Limits of conformal images and conformal images of limits for planar random curves

This paper establishes that for planar random curves in domains with rough boundaries, the operations of taking weak limits and applying conformal maps commute, thereby confirming that the conformal image of a limit curve is the limit of the conformal images under minimal boundary regularity assumptions.

The Gist

Imagine you are trying to understand the shape of a coastline. You have a very detailed map (the "real" domain) and a series of rougher, pixelated maps (the "approximating domains") that get better and better as you zoom in.

In the world of physics and mathematics, scientists study random curves—think of them as the jagged, unpredictable paths of a drunkard walking on a grid, or the boundary between two colors in a melting ice cube. These paths often appear in models of statistical mechanics (like how magnets work or how fluids flow).

The big question this paper answers is: If we take the limit of these random paths on our rough maps, does it matter if we look at the paths directly, or if we first squish the maps into a perfect circle (using a "conformal map") and then look at the paths there?

Here is the breakdown of the paper's discovery using simple analogies.

1. The Two Ways to Look at the Curve

Imagine you have a random curve drawn on a piece of paper with a weird, jagged edge (a "rough domain").

• Method A: You look at the curve directly on the jagged paper.
• Method B: You use a magical lens (a conformal map) that stretches and squishes the jagged paper until it becomes a perfect circle. You then look at the curve inside that circle.

Mathematicians already knew that if you have a sequence of these curves getting closer and closer to a final shape, they converge (settle down) in both Method A and Method B. But there was a nagging doubt: Do the two methods agree?

In other words: If I take the limit of the curve on the circle and then stretch it back to the jagged paper, do I get the same result as taking the limit of the curve directly on the jagged paper?

The Paper's Answer: Yes. The order doesn't matter.

"Limits of conformal images are conformal images of limits."

2. The "Deep Fjord" Problem

Why was this hard to prove? Because the "jagged paper" (the domain) can be incredibly messy.

Imagine a coastline with deep fjords—narrow, winding inlets that go very deep into the land.

• In the "rough" approximations (the pixelated maps), these fjords might look like deep, narrow canyons.
• A random curve (the drunkard's walk) might wander deep into one of these fjords.
• If the fjord is too narrow and deep, the "magic lens" (the conformal map) has to stretch it out a lot to fit it into the circle. This stretching can be violent and unpredictable.

The author, Alex Karrila, proves that even if the domain has these "deep fjords" and the boundary is extremely rough (not smooth like a circle, but jagged like a fractal), the math still holds up. The random curves simply don't go deep enough into these fjords to break the math.

The Analogy:
Think of the random curve as a hiker. The "deep fjords" are dangerous, narrow canyons. The paper proves that the hiker is statistically unlikely to wander so deep into a canyon that they get lost or that the map breaks. Because the hiker stays in "safe zones," the transformation between the jagged map and the perfect circle remains stable.

3. Why Does This Matter? (The "Multiple SLE" Connection)

The paper mentions SLE (Schramm-Loewner Evolution). Think of SLE as the "universal language" for these random curves at a critical point (like the exact moment ice turns to water).

Sometimes, you have multiple random curves interacting.

• Imagine drawing Curve 1. It cuts the paper in half.
• Now you want to draw Curve 2, but it has to stay in the remaining space.
• Curve 1 might have created a weird, jagged "fjord" in the remaining space.
• Now Curve 2 has to navigate this new, messy shape.

This paper is crucial because it allows mathematicians to study these complex, multi-curve scenarios even when the shapes they are drawing on are incredibly rough and irregular. It confirms that we can safely use the "perfect circle" math to solve problems on "messy real-world" shapes, even when those shapes are being carved up by other random curves.

4. The "Warning" (What Could Go Wrong?)

The paper also points out where this logic could fail if the rules weren't right:

• The "Infinite Spiral": Imagine a domain that looks like a spiral staircase with infinitely many layers getting tighter and tighter at the bottom. A curve trying to reach the bottom would have to twist infinitely. In such a case, the math breaks.
• The Solution: The paper relies on specific "crossing estimates" (rules about how likely a curve is to cross a certain area). These rules act like a safety net, ensuring the curves don't get trapped in these impossible infinite spirals.

Summary

The Big Picture:
This paper is a bridge between the messy, real world (irregular, rough domains) and the clean, theoretical world (perfect circles).

It proves that you can safely translate your problem into a perfect circle, solve it there, and translate it back, even if the original shape was a nightmare of jagged edges and deep canyons. The random curves behave well enough that the "translation" (conformal mapping) doesn't distort the final result.

In one sentence:
No matter how jagged the coastline is, if you zoom in on a random path along it, you can safely view that path through a perfect circular lens without losing any of its essential shape or behavior.

Technical Summary

Here is a detailed technical summary of the paper "Limits of conformal images and conformal images of limits for planar random curves" by Alex M. Karrila.

1. Problem Statement

The paper addresses a fundamental issue in the mathematical theory of scaling limits for planar lattice models in statistical mechanics, specifically regarding the convergence of random curves to Schramm–Loewner Evolutions (SLE).

In the standard approach to proving SLE convergence, one typically:

1. Establishes precompactness of a sequence of discrete random curves $\gamma^{(n)}$ in a domain $\Lambda_n$.
2. Maps these curves to the unit disc $\mathbb{D}$ via conformal maps $\phi_n$ to obtain $\gamma^{(n)}_D = \phi_n(\gamma^{(n)})$.
3. Proves that $\gamma^{(n)}_D$ converges weakly to a limit $\gamma_D$ (often an SLE in $\mathbb{D}$).
4. Identifies the limit in the original domain as $\gamma = \phi^{-1}(\gamma_D)$, where $\phi$ is the limit of the conformal maps.

The Core Difficulty:
While the convergence of $\gamma^{(n)}_D \to \gamma_D$ is well-understood under certain crossing probability estimates (Kemppainen-Smirnov conditions), the commutativity of the limit and the conformal map—i.e., proving that the weak limit of the conformal images equals the conformal image of the weak limit ($\lim \phi_n(\gamma^{(n)}) = \phi^{-1}(\lim \gamma^{(n)})$)—has been an open technical gap.

This gap is particularly acute when:

• The domains $\Lambda_n$ approximate a rough domain $\Lambda$ with irregular boundaries (e.g., fractal boundaries or "deep fjords").
• The random curves touch or intersect the boundary (common in multiple SLEs with $\kappa \in (4, 8)$).
• The conformal maps $\phi_n^{-1}$ do not extend continuously to the boundary, making the definition of $\phi^{-1}(\gamma_D)$ non-trivial.

2. Methodology

The author employs a combination of probabilistic estimates, complex analysis, and metric space topology.

• Precompactness Framework: The paper relies on the precompactness results of Kemppainen and Smirnov (KS17), which guarantee that random curves satisfying specific crossing probability estimates (Conditions C and G) have subsequential limits in the topology of unparametrized curves.
• Radial Limits and Prime Ends: To handle irregular boundaries, the author utilizes prime ends and radial limits of conformal maps. Instead of requiring continuous extension of $\phi^{-1}$ to the boundary, the paper defines the image of a curve via radial limits: $\phi^{-1}(\gamma_D(t)) = \lim_{\epsilon \downarrow 0} \phi^{-1}(P_\epsilon(\gamma_D(t)))$, where $P_\epsilon$ is a radial projection.
• The "Fjord" Argument: The core technical innovation is the analysis of "deep fjords" (narrow, deep indentations in the domain).
  • The author proves that under the KS17 crossing conditions, random curves are unlikely to travel deep into narrow fjords.
  • Analytically, it is shown that the conformal map $\phi_n^{-1}$ has large derivatives (stretching) only inside these deep fjords.
  • By combining the probabilistic exclusion of deep fjord visits with the analytic control of the map's derivative, the author demonstrates that the "error" introduced by the radial projection vanishes in the limit.
• Metric Space Techniques: The proof utilizes the Lévy-Prokhorov metric to establish tightness and precompactness in the space of unparametrized curves $X(\mathbb{C})$.

3. Key Contributions

1. Commutativity of Limits and Conformal Maps: The paper rigorously proves that for random curves satisfying the KS17 crossing conditions, the operation of taking a weak limit commutes with applying a conformal map, even when the limiting domain has rough boundaries and the curves touch the boundary.
   $$\lim_{n \to \infty} \phi_n(\gamma^{(n)}) = \phi^{-1}(\lim_{n \to \infty} \gamma^{(n)})$$
   Here, the equality holds in the sense of radial limits.

2. Minimal Boundary Regularity: Unlike previous results that required locally connected boundaries or excluded boundary visits, this result imposes no a priori boundary regularity assumptions on the limiting domain $\Lambda$. It handles domains with "deep fjords" and fractal boundaries.

3. Handling Multiple SLEs ($\kappa \in (4, 8)$): The results are specifically tailored to support the study of multiple SLEs where curves intersect themselves and each other. In such cases, the domain for a subsequent curve is the original domain slit by previous curves, creating highly irregular boundaries. This paper provides the necessary theoretical foundation to define the scaling limit in such slit domains.

4. Measurability of the Limit: The paper establishes that the limiting curve $\gamma$ in the original domain is a measurable random variable with respect to the sigma-algebra generated by the driving function (or the SLE in the disc), resolving technical issues regarding the existence of the curve as a random variable in rough domains.

4. Main Results

• Theorem 4.2 (Main Theorem): Under the setup of Carathéodory convergence of domains and the KS17 crossing conditions (C and G), the laws of the pairs $(\gamma^{(n)}_D, \gamma^{(n)})$ are precompact. Any subsequential limit $(\gamma_D, \gamma)$ satisfies $\gamma = \phi^{-1}(\gamma_D)$ almost surely, where $\phi^{-1}$ is extended by radial limits.
• Theorem 4.6 (Parametrized Curves): The result extends to capacity-parametrized curves in the stronger topology of $C([0,1], \mathbb{C})$.
• Proposition 5.2 (SLE Stability): The chordal SLE($\kappa$) for $\kappa \in [0, 8)$ is stable under domain perturbations and parameter changes. Specifically, if $\Lambda_n \to \Lambda$ and $\kappa_n \to \kappa$, the SLE curves converge weakly in the space of unparametrized curves.
• Key Lemma 4.4: A crucial technical lemma showing that the distance between a point on the curve and its radially projected image under the conformal map is small with high probability, provided the curve does not enter deep fjords.

5. Significance

• Foundational for Multiple SLEs: This work removes a major barrier in the rigorous construction of multiple SLEs, particularly for $\kappa > 4$ where curves are space-filling or self-intersecting. It allows researchers to define limits in domains slit by other random curves without needing to prove complex boundary regularity properties for the slit domains.
• Robustness of SLE Theory: It confirms that the identification of scaling limits as SLEs is robust even in geometrically complex settings, reinforcing the universality of SLE as the scaling limit for critical lattice models.
• Methodological Advancement: The technique of controlling conformal maps via the probabilistic exclusion of "deep fjords" offers a new tool for analyzing conformal mappings in irregular domains, potentially applicable to other problems in geometric function theory and probability.

In summary, Karrila's paper provides the rigorous "missing link" that allows the standard SLE convergence strategy to be applied to the most challenging geometric scenarios involving rough boundaries and interacting random curves.