A matrix with a repeated eigenvalue may not have enough eigenvectors to be diagonalised. It is then similar to a Jordan form, built from blocks J = ( λ 0 1 λ ) , which is as close to diagonal as possible.
A = P J P − 1 , J = diag ( J k 1 ( λ 1 ) , … )
On this page
Before you start
This page is easier if you are comfortable with:
What Jordan normal form means
Suppose λ is a double eigenvalue but A − λ I is not the zero matrix. Then there is only one eigenvector direction. A second vector w with ( A − λ I ) w = v completes a basis, and in that basis the matrix is a Jordan block.
Powers of a block are easy: J n has λ n on the diagonal and n λ n − 1 above it. This extends diagonalisation to the cases that fail, and it is the key to the matrix exponential of such matrices.
The discriminant of the characteristic equation being zero signals a repeated eigenvalue, but you need the eigenvector equation to decide whether the matrix is diagonalisable or needs a Jordan block.
What is it used for? The Jordan form handles matrices that cannot be diagonalised. It puts them in a form that is almost diagonal and still easy to use.
Powers and exponentials of defective matrices. Solving systems of differential equations with repeated roots. Understanding when diagonalisation is impossible.
How to use jordan normal form
Find the repeated eigenvalue. Check whether A − λ I is zero; if so the matrix is already diagonal. Otherwise find the one eigenvector and a generalised eigenvector. Write the Jordan block with 1 above the diagonal.
Graph Entries of Jⁿ for J with λ = 2: λⁿ and nλⁿ⁻¹ 12 at n = 3 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 20 40 60 80 Entries of Jⁿ for J with λ = 2: λⁿ and nλⁿ⁻¹ x y λⁿ (diagonal) nλⁿ⁻¹ (above diagonal) Graph with x from 0 to 5 and y from 0 to 90. Shows the curve λⁿ (diagonal); the curve nλⁿ⁻¹ (above diagonal); point 12 at n = 3 at (3, 12). The off-diagonal entry grows faster than the diagonal, which is the signature of a Jordan block.
Worked examples
These examples are worked by hand, step by step, so you can compare your own working.
Example 1: a defective matrix Find the Jordan form of ( 3 − 1 1 1 ) .
Worked example
Answer
( 2 1 0 2 ) \begin{pmatrix}2&1\\0&2\end{pmatrix} ( 2 0 1 2 ) Trace 4, determinant 4
λ = 2 (repeated)
A − 2I
( 1 − 1 1 − 1 ) = 0 (one eigenvector)
Example 2: powers of a block Find J 3 for J = ( 2 0 1 2 ) .
Worked example
Answer
( 8 12 0 8 ) \begin{pmatrix}8&12\\0&8\end{pmatrix} ( 8 0 12 8 ) J²
( 4 0 4 4 )
J³
( 8 0 12 8 )
Example 3: a diagonalisable repeated eigenvalue Is ( 2 0 0 2 ) in Jordan form?
Worked example
Answer
yes, it is 2 I \text{yes, it is } 2I yes, it is 2 I Already diagonal, so no Jordan block is needed
Special cases
A few situations come up again and again. Knowing them saves time and helps you spot answers that cannot be right.
J = (3 1; 0 3) trace
6.
same det
9.
J² entry (1,2) = 2λ
6.
Jⁿ entry (1,2) = nλⁿ⁻¹, n = 4
108.
Word problems
A defective matrix (too few eigenvectors) is similar to a Jordan form with blocks (λ 1; 0 λ). Jⁿ = (λⁿ nλⁿ⁻¹; 0 λⁿ).
Trace λ = 3 twice.
3 + 3 Check: ✓ 6
Determinant λ² = 9.
3² Check: ✓ 9
J² off-diagonal Higher 2λ.
2 × 3 Check: ✓ 6
J⁴ off-diagonal Higher 4λ³.
4 × 27 Check: ✓ 108
Common mistakes
Jordan block is diagonal diagonal ✗
has a 1 above ✓
Defective.
Jⁿ off-diagonal λⁿ λⁿ ✗
nλⁿ⁻¹ ✓
Binomial.
Every matrix diagonalisable always ✗
not if defective ✓
Jordan form.
Eigenvalues change change ✗
unchanged by similarity ✓
Same.
Exam tips
Repeated eigenvalue, one eigenvector ⇒ Jordan block.
Jⁿ off-diagonal nλⁿ⁻¹.
tr and det unchanged.
Diagonalisable if enough eigenvectors.
Block size shows defect.
Practice questions
Try each one on paper, then open the answer. Use the solver above to see the full working.
1. J² off-diagonal for λ = 5 Foundation 10
2. J³ off-diagonal for λ = 2 Foundation 12
3. det J for λ = 4 Foundation 16
4. tr J for λ = 7 Higher 14
5. J⁵ off-diagonal for λ = 1 Higher 5
6. J³ diagonal for λ = 3 Higher 27
Practice generator Unlimited fresh questions with instant marking. Choose a level, work it out on paper, then type your answer.
Foundation Higher Mixed New question
Solve: …
The practice generator needs JavaScript.
Key terms
Jordan normal form Block-diagonal canonical form. Jordan block (λ 1; 0 λ). Defective matrix Too few eigenvectors. Generalised eigenvector Solves (A − λI)w = v. Similar matrices Same Jordan form. Nilpotent Nⁿ = 0.
Frequently asked questions
What is Jordan normal form? The closest to diagonal form for a defective matrix.
What is a Jordan block? (λ 1; 0 λ).
What is Jⁿ? (λⁿ nλⁿ⁻¹; 0 λⁿ).
When is a matrix diagonalisable? When it has a full set of eigenvectors.
Which page covers diagonalisation? diagonalisation.
Which page covers the matrix exponential? matrix exponential.