🔢 mathematics

Jordan types commuting with a hook partition

This paper provides a complete classification of the Jordan types found in the nilpotent commutator of a matrix with a hook partition, demonstrating that sharing the same generic commuting Jordan type does not guarantee that two partitions commute.

Original authors: Leila Khatami, Tomaž Košir

Published 2026-05-28

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Leila Khatami, Tomaž Košir

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 box of building blocks. In the world of mathematics, these blocks are arranged in specific shapes called partitions. A partition is just a way of breaking a number down into smaller chunks, like breaking the number 6 into 4 + 2, or 3 + 3, or 2 + 2 + 2.

Now, imagine these blocks are also machines (matrices) that can spin and move. Some of these machines are "nilpotent," which is a fancy way of saying they are like a clock that eventually stops ticking completely after a certain number of turns.

The big question this paper asks is: Which shapes of blocks can work together without crashing?

In math terms, if you have two machines, $A$ and $B$ , they "commute" if you can run $A$ then $B$ , or $B$ then $A$ , and get the exact same result. The authors are looking at a very specific, special shape of machine called a "hook partition."

The "Hook" Shape

Think of a hook partition like a fishing hook or a lollipop. It has one long, straight stick (the main part of the number) and a bunch of tiny, single blocks hanging off the bottom (the "1"s).

Example: If you have 6 blocks, a hook shape might look like a long stick of 5 blocks with 1 little block hanging off the end: (5, 1).
Another hook: A stick of 4 with two little blocks: (4, 1, 1).

The authors wanted to know: If I have a machine shaped like a hook, what other shapes of machines can I pair it with so they work together perfectly?

The Discovery: The "Almost Rectangular" Rule

The paper provides a complete "rulebook" for this. They found that for a hook machine to work with another machine, the other machine must contain a specific "core" shape inside it.

Think of it like a puzzle. If your hook is the frame, the other piece must have a core that is "almost rectangular."

An "almost rectangular" shape is one where the pieces are all roughly the same size, like a square or a rectangle that is just one block off from being perfect.
The authors proved that the other machine must contain a chunk that looks like a rectangle of size $n$ , $n-1$ , or $n-2$ (where $n$ is the length of your hook's stick).

If the other machine doesn't have this "almost rectangular" core, it's like trying to fit a round peg in a square hole—they simply won't commute. They will clash.

The Surprising Twist: "Generic" vs. "Specific"

Here is the most interesting part of the paper, which the authors highlight as a major consequence.

In math, sometimes we look at the "average" or "generic" behavior of these machines. It's like saying, "On average, these two shapes get along." The paper shows that just because two shapes get along with a third shape on average, it doesn't mean they get along with each other.

The Analogy:
Imagine a party.

Person A (Hook Shape) gets along with Person B (Shape X).
Person C (Shape Y) also gets along with Person A.
You might assume B and C are friends because they both like A.
The Paper's Finding: B and C might actually hate each other!

The authors give specific examples where two different shapes both work perfectly with a hook shape, but when you put those two shapes together, they crash and burn. They prove that "getting along with the same person" does not guarantee "getting along with each other."

Why This Matters (In the Paper's Context)

The paper doesn't talk about building bridges or curing diseases. Instead, it solves a pure logic puzzle in the world of algebra.

Classification: They finally listed every single possible shape that can work with a hook shape. No more guessing.
Debunking a Myth: They showed that a common assumption in this field (that if two shapes share a "generic" partner, they must be compatible) is false.

Summary

The Hook: A specific, lollipop-shaped arrangement of numbers.
The Rule: To work with a hook, another shape must hide a "nearly rectangular" block inside it.
The Twist: Two shapes that both work with a hook don't necessarily work with each other.
The Goal: To map out exactly who plays nice with whom in the universe of these mathematical machines.

The authors, Leila Khatami and Tomaž Košir, have essentially drawn the final map for this specific corner of the mathematical universe, showing us exactly which shapes fit together and which ones don't.

Technical Summary: Jordan Types Commuting with a Hook Partition

Problem Statement
The paper addresses the classification of partitions that can occur as the Jordan types of nilpotent matrices commuting with a fixed nilpotent matrix. Specifically, it focuses on the case where the fixed matrix has a hook partition Jordan type, denoted as $\Gamma_{n,m} = (n, 1^m)$ , where $n \ge 3$ and $m \ge 1$ . The central question is to determine the complete set of partitions $Q$ of $N = n+m$ such that the nilpotent orbit $\mathcal{O}_Q$ intersects the nilpotent commutator $\mathcal{N}_B$ of a matrix $B$ with Jordan type $\Gamma_{n,m}$ .

Methodology
The authors employ a structural analysis of the nilpotent commutator algebra. Their approach involves:

Algebraic Setup: Utilizing the maximal nilpotent subalgebra $U_B$ of the commutator $\mathcal{N}_B$ . They establish that $U_B$ intersects every nilpotent orbit that intersects $\mathcal{N}_B$ .
Matrix Representation: Representing elements $A \in U_B$ in a specific block matrix form involving a strictly upper-triangular block $H$ , a strictly lower-triangular block $V$ , and coupling vectors $u$ and $d$ .
Rank Analysis: To determine the Jordan type of a generic element $A$ , the authors compute the ranks of its powers $A^l$ . They analyze how the off-diagonal blocks interact with the diagonal blocks ( $J_n^k$ and $J_\nu$ ) to alter the rank sequence, which dictates the resulting partition.
Case Decomposition: The proof is structured by decomposing the possible Jordan types of the sub-blocks and analyzing specific conditions on the parameters (such as the relationship between $n$ , $k$ , and the parts of the auxiliary partition $\mu$ ) to derive necessary and sufficient conditions for commutativity.

Key Contributions and Results
The paper provides a complete classification of all partitions $Q$ that commute with a hook partition $\Gamma_{n,m}$ . The main result (Theorem 3) states that a partition $Q$ commutes with $(n, 1^m)$ if and only if it falls into one of three specific structural categories derived from an auxiliary partition $\mu$ :

Type (a): $Q$ is obtained from $(\mu, [n]_k)$ , where $\mu$ is an arbitrary partition of $m$ and $1 \le k \le n$ . Here, $[n]_k$ denotes the Jordan type of the $k$ -th power of an $n \times n$ Jordan block.
Type (b): $Q$ is obtained from $(\mu, [n-1]_k)$ , where $\mu$ is a partition of $m+1$ and $k$ satisfies $\lceil \frac{n-1}{\mu_1} \rceil \le k \le n-1$ .
Type (c): $Q$ $Q$ is obtained from $(\mu, [n-2]_k)$ $(μ, [n - 2]_{k})$ , where $\mu$ $μ$ is a partition of $m+2$ $m + 2$ . This case splits into two sub-cases:
- $\mu = (m+2)$ with $\lceil \frac{n-1}{m+1} \rceil < k \le n-2$ .
- $\mu = (\mu_1, \mu_2, \dots)$ with $\mu_2 \ge 2$ and $\lceil \frac{n-1}{\mu_2} \rceil < k \le n-2$ .

The authors prove that a necessary condition for a partition to commute with $(n, 1^m)$ is that it must contain a subpartition that is an almost rectangular partition of size $n$ , $n-1$ , or $n-2$ .

Corollaries and Examples

Corollary 5: For the specific case of $\Gamma_{3,m} = (3, 1^m)$ , a partition $P$ of $m+3$ commutes with it if and only if the smallest part of $P$ is at most 3. Consequently, partitions with all parts $\ge 4$ do not commute with $(3, 1^m)$ .
Non-Commutativity of Inverse Images: The paper provides examples demonstrating that two partitions with the same generic commuting Jordan type (i.e., they map to the same Rogers-Ramanujan partition $Q$ under the map $D$ ) do not necessarily commute with each other. Specifically, for a Rogers-Ramanujan partition $Q=(p,q)$ , the two hook partitions in its inverse image, $\Gamma_{p,q}$ and $\Gamma_{q+2, p-2}$ , may fail to commute with certain other partitions in $D^{-1}(Q)$ , and in some cases, fail to commute with each other.

Significance
The paper claims significance in providing a definitive solution to the classification problem for hook partitions, a specific but non-trivial class of partitions. By characterizing the exact structure of commuting orbits, the authors:

Clarify the relationship between hook partitions and the map $D$ (which maps a partition to the generic Jordan type of its commutator).
Refute the intuitive assumption that partitions sharing a generic commuting type must commute with one another, thereby refining the understanding of the structure of the inverse image $D^{-1}(Q)$ .
Extend previous results, such as those by Britnell and Wildon for the case $m=1$ , to the general case $m \ge 1$ .

The work relies on and synthesizes prior results regarding universally commuting partitions (e.g., $(2k, 1^l)$ ) and Rogers-Ramanujan partitions, placing the hook partition case within the broader context of the "Box Conjecture" and the geometry of nilpotent orbits.