Generic regularity of intermediate complex structure limits

This paper establishes that for certain polarized degenerations of Calabi-Yau manifolds near an intermediate complex structure limit, the potential $C^0$-convergence of the corresponding collapsing Ricci-flat Kähler metrics can be improved to metric convergence on a generic region.

The Gist

Imagine you have a giant, complex, multi-dimensional balloon (a Calabi-Yau manifold). Inside this balloon, there is a perfect, invisible fabric stretched out in a way that makes the surface perfectly balanced and smooth (a Ricci-flat Kähler metric). This is the "shape" of the universe in certain string theory models.

Now, imagine you start slowly deflating this balloon. As it shrinks, the fabric gets crumpled, stretched, and distorted. Mathematicians want to know: What does this fabric look like right before the balloon disappears?

This paper by Yang Li and Valentino Tosatti is about figuring out exactly how that fabric behaves when the balloon is deflating in a specific, tricky way called an "intermediate complex structure limit."

Here is the breakdown using everyday analogies:

1. The Two Types of Deflation

When these mathematical balloons shrink, they usually do one of two things:

• The "Flat" Collapse (m=0): The balloon shrinks down to a single point or a very simple shape. This was already well understood. It's like a balloon popping and leaving a tiny, crumpled piece of rubber on the floor.
• The "Long" Collapse (m=n): The balloon stretches out into a long, thin tube that looks like a donut (a torus) wrapped around a circle. This is the famous "Large Complex Structure Limit," which mathematicians have been studying for decades.

The Mystery: This paper tackles the "Intermediate" case (where $0 < m < n$). Imagine the balloon is shrinking, but it's not just becoming a point, nor is it becoming a simple long tube. It's becoming a tube wrapped around a smaller, shrinking balloon. It's a "tube inside a tube" situation. This is the "intermediate" zone that was harder to understand.

2. The "Generic Region" (The Safe Zone)

When you look at the deflating balloon, most of the fabric is behaving nicely, but there are a few "bad spots" (singularities) where the fabric is knotted or tearing.

• The authors focus on the "Generic Region." Think of this as the 99% of the balloon that is far away from the knots.
• In this safe zone, the fabric is smooth and predictable. The paper proves that in this safe zone, we can predict exactly what the fabric looks like as it shrinks.

3. The "Ansatz" (The Blueprint)

Before this paper, mathematicians had a blueprint (called an ansatz) for what the fabric should look like in this intermediate zone. They knew the rough shape, but they didn't have a precise map of the texture.

• The Old Result: They knew the blueprint was "close" to the real fabric (within a certain distance).
• The New Result: Li and Tosatti prove that the blueprint is perfectly accurate in the safe zone. The real fabric and the blueprint are essentially identical as the balloon shrinks.

4. The Tricky Part: The "Unwrapping" Problem

Why was this so hard?

• In the "Long Collapse" case (the simple tube), you could imagine unwrapping the tube like a roll of toilet paper to see the flat sheet underneath. Once unwrapped, standard math tools worked perfectly.
• In this "Intermediate" case, the structure is more complex. You can unwrap the outer layer (the tube), but the inner layer (the smaller balloon inside) is still shrinking.
• It's like trying to smooth out a rug that is being rolled up while you are trying to iron it. The math gets messy because the "ironing board" itself is changing size.

5. The Solution: A "Truncated" Ironing

The authors used a clever trick inspired by a mathematician named Savin.

• They realized that even though they couldn't iron the whole rug perfectly, they could iron a small patch of it perfectly, as long as they stopped before the rug got too small.
• They proved that if you zoom in close enough (but not too close), the fabric behaves like a smooth, flat sheet.
• They used a "stop-and-go" method:
  1. Step 1: Prove the fabric is roughly smooth (like checking the rug is flat).
  2. Step 2: Use a "Harnack Inequality" (a fancy way of saying "if one part is smooth, the neighbors must be smooth too") to show the smoothness spreads.
  3. Step 3: Use a "De Giorgi Iteration" (a process of zooming in and refining the view) to prove that the fabric isn't just smooth, but perfectly smooth (like a mirror).

The Big Takeaway

This paper is a major step forward in understanding the geometry of the universe in string theory. It tells us that even in these complex, "in-between" shrinking scenarios, the universe doesn't get chaotic. Instead, in the vast majority of places (the generic region), the geometry settles down into a very specific, predictable, and smooth pattern.

In short: They took a messy, shrinking, multi-layered mathematical balloon and proved that, for almost all of it, the surface becomes perfectly smooth and matches a known blueprint. They solved the puzzle of how the "middle" of the collapse behaves.

Technical Summary

Here is a detailed technical summary of the paper "Generic Regularity of Intermediate Complex Structure Limits" by Yang Li and Valentino Tosatti.

1. Problem Statement and Context

The paper investigates the asymptotic behavior of Ricci-flat Kähler metrics ($\omega_{CY,t}$) on a family of compact Calabi-Yau manifolds ($X_t$) as the complex structure degenerates. This is a central problem in the study of the Strominger-Yau-Zaslow (SYZ) conjecture and the geometry of collapsing Calabi-Yau manifolds.

The degenerations are classified by an integer $m \in \{0, \dots, n\}$, representing the real dimension of the essential skeleton $Sk(X)$ of the family:

• Case $m=0$: The metrics converge (without rescaling) to a singular Calabi-Yau metric on a variety with klt singularities. This case is well-understood.
• Case $m=n$ (Large Complex Structure Limit): The metrics must be rescaled by $|\log|t||^{-1}$. It was previously shown (by Li [9] and others) that on a "generic region" (filling almost the entire volume), the potentials converge smoothly to a solution of a real Monge-Ampère equation. This relies on Savin's small perturbation theorem.
• **Case $0 < m < n$(Intermediate Complex Structure Limits):** This is the focus of the current paper. Here, the geometry is more complex. The rescaled metrics are volume-collapsed, and the Gromov-Hausdorff limit is expected to be an$m$-dimensional simplicial complex. While previous work established$C^0$ convergence of the potentials to a non-Archimedean potential, the convergence of the actual metrics (i.e., higher-order regularity) remained an open problem.

The Core Question: Can the $C^0$ convergence of potentials be upgraded to $C^0$ (and higher) convergence of the metrics on the generic region for intermediate limits ($0 < m < n$)?

2. Methodology

The authors adapt the proof of Savin's small perturbation theorem (originally for elliptic PDEs) to the specific geometric setting of intermediate complex structure limits. The methodology involves several sophisticated steps:

A. Geometric Setup and Ansatz

The authors consider a specific family of Calabi-Yau hypersurfaces defined as transverse complete intersections inside a Fano manifold $M$.

• They construct a reference ansatz metric $\omega_t$ based on a convex potential $u$ solving a real Monge-Ampère equation on the skeleton.
• The actual Calabi-Yau metric is written as $\omega_{CY,t} = \omega_t + |\log|t|| i\partial\bar{\partial}\psi_t$.
• The goal is to prove $\|\omega_{CY,t} - \omega_t\|_{C^0} \to 0$ on a generic region $U_t$.

B. The Model Case and "Unwrapping"

To handle the analysis, the authors work in a "model case" where the geometry is a $T^m$-fibration over a base, with fibers being a compact Calabi-Yau $(n-m)$-fold $Y$.

• Key Difference from $m=n$: In the large limit ($m=n$), one can "unwrap" the $T^n$ fibers to apply standard elliptic regularity. In the intermediate case ($0 < m < n$), the base$Y$ is still collapsing (at a slower rate than the torus fibers).
• Strategy: The authors "unwrap" only the $T^m$ fibers. The remaining geometry involves a base $Y$ that is still collapsing. This prevents a direct application of Savin's theorem.

C. Truncated Estimates

The authors develop a truncated version of Savin's arguments:

1. Harnack Inequality: They prove a truncated Harnack inequality for the oscillation of the potential $\psi$.
   • They observe that the fiberwise maximum and minimum of $\psi$ act as viscosity subsolutions and supersolutions to a limiting real Monge-Ampère equation.
   • They apply "half" of Savin's results to these functions on the base.
   • Crucially, the estimates hold only up to the scale of the torus fibers ($|\log|t||^{-1}$), not down to the scale of the base $Y$.
2. De Giorgi Iteration: They perform a blow-up argument to improve the Harnack estimate to a quadratic approximation.
   • They use a contradiction argument involving sequences of rescaled solutions.
   • A critical point is showing that the fiberwise maximum and minimum of the potential coincide in the limit, allowing the limit function to be harmonic on the base.
   • This yields a truncated Hölder bound valid only up to the fiber scale.

D. Bootstrapping to Smoothness

Once the truncated Hölder bound is established, the authors rescale the torus fibers. In this rescaled coordinate system, the metric becomes uniformly equivalent to a standard Euclidean metric.

• They apply Savin's original theorem (which requires $C^0$ smallness and uniform ellipticity) in these rescaled coordinates.
• This allows them to "bootstrap" the $C^0$ estimate into $C^\infty$ estimates for the potential $\psi$ with respect to the stretched coordinates.

3. Key Contributions and Results

Main Theorem (Theorem 1.1):
On the generic region $U_t \subset X_t$ (where the normalized Calabi-Yau measure tends to 1 as $t \to 0$), the Calabi-Yau metric converges in the $C^0$ sense to the ansatz metric:
$$\lim_{t \to 0} \|\omega_{CY,t} - \omega_t\|_{C^0(U_t, \omega_t)} = 0$$

Corollary 5.1 (Higher Order Estimates):
The authors prove that after stretching the base coordinates and the $T^m$-fibers by a factor of $|\log|t||^{1/2}$, the Calabi-Yau metric satisfies $C^\infty$ estimates. Specifically, the difference between the true metric and the ansatz metric vanishes in all derivatives on the generic region.

Technical Innovations:

• Handling Mixed Scales: Successfully adapting elliptic regularity theory to a setting where two different collapsing scales exist (torus fibers vs. base manifold $Y$).
• Truncated Regularity: Proving that regularity holds up to the fiber scale, which is sufficient to apply Savin's theorem after rescaling, even though the full geometry is not uniformly non-collapsed.
• Viscosity Solutions: Utilizing the fiberwise max/min as viscosity solutions to bridge the gap between the complex Monge-Ampère equation on the total space and the real Monge-Ampère equation on the base.

4. Significance

• Resolution of Intermediate Limits: This paper fills a major gap in the understanding of Calabi-Yau degenerations, extending the smooth convergence results from the "Large Complex Structure Limit" ($m=n$) to the "Intermediate" case ($0 < m < n$).
• SYZ Conjecture: The results provide strong evidence for the metric version of the SYZ conjecture in intermediate cases, suggesting that the geometry is indeed modeled by a fibration over a lower-dimensional base with specific singular fibers.
• Future Applications:
  • Coisotropic Fibrations: The authors suggest that these higher-order estimates could be used (via implicit function theorems) to construct coisotropic fibrations on the generic region, analogous to the special Lagrangian torus fibrations constructed in the $m=n$ case.
  • General Collapsing: The techniques are applicable to other collapsing scenarios, such as families of Ricci-flat metrics on a fixed manifold fibering over a lower-dimensional base (as discussed in Remark 1.5), though the estimates obtained are slightly weaker than those in specific prior works due to the different scaling regimes.

In summary, Li and Tosatti have successfully generalized the regularity theory for collapsing Calabi-Yau metrics, demonstrating that despite the complex mixed-scale geometry of intermediate limits, the metrics behave smoothly and predictably on the generic part of the manifold.