🔢 mathematics

Lie and Jordan Isomorphisms of Algebras of Triangular Matrices over Associative Rings

This paper characterizes the Lie and Jordan isomorphisms of algebras consisting of triangular matrices defined over associative rings.

Original authors: Oksana Bezushchak

Published 2026-07-08

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Oksana Bezushchak

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 machine made of many smaller gears and levers. In the world of mathematics, this machine is called an algebra of triangular matrices. Think of it as a grid of numbers where everything below the main diagonal is zero. It's a specific, structured way of organizing information.

The author of this paper, Oksana Bezushchak, is asking a very specific question: "If I have two of these machines, and I can translate the rules of one into the rules of the other without breaking anything, what does that translation look like?"

To understand the answer, we need to look at two different ways of "translating" these machines, which the paper calls Lie and Jordan isomorphisms.

The Two Ways to Translate

Think of the machine as a set of instructions for moving parts around.

The "Lie" Translation (The Difference Maker):
Imagine you have two people, Alice and Bob. You ask them to swap places.
- In the Lie view, we only care about the difference between what happens when Alice goes first and Bob goes second, versus Bob first and Alice second. It's all about the "order of operations" and the friction (or commutator) created when things don't commute.
- The paper asks: If I have a rule that perfectly preserves these "differences" between two machines, what does that rule look like?
The "Jordan" Translation (The Symmetry Keeper):
Now, imagine you are squaring a number (multiplying it by itself) or doing a special three-step dance: $x \cdot y \cdot x$ .
- In the Jordan view, we care about preserving these specific patterns and symmetries. It's like saying, "If you do this specific move, the result must look the same in the new machine."
- The paper asks: If I have a rule that perfectly preserves these "squaring" and "three-step dance" patterns, what does that rule look like?

The Big Discovery: The "Split Personality"

The paper's main discovery is that these translations aren't random. They are surprisingly structured.

Imagine you have a large, complex puzzle. The author proves that if you can translate one triangular matrix machine to another, your translation is actually a mixture of two very simple, opposite types of translations:

The "Mirror" (Isomorphism): This is a standard, direct translation. It's like looking at the machine in a mirror that doesn't flip left and right; it just copies the structure perfectly.
The "Anti-Mirror" (Anti-isomorphism): This is a translation that flips the order. It's like looking in a mirror that reverses everything. If the original machine says "Do A then B," the anti-mirror says "Do B then A."

The "Split Personality" Metaphor:
The paper shows that your translation machine is like a person with a split personality.

Part of the machine (let's call it the "Left Side") acts like a Mirror, copying things directly.
The other part (the "Right Side") acts like an Anti-Mirror, flipping the order of things.

Crucially, the paper proves that the machine is split by a specific switch (an "idempotent" element).

If the switch is set to "1" (On), the whole machine acts like a Mirror.
If the switch is set to "0" (Off), the whole machine acts like an Anti-Mirror.
If the switch is set to something in between, the machine is a hybrid: some parts are Mirrors, and other parts are Anti-Mirrors.

The "Twist" (Conjugation)

There is one more ingredient. Even after you decide which parts are Mirrors and which are Anti-Mirrors, the translation might still be "twisted."

Imagine you have a perfect translation, but the machine is rotated or shifted slightly. The paper shows that any valid translation is just a Mirror/Anti-Mirror mix that has been twisted by a specific type of "gimmick" (mathematically called conjugation by invertible elements).

Think of it like this: You have a recipe (the translation).

You can follow the recipe normally (Mirror).
You can follow the recipe backwards (Anti-Mirror).
You can split the recipe: follow the first half normally and the second half backwards.
Finally, you can shake the ingredients a bit (the "twist") before you start cooking, but the core structure of the recipe remains one of the three options above.

Summary of the Paper's Claims

The Scope: The paper looks at "triangular matrix algebras" over "associative rings." In plain English, these are structured grids of numbers with specific rules, built over a foundation of other numbers.
The Result: The author provides a complete "blueprint" for every possible way to translate these machines while preserving their Lie or Jordan rules.
The Blueprint: Every such translation is a combination of:
1. A direct copy (Isomorphism).
2. A reversed copy (Anti-isomorphism).
3. A split between the two (determined by a switch in the underlying number system).
4. A geometric twist (conjugation by specific matrices).

The paper does not claim to solve problems in physics, engineering, or medicine. It is a pure mathematics paper that maps out the "DNA" of these specific mathematical structures, showing that despite their complexity, their translations are always built from these simple, fundamental blocks.

Technical Summary: Lie and Jordan Isomorphisms of Algebras of Triangular Matrices over Associative Rings

Problem Statement

The paper addresses the structural classification of Lie and Jordan isomorphisms between algebras of upper triangular matrices, denoted as $T(R)$ and $T(S)$ , where $R$ and $S$ are associative unital $\Phi$ -algebras over a commutative associative ring $\Phi$ containing 1.

While the structure of such isomorphisms is well-understood for semiprime rings (e.g., full matrix rings) and specific cases of triangular matrices over commutative rings without proper idempotents, a general description for triangular matrix algebras over arbitrary associative rings (which are generally not semiprime and may lack full proper idempotents) remained incomplete. Specifically, the paper seeks to describe isomorphisms $\phi: T(R) \to T(S)$ without assuming that $\phi$ is $R$ -linear or that the base ring $\Phi$ is free of 2-torsion or proper idempotents.

Methodology

The author employs a decomposition strategy based on the properties of idempotents in the base ring $\Phi$ . The methodology proceeds through the following logical steps:

Decomposition via Idempotents: The core technique involves identifying an idempotent $h \in \Phi$ that splits the problem into two distinct components: one where the isomorphism behaves as a standard homomorphism (associated with $h$ ) and another where it behaves as an anti-homomorphism (associated with $1-h$ ).
Analysis of Adjoint Lie Algebras: For Lie isomorphisms, the analysis is conducted on the adjoint Lie algebras $T(R)(-)$ and $T(S)(-)$ , where multiplication is replaced by the commutator $[x, y] = xy - yx$. The study focuses on the quotient algebras modulo the central ideals $\Phi \cdot I_R$ and $\Phi \cdot I_S$ .
Inductive Normalization of Idempotents: A critical technical component is the inductive proof (Lemmas 3.2 and 4.2) demonstrating that, up to conjugation by specific invertible matrices (from the upper unitriangular group $UT(n, S)$ and the diagonal group $Diag(S^*)$ $D ia g (S^{*})$ ), the isomorphism can be normalized to map the standard matrix units $e_{ii}(h)$ $e_{ii} (h)$ to $e_{ii}(h)$ $e_{ii} (h)$ .
- This involves analyzing the centralizers of sets of orthogonal idempotents and utilizing the structure of the strictly upper triangular part $A(R)$ .
- The proofs rely on the commutator relations (for Lie) and the Jordan product relations $x \circ y = xy + yx$ (for Jordan) to constrain the images of matrix units.
Reduction to Base Ring Isomorphisms: Once the matrix units are normalized, the isomorphism is shown to induce a mapping on the underlying ring entries. The linearity and multiplicative (or anti-multiplicative) properties of this induced map are derived from the preservation of the Lie bracket or Jordan product on the matrix units.
Handling Anti-isomorphisms: For the component corresponding to $1-h$ , the paper utilizes the anti-isomorphism $\tau_S: T(S) \to T(S^{op})$ to transform the problem into one solvable by the homomorphism case, effectively showing that the "anti" part of the isomorphism corresponds to an anti-isomorphism of the underlying rings.

Key Contributions and Results

1. Classification of Lie Isomorphisms (Theorem 1)

The paper establishes that any Lie isomorphism $\phi$ between the adjoint Lie algebras of triangular matrix algebras (modulo central ideals) can be decomposed. Specifically, there exists an idempotent $h \in \Phi$ such that:

On the component $hR$, $\phi$ lifts to a standard isomorphism $\psi_1: hR \to hS$ .
On the component $(1-h)R$ , $\phi$ lifts to an anti-isomorphism $\psi_2: (1-h)R \to (1-h)S$ .
The total map is a composition of these ring maps, lifted to the matrix level, and conjugated by invertible elements $g_1 \in UT(n, S)$ and $g_2 \in Diag(S^*)$ .
Formally, $\phi$ lifts to $b_{g_1} b_{g_2} (\widehat{\psi_1} - \widehat{\psi_2})$ .

2. Classification of Jordan Isomorphisms (Theorem 2)

Similarly, for Jordan isomorphisms $\phi: T(R) \to T(S)$ , the paper proves that there exists an idempotent $h \in \Phi$ such that $\phi$ lifts to:

A sum of a homomorphism part and an anti-homomorphism part, conjugated by the same types of invertible matrices.
Formally, $\phi = b_{g_1} b_{g_2} (\widehat{\psi_1} + \widehat{\psi_2})$ .

3. Generalization of Previous Work

The results generalize previous descriptions found in the literature:

They extend the work of Dokovi´c [14] and Cao [12] on $T(\Phi)$ (where the base ring is commutative) to the case where $R$ is an arbitrary associative $\Phi$ -algebra.
They extend the results of Beidar, Breˇsar, and Chebotar [8] and Liu and Tsai [17, 18] regarding rings without proper idempotents or 2-torsion to the general case where such restrictions are not assumed.

Significance and Claims

The paper claims that Theorems 1 and 2 are new even in the specific case of automorphisms of $T(R)$ where $R$ is a commutative $\Phi$ -algebra. The novelty lies in the removal of the assumption that the Lie or Jordan automorphism $\phi$ must be $R$ -linear. By allowing $\phi$ to be merely $\Phi$ -linear, the paper provides a more comprehensive structural description that accounts for the potential splitting of the algebra via idempotents in the base ring $\Phi$ .

The work completes the description of these isomorphisms for the class of triangular matrix algebras over associative rings, filling a gap left by earlier studies that were restricted to semiprime rings, 2-torsion-free rings, or rings without proper idempotents. The results confirm that the structure of such isomorphisms is fundamentally determined by the decomposition of the base ring into parts supporting homomorphisms and parts supporting anti-homomorphisms, mediated by standard matrix conjugations.