Why use the rational canonical form when the Jordan form is more familiar?

The Jordan form requires the eigenvalues to lie in the field, so it needs an algebraically closed field. The rational canonical form works over any field, including the rationals, by using companion matrices of the invariant factors instead of eigenvalues.

How are canonical forms related to module theory?

A vector space with a fixed operator is a module over the polynomial ring in one variable, a principal-ideal domain. The structure theorem for such modules decomposes it into cyclic pieces, and reading off those pieces gives exactly the rational and Jordan canonical forms.

Canonical Form

A canonical form is a standard matrix representative of a linear operator under similarity, providing a complete and computable invariant that classifies operators up to change of basis.

Definition

A canonical form is a distinguished matrix to which every operator in a similarity class is similar, so that two operators are conjugate exactly when they share the same canonical form; the principal examples are the rational and Jordan canonical forms.

Scope

This topic covers similarity of matrices, invariant factors and elementary divisors, the rational canonical form valid over any field, the Jordan canonical form over an algebraically closed field, and their derivation from the structure theorem for modules over a principal-ideal domain.

Core questions

When are two matrices similar?
What complete set of invariants classifies an operator up to similarity?
How are the rational and Jordan canonical forms constructed?
How does the module structure theorem produce canonical forms?

Key theories

Rational canonical form: Over any field every operator is similar to a unique block-diagonal matrix built from companion matrices of its invariant factors, so the invariant factors form a complete similarity invariant.
Jordan canonical form: Over an algebraically closed field every operator is similar to a unique Jordan matrix, a block-diagonal arrangement of Jordan blocks indexed by eigenvalues and elementary divisors, refining the rational form.
Canonical forms from the PID structure theorem: Viewing a vector space with an operator as a module over the polynomial ring, the structure theorem for finitely generated modules over a principal-ideal domain yields both canonical forms as its concrete manifestation.

Clinical relevance

Canonical forms make the classification of operators effective: the Jordan form reveals how an operator acts even when not diagonalizable, which is essential for solving linear systems of differential equations, computing matrix exponentials, and analyzing the long-term behavior of linear dynamical systems.

History

Weierstrass introduced elementary divisors and Jordan gave his canonical form in the 1870s, classifying operators by their behavior on generalized eigenspaces. Frobenius developed the rational canonical form valid over any field, and the modern derivation unifies them through module theory.

Key figures

Camille Jordan
Karl Weierstrass
Ferdinand Georg Frobenius

Seminal works

hoffman1971
dummit2004
roman2008

Frequently asked questions

Why use the rational canonical form when the Jordan form is more familiar?: The Jordan form requires the eigenvalues to lie in the field, so it needs an algebraically closed field. The rational canonical form works over any field, including the rationals, by using companion matrices of the invariant factors instead of eigenvalues.
How are canonical forms related to module theory?: A vector space with a fixed operator is a module over the polynomial ring in one variable, a principal-ideal domain. The structure theorem for such modules decomposes it into cyclic pieces, and reading off those pieces gives exactly the rational and Jordan canonical forms.