[Paper Review] Classification of Finite Alexander Quandles

Sam Nelson|ArXiv.org|Feb 26, 2002

Rings, Modules, and Algebras26 citations

TL;DR

This paper classifies finite Alexander quandles by showing that two such quandles of the same size are isomorphic if and only if their $\mathbb{Z}[t^{\pm 1}]$-submodules $\mathrm{Im}(1-t)$ are isomorphic as modules. The key contribution is a complete classification procedure for finite Alexander quandles, including explicit conditions for isomorphism, duality, and connectivity of linear quandles of the form $\mathbb{Z}_n[t^{\pm 1}]/(t-a)$ with $\gcd(n,a)=1$, and a full enumeration of distinct and connected quandles up to 15 elements.

ABSTRACT

Two finite Alexander quandles with the same number of elements are isomorphic iff their Z[t,t^-1]-submodules Im(1-t) are isomorphic as modules. This yields specific conditions on when Alexander quandles of the form Z_n[t,t^-1]/(t-a) where gcd(n,a)=1 (called linear quandles) are isomorphic, as well as specific conditions on when two linear quandles are dual and which linear quandles are connected. We apply this result to obtain a procedure for classifying Alexander quandles of any finite order and as an application we list the numbers of distinct and connected Alexander quandles with up to fifteen elements.

Motivation & Objective

• To provide a complete classification of finite Alexander quandles by reducing isomorphism problems to module-theoretic comparisons.
• To resolve open questions on when linear Alexander quandles $\mathbb{Z}_n[t^{\pm 1}]/(t-a)$ are isomorphic or dual, particularly for $\gcd(n,a)=1$.
• To determine which linear quandles are connected, extending prior results on quandle classification.
• To enumerate all distinct and connected Alexander quandles of order up to 15, offering a complete computational reference.

Proposed method

• The paper establishes that two finite Alexander quandles are isomorphic if and only if their $\mathrm{Im}(1-t)$ submodules are isomorphic as $\mathbb{Z}[t^{\pm 1}]$-modules.
• It uses module-theoretic techniques to analyze the structure of $\mathrm{Im}(1-t)$ for quandles over $\mathbb{Z}_n[t^{\pm 1}]/(h)$, particularly for linear quandles with $h = t - a$.
• The method involves analyzing the action of the variable $t$ as an automorphism on the underlying abelian group, reducing the problem to classifying $\mathbb{Z}[t^{\pm 1}]$-module structures.
• It applies the isomorphism criterion to enumerate quandles by computing $\mathrm{Im}(1-t)$ for various quotient rings and comparing their module types.
• For non-cyclic abelian groups like $\mathbb{Z}_4 \oplus \mathbb{Z}_2$, it classifies quandle structures via automorphism conjugacy classes and module isomorphisms.
• It leverages the fact that conjugate automorphisms yield isomorphic quandle structures, bounding the number of distinct quandles via conjugacy class counts.

Experimental results

Research questions

• RQ1 When are two finite Alexander quandles of the same order isomorphic as quandles?
• RQ2 What conditions on $a$ and $b$ ensure that the linear quandles $\mathbb{Z}_n[t^{\pm 1}]/(t-a)$ and $\mathbb{Z}_n[t^{\pm 1}]/(t-b)$ are isomorphic?
• RQ3 When is a linear Alexander quandle $\mathbb{Z}_n[t^{\pm 1}]/(t-a)$ connected, and when are two such quandles dual?
• RQ4 How many distinct and connected Alexander quandles exist of order $n$ for $n \leq 15$?
• RQ5 What is the role of $\mathrm{Im}(1-t)$ as a $\mathbb{Z}[t^{\pm 1}]$-module in determining quandle isomorphism?

Key findings

• Two finite Alexander quandles of the same cardinality are isomorphic if and only if their $\mathrm{Im}(1-t)$ submodules are isomorphic as $\mathbb{Z}[t^{\pm 1}]$-modules.
• For linear quandles $\mathbb{Z}_n[t^{\pm 1}]/(t-a)$ with $\gcd(n,a)=1$, isomorphism holds iff $a$ and $b$ generate the same ideal in $\mathbb{Z}_n$ under the action of $t$.
• $\Lambda_9/(t-4) \cong \Lambda_9/(t-7) \cong \Lambda_9/(t^2 + t + 1)$, showing that non-isomorphic polynomials can yield isomorphic quandles.
• Among order 9 quandles, 8 are distinct and 5 are connected; specifically, $\Lambda_9/(t-2)$, $\Lambda_9/(t-5)$, and $\Lambda_9/(t-8)$ are connected.
• For $n \leq 15$, the number of distinct Alexander quandles ranges from 1 at $n=2$ to 12 at $n=13$, with 11 connected quandles at $n=13$.
• The only connected Alexander quandle of order 4 is $\Lambda_2/(t^2 + t + 1)$, and no linear quandle of order 8 is connected.