Proto-exact categories and injective Banach modules

This paper establishes the foundational theory of covers and envelopes within proto-exact categories and applies these results to prove that categories of Banach modules over arbitrary Banach rings possess enough injectives.

The Gist

Here is an explanation of Jack Kelly's paper, "Proto-Exact Categories and Injective Banach Modules," translated into simple language with creative analogies.

The Big Picture: Building a Safety Net for Math

Imagine you are a mathematician working with Banach modules. Think of these not as abstract numbers, but as flexible, stretchy fabrics (like rubber sheets) that have a specific "tension" or "size" (a norm) attached to them. These fabrics can be stretched, cut, and glued together, but they must follow strict rules about how much they can stretch without breaking.

The main goal of this paper is to prove a very important safety property: No matter what stretchy fabric you have, you can always wrap it inside a "super-fabric" that is perfectly safe and unbreakable. In math terms, this is called having "enough injectives."

However, the author runs into a problem. The usual tools mathematicians use to prove this safety net exists (called "Grothendieck categories") are like heavy-duty cranes. They work great for standard, rigid structures, but they are too heavy and clumsy for these stretchy, non-standard fabrics. The fabrics don't play nice with the heavy cranes.

So, the author invents a new, lighter, more flexible tool called a Proto-Exact Category.

---

Part 1: The New Tool (Proto-Exact Categories)

To understand the new tool, let's look at the old one.

• Standard Math (Abelian Categories): Imagine a world of Lego bricks. Everything snaps together perfectly. If you take a piece off, you know exactly what's left. It's very rigid and predictable.
• The Author's World (Proto-Exact Categories): Imagine a world of playdough. You can squish it, pull it, and stretch it. It doesn't snap back perfectly like Lego. Sometimes, when you pull a piece off, the edges get messy.

The author realizes that while playdough is messier, it still has rules. He defines a set of rules called Proto-Exact Categories. These rules allow us to talk about "short exact sequences" (cutting a piece of playdough) even when the edges aren't perfectly straight.

The "Obscure" Secret Sauce:
The paper introduces a rule called the "Obscure Axiom."

• The Analogy: Imagine you have a chain of people passing a bucket of water. If the person at the end of the line successfully passes the bucket to the next person, and the person at the very start successfully passed it to the second person, the "Obscure Axiom" guarantees that the middle person must have been doing their job correctly, even if you couldn't see them.
• In standard math, this is obvious. In the messy "playdough" world, it's not always true. The author proves that for these specific Banach modules, this "Obscure Axiom" does hold. This is the key that unlocks the door to the rest of the proof.

---

Part 2: The Construction Site (Covers and Envelopes)

Now that we have our flexible rules, the author needs to build the safety net. He uses a construction technique called Covers and Envelopes.

• The Cover: Imagine you have a messy pile of rocks (your module). You want to cover it with a smooth, perfect blanket (a "projective" object) so you can move it easily.
• The Envelope: Imagine you have a fragile, delicate glass sculpture (your module). You want to put it inside a strong, unbreakable box (an "injective" object) so it never breaks.

The paper proves that in this "playdough" world, you can always find a perfect box for any sculpture.

How do we build these boxes?
The author uses a concept called Deconstructibility.

• The Analogy: Imagine you want to build a giant castle out of Lego, but you don't have a blueprint for the whole thing. However, you know that every single brick in the castle is made of a specific set of smaller, standard bricks. If you can prove that the whole castle is just a giant stack of these standard bricks, you can build it piece by piece.
• The author proves that these Banach modules are "deconstructible." They are built from a small, manageable set of basic building blocks. Because they are built from these blocks, we can use a mathematical "small object argument" (a step-by-step construction process) to build the perfect safety net (the injective envelope) around any object.

---

Part 3: The Final Result (Banach Modules)

Finally, the author applies this new theory to the specific world of Banach Modules (the stretchy fabrics).

1. The Problem: The category of Banach modules is "messy" (non-additive in the strict sense). You can't just add two maps together if they are too big (norm > 1).
2. The Fix: The author switches to a slightly different version of the category where all maps are "non-expanding" (norm $\le$ 1). In this version, the rules are cleaner, and the "Obscure Axiom" works perfectly.
3. The Proof:
   • He shows these categories are Locally Presentable (they are built from small, manageable pieces).
   • He shows they satisfy the Obscure Axiom.
   • He shows they are Coherent (the pieces fit together logically).
   • Therefore, by the construction method developed in the first half of the paper, every object has an injective envelope.

The "So What?"
The paper concludes with Theorem 1.1:

"For any Banach ring $R$, the category of Banach $R$-modules has enough injectives."

In plain English:
No matter what complex, stretchy mathematical structure you are working with in this field, you can always embed it into a "perfectly safe" structure. This is a fundamental property that allows mathematicians to solve equations, extend functions, and prove other deep theorems that were previously impossible because they didn't know if a "safety net" existed.

Summary Metaphor

Think of the author as an architect.

• Old Architects tried to build a safety net for "stretchy fabrics" using steel girders (standard abelian category theory). The girders were too heavy; the fabric tore.
• Jack Kelly realized the fabric needed spider silk (proto-exact categories).
• He proved that spider silk has a special property (the Obscure Axiom) that makes it strong enough to hold the fabric.
• He then showed that you can weave this silk into a net (the injective envelope) using a specific weaving pattern (deconstructibility).
• Result: The fabric is now safe, secure, and ready for any mathematical adventure.

Technical Summary

Here is a detailed technical summary of the paper "Proto-exact Categories and Injective Banach Modules" by Jack Kelly.

1. Problem Statement

The central problem addressed in this paper is the existence of enough injectives in the category of Banach modules over an arbitrary Banach ring $R$, denoted as $\text{Ban}_R$.

• Context: In classical functional analysis, if $K$ is a spherically complete Banach field, the Hahn-Banach theorem implies that $K$ is an injective object in the category of Banach spaces over $K$ ($\text{Ban}_K$), and the category has enough injectives.
• The Obstacle: For a general Banach ring $R$, the category $\text{Ban}_R$ (with bounded morphisms) is quasi-abelian but lacks finite limits and colimits in a way that prevents the application of standard techniques from Grothendieck abelian categories or exact categories of Grothendieck type (e.g., the Gabriel-Popescu theorem or standard injective hull constructions).
• The Trade-off: Restricting morphisms to have norm $\leq 1$ yields the category $\text{Ban}^{\leq 1}_R$. This category is locally presentable (a desirable property for existence proofs) but loses additivity (it is not an additive category). Consequently, standard homological algebra tools for additive categories cannot be directly applied.

2. Methodology

Kelly develops a non-additive homological framework to bridge the gap between the structural properties of $\text{Ban}^{\leq 1}_R$ and the existence of injectives. The methodology proceeds in three main stages:

A. Generalizing to Proto-exact Categories

The paper utilizes the theory of proto-exact categories (introduced by B. Keller and further developed by others), which are non-additive generalizations of Quillen exact categories.

• Structure: The category $\text{Ban}^{\leq 1}_R$ is identified as a parabelian pointed category equipped with a canonical proto-exact structure.
• The "Obscure Axiom": A critical technical hurdle is the obscure axiom (a condition regarding the stability of strict monomorphisms/epimorphisms under composition and pullbacks/pushouts). While many proto-exact categories (like pointed sets) fail this, Kelly proves that $\text{Ban}^{\leq 1}_R$ satisfies a strong version of this axiom. This property is essential for the subsequent lifting arguments.

B. Deconstructibility and Covering Theory

To prove the existence of injectives, the author adapts the Eklof-Trlifaj strategy (originally for abelian categories) to the proto-exact setting.

• Deconstructibility: The paper generalizes the notion of deconstructibility (from [27], [29]) to proto-exact categories. A class of objects is deconstructible if its admissible monomorphisms can be built via transfinite compositions of pushouts of a small set of "building block" morphisms.
• Coherence: The author introduces local pre-coherence and coherence conditions for proto-exact categories. These conditions ensure that the class of admissible monomorphisms is generated by a set of morphisms with small domains/codomains.
• Small Object Argument: By establishing that the category is purely locally presentable and satisfies coherence, the author applies a small object argument to construct covers and envelopes (specifically, special preenvelopes) for any object.

C. Scaling and Transfer

The final step involves transferring results from the non-additive category $\text{Ban}^{\leq 1}_R$ back to the additive category $\text{Ban}_R$.

• Scalable Cotorsion Theory: The author defines scalable classes of objects. A key insight is that if a cotorsion pair is complete in the non-expanding category ($\leq 1$), it induces a complete cotorsion pair in the bounded category ($\text{Ban}_R$) via rescaling of norms.
• Injective Transfer: Since $\text{Ban}^{\leq 1}_R$ has enough injectives (proven via the covering theory), the scaling argument implies that $\text{Ban}_R$ also has enough injectives.

3. Key Contributions

1. Development of Proto-exact Covering Theory:
   The paper establishes a robust theory of covers and envelopes in proto-exact categories. It proves that if a proto-exact category is locally presentable, satisfies the obscure axiom, and possesses a coherent structure (specifically, if the class of admissible monomorphisms is generated by a set), then it has enough injectives.

2. Verification of the Obscure Axiom for Banach Modules:
   A significant technical contribution is the proof that the category of Banach modules with non-expanding morphisms ($\text{Ban}^{\leq 1}_R$) satisfies the strong obscure axiom. This distinguishes it from other non-additive categories (like pointed sets) and validates the use of advanced homological tools.

3. Generalization of Deconstructibility:
   The paper successfully generalizes the concept of deconstructibility and coherence (previously defined for exact/abelian categories) to the non-additive proto-exact setting. This allows the application of the "small object argument" in a non-additive context.

4. Proof of Enough Injectives for $\text{Ban}_R$:
   The main theorem (Theorem 4.22) resolves the open problem of whether $\text{Ban}_R$ has enough injectives for an arbitrary Banach ring $R$. The result holds for:

   • Semi-normed modules ($\text{SNrm}_R$)
   • Normed modules ($\text{Nrm}_R$)
   • Banach modules ($\text{Ban}_R$)
   • Their non-Archimedean variants.

4. Key Results

• Theorem 1.1 (Main Theorem): Let $R$ be a Banach ring. The quasi-abelian category $\text{Ban}_R$ of Banach $R$-modules has enough injectives.
• Theorem 2.32: A proto-exact category satisfying the strong right obscure axiom, having a set of admissible generators, and being cocomplete is locally presentable.
• Corollary 3.42: A purely locally $\lambda$-presentable, totally proto-exact category satisfying the obscure axiom and weakly elementary conditions has enough injectives.
• Theorem 4.22: The injective cotorsion pair in the non-expanding category ($\text{SNrm}^{\leq 1}_R$) is scalable and complete, which implies the completeness of the cotorsion pair in the bounded category ($\text{SNrm}_R$), ensuring enough injectives.

5. Significance

• Foundational for Non-Archimedean Analysis: The result provides a solid homological foundation for the study of Banach modules over arbitrary Banach rings, which is crucial for non-Archimedean functional analysis and arithmetic geometry (e.g., in the study of perfectoid spaces and rigid analytic geometry).
• Bridging Additive and Non-Additive Worlds: The paper demonstrates how to perform deep homological algebra (existence of injectives) in categories that are not additive but are "close" to being so (parabelian/proto-exact). This opens the door for applying model category techniques and cotorsion theory to a broader range of functional analytic categories.
• Resolution of a Long-standing Gap: Prior to this work, the existence of enough injectives in $\text{Ban}_R$ was known only for specific cases (like fields). This paper generalizes the result to the most general setting of Banach rings, removing the reliance on the Hahn-Banach theorem for the specific field $K$ and replacing it with categorical structural properties.

In summary, Jack Kelly successfully constructs a new homological machinery tailored for non-additive functional analysis, proving that the category of Banach modules over any Banach ring possesses the necessary injective objects to support a rich homological theory.