🔢 mathematics

Murphy's law in non-abelian Hodge theory

This paper constructs explicit examples of non-integral variations of $\mathbb{Q}$ -Hodge structures using Fenchel--Nielsen parameterizations to demonstrate that, in the absence of integral structures, the Hodge locus exhibits pathological behaviors such as the failure of the Cattani--Deligne--Kaplan theorem and the André--Oort conjecture.

Original authors: Gregorio Baldi, Yeuk Hay Joshua Lam

Published 2026-08-05

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Gregorio Baldi, Yeuk Hay Joshua Lam

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

Imagine you are a detective trying to solve a mystery about the hidden shapes of the universe. In the world of mathematics, there is a famous rule called the "Hodge Conjecture." Think of it like a treasure map: it tells you that if you find a certain kind of hidden pattern (a cohomology class) in a complex geometric shape, that pattern must have been built by stacking up simpler, tangible blocks (algebraic cycles). It's a way of saying that the most mysterious parts of these shapes are actually made of something you can count and build with.

Now, mathematicians have been trying to apply this same logic to a much stranger, more twisted kind of geometry called "non-abelian Hodge theory." Here, instead of simple blocks, the "patterns" are like invisible forces or vibrations traveling through the shape, described by things called "local systems." The big question is: If we find one of these invisible patterns that follows the rules of Hodge theory, does it have to be built from "geometric" blocks, or can it be something wild and random? Usually, mathematicians hope the answer is "yes, it must be geometric." But this paper asks a cheeky question: What if the rules break? What if we can build a pattern that looks perfect on the outside but is made of "broken" pieces that don't fit the standard integer grid? This is where the title comes in: "Murphy's Law," the idea that if something can go wrong, it will. The authors are hunting for these "broken" patterns to see just how badly the rules can fail.

The Paper: Hunting for the "Broken" Patterns

In this paper, Gregorio Baldi and Yeuk Hay Joshua Lam set out to prove that in the world of non-abelian Hodge theory, things don't just go wrong occasionally—they go wrong spectacularly. They construct explicit examples of "non-integral" variations of Hodge structures. To understand what that means, imagine you are building a wall. A "standard" wall is made of whole bricks (integers). A "non-integral" wall is made of bricks that have been chopped into weird fractions that don't fit together nicely on a standard grid.

The authors show that you can create these weird, fraction-based walls (called $\mathbb{Q}$ -VHS) that still look like perfect, smooth structures from a distance, but they are fundamentally "broken" because they aren't made of whole numbers. They prove that for any curve with a certain amount of twisting (genus $g \ge 2$ ), there are infinitely many different shapes that can support these broken, non-integral patterns. These patterns have a special property: they are "Zariski dense," which is a fancy way of saying they wiggle around so much that they fill up the entire space they are in, refusing to stay in any small, neat corner.

The "Murphy's Law" Discovery

The most exciting part of their work is showing that when you switch from whole-number bricks to these fraction-based ones, the usual safety nets of mathematics disappear. In the world of whole numbers, there are famous theorems (like the Cattani–Deligne–Kaplan theorem) that guarantee the "Hodge locus"—the set of points where these patterns behave in a special, "magic" way—looks like a neat collection of algebraic curves or surfaces.

But the authors prove that for their non-integral examples, this neatness vanishes. They show that the Hodge locus can become a messy, chaotic cloud that cannot be described as a union of algebraic shapes. It's like expecting a flock of birds to fly in a perfect V-formation, only to find them scattered randomly across the sky in a way that defies any geometric description.

Furthermore, they tackle a famous idea called the "André–Oort conjecture," which suggests that if you find a lot of "special" points (called CM points) in a shape, the shape itself must be a very special, symmetric kind of object (a Shimura variety). The authors find examples where you have a dense cloud of these special points, yet the shape is not a Shimura variety. It's as if you found a room full of golden coins, but the room itself is just a regular, boring shed. This proves that for non-integral structures, "what can go wrong, must go wrong."

How They Did It: The Map and the Key

To find these examples, the authors used a clever map-making technique. They looked at the "Teichmüller component," which is a specific region in the vast landscape of all possible shapes. They used a method developed by mathematicians Kabaya and Maskit, which is like having a special ruler that can measure the "twist" and "turn" of a shape using simple numbers.

They focused on a specific shape: a sphere with four holes punched in it (a 4-punctured sphere). They showed that if you allow your numbers to be fractions involving certain prime numbers (called $S$ -integral points), you can find infinitely many different shapes that fit the bill. They even gave a new, simpler proof of a classic theorem by Beauville, showing that if you strictly stick to whole numbers (integers), there are only four such shapes. But the moment you relax the rules to allow fractions, the number of possibilities explodes into infinity.

What They Rule Out

The paper explicitly rules out the optimistic hope that the "Higgs standard conjecture" (a version of the Hodge conjecture for these twisted shapes) would hold true for these non-integral examples. They show that these non-integral patterns are not of "geometric origin," meaning they cannot be explained as coming from a family of simpler, standard shapes like families of elliptic curves. They also demonstrate that two different definitions of what it means to be a "K-variation of Hodge structure" (a technical way of defining these shapes over number fields) do not always agree, creating further confusion and "Murphy's Law" scenarios.

The Bottom Line

The authors have not just found a few oddities; they have proven that for non-integral Hodge structures, the usual laws of geometry and number theory break down completely. They have shown that the Hodge locus can be messy, the André–Oort conjecture can fail, and that there are infinitely many ways to build these "broken" structures. Their work serves as a warning: in the world of non-abelian Hodge theory, if you try to cut the corners and use fractions instead of whole numbers, the universe will not just tolerate it—it will throw a chaotic party where the rules no longer apply.

Technical Summary: Murphy's Law in Non-Abelian Hodge Theory

Problem Statement
The paper investigates the validity of non-abelian analogues of classical Hodge conjectures and standard conjectures in the context of rational ( $\mathbb{Q}$ ) variations of Hodge structures (QVHS), as opposed to integral ( $\mathbb{Z}$ ) variations. While the classical Hodge conjecture fails for $\mathbb{Z}$ -coefficients, the non-abelian version (Conjecture 1.1) posits that a $\mathbb{Z}$ -local system underlying a polarized complex variation of Hodge structure (VHS) is of "geometric origin." The authors question whether this holds for $\mathbb{Q}$ -local systems. Specifically, they seek explicit examples of non-integral QVHS that are not of geometric origin, thereby testing the limits of the "Higgs standard conjecture" (that bi-algebraic points of the Simpson correspondence are of geometric origin) and the stability of theorems like Cattani–Deligne–Kaplan and André–Oort when integrality is dropped.

Methodology
The authors employ a combination of non-abelian Hodge theory, Teichmüller theory, and Diophantine geometry. Their primary tools include:

Character Varieties and Teichmüller Components: They analyze the character varieties of surfaces, specifically focusing on the Teichmüller component, which corresponds to uniformizing local systems of Riemann surfaces.
Explicit Parametrizations:
- Maskit's Parametrization: Used for the signature $(0; 4)$ case (four-punctured sphere), providing rational coordinates for the deformation space of Fuchsian groups.
- Kabaya's Fenchel–Nielsen Parametrization: Used for arbitrary genus $g \geq 2$ , offering a rational parametrization of the Teichmüller component via eigenvalue and twist parameters associated with a pants decomposition.
Diophantine Analysis: They study $S$ -integral points on these character varieties. By leveraging Dirichlet's $S$ -unit theorem, they construct infinite families of points with specific trace rings ( $\mathbb{Z}[S^{-1}]$ ) to generate the required local systems.
Trace Ring Analysis: They utilize the relationship between the trace field of a representation and the existence of a $\mathbb{Q}$ -structure (via quaternion algebras and adjoint representations) to ensure the constructed local systems underlie QVHS.

Key Contributions and Results

Construction of Non-Integral QVHS (Theorem 1.2 & 1.6):
The authors construct explicit examples of rank two local systems on smooth projective curves of genus $g \geq 2$ that:
- Have trivial determinant and Zariski dense monodromy in $SL_2$ .
- Are complex direct factors of non-integral QVHS.
- Are not of geometric origin (since they are not integral).
- For any finite set of primes $S$ , there exist infinitely many non-isomorphic curves supporting such systems with trace ring $\mathbb{Z}[S^{-1}]$ (property $\tilde{P}_S$ ).
Classification of Signature $(0; 4)$ Cases (Theorem 1.8 & 4.5):
- Integral Case ( $S = \emptyset$ ): They provide a new proof of Beauville's classical theorem, showing there are exactly four Riemann surfaces of signature $(0; 4)$ admitting a modular embedding (i.e., supporting a $\mathbb{Z}$ -VHS). These correspond to specific points in the Teichmüller space.
- Non-Integral Case ( $S \neq \emptyset$ ): They prove that if $S$ is non-empty, there are infinitely many non-isomorphic curves of signature $(0; 4)$ supporting rank two local systems with property $P_S$ (trace ring $\mathbb{Z}[S^{-1}]$ ).
Diophantine Results on Character Varieties (Theorem 5.1):
The set of $\mathcal{O}_{K,S}$ -Teichmüller points (points on the Teichmüller component with coordinates in the ring of $S$ -integers of a number field $K$ ) is finite up to the mapping class group action if and only if $(K, S) = (\mathbb{Q}, \emptyset)$ . Otherwise, there are infinitely many distinct orbits. This contrasts with the finiteness of $\mathbb{Z}$ -points.
Failure of Standard Conjectures for Non-Integral QVHS (Section 7):
The paper demonstrates that "Murphy's Law" applies to non-integral QVHS, meaning pathological behaviors occur:
- Cattani–Deligne–Kaplan Failure: The Hodge locus of a non-integral QVHS is not necessarily a countable union of algebraic subvarieties (Theorem 7.2).
- André–Oort Failure: There exist non-integral QVHS with discrete monodromy that possess a dense set of CM (Complex Multiplication) points but are not of Shimura type (Proposition 7.3).
- Definition Discrepancy: They show that two common definitions of $K$ -VHS (one requiring the local system to be a $K$ -factor of a $\mathbb{Q}$ -VHS, the other requiring underlying VHS for all embeddings) do not generally agree (Theorem 1.11).

Significance and Claims
The paper claims that for QVHS which are not $\mathbb{Z}$ -VHS, "what can go wrong must go wrong." The significance lies in:

Refuting Optimism: It provides concrete counterexamples to the hope that non-abelian Hodge conjectures might hold for $\mathbb{Q}$ -coefficients, suggesting that integrality is a crucial structural constraint.
Testing Grounds: The constructed examples serve as test cases for the "Higgs standard conjecture," indicating that if the Higgs standard conjecture holds, then the associated curves in Theorem 1.2 are not defined over $\mathbb{Q}$ (Conjecture 1.4).
Pathological Behavior: It highlights that the Hodge locus behaves fundamentally differently in the non-integral setting, failing to retain the algebraic structure and distribution properties (like the André–Oort conjecture) that hold for integral variations.
New Proof: It offers a new, explicit proof of Beauville's classification of modular curves of signature $(0; 4)$ using character variety parametrizations.

The authors conclude by suggesting a refined notion of the "algebraic Hodge locus" to recover some structure in the non-integral setting, though they note that points remain elusive without further constraints.

More like this