2$ convergence of the approximating gradients against a limiting measure, which is crucial for passing to the limit in the Euler-Lagrange equations.
Generalized Mass Functionals: For general norms $F$ , the standard mass functional of currents is replaced by a generalized mass functional $M_F$ , defined via the Legendre transform of $F$ .

Key Contributions and Results

Characterization via Currents: The paper establishes that an $\infty$ -harmonic map $u$ (with $E_\infty(u) > 0$ ) is associated with a non-trivial, finite-mass, $\mathbb{R}^n$ -valued 1-current $T$ satisfying two main conditions:
1. Weak u-boundary condition: $\partial_u T = 0$ in a weak sense. This condition encodes the interaction between the current and the geometry of the map $u$ .
2. Gradient Representation: On the support of the current, the derivative $du$ (or its mollified version) is determined by the current's direction vector $\vec{T}$ and the energy level $e_\infty$ . Specifically, for regular $F$ , the relation is:
  $\text{grad } u = e_\infty \frac{\nabla F^*(\vec{T}^\dagger)}{\sqrt{F^*(\vec{T}^\dagger)}}$
  where $F^*$ is the Legendre transform of $F$ .
Equivalence Theorems:
- Theorem 12: If $u$ is $\infty$ -harmonic, such a current $T$ exists satisfying the above conditions.
- Theorem 14: Conversely, if a map $u$ and a current $T$ satisfy the boundary condition and the gradient representation, then $u$ is $\infty$ -harmonic and the energy level matches. This provides a necessary and sufficient condition for criticality, analogous to an Euler-Lagrange equation.
Criticality of the Current: The current $T$ itself is shown to be a critical point of the generalized mass functional $M_F$ under variations induced by diffeomorphisms of the domain. This is expressed as the vanishing of the first variation of $M_F$ with respect to vector fields on $M$ .
Handling Non-Regular Norms: For cases where the norm is not "regular" (lacking strong convexity, such as the operator norm $|\cdot|_\infty$ ), the paper provides a weaker characterization (Theorem 15). Here, the current is decomposed into a family of curves (via Smirnov's decomposition), and the characterization involves integral identities along these curves rather than a pointwise gradient formula.
Compactness: The paper proves that the set of $\infty$ -harmonic maps with bounded energy is compact in the uniform topology, a property derived from the quadratic energy bound established in Proposition 2.

Significance and Claims
The paper claims to extend the theory of $\infty$ -harmonic maps from scalar-valued functions and Euclidean domains to maps between general Riemannian manifolds. It addresses the specific difficulties arising from the non-differentiability of the $L^\infty$ functional and the geometric constraints of the target manifold $N$ .

The significance lies in providing a geometric characterization of critical points that does not rely on the existence of a classical PDE. Instead, it identifies a geometric object (the 1-current) that encapsulates the essential variational properties of the map. The author notes that while previous works (e.g., by Katzourakis and Moser) focused on global minimizers in Euclidean settings, this work generalizes the framework to critical points in a broader geometric context, including various choices of the functional $F$ (such as Schatten norms). The results are presented as a refinement and extension of existing measure-theoretic approaches to $L^\infty$ variational problems, offering a unified framework for understanding the structure of $\infty$ -harmonic maps.