🔢 mathematics

Pullback Method with Applications to Severi--Brauer Fibrations

This paper introduces a general pullback construction for generating varieties with Brauer–Manin obstructions and applies it to Severi–Brauer fibrations to demonstrate the existence of such fibrations with index one that fail the Hasse principle.

Original authors: Mridul Biswas, Divyasree C Ramachandran, Biswanath Samanta

Published 2026-07-30

📖 7 min read🧠 Deep dive

CC BY 4.0

Original authors: Mridul Biswas, Divyasree C Ramachandran, Biswanath Samanta

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 Great Treasure Hunt of Math

Imagine you are a detective trying to find a hidden treasure on a map. In the world of mathematics, specifically a field called arithmetic geometry, the "treasure" is a rational point—a solution to an equation that uses only simple, whole-number-like fractions. The map is a shape called a variety, and the clues are the local conditions: does the treasure exist in every single neighborhood (or "place") of the map?

For a long time, mathematicians believed that if you found the treasure in every neighborhood, you were guaranteed to find it in the center of the map too. This idea is called the Hasse Principle. It's like saying, "If I can find a lost dog in every park in the city, the dog must be in the city center." But sometimes, the dog is nowhere to be found, even though it seems to be everywhere. This is a "failure of the Hasse Principle."

To explain these missing treasures, mathematicians use a special tool called the Brauer group. Think of the Brauer group as a set of invisible "ghosts" or "magic spells" that float over the shape. These spells can create a Brauer–Manin obstruction. It's like a magical force field that allows the treasure to appear in every neighborhood but makes it vanish the moment you try to bring all the clues together to find the real location. The big question in this field is: Can we predict exactly when these ghosts will hide the treasure, and what kind of ghosts are doing the hiding?

The Paper's Discovery: A Magic Trick with Polynomials

In this paper, the authors—Mridul Biswas, Divyasree C Ramachandran, and Biswanath Samanta—introduce a clever new way to build shapes that hide their treasures, even when they look like they should be easy to find. They call this the "Pullback Method."

Imagine you have a shape (a variety) that has a few "ghosts" (Brauer classes) floating around it. Sometimes, these ghosts are weak; they might stop you from finding the treasure in the center, but they don't stop you from finding a "zero-cycle of degree one." In math-speak, a zero-cycle is like a collection of points that, when you add up their "weights" (degrees), equals one. If you can find a collection of points that adds up to one, the shape is said to have an index of one. Usually, having an index of one is a strong hint that a real treasure (a rational point) exists.

The authors' main trick is to take a shape that already has a treasure (a rational point) and a ghost that causes trouble, and then "pull it back" using a special polynomial (a fancy algebraic recipe). This is like taking a piece of fabric with a pattern and stretching it over a new frame. The new shape looks different, but it inherits the ghosts from the original.

Here is the magic: By carefully choosing this polynomial, they can stretch the shape in a way that:

Keeps the local clues: The new shape still has treasures in every single neighborhood (it satisfies the local conditions).
Activates the ghost: The ghost becomes strong enough to block the final treasure hunt, making the set of possible global solutions empty.
Keeps the index at one: Even though the treasure is hidden, the shape still has that "collection of points adding up to one" property.

This is a big deal because it proves that having an index of one does not guarantee a rational point, even when the only thing stopping you is this specific type of ghost.

The Specifics: Odd Primes and Cyclic Algebras

The authors don't just say this is possible; they build specific examples to prove it. They focus on shapes called Severi–Brauer fibrations. You can think of these as a bundle of smaller shapes (like a stack of pancakes) where each pancake is a "Severi–Brauer variety."

They show that for any odd prime number (like 3, 5, 7, etc.), they can construct a shape where the ghost blocking the treasure is a "p-torsion" class. In simple terms, this means the ghost has a specific "order" or "power" related to that prime number. If you try to cancel the ghost out by adding it to itself $p$ times, it disappears.

Their main result, Theorem 1.5, states that if you have a number field (a type of number system) that contains a special root of unity (a complex number that cycles back to 1), you can build a Severi–Brauer fibration with:

An index of one (it has a "degree one" collection of points).
Local solutions everywhere (it looks like it has a treasure in every neighborhood).
No global solution (the treasure is actually missing).
The obstruction is captured by a group of size $p$ (where $p$ is your chosen odd prime).

This leads to Corollary 1.6, which answers a big question in the field: Which groups of ghosts can hide a treasure? The authors prove that the group $\mathbb{Z}/p\mathbb{Z}$ (the integers modulo $p$ ) can capture the Brauer–Manin obstruction for any odd prime $p$ . This means there is no upper limit to how complex these "ghost groups" can be; they can be as large as you want, depending on the prime number you choose.

How They Did It (Without the Heavy Math)

The authors didn't just guess; they used a step-by-step construction.

The Setup: They started with a cyclic algebra (a specific type of mathematical object) defined over a field of rational functions. This algebra acts as the "seed" for the ghost.
The Construction: They built a variety (a shape) over a line ( $\mathbb{P}^1$ ) where the "ghost" is non-trivial. They used a theorem called the Purity Theorem to ensure the ghost stays "pure" and doesn't get diluted or disappear when they look at the whole shape.
The Pullback: They applied their "Pullback Method" (Theorem 1.1). They found a polynomial $g(t)$ that, when used to stretch the shape, ensured that the "ghost" would block the global solution while keeping the local solutions intact.
The Proof: They showed that for their specific example (using the prime $p$ and a number field with a primitive $p$ -th root of unity), the resulting shape has an index of one but fails the Hasse principle due to the Brauer–Manin obstruction.

Why This Matters

Before this paper, it was known that the Brauer–Manin obstruction could hide treasures. But the relationship between the "index" (the degree of point collections) and the "obstruction" was a bit murky. Some thought that if the index was one, the treasure must exist. This paper shatters that hope. It shows that even when the index is one, the Brauer–Manin obstruction can still be the only thing standing between you and the treasure.

Furthermore, they answer the question of which groups can be the "bad guys." They prove that for any odd prime, the cyclic group of that size can be the culprit. This expands our understanding of the "menu" of obstructions available in arithmetic geometry.

In short, the authors have built a mathematical machine that takes a shape with a treasure, stretches it with a polynomial, and creates a new shape that looks like it should have a treasure, has a "degree one" signature, but is actually empty. They did this for any odd prime, showing that the universe of these mathematical ghosts is vast and varied.

Technical Summary: Pullback Method with Applications to Severi–Brauer Fibrations

Problem Statement
The paper addresses the failure of the Hasse principle for rational points on smooth projective geometrically integral varieties (referred to as "nice" varieties) defined over global fields. While the existence of local points ( $X(k_v) \neq \emptyset$ for all places $v$ ) is a necessary condition for the existence of a global rational point ( $X(k) \neq \emptyset$ ), it is not sufficient. The primary obstruction studied is the Brauer–Manin obstruction, defined via the pairing between the adelic points $X(\mathbb{A}_k)$ and the Brauer group $\text{Br}(X)$ .

A significant gap in the literature is the reliance on explicit defining equations to compute Brauer groups and demonstrate such obstructions. Furthermore, while it is known that the Brauer–Manin obstruction can prevent the existence of rational points even when the index of the variety is one (i.e., there exists a zero-cycle of degree one), constructing explicit examples of Severi–Brauer fibrations with index one that fail the Hasse principle without relying on specific polynomial models remains a challenge. Additionally, the question of which finite abelian groups can capture the Brauer–Manin obstruction (i.e., be the minimal subgroup of $\text{Br}(X)$ responsible for the obstruction) is not fully resolved.

Methodology
The authors develop a general "pullback method" that constructs varieties violating the Hasse principle without requiring explicit defining equations. The core technique involves:

Pullback Construction: Starting with a variety $X$ over a global field $k$ equipped with a dominant morphism $\pi: X \to \mathbb{P}^1$ having a geometrically irreducible generic fiber, and a Brauer class $\alpha \in \text{Br}(X)$ that obstructs weak approximation, the authors construct a base change $X_g$ via a carefully chosen non-constant polynomial $g \in k[t]$ .
Control of Local Images: By leveraging weak approximation on $\mathbb{P}^1$ , the polynomial $g$ is selected to manipulate the images of the evaluation maps of the Brauer class at specific places. Specifically, $g$ is chosen such that the pullback variety $X_g$ retains local points everywhere (ensuring $X_g(\mathbb{A}_k) \neq \emptyset$ ) but forces the sum of local invariants of the pulled-back Brauer class to be non-zero for all adelic points, thereby creating a Brauer–Manin obstruction to the existence of rational points.
Severi–Brauer Fibrations: The method is applied to Severi–Brauer fibrations. The authors construct these fibrations using cyclic algebras $(a, f)$ over the function field $k(t)$ . By utilizing the purity theorem for the Brauer group and base changing to Henselian local rings, they analyze the Brauer group of the resulting fibration $X_{a,f}$ purely through its generic fiber properties, avoiding explicit global equations.
Index One Construction: To achieve index one, the authors construct a specific fibration where the existence of a zero-cycle of degree one is guaranteed by the presence of closed points of various degrees (derived from the factorization of the defining polynomial $f$ ) and the existence of fibers of degree $p$ .

Key Contributions and Results

General Pullback Theorem (Theorem 1.1): The paper establishes that if a variety $X$ has a rational point and a Brauer class $\alpha$ obstructing weak approximation, there exists a base change $X_g$ that admits a Brauer–Manin obstruction to the existence of rational points. This theorem relies only on the existence of a suitable Brauer class and a rational point, not on explicit models.
Brauer Groups of Severi–Brauer Fibrations (Theorem 4.3): The authors characterize the Brauer group of Severi–Brauer fibrations associated with cyclic algebras $(a, f)$ . They prove that under specific conditions on the polynomial $f$ and the constant $a$ , the vertical Brauer group is $p$ -torsion and non-trivial, generated by classes pulled back from the base.
Existence of Index One Counterexamples (Theorem 1.5): The paper constructs Severi–Brauer fibrations over number fields containing a primitive $p$ -th root of unity (for an odd prime $p$ ) that have index one but fail the Hasse principle due to a Brauer–Manin obstruction captured by a $p$ -torsion class.
Capturing Groups (Corollary 1.6): As a consequence of the construction, the authors demonstrate that for any odd prime $p$ , the cyclic group $\mathbb{Z}/p\mathbb{Z}$ can capture the Brauer–Manin obstruction. This implies that the exponent of a finite group capturing the obstruction has no upper bound.

Significance and Claims
The paper claims to provide a systematic, equation-free approach to generating examples of varieties with Brauer–Manin obstructions. By shifting the focus from explicit models to the properties of the generic fiber and the functoriality of the Brauer group, the method extends applicability to a broader class of varieties.

The authors position their work as a step toward understanding the structure of obstructions in rational connected varieties, supporting the Colliot-Thélène conjecture that the Brauer–Manin obstruction is the only obstruction for such varieties. Specifically, they show that the failure of the Hasse principle can occur even when the index is one, a phenomenon where the obstruction to rational points is distinct from the obstruction to zero-cycles of degree one.

The results offer a partial answer to the question of which finite abelian groups can capture the Brauer–Manin obstruction, confirming that $\mathbb{Z}/p\mathbb{Z}$ is sufficient for any odd prime $p$ . The authors acknowledge that a preprint by Liang and Liu suggests any finite abelian group can capture the obstruction, but their work provides a concrete construction for the cyclic case using the pullback method on Severi–Brauer fibrations.

The Great Treasure Hunt of Math

The Paper's Discovery: A Magic Trick with Polynomials

The Specifics: Odd Primes and Cyclic Algebras

How They Did It (Without the Heavy Math)

Why This Matters

More like this