Diagonalizable matrix Created 2026-09-24 Updated 2026-10-14
A finite-dimensional linear operator is diagonalizable over a specified field when it has an eigenbasis. Equivalently, its matrix admits
where is an invertible matrix whose columns are eigenvectors, and is a diagonal matrix of the corresponding eigenvalues. This change of basis reveals a linear map as independent one-dimensional scalings. The matrix may look coupled in the original coordinates, but each coordinate in the eigenbasis evolves on its own.
For the symmetric matrix computed in eigenvector, take
Then , hence . The orthonormal basis version is , an orthogonal matrix satisfying . Every real symmetric matrix has an orthonormal eigenbasis by the spectral theorem; a general diagonalizable matrix need not have orthogonal eigenvectors.
The immediate payoff is repeated application: adjacent factors cancel, giving . In this example,
The component in grows while the component in stays fixed. After division by its Euclidean norm, this particular sequence approaches . This is the mechanism behind the power method; an initial vector entirely in would stay there and never find the dominant eigenvalue .
Figure 1.
Repeated maps: eigenvalue growth, alignment, and the exceptional starting direction
. Details.
Two different obstructions deserve comparison. The planar rotation in eigenvector is diagonalizable over but not over , because its eigenvalues are and . By contrast, the shear mapping
has only , of algebraic multiplicity two, and , of geometric multiplicity one. There is no second independent eigenvector, even over . This is a defective matrix, already a two-dimensional Jordan block. Its powers are
Thus even though both eigenvalues are one, repeated application can produce growth. Eigenvalues alone miss the coupling recorded by a Jordan block. A repeated eigenvalue is harmless when it has enough independent eigenvectors, as demonstrates. More generally, a matrix is diagonalizable exactly when its characteristic polynomial splits over the chosen field and every eigenvalue has equal geometric multiplicity and algebraic multiplicity. The proof outline is to join bases of the distinct eigenspaces; their direct sum fills the space exactly under these conditions.
A physical motivation is the separation of normal modes. Consider two equal masses between three identical springs of stiffness , with the outer ends fixed. By Hooke's law and Newton's second law, the displacement vector obeys
The eigenvectors and have eigenvalues and : the masses move together in the first normal mode and oppositely in the second. Writing separates the coupled equations into two harmonic oscillators,
Their angular frequencies are and ; the eigenvalues give squared angular frequencies after multiplication by . This is the beauty of diagonalization: finding the right directions turns interacting motions into independent patterns. For a physical treatment, see Yen-Jie Lee's "Coupled Oscillators, Normal Modes" (MIT 8.03SC, 2016), and the local article normal modes of two equal masses between three springs.[ref]
Free two-mass spring 2026-10-14
Take two point masses joined by a massless spring of spring constant and equilibrium length . Assume one-dimensional motion along the spring, no external force, no damping, and an extension satisfying Hooke's law; keep . For a real spring this is a model within its elastic range. Its potential energy is , so
Here is angular frequency and is frequency. Choose so the model bodies do not cross. The internal normal mode has displacements and about equilibrium offsets from the centre of mass; its amplitude ratio is . The other normal mode is collective translation with zero frequency and no spring extension. This free pair has different boundary conditions from normal modes of two equal masses between three springs.
For and two masses, and . Replacing the second mass by gives and . Fixing the second endpoint instead would give : the free heavy partner still moves. The free pair's angular frequency exceeds that fixed-end limit by the factor , about at .
Figure 1.
Reduced mass and free two-mass spring angular frequency as the partner becomes heavier
. Original dimensionless model with and held fixed, , and . The marks at and correspond to the computed examples; curves are predictions, not measured data. Details.
The Euler--Lagrange equations for the quadratic Lagrangian are
For a normal mode , they become the generalized eigenvalue problem for small oscillations
Put . With the stated spring constants,
The final matrix has characteristic polynomial
so its eigenvalue are . Hence
The two angular frequencies therefore satisfy
This is the spectrum recorded by normal modes of two equal masses between three springs.
Expand about the equilibrium and keep only terms quadratic in displacement and velocity. Writing and , the linearized Lagrangian is , up to a constant. Its equations are . With positive-definite mass matrix , the normal frequencies and mode vectors solve
For a nondegenerate stable equilibrium, is positive definite and all these frequencies are positive. A zero eigenvalue indicates a zero-frequency linearized mode requiring higher-order analysis if nonlinear stability is in question.
The three spring extensions are , and . Hence
The matrices are and . The two independent modes are
In the first mode both masses move together with equal displacement, leaving the middle spring unchanged. In the second they move in opposite directions with equal magnitude; the midpoint stays fixed and the middle spring changes length by twice either displacement. Arbitrary small motion is a superposition of sinusoidal motions in these two modes.
Figure 1.
The two normal modes of equal masses joined to fixed walls by three equal springs
. The in-phase mode leaves the middle spring unchanged; the opposite-phase mode stretches or compresses it by twice either mass displacement. Details.