🔢 mathematics

Nash's theorem via Günther's trick

This paper presents an accessible and rigorous proof of Nash's smooth embedding theorem by utilizing Günther's trick.

Original authors: Anton Petrunin

Published 2026-06-12

📖 6 min read🧠 Deep dive

Original authors: Anton Petrunin

Original paper dedicated to the public domain under CC0 1.0 (http://creativecommons.org/publicdomain/zero/1.0/). ✨ 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

The Big Picture: The "Shape-Shifting" Puzzle

Imagine you have a piece of fabric with a specific pattern of wrinkles and folds. In mathematics, this fabric is a Riemannian manifold (a curved surface), and the pattern of wrinkles is its metric (the rule that tells you how to measure distance and angles on that surface).

Nash's Theorem asks a seemingly impossible question: Can you take this crumpled, curved fabric and lay it perfectly flat onto a giant, smooth table (Euclidean space) without stretching, tearing, or distorting the wrinkles?

The answer is yes. You can embed any curved shape into a high-dimensional flat space while keeping all the distances exactly the same.

However, finding the exact way to do this is incredibly hard. The paper by Anton Petrunin explains a proof of this theorem that is easier to understand than the original, thanks to a "magic trick" discovered by a mathematician named Günther.

The Problem: The "Loss of Derivatives" Trap

To solve this puzzle, Nash originally used a method similar to Newton's method (a way of guessing and correcting).

The Analogy: Imagine trying to balance a broom on your hand. You guess a position, see it's off, and make a correction.
The Trap: In Nash's original math, every time you made a correction to fix the shape, you needed to know more about the smoothness of the fabric than you started with. It was like trying to fix a blurry photo, but every time you sharpened it, the image became more blurry in other ways. This is called the "loss of derivatives" problem. To fix it, Nash had to invent a very complex, delicate smoothing process.

The Solution: Günther's "Magic Trick"

Matthias Günther found a way to make that "loss of derivatives" problem disappear. Petrunin's paper walks us through Günther's approach, which is much cleaner.

Instead of trying to fix the whole shape at once with complex smoothing, Günther's method uses a contraction argument.

The Analogy: Imagine you are trying to fit a large, squishy pillow into a tight box. Instead of forcing it, you have a machine that gently squeezes the pillow. Every time the machine squeezes, the pillow gets slightly smaller and fits a bit better. Eventually, after many squeezes, the pillow fits perfectly.
The Math: Günther's trick turns the problem into a loop where you make a small guess, check the error, and make a correction that is guaranteed to be smaller than the error. Because the corrections get smaller and smaller, the process naturally settles on the perfect solution without needing those messy smoothing steps.

The Steps of the Proof (Simplified)

Petrunin breaks the proof down into four main stages:

1. The "Approximate" Version (The Rough Draft)
First, the paper shows that you can get close to the right shape.

The Trick: Nash invented a "twist." Imagine you have a flat sheet. You can twist it into a spiral or a wave. By combining many of these twists (some big, some small), you can create a surface that looks almost exactly like the curved fabric you want, even if it's not perfect yet.
The Result: This proves that you can get arbitrarily close to the target shape.

2. The Reduction (The "Torus" Shortcut)
Proving this for every possible shape is too hard. So, the paper reduces the problem to a specific, manageable shape: a Torus (a donut shape).

The Logic: If you can prove you can flatten a donut with any pattern of wrinkles, you can prove you can flatten anything. This is because any complex shape can be broken down into pieces that look like a donut, and you can stitch the solutions together.

3. The "Free Map" and The Linear Fix
To solve the donut problem, the authors use a "Free Map."

The Analogy: Think of a "Free Map" as a very flexible, stretchy net that is already laid out in a high-dimensional space. It's so flexible that it has no "kinks" or hidden constraints.
The Linear Step: The authors show that if you have a small error (a tiny wrinkle), you can fix it using a simple linear equation (like solving $Ax = B$). This is the easy part.

4. The "Günther's Lemma" (The Heavy Lifting)
This is the core of the paper. The authors need to solve a non-linear equation (where the variables interact in complex ways).

The Problem: The simple linear fix isn't enough for the whole puzzle.
The Magic: Günther's Lemma introduces a special "correction term."
- The Tangential Trick: Usually, when you move a surface, you move it "outward" (normal direction). But Günther's trick allows you to slide the surface "sideways" (tangential direction).
- Why it works: Sliding sideways doesn't change the distances on the fabric much (it's like shuffling a deck of cards; the cards are still in the same order, just moved slightly). By adding this "sideways slide" to the math, they can cancel out the messy parts of the equation that cause the "loss of derivatives."
The Smoothing Operator: The paper uses a mathematical tool (based on the Laplacian, which is like a heat equation) to smooth out the rough edges of the solution. Because of the "sideways slide" trick, this smoothing works perfectly without breaking the math.

The Conclusion

By combining these steps, the paper proves that:

You can get close to the shape using Nash's twists.
You can reduce the problem to a donut.
You can use Günther's "sideways slide" trick to fix the remaining errors without losing control of the math.
Therefore, any curved shape can be perfectly embedded into a flat, high-dimensional space.

In short: The paper takes a famous, notoriously difficult mathematical proof and re-arranges the furniture so that the "magic trick" (Günther's approach) does the heavy lifting, making the logic much clearer and more rigorous than before.

Technical Summary: Nash's Theorem via Günther's Trick

Problem Formulation
The paper addresses John Nash's smooth embedding theorem, which asserts that any $n$ -dimensional Riemannian manifold $(\Omega, g)$ admits a smooth isometric embedding into a Euclidean space $\mathbb{R}^d$ for some sufficiently large $d$ . The author, Anton Petrunin, focuses on providing a proof that balances accessibility with rigor, specifically for compact manifolds, while indicating the reduction to the non-compact case. The central difficulty in the original proof by Nash involves the "loss-of-derivatives" problem inherent in applying Newton's method to the nonlinear system of partial differential equations governing isometric embeddings.

Methodology and Approach
The proof strategy diverges from Nash's original use of a modified Newton method with delicate smoothing procedures. Instead, it employs a contraction argument within Banach spaces, a technique facilitated by a "magic way" discovered by Matthias Günther that eliminates the loss-of-derivatives issue. The paper follows Nash's structural framework but integrates Günther's lemma to handle the realization of small metric perturbations.

The methodology proceeds through the following logical stages:

Approximate Theorem and Nash's Twist: The paper first establishes an approximate version of the embedding theorem. It utilizes "Nash's twist," a specific construction involving a composition of maps to a circle, to generate a family of maps $w_t$ that induce metrics of the form $g + t \cdot h$ . This demonstrates that the set of metrics induced by smooth maps is dense in the space of Riemannian metrics.
Reduction to Torus Perturbations: The global embedding problem is reduced to a local problem on the $n$ -dimensional torus $T^n$ . Specifically, the proof reduces the main theorem to showing that any metric $g$ sufficiently close to a fixed metric $g_0$ on $T^n$ can be realized as the induced metric of a smooth map. This reduction relies on the Whitney embedding theorem and partition of unity arguments to extend results from the torus to general compact manifolds.
Free Maps and Linearization: A fixed "free" smooth embedding $f: T^n \to \mathbb{R}^d$ is introduced. A map is "free" if its first and second-order partial derivatives are linearly independent at every point. This property allows the linearization of the metric realization problem. The induced metric is expressed via a symmetric bilinear form $Q$ , and the problem of finding a perturbation $w$ such that $Q(f+w, f+w) = g$ is linearized to $L(w) + Q(w, w) = h$ , where $L$ is a linear operator derived from $f$ .
Nash's Lemma: The paper proves that the linear operator $L$ admits a right inverse $M$ that preserves regularity (mapping Hölder spaces to Hölder spaces without loss of derivatives). This is possible because $f$ is free, allowing the unique definition of a normal vector field satisfying specific orthogonality conditions.
Günther's Lemma and Contraction: The core innovation is Günther's lemma, which constructs a symmetric bilinear form $\tilde{Q}$ such that $L(\tilde{Q}(w, w)) = Q(w, w)$ . Crucially, $\tilde{Q}$ satisfies specific Hölder norm estimates that allow the equation $w = Mh - \tilde{Q}(w, w)$ to be solved via a contraction mapping argument in the Banach space $C^{2,\alpha}$ .
Smoothing and Regularity: To ensure the solution $w$ is smooth ( $C^\infty$ ) rather than just Hölder continuous, the proof utilizes a smoothing operator based on the inverse of the Laplacian minus the identity, $(\Delta - 1)^{-1}$ . This operator commutes with partial derivatives and gains regularity, allowing the authors to bootstrap the solution from $C^{2,\alpha}$ to $C^\infty$ .

Key Contributions and Results

Günther's Trick: The paper provides a detailed exposition of Günther's method, which replaces the complex smoothing procedures of the Nash-Moser theorem with a standard contraction argument in Banach spaces. This simplifies the analytical machinery required to prove the embedding theorem.
Explicit Bounds: While the author notes that optimizing the dimension $d$ is not the primary goal, the argument yields a bound of $d = 10 \cdot n^4$ . For compact surfaces, the bound $d=5$ is noted as achievable.
Structural Decomposition: The paper systematically breaks down the proof into manageable components: the approximate theorem, the reduction to the torus, the properties of free maps, and the realization of small perturbations via Günther's lemma.
Tangential Correction: A specific technical contribution is the introduction of a tangential correction term $x$ in the definition of $\tilde{Q}$ . This term, which depends quadratically on $w$ , is necessary to satisfy the required estimates and is analogous to techniques used in the DeTurck trick for Ricci flow.

Significance and Claims
The paper claims to present a proof of Nash's smooth embedding theorem that emphasizes accessibility and rigor without sacrificing mathematical depth. By following Nash's proof structure but substituting the solution of the nonlinear equation with Günther's contraction argument, the author aims to make the theorem more approachable for readers familiar with smooth manifolds and standard analysis tools (like Schauder estimates and the Whitney embedding theorem), treating the latter as "black boxes."

The author explicitly states that the paper does not optimize the dimension bound $d$ and does not claim to improve upon the known optimal bounds for specific cases (like $d=5$ for surfaces). The significance lies in the pedagogical clarity of the proof and the demonstration of how Günther's approach simplifies the historical "loss-of-derivatives" obstacle. The paper serves as a polished lecture-note style derivation, incorporating minor cosmetic tweaks from other mathematicians (Deane Yang, Terence Tao, Ralph Howard) to streamline the presentation.