🔢 mathematics

Decompositions into a direct sum of projective and stable submodules

This paper investigates the conditions under which $R$ -modules decompose into a direct sum of projective and stable submodules, proving the existence of such decompositions for modules with finite uniform or hollow dimensions and finitely presented modules over left semihereditary rings, while providing counterexamples of modules where this decomposition fails.

Original authors: Gulizar Gunay, Engin Mermut

Published 2026-04-03

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Gulizar Gunay, Engin Mermut

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 have a giant, complex Lego structure (a module). You want to take it apart to understand how it works. The mathematicians in this paper are asking a very specific question:

"Can we always split any Lego structure into two distinct piles: one pile made entirely of 'perfect, standard bricks' (Projective), and another pile made of 'weird, custom-shaped bricks' that can never be broken down into standard bricks (Stable)?"

Here is the breakdown of their findings, using simple analogies.

1. The Goal: The Perfect Split

In the world of algebra, a Projective module is like a "standard Lego brick." It's versatile, easy to build with, and acts as a solid foundation. A Stable module is like a "weird, custom-shaped piece" that has no standard bricks inside it; it's purely unique and cannot be simplified into the standard kind.

The authors wanted to know: Can we always separate any complex structure into a pile of standard bricks and a pile of pure weirdness?

2. The Good News: When It Works

The paper starts by showing that for many common types of structures, the answer is YES.

Finite Size: If your Lego structure isn't infinitely huge (it has "finite uniform dimension"), you can always find that perfect split.
The "Local" Rule: If you are working in a "local" environment (like a small, self-contained workshop), any structure you build can be split this way.
The "Semi-Hereditary" Rule: If the rules of your Lego world are very nice and orderly (mathematically called a "left semihereditary ring"), then even complex, finitely described structures can be split perfectly.

The Analogy: Think of a well-organized factory. If the factory follows strict, simple rules, every product coming off the line can be easily separated into "standard parts" and "unique custom parts."

3. The Bad News: When It Fails

The main focus of the paper is to show that this is not always true. In some chaotic or infinite Lego worlds, you cannot split the structure. You get a mess where standard bricks and weird shapes are fused together so tightly that you can't pull them apart.

The authors provide three main scenarios where the split fails:

A. The Infinite Tower

If you have a structure that is infinitely tall (infinitely generated), you might run into a problem.

The Analogy: Imagine a tower of bricks that goes up forever. If you try to pull out the "standard" bricks, you might find that the "weird" bricks are holding the whole infinite tower together. You can't separate them without the tower collapsing.
The Result: There are infinite structures that are a mix of standard and weird, but they refuse to split.

B. The "Finitely Generated" Trap

You might think, "Okay, if I build a structure using only a finite number of bricks, it should be easy to split."

The Analogy: Imagine a small, intricate puzzle made of 100 pieces. You'd think it's easy to sort. But the authors found specific rings (rulesets) where even a small, finite puzzle has pieces that are fused together. You can't separate the "standard" pieces from the "weird" ones because the rules of the puzzle force them to stick.
The Result: Even small, finite structures can fail to split if the underlying rules of the world are slightly "wrong" (specifically, if the ring isn't "perfect" or "coherent").

C. The Ultimate Failure (The Cyclic Example)

The paper's "grand finale" is a specific example of a structure that is:

Finitely Presented: It's described by a simple, finite set of rules.
Not Projective: It's not made of standard bricks.
No Stable Part: It has zero "weird" parts that can be pulled out on their own.

The Analogy: Imagine a single, solid block of clay. It's not a standard brick, but it's also not a "weird shape" that can be separated. It is a "hybrid" that is stuck in the middle.
The Twist: The authors prove that this block isn't just a failed split; it's fundamentally different from any "weird" block. You can't even pretend it's a weird block by adding some standard bricks to it. It exists in a category all its own.

4. Why Does This Matter?

In mathematics, being able to split things into "standard" and "weird" parts is like having a universal translator. It allows mathematicians to ignore the messy "weird" parts and focus on the clean "standard" parts, or vice versa.

If the split works: Math is easy. We can solve problems by looking at just one half.
If the split fails (as shown in this paper): Math gets messy. We have to deal with the whole complex object at once.

Summary

The paper is a detective story.

The Theory: "We think we can always separate the standard from the weird."
The Confirmation: "Yes, we can, if the structure is finite or the rules are nice."
The Discovery: "But wait! If the world is infinite, or if the rules are slightly broken, we found structures that are hybrids. They are a tangled knot of standard and weird that cannot be untangled."

The authors spent their time building these "tangled knots" (counterexamples) to prove that the beautiful, clean theory of splitting modules doesn't work everywhere. They showed us the limits of our mathematical tools.

1. Problem Statement

The central problem addressed in this paper is the decomposition of modules over an arbitrary associative ring $R$ . Specifically, the authors investigate conditions under which an $R$ -module $M$ can be decomposed as a direct sum:
$M = P \oplus N$
where:

$P$ is a projective submodule.
$N$ is a stable submodule (defined as a module having no nonzero projective direct summand).

While such decompositions are known to exist for specific classes of rings (e.g., semiperfect rings, left hereditary rings) and specific module classes (e.g., finitely generated modules over semiperfect rings), the paper aims to:

Establish general conditions for the existence of such decompositions.
Identify and construct explicit counterexamples where such decompositions fail, even for finitely generated or finitely presented modules.
Determine if a module failing this decomposition is "projectively equivalent" to a stable module (i.e., if $M \oplus Q \cong N' \oplus P'$ for some stable $N'$ and projective $Q, P'$ ).

2. Methodology

The authors employ a combination of homological algebra, lattice theory, and ring-theoretic constructions:

Dimension Theory: They utilize uniform dimension (Goldie dimension) and hollow dimension to analyze the structural constraints of modules. They prove that if a module has finite uniform or hollow dimension, the decomposition must exist.
Auslander-Bridger Transpose: For finitely presented modules, the authors utilize the Auslander-Bridger transpose ( $\text{Tr}$ ). This functor maps finitely presented right modules to finitely presented left modules. They analyze the projective dimension of $\text{Tr}(M)$ and its relationship to the dual module $M^* = \text{Hom}_R(M, R)$ .
Torsionless Modules: The paper investigates the relationship between torsionless modules (modules embeddable in a direct product of copies of $R$ ) and projective modules. They leverage the result that over a right perfect and left coherent ring, all torsionless modules are projective.
Counterexample Construction: To demonstrate failures, the authors construct specific rings and modules using:
- Infinite products of fields.
- Triangular matrix rings.
- Semiartinian right $V$ -rings.
- Idealization of rings (Nagata's construction).

3. Key Contributions and Results

A. Existence of Decompositions

The paper establishes sufficient conditions for the decomposition $M = P \oplus N$ to hold:

Finite Dimensions: If a module $M$ $M$ has finite uniform dimension ( $u.\dim(M) < \infty$ $u . dim (M) < \infty$ ) or finite hollow dimension ( $h.\dim(M) < \infty$ $h . dim (M) < \infty$ ), it admits such a decomposition.
- Corollary: This implies the decomposition holds for all Noetherian, Artinian, and finitely generated modules over semilocal rings.
Finitely Presented Modules over Semihereditary Rings:
- Theorem 3.10: If $R$ is a left semihereditary ring and $M$ is a finitely presented right $R$ -module, then $M = P \oplus N$ .
- Mechanism: The proof relies on the fact that for such rings, $\text{pd}(\text{Tr}(M)) \leq 1$ . This leads to the decomposition $M \cong M^{**} \oplus \text{Ext}^1_R(\text{Tr}(M), R)$ , where $M^{**}$ is projective and the Ext term is stable.

B. Failure of Decompositions (Counterexamples)

The core contribution of the paper is the construction of modules that cannot be decomposed into a projective and a stable part.

General Failure Condition:
- Theorem 4.1: If $R$ is a right hereditary ring that is not right perfect (or not left coherent), there exists a torsionless module that is not projective. Such a module cannot be decomposed into $P \oplus N$ .
- Example: The infinite product of integers $\prod_{i=1}^\infty \mathbb{Z}$ over the ring $\mathbb{Z}$ .
Finitely Generated Failures:
- Theorem 4.3: If $R$ is a right semihereditary ring that is not a right Baer ring, there exists a cyclic (finitely generated) torsionless module that fails the decomposition.
- Example: Modules over triangular matrix rings derived from von Neumann regular rings.
Finitely Presented Failures (The Main Result):
- Theorem 4.6: Over a right semiartinian right $V$ -ring that is not semisimple, there exists a cyclic module $M$ that cannot be decomposed.
- Example 4.8 (The Primary Counterexample):
  - Construction: Let $R = \prod_{i=1}^\infty \mathbb{Z}_2$ and $S = R/T$ (a simple bimodule). Construct the commutative ring $A = \begin{pmatrix} R & S \\ 0 & R \end{pmatrix}$ (idealization).
  - Module: Let $J$ be the Jacobson radical of $A$ , and $M = A/J$ .
  - Properties:
    - $M$ is cyclically presented (hence finitely presented).
    - $M$ is not projective.
    - $M$ has no nonzero stable submodule.
  - Significance: Because $M$ has no nonzero stable submodule, it cannot be written as $P \oplus N$ (unless $N=0$ , which would imply $M$ is projective, a contradiction).
  - Projective Equivalence: The authors prove (Lemma 4.7) that if a module is not projective and has no nonzero stable submodule, it is not projectively equivalent to any stable module. Thus, $M$ represents a "pure" failure of the decomposition concept, not just a failure of a specific representation.

4. Significance

Refining Representation Theory: The paper clarifies the boundaries of decomposition theorems in module theory. While decomposition into projective and stable parts is a standard tool in the representation theory of Artin algebras (where rings are semiperfect), this work demonstrates that such tools are not universally applicable, even for well-behaved classes like finitely presented modules over commutative rings.
Stable Category Implications: The existence of a decomposition $M = P \oplus N$ allows one to study the stable module $N$ in the stable category. The counterexamples show that for certain modules, the "stable part" cannot be isolated in this manner, and the module is not even equivalent to a stable object in the stable category.
Ring Characterization: The results provide new characterizations of ring properties. For instance, the failure of decomposition for finitely presented modules is linked to the ring failing to be right perfect or left coherent.
Methodological Advancement: The use of the Auslander-Bridger transpose to derive decomposition criteria for semihereditary rings offers a powerful homological technique that connects projective dimension of the transpose to the stability of the original module.

In summary, the paper successfully delineates the precise algebraic conditions under which the decomposition $M = P \oplus N$ holds and provides rigorous, constructive proofs of its failure in specific, non-trivial cases, thereby deepening the understanding of the structure of modules over general rings.