A Knebusch trace formula for Azumaya algebras with involution

This paper establishes a Knebusch-type trace formula for signatures of hermitian forms over Azumaya algebras with involution, extending Knebusch's results on symmetric bilinear forms and yielding an exact sequence for total signatures over semilocal base rings that relates to Pfister's local-global principle and the stability index.

The Gist

Here is an explanation of the paper "A Knebusch Trace Formula for Azumaya Algebras with Involution" using everyday language and creative analogies.

The Big Picture: A Mathematical "Translation" Machine

Imagine you are a translator working in a very complex, high-stakes environment. You have a document written in a secret, encrypted code (let's call it Azumaya Algebras). This code is used to describe shapes and symmetries in a world that is slightly more complicated than our own.

The authors, Vincent Astier and Thomas Unger, have built a universal translator (the Trace Formula) that allows them to take a complex message written in this secret code and translate it into a language everyone understands: Signatures.

In this paper, they prove that no matter how complicated the secret code is, you can always break it down into simple "signatures" (like positive or negative numbers) that tell you the essential nature of the shape, just by looking at it through different "lenses."

---

The Cast of Characters

To understand the paper, we need to meet the main characters:

1. The Base Ring ($R$): Think of this as the ground or the foundation of a city. It's where everything lives. In math, it's a set of numbers with rules for adding and multiplying.
2. Azumaya Algebras ($A$): These are like specialized skyscrapers built on that foundation. They are complex structures that hold data. Unlike a normal building, these skyscrapers have a hidden "mirror" inside them called an Involution ($\sigma$).
   • The Mirror: If you look at a number in the building through this mirror, it flips or changes in a specific way. This mirror is crucial because it defines the "shape" of the building.
3. Hermitian Forms ($h$): These are the blueprints or furniture inside the skyscrapers. They describe how things fit together. The authors are interested in measuring these blueprints.
4. Signatures: These are the measurements or scores we get when we look at the blueprints.
   • Imagine looking at a blueprint through a specific pair of glasses (an Ordering or Real Spectrum). Through these glasses, a shape might look "positive" (upright) or "negative" (upside down). The "signature" is the count of how many things are up vs. down.
5. Finite Étale Extensions ($T$): Think of these as expansion packs or new districts built next to the original city. They are connected to the original foundation but have their own local rules.

---

The Problem: The "Lost in Translation" Dilemma

In the 1970s, a mathematician named Knebusch discovered a rule for simple buildings (symmetric bilinear forms). He found that if you take a blueprint from a small district and "trace" it back to the main city, the total score is just the sum of the scores of all the districts it came from.

The Problem:
The authors wanted to do this for the complex skyscrapers (Azumaya Algebras) with their mirrors (Involutions).

• The Challenge: These skyscrapers are tricky. Sometimes, when you look at them through a specific lens, the measurement is zero (the blueprint disappears). Sometimes, the "mirror" flips the sign in a way that makes the measurement ambiguous.
• The Question: Can we still sum up the scores from the different districts to get the total score for the main city, even with these complex, mirrored skyscrapers?

---

The Solution: The Trace Formula

The authors say: "Yes, we can!"

They established a Trace Formula. Here is how it works in plain English:

1. The Setup: You have a complex blueprint ($h$) inside a skyscraper ($A$) in a new district ($T$).
2. The Trace: You use a special machine (the Trace Map) to pull that blueprint back to the main city ($R$). This machine effectively "collapses" the complex structure down to something simpler.
3. The Formula: The signature (score) of the collapsed blueprint in the main city is exactly equal to the sum of the signatures of the original blueprint, viewed through every possible "lens" (ordering) in the new district that connects to the main city.

The Analogy:
Imagine you have a giant, multi-colored mosaic ($h$) in a new wing of a museum ($T$). You want to know the "vibe" of the whole thing.

• You can't just look at the whole thing at once because it's too big.
• Instead, you send a team of inspectors to look at the mosaic through different colored filters (the Orderings).
• Each inspector writes down a score (+1 or -1) based on what they see.
• The Trace Formula says: If you take the "shadow" of the mosaic cast back onto the main hall, its score is simply the sum of all the scores your inspectors wrote down.

Why is this hard? (The "Reference Form" Issue)

There is a catch. In the complex skyscrapers, the "mirror" (involution) can sometimes make the measurement flip signs unpredictably. It's like if one inspector sees a red apple as "positive" and another sees it as "negative" just because they are standing on different sides of the mirror.

To fix this, the authors introduce a Reference Form (a "Standard Ruler").

• They pick one specific, well-behaved blueprint ($\eta$) that they know is always "positive" in the places that matter.
• They use this ruler to calibrate all the other measurements. If an inspector's measurement is negative, but the ruler says it should be positive, they flip the sign.
• This ensures that everyone is counting in the same direction, making the sum valid.

The "Semilocal" Bonus: A Perfect Sequence

The paper also looks at a special case where the foundation is Semilocal (think of a city with only a few distinct neighborhoods).

In this case, they prove something beautiful:

• There is a perfect Exact Sequence.
• Imagine a conveyor belt.
  • Input: All possible blueprints.
  • Process: Measuring them with the "Standard Ruler" to get their signatures.
  • Output: The list of all possible scores.
• The authors prove that the "waste" (the blueprints that measure to zero) and the "gaps" (scores you can't reach) are perfectly controlled. Specifically, the "waste" and the "gaps" are always related to powers of 2 (like 2, 4, 8, 16).

This connects to a famous idea called Pfister's Local-Global Principle, which basically says: "If you understand the local neighborhoods, you understand the whole city."

Summary: What did they actually do?

1. They generalized a classic rule: They took a rule that worked for simple shapes and proved it works for complex, mirrored skyscrapers.
2. They fixed the sign problem: They created a method (using a Reference Form) to ensure that measurements don't get flipped upside down by the mirrors.
3. They connected the dots: They showed that in certain types of cities (semilocal rings), the relationship between the local measurements and the global total is perfectly structured and predictable.

In a nutshell: They built a reliable calculator that tells you how to add up the "vibes" of a complex mathematical structure by looking at its parts, ensuring that the mirrors and twists don't break the math.

Technical Summary

Here is a detailed technical summary of the paper "A Knebusch Trace Formula for Azumaya Algebras with Involution" by Vincent Astier and Thomas Unger.

1. Problem Statement

The paper addresses the extension of Knebusch's trace formula from the classical setting of symmetric bilinear forms over commutative rings to the more complex setting of hermitian forms over Azumaya algebras with involution.

• Context: In the 1970s, Manfred Knebusch established a trace formula relating the signature of a symmetric bilinear form over a base ring $R$ to the signatures of its extension over a finite étale algebra $T$. This was a cornerstone in the study of the Witt ring $W(R)$ and real algebraic geometry.
• Gap: While signatures of hermitian forms over Azumaya algebras (generalizations of central simple algebras over fields) had been defined using Morita equivalence and the real spectrum ($\text{Sper } R$), a general trace formula connecting the signature of a form over $A \otimes_R T$ to the signature of its trace over $A$ was missing.
• Objective: To establish a trace formula for signatures of hermitian forms over Azumaya algebras with involution, specifically utilizing signatures defined at orderings $\alpha \in \text{Sper } R$.

2. Methodology and Preliminaries

The authors employ a blend of real algebra, Morita theory, and Azumaya algebra theory.

• Signatures and the Real Spectrum: The paper utilizes the real spectrum $\text{Sper } R$ (the set of orderings on $R$). Signatures are defined as ring homomorphisms from the Witt ring to $\mathbb{Z}$.
• $\eta$-Signatures: A major technical hurdle in hermitian forms is the lack of a canonical choice for the Morita equivalence used to define signatures (the "M-signature"). The authors rely on their previous work introducing $\eta$-signatures, which fix the sign ambiguity by using a reference form $\eta$. This ensures the signature is a continuous function on $\text{Sper } R$.
• Trace Maps: The authors utilize the trace map $\text{Tr}_{T/R}$ for finite étale extensions $T/R$ and extend it to Azumaya algebras via $\text{Tr}_{A \otimes T / A} = \text{id}_A \otimes \text{Tr}_{T/R}$.
• Base Change to Real Closed Fields: The core of the proof involves analyzing the behavior of forms after scalar extension to the real closure $k(\alpha)$ of the residue field at an ordering $\alpha$. This reduces the problem to the case of algebras over real closed fields, where the structure of the algebra (split, quaternion, or product of fields) is well-understood.

3. Key Contributions and Results

A. The Knebusch Trace Formula (Theorem 3.4)

The central result is the establishment of the trace formula for hermitian forms.

• Statement: Let $R$ be a commutative ring, $T$ a finite étale $R$-algebra, and $(A, \sigma)$ an Azumaya algebra with involution over $R$. For any hermitian form $h$ over $A \otimes_R T$, the signature of its trace over $A$ at an ordering $\alpha \in \text{Sper } R$ is the sum of the signatures of $h$ at the extensions of $\alpha$ to $T$.
  $$\text{sign}^\eta_\alpha (\text{Tr}^*_{A \otimes T / A} h) = \sum_{i=1}^r \text{sign}^{\eta \otimes T}_{\gamma_i} (h)$$
  Here, $\gamma_1, \dots, \gamma_r$ are the specific extensions of $\alpha$ to $T$ (preimages of the unique ordering on the real closure of $\alpha$).
• Semilocal Case: If $R$ is semilocal, the number of such extensions $r$ is exactly equal to $\text{sign}_\alpha (\text{Tr}^*_{T/R} \langle 1 \rangle)$, and the extensions are distinct maximal orderings.

B. Existence of Reference Forms with $2^k$ Signatures (Section 4)

The authors prove the existence of "nonsingular reference forms" with constant signatures that are powers of 2.

• Theorem 4.5: If $R$ is semilocal, there exists a nonsingular hermitian form $h_0$ over $(A, \sigma)$ such that its absolute signature $|\text{sign}^\eta_\bullet h_0|$ is constant and equal to $2^m$on the set$\text{Sper } R \setminus \text{Nil}[A, \sigma]$.
• Significance: This result is crucial for the next step, as it provides a "unit" in the Witt group that behaves predictably under the trace map, allowing for the construction of exact sequences.

C. Exact Sequence for Total Signatures (Section 5)

Using the trace formula and the existence of $2^m$-signature forms, the authors derive an exact sequence relating the Witt group of the Azumaya algebra to the group of continuous integer-valued functions on the real spectrum.

• Theorem 5.3: For a semilocal ring $R$, the following sequence is exact:
  $$0 \to W_t(A, \sigma) \to W(A, \sigma) \xrightarrow{\text{sign}^\eta_\bullet} C(\text{Sper } R, \mathbb{Z})_{[A, \sigma]} \to S_\eta(A, \sigma) \to 0$$
  • $W_t(A, \sigma)$ is the torsion subgroup of the Witt group.
  • $C(\text{Sper } R, \mathbb{Z})_{[A, \sigma]}$ is the group of continuous functions vanishing on the "nil" set (where signatures are zero).
  • $S_\eta(A, \sigma)$ is the cokernel, which is a 2-primary torsion group.
• Connection to Stability Index: This sequence generalizes the classical relationship between the Witt group of a field and its stability index. The group $S_\eta(A, \sigma)$ is identified as the stability group of the algebra with involution.

D. Hermitian Morita Equivalence (Appendix A)

The paper provides a rigorous technical appendix proving that Brauer equivalent Azumaya algebras with involution over semilocal connected rings are Hermitian Morita equivalent. This equivalence preserves signatures and is compatible with base change, which is a foundational tool used throughout the main proofs.

4. Significance and Impact

• Generalization of Classical Theory: The paper successfully lifts Knebusch's seminal work from commutative rings (symmetric bilinear forms) to the non-commutative setting of Azumaya algebras with involution.
• Unification of Real Algebra and Non-Commutative Algebra: It bridges the gap between the theory of orderings on rings (real algebra) and the classification of hermitian forms over central simple algebras.
• Pfister's Local-Global Principle: The derived exact sequence provides a non-commutative version of Pfister's local-global principle. It characterizes when a hermitian form is determined by its local signatures (at orderings) and quantifies the obstruction (the 2-primary torsion group).
• Stability Index for Algebras: The work defines and computes the "stability index" for Azumaya algebras with involution, a concept previously restricted to fields or commutative rings. This has implications for understanding the structure of Witt groups in non-commutative settings.

In summary, Astier and Unger provide a robust framework for analyzing hermitian forms over Azumaya algebras using the tools of real algebra, culminating in a trace formula and an exact sequence that generalize fundamental results from the commutative case to the non-commutative realm.