🔢 mathematics

The motivic Lie algebra embeds into the cohomology of the general linear group

This paper establishes a canonical embedding of the motivic Lie algebra of mixed Tate motives over $\mathbb{Z}$ into the cohomology of locally symmetric spaces for $\mathrm{GL}_g(\mathbb{Z})$ and the moduli space of principally polarized abelian varieties by utilizing tropical geometry and graph complexes to link compactly-supported Borel classes to graph cocycles that generate the motivic Lie algebra.

Original authors: Francis Brown, Erik Panzer, Jean-Luc Portner

Published 2026-08-03

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Francis Brown, Erik Panzer, Jean-Luc Portner

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 the universe of mathematics as a vast, interconnected archipelago. On one distant island sits the world of Algebraic K-Theory, a realm where mathematicians study the hidden symmetries of numbers and shapes, specifically looking at how things can be built up or taken apart. On another far-away island lies Topology, the study of shapes that can be stretched and twisted without tearing, where the focus is often on the "holes" inside objects. For a long time, these two islands seemed to speak different languages. The "Motivic Lie Algebra" is like a secret, universal translator that mathematicians believe connects these two worlds. It is a special mathematical structure made of building blocks (generators) that appear in specific, odd-numbered "weights" (like 3, 5, 7, etc.), acting as a bridge between the abstract algebra of numbers and the geometric shapes of space.

To understand why this matters, think of a Lie Algebra not as a scary equation, but as a set of instructions for how to twist and turn a shape. The "Motivic" part refers to a deep, unifying theory that tries to find the "DNA" shared by all algebraic varieties (shapes defined by polynomial equations). The big mystery has been: Where exactly does this secret DNA live? We know it exists in the algebra of numbers, but can we find it hiding inside the geometry of spaces we can actually visualize? If we can pin it down to a specific geometric location, we can finally see how the number-theory island and the geometry island are actually the same continent.

The Great Mathematical Heist: Finding the Hidden DNA

In this paper, Francis Brown, Erik Panzer, and Jean-Luc Portner pull off a mathematical heist. They show that the mysterious Motivic Lie Algebra isn't just floating in the abstract ether; it is actually hiding inside the "cohomology" (a fancy word for counting the holes and twists) of a very specific, somewhat chaotic geometric space: the space of General Linear Groups (think of it as the space of all possible ways to stretch and rotate grids of numbers).

Here is the story of how they found it, using a map that passes through a strange, pixelated world called Tropical Geometry.

The Map: From Smooth Curves to Pixelated Graphs

The authors start with a smooth, continuous space of matrices (grids of numbers) called $P_g$ . Inside this space, there are special "forms" (mathematical waves) that swirl around. These are the Borel classes, which are like the standard, well-known fingerprints of this space. However, the authors needed something slightly different: compactly-supported versions of these forms. Imagine taking a smooth wave and forcing it to vanish completely at the edges of the universe; that's what "compactly supported" means.

The trickiest part of the journey is getting from this smooth world to a world made of graphs (dots connected by lines). To do this, the authors use a "Tropical Torelli map." Think of this as a magical translator that takes a smooth, flowing shape (a tropical curve) and turns it into a rigid, pixelated skeleton (a graph).

The Tropical Curve: A shape made of straight lines and angles, like a stick-figure drawing of a curve.
The Graph: A collection of dots and edges.
The Translation: The map takes the smooth "waves" from the matrix space and projects them onto these stick-figure graphs.

The Discovery: The Graphs Sing the Same Song

Once the waves are projected onto the graphs, the authors calculate a specific number for each graph. These numbers are called canonical graph integrals. It's like asking, "If I run a current through this specific wireframe shape, how much energy does it take?"

Here is the big reveal: The authors prove that these numbers are exactly the same as numbers calculated by a completely different method developed by Rossi and Willwacher, which involves complex physics-like integrals over "position spaces."

The Analogy: Imagine two musicians playing different instruments in different rooms. One plays a smooth violin melody (the matrix forms), and the other plays a complex drum rhythm (the graph integrals). The authors prove that if you listen closely, they are playing the exact same song. The "drum beats" (graph integrals) are actually just a different way of hearing the "violin melody" (the motivic algebra).

The Verdict: The DNA is Found

The paper proves that these graph-based numbers (the "drum beats") are not random. They correspond perfectly to the generators of the Motivic Lie Algebra.

The Result: The authors show a direct, one-to-one line connecting the abstract algebraic generators (the "DNA") to the geometric cohomology of the General Linear Group.
The Proof: They use a tool called single-valued periods (a way of measuring shapes that ignores the confusing "twists" of complex numbers) to confirm that the graph numbers are indeed "motivic." This means they belong to the specific family of numbers that define the Motivic Lie Algebra.

What This Means for the Reader

The paper does not just suggest this connection; it proves it. They have constructed a "canonical embedding," which is a fancy way of saying they built a sturdy, unbreakable bridge.

The Bridge: The Motivic Lie Algebra (the secret DNA) is now proven to live inside the cohomology of the space of matrices ( $GL_g(\mathbb{Z})$ ) and the moduli space of abelian varieties (spaces of donut-shaped surfaces).
The Surprise: This happens in the "unstable" range. Usually, mathematicians look at these spaces when they get very large (stable range) to find simple patterns. This paper finds the complex, rich patterns in the "unstable" range, where things are messy and unpredictable.

The "Wheel" and the "Zigzag"

To make their case, the authors computed these graph integrals for all graphs with up to 14 edges. They found that for simple shapes like "wheels" (a central dot with spokes) and "zigzags" (a snake-like line), the numbers they calculated matched the expected values of the Motivic Lie Algebra generators perfectly.

Example: For a graph with 5 edges (a wheel), the calculation yields a number related to $\zeta(5)$ (a famous mathematical constant). This confirms that the "graph song" is indeed the "motivic song."

The Bottom Line

This paper solves a long-standing puzzle: How do we reconcile the projective line minus three points (a simple shape from algebraic geometry) with the General Linear Group (a complex group of matrices)?
The answer is: They are connected through Tropical Geometry and Graphs. The Motivic Lie Algebra, which was thought to be a ghostly structure existing only in theory, is now shown to be a concrete, calculable part of the geometry of these matrix spaces. The authors have successfully mapped the "DNA" of mixed Tate motives directly onto the "skeleton" of graphs, proving that the two islands are, in fact, connected by a hidden land bridge.

The paper concludes by noting that while they have found the bridge, there are still many other "classes" (other types of holes and twists) in these spaces that we haven't fully explored yet. But for the specific generators of the Motivic Lie Algebra, the mystery is solved: they are here, they are real, and they are hiding in the graphs.

Technical Summary: The Motivic Lie Algebra Embeds into the Cohomology of the General Linear Group

Problem Statement
The paper addresses the longstanding mathematical mystery of reconciling two distant objects: the motivic Lie algebra $\mathfrak{g}_m$ associated with the category of mixed Tate motives over $\mathbb{Z}$ , and the cohomology of the general linear group $GL_g(\mathbb{Z})$ . The algebra $\mathfrak{g}_m$ is a free graded Lie algebra generated by elements $\sigma_{2k+1}$ in odd degrees $2k+1 \geq 3$ , originating from the algebraic K-theory of $\mathbb{Z}$ and the motivic fundamental groupoid of $\mathbb{P}^1 \setminus \{0, 1, \infty\}$ . Conversely, the stable cohomology of $GL_g(\mathbb{Z})$ is generated by Borel classes. While both structures are known to relate to odd zeta values, a direct, canonical geometric embedding of $\mathfrak{g}_m$ into the unstable compactly-supported cohomology of $GL_g(\mathbb{Z})$ had not been established. Previous partial results relied on the Grothendieck–Teichmüller Lie algebra ( $\mathfrak{grt}_1$ ) and computer calculations in low degrees.

Methodology
The authors construct a canonical embedding through a multi-step geometric and algebraic pathway involving tropical geometry, graph complexes, and the theory of single-valued periods:

Compactly-Supported Borel Classes: The authors begin with the Borel classes $\omega_{4k+1} = \text{tr}((X^{-1}dX)^{4k+1})$ on the space of positive definite symmetric matrices $P_g$ . They utilize compactly-supported analogues $[\omega_{4k+1}^c]$ defined on the Borel–Serre compactification, which shift the degree by one (mapping to $H^{4k+2}_c$ ).
Tropical Torelli Map: These classes are pulled back via the tropical Torelli map $\lambda: M_g^{\text{trop}} \to A_g^{\text{trop}}$ (from the moduli space of tropical curves to tropical abelian varieties). This maps the differential forms to the link of the moduli space of tropical curves, $LM_g^{\text{trop}}$ .
Graph Integrals and Cocycles: The pullback forms are integrated over cells $\sigma_G$ indexed by stable weighted graphs $G$ . This yields "canonical graph integrals" $I_G^{\text{can}}$ . These integrals define a closed cocycle $C_{4k+1}^{\text{can}}$ in the graph complex $GC_2$ .
Identification with Rossi–Willwacher Integrals: A crucial step utilizes recent work by the third author [Por26] to prove that the canonical integrals $I_G^{\text{can}}$ coincide with the Rossi–Willwacher integrals $I_G^{\text{RW}}$ (defined via configuration spaces of points on $\mathbb{C}$ ). This identification relies on the Schwinger trick from quantum field theory to relate parametric and position space representations.
Motivic Interpretation: The graph cocycles are mapped to Lie words in the Ihara Lie algebra $(\widehat{L}(e_0, e_1), \{,\})$ via the morphism $\phi$ . By relating these words to single-valued periods of mixed Tate motives and utilizing the single-valued Drinfeld associator, the authors demonstrate that these Lie words lie within the motivic Lie algebra $\mathfrak{g}_m$ .

Key Contributions and Results

Theorem 1 (Embedding into $GL_g(\mathbb{Z})$ ): The paper establishes a canonical embedding of the free Lie algebra generated by $\omega_{4k+1}^c$ into the direct sum of compactly-supported cohomology groups of the locally symmetric spaces $P_g/GL_g(\mathbb{Z})$ :
$L(\omega_5^c, \omega_9^c, \dots) \hookrightarrow \bigoplus_{g>0} H^{2g}_c(P_g/GL_g(\mathbb{Z}); \mathbb{R})$
The generators map to classes that are compactly-supported analogues of the Borel classes, shifted by one degree.
Theorem 2 (Embedding into $\mathcal{A}_g$ ): A parallel canonical embedding is constructed into the weight-zero compactly-supported cohomology of the moduli space of principally polarized abelian varieties $\mathcal{A}_g$ :
$L(\omega_5^c, \omega_9^c, \dots) \hookrightarrow \bigoplus_{g>0} W_0 H^{2g}_c(\mathcal{A}_g; \mathbb{R})$
The image lands in the subspace of primitive elements with respect to the Hopf algebra structure on these cohomology groups.
Theorem 3 (Graph Cocycles and $\mathfrak{g}_m$ ): The authors prove that the canonical graph cocycles $C_{4k+1}^{\text{can}}$ generate a free Lie subalgebra isomorphic to $\mathfrak{g}_m \otimes \mathbb{R}$ inside the degree-zero cohomology of the graph complex $H^0(GC_2)$ . Specifically, the map $\phi$ sends the graph cocycles to elements in the motivic Lie subalgebra of the Ihara Lie algebra.
$L(\omega_5^c, \omega_9^c, \dots) \otimes \mathbb{R} \cong \mathfrak{g}_m \otimes \mathbb{R}$
Explicit Computations: The paper provides explicit forms for the canonical graph cocycles for all graphs with up to 14 edges, including new computations for $C_{13}^{\text{can}}$ .

Significance and Claims
The paper claims to resolve the "enduring mystery" of connecting the motivic fundamental group of $\mathbb{P}^1 \setminus \{0, 1, \infty\}$ with the cohomology of $GL_g(\mathbb{Z})$ through natural geometric constructions.

Independence from $\mathfrak{grt}_1$ : Unlike previous approaches (e.g., [Bro+24]), this work circumvents unresolved conjectures regarding the Grothendieck–Teichmüller Lie algebra ( $\mathfrak{grt}_1$ ). Instead of relying on the identification $H^0(GC_2) \cong \mathfrak{grt}_1$ and low-degree computer calculations, the authors derive the freeness of the cocycles and their motivic nature directly from the theory of single-valued periods and the Rossi–Willwacher results.
Direct Geometric Realization: The result exhibits the motivic Lie algebra directly inside the unstable compactly-supported cohomology of arithmetic groups, suggesting its appearance in other unexpected arithmetic settings (e.g., symplectic groups).
Hopf Algebra Structure: The authors highlight that the target spaces carry Hopf algebra structures, allowing the generation of large families of additional cohomology classes from the motivic generators via the Milnor–Moore theorem.
Answer to Open Questions: Theorem 2 answers one half of Question 1.17 in [Bro+24] by establishing the embedding. The authors note that the surjectivity of this map onto the primitive elements remains an open question (though it holds for $g < 8$ ).

The paper concludes that the correspondence between canonical integrals on tropical moduli spaces and single-valued integrals on moduli spaces of curves provides the "missing link" connecting Borel forms to the motivic fundamental group.