🔢 mathematics

Real algebraic surfaces biholomorphically equivalent but not algebraically equivalent

This paper resolves a long-standing open question by providing a counterexample of two germs of real algebraic surfaces in $\mathbb{C}^2$ that are biholomorphically equivalent but not algebraically equivalent, thereby proving that biholomorphic equivalence does not imply algebraic equivalence for such germs.

Original authors: Guillaume Rond

Published 2026-05-27

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Guillaume Rond

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

The Shape-Shifting Surfaces: A Story of Two Twins

Imagine you are an architect who designs buildings. In your world, there are two ways to say two buildings are "the same":

The Smooth Way (Biholomorphic): You can stretch, twist, and mold the first building into the shape of the second using a perfectly smooth, continuous motion. No tearing, no folding, just a fluid transformation.
The Algebraic Way (Algebraically Equivalent): You can do the same transformation, but with a strict rule: every step of your stretching and twisting must follow a specific set of mathematical "recipes" (polynomials and equations). You can't just use any smooth motion; it has to be one that fits a strict algebraic formula.

For a long time, mathematicians wondered: If you can turn Building A into Building B using any smooth motion, can you always do it using one of those strict algebraic recipes?

Most experts thought the answer was "Yes." They knew that for many types of surfaces, if a smooth transformation exists, an algebraic one must exist too. But this paper by Guillaume Rond says: No, not always.

The Cast of Characters: The "Bishop Surfaces"

The author introduces a special family of shapes called Bishop surfaces. Think of these as delicate, curved sheets floating in a 4-dimensional space (which we can imagine as two complex dimensions).

These surfaces have a unique quirk:

At the very center (the origin), they look like a flat, complex line.
But as soon as you move even a tiny bit away from the center, they twist into a "totally real" shape, losing their complex symmetry.

It's like a piece of paper that is perfectly flat at the exact center point but immediately crumples into a chaotic, non-complex shape the moment you touch it.

The Experiment

The author picks two specific versions of these surfaces:

Surface A: A Bishop surface defined by a specific equation involving a number called $\gamma$ (gamma).
Surface B: A "Normal Form" of that surface. This is the simplest, most standardized version of Surface A that mathematicians know how to describe.

The Good News:
The author proves that Surface A and Surface B are smoothly equivalent. There exists a perfect, fluid way to morph Surface A into Surface B. They are twins in the eyes of smooth geometry.

The Bad News (The Twist):
The author then asks: Is there an algebraic recipe to do this?

He chooses a very special number for $\gamma$ . It's not just any number; it's a transcendental number (like $\pi$ or $e$ , but even more "random" in a mathematical sense). It doesn't fit into any neat polynomial equation.

Using a clever logical trap, the author shows that:

If you try to use an algebraic recipe to turn Surface A into Surface B, the math breaks down.
The "ingredients" required for the algebraic recipe would have to be made of the same transcendental number $\gamma$ .
But algebraic recipes are built from simpler, rational ingredients. You can't build a complex, transcendental structure out of simple, rational blocks in this specific context.

The Conclusion: The "Impossible" Transformation

The paper concludes with a counter-example that settles a long-standing debate.

The Reality: Surface A and Surface B are biholomorphically equivalent (they can be morphed into each other smoothly).
The Limitation: They are not algebraically equivalent. There is no "recipe" that follows the strict rules of algebraic power series to perform this morph.

A Simple Analogy

Imagine you have a lump of clay (Surface A) and you want to turn it into a perfect sphere (Surface B).

The Smooth Way: You can easily mold the clay with your hands. It's a smooth, continuous process.
The Algebraic Way: Imagine you are a robot programmed with a specific set of instructions (algebraic equations). The robot can only move the clay in ways that follow its code.

The author's discovery is like finding a lump of clay that can be turned into a sphere by a human hand, but cannot be turned into a sphere by any robot, no matter how complex the robot's code is, because the shape of the clay requires a "secret ingredient" (the transcendental number) that the robot's code simply cannot generate.

Why This Matters (According to the Paper)

This paper doesn't claim to solve engineering problems or predict physical phenomena. Its impact is purely mathematical. It answers a specific, decades-old question in the field of complex geometry: "Does smooth equivalence always imply algebraic equivalence?"

The answer is a definitive No.

By constructing this specific example of two "Bishop surfaces," the author proves that the world of smooth shapes is richer and more flexible than the world of algebraic shapes. Sometimes, you need a fluid, non-algebraic touch to make two things match, and no amount of algebraic formula can replicate that touch.

Technical Summary: Real Algebraic Surfaces Biholomorphically Equivalent but Not Algebraically Equivalent

Problem Statement
This paper addresses a long-standing open question in complex geometry and real algebraic geometry: whether biholomorphic equivalence implies algebraic equivalence for germs of real algebraic manifolds in $\mathbb{C}^n$ . Specifically, given two germs of smooth real algebraic manifolds $(M, 0)$ and $(M', 0)$ in $\mathbb{C}^n$ , if there exists a biholomorphism $\Phi$ mapping $M$ to $M'$ , does there necessarily exist a biholomorphism $\Psi$ defined by algebraic power series (i.e., power series satisfying a non-trivial polynomial equation with polynomial coefficients) such that $\Psi(M) = M'$ ?

While the answer is known to be positive for several important classes of manifolds—including Levi non-degenerate real algebraic hypersurfaces (Webster, 1977) and real algebraic generic manifolds of finite type that are holomorphically non-degenerate (Berteloot, Ebenfelt, and Rakhimov, 1996)—the general case for real algebraic manifolds remained unresolved. Previous work by Moser and Webster (1983) demonstrated that formal equivalence does not always imply convergent (biholomorphic) equivalence for Bishop surfaces, but the specific question of whether convergent equivalence implies algebraic equivalence had not been settled.

Methodology
The author constructs a counterexample using Bishop surfaces, a class of real surfaces in $\mathbb{C}^2$ characterized by having a totally real tangent space everywhere except at the origin, where the tangent space is a complex line.

Construction of the Counterexample:
The paper considers the family of Bishop surfaces $M_\gamma$ defined by:
$M_\gamma := \{(z, w) \in \mathbb{C}^2 \mid w = |z|^2 + \gamma(z^2 + \bar{z}^2)\}$
where $\gamma \in (0, 1/2)$ is a transcendental number.
Existence of Biholomorphic Equivalence:
Citing Moser and Webster (1983), the paper notes that for $\gamma \in (0, 1/2)$ , the surface $M_\gamma$ admits a convergent biholomorphic map $\Phi$ to a normal form $N_{\gamma, \epsilon, s}$ (specifically, a surface defined by $w = |z|^2 + (\gamma + \epsilon w^s)(z^2 + \bar{z}^2)$ with $\text{Im}(w)=0$ ). Thus, $M_\gamma$ is biholomorphically equivalent to its normal form.
Proof of Non-Algebraic Equivalence:
The core of the proof proceeds by contradiction. The author assumes there exists a biholomorphism $\Phi = (\phi_1, \phi_2)$ with components in the ring of algebraic power series $\mathbb{C}\langle z, w \rangle$ that maps $M_\gamma$ to the normal form (specifically, mapping it into the real hyperplane $\text{Im}(w)=0$ ).
- Algebraic Structure Analysis: The proof utilizes properties of the ring of algebraic power series, including the Implicit Function Theorem, faithful flatness, and the behavior of coefficients under field extensions.
- Coefficient Field Reduction: Using Lemma 6 and Lemma 5, the author reduces the coefficients of the assumed algebraic map to a finitely generated field extension $L$ of $\mathbb{Q}(\gamma)$ . By further analysis (Lemma 7, an Eisenstein-type theorem), the coefficients are shown to be rational functions of the transcendental parameters $\gamma_1, \dots, \gamma_r$ .
- Parameter Deformation: The author constructs a family of surfaces $M_\delta$ parameterized by $\delta \in \mathbb{R}$ . By exploiting the algebraic independence of the transcendental $\gamma$ , the proof shows that if an algebraic equivalence exists for the transcendental $\gamma$ , then an algebraic equivalence must exist for a "nearby" parameter $\delta_1$ chosen from a dense set of real numbers.
- Contradiction via Exceptional Parameters: The author selects $\delta_1 > 1/2$ to be an "exceptional" parameter (where the solutions to $\gamma T^2 - T + \gamma = 0$ are not roots of unity). Moser and Webster (1983, Theorem 6.1) proved that for such exceptional parameters, no convergent biholomorphism exists that maps the surface to the normal form. Since algebraic power series are convergent, the existence of an algebraic map for $\delta_1$ would imply the existence of a convergent map, which is a known impossibility.

Key Contributions and Results

Theorem 1: The paper establishes that if $\gamma \in (0, 1/2)$ is a transcendental number, the Bishop surface $M_\gamma$ is biholomorphically equivalent to its normal form, but not by means of an algebraic biholomorphism.
Negative Answer to the General Question: This result provides a definitive negative answer to the question posed in [BER00] and [Mir13]: biholomorphic equivalence does not imply algebraic equivalence for germs of real algebraic manifolds in general.
Refinement of Bishop Surface Theory: The work clarifies the distinction between formal, convergent, and algebraic equivalences specifically within the context of Bishop surfaces, highlighting the role of transcendental invariants.

Significance
The paper resolves a specific, open problem regarding the rigidity of real algebraic structures under biholomorphic mappings. It demonstrates that while the analytic structure of these surfaces is rigid enough to admit a convergent normal form (for $\gamma \in (0, 1/2)$ ), the algebraic structure is not preserved by this equivalence. The result underscores that the class of algebraic power series is strictly smaller than the class of convergent power series in the context of mapping real algebraic manifolds, even when the manifolds themselves are algebraic. The author modestly frames the work as providing a specific counterexample to a general conjecture, relying on the established non-convergence results for exceptional parameters by Moser and Webster to drive the contradiction.