Trace formalism for motivic cohomology

Here is an explanation of Tomoyuki Abe's paper, "Trace formalism for motivic cohomology," translated into simple, everyday language with creative analogies.

The Big Picture: Mapping the Invisible

Imagine you are an architect trying to understand a complex, invisible city. You can't see the buildings directly, but you have a special set of tools (mathematical theories) that let you measure the "shape" and "weight" of the city's structure.

In the world of algebraic geometry, mathematicians study shapes called schemes (which are like generalized geometric spaces). They use a tool called Motivic Cohomology to measure these spaces. Think of Motivic Cohomology as a high-tech scanner that takes a picture of a geometric shape and tells you its hidden properties.

The paper is about building a specific feature for this scanner: a "Trace Map."

The Problem: The "Ghost" in the Machine

In the past, mathematicians (specifically in the 1970s with the SGA4 book) figured out how to build a Trace Map for a simpler type of scanner called Étale Cohomology. This map acts like a translator.

The Input: A "Cycle." Imagine a cycle as a physical object, like a specific road or a building in our invisible city. It's concrete and tangible.
The Output: A number or a cohomology class. This is the abstract measurement of that object.

The Trace Map takes the concrete object (the cycle) and translates it into the abstract language of the scanner so the math can work with it.

The Defect: The old way of doing this was a bit rigid. It treated the object as a "scheme" (a rigid mathematical structure). But sometimes, a scheme has "ghosts" or "fuzz" (mathematicians call these nilpotents). The scanner doesn't care about the fuzz; it only sees the core shape. The old method was sometimes confused by the fuzz, even though the result should have been the same.

Abe's Insight: "We shouldn't be translating the building (the scheme); we should be translating the blueprint (the cycle)."

The Solution: The "Cycle" Translator

Abe introduces a new way to build this translator using Relative Cycle Groups (developed by Suslin and Voevodsky).

The Analogy: Imagine you have a pile of Lego bricks.
- The Old Way: You try to translate the instructions for a specific, slightly messy Lego castle. If the castle has a loose brick (a nilpotent), the instructions get confused.
- Abe's Way: You ignore the messy castle and look at the list of bricks (the cycle). You translate the list of bricks directly. Since the list of bricks is the same whether the castle is messy or clean, your translation is perfect and robust.

Abe constructs a map that takes these "lists of bricks" (cycles) and converts them directly into the language of the Motivic scanner. This is the Trace Map for Motivic Cohomology.

The Secret Weapon: "Vanishing"

How did he prove this map works? He used a clever trick involving Higher Homotopies.

The Metaphor: Imagine you are trying to send a message across a noisy radio channel. Usually, there is static (noise) that distorts the message.
The Discovery: Abe realized that for this specific type of math, the "static" (higher homotopies) completely vanishes (disappears) in certain situations.
The Result: Because the static is gone, he didn't need to build the whole radio tower at once. He could build the message locally (in small, smooth neighborhoods) and then stitch them together perfectly. It's like building a puzzle where the pieces are so perfectly smooth that they snap together without any gaps.

This "vanishing" property allowed him to reduce a incredibly difficult, global problem to a much simpler local problem, which he could solve easily.

The "Infinity" Upgrade: The 4D Movie

The paper doesn't just stop at building the map; it also gives it an $\infty$ -enhancement (Infinity Enhancement).

The Analogy:
- Standard Math (1-Category): This is like a photograph. It captures a single moment. You can see the cycle and the cohomology, and the arrow connecting them. But it's static.
- Infinity Math ( $\infty$ -Category): This is like a 4D movie. It captures not just the connection, but all the possible ways the connection could wiggle, twist, or deform. It captures the "history" of the relationship.

Abe shows that his Trace Map isn't just a static arrow; it's a living, breathing structure that fits perfectly into this 4D movie world. This is crucial because modern mathematics is moving toward this "movie" style of thinking (using $\infty$ -categories) to handle very complex systems.

Why Does This Matter?

Universality: This new map works for any geometric shape, even the messy ones with "ghosts," because it focuses on the underlying cycles (the bricks) rather than the messy structure.
Bridging Worlds: It connects the world of "actual cycles" (things you can count and draw) with the world of "infinity-enhanced cohomology" (the most advanced, abstract math available).
Future Proofing: By building this in the "Infinity" framework, Abe has created a tool that other mathematicians can use to solve even harder problems in the future, specifically in a follow-up paper mentioned in the text ([Abe22b]).

Summary in One Sentence

Tomoyuki Abe built a robust, "ghost-proof" translator that turns concrete geometric shapes (cycles) into abstract mathematical data, proving that the translation is perfect by showing that the "noise" in the system disappears, and then upgraded this translator to work in the most advanced, multi-dimensional language of modern mathematics.

Here is a detailed technical summary of the paper "Trace formalism for motivic cohomology" by Tomoyuki Abe, published in Épijournal de Géométrie Algébrique (2023).

1. Problem Statement

The paper addresses the construction of trace maps within the six-functor formalism of motivic cohomology (developed by Voevodsky, Ayoub, and Cisinski–Déglise).

Context: In classical étale cohomology (SGA 4), for a flat morphism $f: X \to S$ of relative dimension $d$ , there exists a trace map $\text{Tr}_f: Rf_! f^* \Lambda(d)[2d] \to \Lambda$ . This map is fundamental for defining cycle class maps and transferring cycle-theoretic information into cohomological frameworks.
The Gap: While the six-functor formalism for motivic cohomology is well-established, a rigorous construction of the trace map analogous to the étale case was missing, particularly one that fits naturally into an $\infty$ -categorical enhancement.
Specific Challenges:
1. Constructing the map for general (possibly singular) schemes, not just smooth ones.
2. Reconciling the fact that motivic cohomology is insensitive to nil-immersions (non-reduced structures), suggesting the trace should be associated with "cycles" rather than "schemes."
3. Lifting the construction to the level of $\infty$ -categories to interface with "actual cycles" and $\infty$ -enhanced motivic cohomology.

2. Methodology

Abe employs a strategy combining vanishing theorems, descent theory, and Suslin–Voevodsky's relative cycle groups.

A. Vanishing of Higher Homotopies

The core technical insight is the vanishing of negative homotopy groups for the relevant mapping spectra. Specifically, for a morphism $f: X \to S$ of dimension $d$ :
$\text{R}^i\text{Hom}(Rf_! f^* \Lambda(d)[2d], \Lambda) = 0 \quad \text{for } i < 0.$
This vanishing allows the construction of the trace map to be reduced to a local problem. It implies that constructing the global trace map is equivalent to constructing a morphism of sheaves $R^{2d}f_! f^* \Lambda(d) \to \Lambda$ . This property is crucial for reducing the problem to the case where the base scheme $S$ is smooth.

B. Reduction to Smooth Base via pdh-Topology

To handle singular schemes, the author utilizes the pdh-topology (generated by finite surjective flat morphisms of degree a power of $p$ and cdh-coverings).

Using Temkin's alteration theorem, any scheme $S$ admits a pdh-hypercovering by smooth schemes.
The author proves that motivic cohomology with coefficients in $\Lambda = \mathbb{Z}[1/p]$ satisfies pdh-descent. This allows the construction of the trace map on smooth bases to be extended to arbitrary bases via descent.

C. Reinterpretation via Cycle Groups

To address the "cycle vs. scheme" issue, the paper reinterprets the trace map using Suslin–Voevodsky's relative cycle groups, denoted $z_{\text{equi}}(X/S, d)$ .

Instead of defining the trace for a morphism $f$ , the author constructs a morphism of bivariant theories:
$\tau: z_{\text{equi}}(-, *) \to H^{\text{BM}}_{2*}(-, \Lambda(*))$
where $H^{\text{BM}}$ denotes Borel–Moore homology.
When $f$ is flat of dimension $d$ , the fundamental cycle $[X] \in z_{\text{equi}}(X/S, d)$ maps to the traditional trace map $\text{Tr}_f$ .

D. $\infty$ -Enhancement

The final step involves lifting the 1-categorical trace map to an $\infty$ -functor.

The author defines a category $\widehat{\text{Ar}}$ of morphisms in the category of schemes.
Using the language of $\infty$ -categories (specifically presentable $\infty$ -categories and left Kan extensions), the trace map is formulated as a natural transformation between $\infty$ -functors from $\widehat{\text{Ar}}^{\text{op}}$ to the $\infty$ -category of spectra ( $\text{Sp}$ ).

3. Key Contributions and Results

Main Theorem (Theorem 3.5)

There exists a unique morphism of bivariant theories compatible with $\mathbb{A}^1$ -orientation:
$\tau: z_{\text{equi}}(-, *) \to H^{\text{BM}}_{2*}(-, \Lambda(*))$
where $\Lambda = \mathbb{Z}[1/p]$ .

Significance: This generalizes the trace map from [SGA 4] (étale cohomology) to the motivic setting.
Universality: The result holds for the motivic Eilenberg–MacLane spectrum, which is universal among absolute SH-spectra with additive formal group laws. Consequently, trace maps for other theories (e.g., $\ell$ -adic étale cohomology) can be derived from this motivic trace.

Construction of the Trace Map (Sections 4 & 5)

Smooth Base Case: When the base $S$ is smooth, the trace map is constructed using the fundamental class and relative Poincaré duality (isomorphism $f^*(d)[2d] \cong f^!$ ). The author proves that the constructed map coincides with the fundamental class $\eta_f$ .
General Case: Using the vanishing of higher homotopies and pdh-descent, the map is extended to arbitrary bases. The construction relies on the fact that $z_{\text{equi}}$ and $H^{\text{BM}}$ are sheaves in the pdh-topology.

$\infty$ -Enhancement (Section 6)

Theorem 6.4: There exists an essentially unique morphism of spectra-valued sheaves $\tau^\dagger: z(d) \to H^{\text{BM}}(d)$ on the category $\widehat{\text{Ar}}$ such that its $\pi_0$ recovers the classical trace map $\tau$ .

This provides a rigorous $\infty$ -categorical interface between algebraic cycles and motivic cohomology, which is a prerequisite for further applications in the author's subsequent work (e.g., [Abe22b]).

4. Significance

Foundational: It completes the six-functor formalism for motivic cohomology by providing the missing trace map, which is essential for duality theorems and cycle class maps.
Methodological Innovation: The paper demonstrates how vanishing of higher homotopies can be used to simplify the construction of global maps in derived categories, avoiding the circular arguments often found in classical étale cohomology proofs (where trace is used to prove duality, and duality is used to construct trace).
Categorical Unification: By formulating the trace map in the language of $\infty$ -categories and bivariant theories, the paper bridges the gap between classical algebraic geometry (cycles) and modern homotopical algebra (spectra).
Generality: The results hold for perfect fields of characteristic $p > 0$ with coefficients in $\mathbb{Z}[1/p]$ , covering the most general setting for motivic cohomology without requiring resolution of singularities (though the paper notes that with resolution, $\Lambda=\mathbb{Z}$ is possible).

In summary, Abe successfully constructs the motivic trace map by leveraging the vanishing of higher homotopies to reduce the problem to smooth bases, utilizing Suslin–Voevodsky cycle groups to handle singularities, and elevating the entire formalism to the $\infty$ -categorical level.