TheoremComplete

The Jordan Canonical Form via Modules

The Jordan canonical form of a linear operator is a direct application of the structure theorem for modules over a PID, specifically applied to k[x]k[x]-modules.


Setup

Definition9.7$k[x]$-module structure

Let VV be a finite-dimensional vector space over a field kk and T:VVT: V \to V a linear operator. Then VV becomes a k[x]k[x]-module via f(x)v=f(T)(v)f(x) \cdot v = f(T)(v). Two operators are similar (TST \sim S iff S=PTP1S = PTP^{-1}) if and only if the corresponding k[x]k[x]-modules are isomorphic.


The Jordan Form

Theorem9.3Jordan canonical form

Let VV be a finite-dimensional vector space over an algebraically closed field kk, and T:VVT: V \to V a linear operator. By the structure theorem for k[x]k[x]-modules:

Vk[x]/(xλ1)e1k[x]/(xλs)es,V \cong k[x]/(x - \lambda_1)^{e_1} \oplus \cdots \oplus k[x]/(x - \lambda_s)^{e_s},

where the λi\lambda_i are eigenvalues of TT (not necessarily distinct) and ei1e_i \geq 1. Each summand k[x]/(xλ)ek[x]/(x-\lambda)^e corresponds to a Jordan block:

Je(λ)=(λ1000λ100λ10λ)ke×e.J_e(\lambda) = \begin{pmatrix} \lambda & 1 & 0 & \cdots & 0 \\ 0 & \lambda & 1 & \cdots & 0 \\ \vdots & & \ddots & \ddots & \vdots \\ 0 & & & \lambda & 1 \\ 0 & & & & \lambda \end{pmatrix} \in k^{e \times e}.

The Jordan form J=diag(Je1(λ1),,Jes(λs))J = \mathrm{diag}(J_{e_1}(\lambda_1), \ldots, J_{e_s}(\lambda_s)) is unique up to permutation of blocks.


Rational Canonical Form

Theorem9.4Rational canonical form

Over any field kk (not necessarily algebraically closed), the invariant factor decomposition gives:

Vk[x]/(f1)k[x]/(ft),V \cong k[x]/(f_1) \oplus \cdots \oplus k[x]/(f_t),

with f1f2ftf_1 \mid f_2 \mid \cdots \mid f_t monic polynomials. Each k[x]/(fi)k[x]/(f_i) has a basis in which TT acts as the companion matrix of fif_i. The last invariant factor ftf_t is the minimal polynomial, and fi\prod f_i is the characteristic polynomial.

ExampleComputing Jordan form

T:C4C4T: \mathbb{C}^4 \to \mathbb{C}^4 with characteristic polynomial (x2)3(x5)(x-2)^3(x-5) and minimal polynomial (x2)2(x5)(x-2)^2(x-5).

Elementary divisors: (x2)2,(x2),(x5)(x-2)^2, (x-2), (x-5). Jordan form:

J=(2100020000200005).J = \begin{pmatrix} 2 & 1 & 0 & 0 \\ 0 & 2 & 0 & 0 \\ 0 & 0 & 2 & 0 \\ 0 & 0 & 0 & 5 \end{pmatrix}.

Invariant factors: f1=(x2)f_1 = (x-2) and f2=(x2)2(x5)f_2 = (x-2)^2(x-5).

RemarkWhy modules unify linear algebra

The module-theoretic approach unifies: diagonalization (all elementary divisors are linear), Jordan form (algebraically closed field), rational canonical form (arbitrary field), Smith normal form (matrices over PIDs), and the classification of finitely generated abelian groups (R=ZR = \mathbb{Z}). All are instances of the single structure theorem.