Eigenspaces and Diagonal Matrices
Diagonal Matrices
Definition: diagonal matrix
A diagonal matrix is a square matrix that is 0 everywhere except
possibly along the diagonal.
Diagonal Matrices
Definition: diagonal matrix
A diagonal matrix is a square matrix that is 0 everywhere except
possibly along the diagonal.
8 0 0
Example: 0 5 0
0 0 5
is a diagonal matrix.
Diagonal Matrices
Definition: diagonal matrix
A diagonal matrix is a square matrix that is 0 everywhere except
possibly along the diagonal.
8 0 0
Example: 0 5 0
0 0 5
is a diagonal matrix.
Every diagonal matrix
is upper triangular.
Diagonal Matrices
Definition: diagonal matrix
A diagonal matrix is a square matrix that is 0 everywhere except
possibly along the diagonal.
8 0 0
If an operator has a diagonal
Example: 0 5 0
matrix with respect to some
0 0 5
basis, then the entries along the
is a diagonal matrix.
diagonal are precisely the
Every diagonal matrix eigenvalues of the operator.
is upper triangular.
Eigenspaces
Definition: eigenspace, E(λ, T)
Suppose T ∈ L(V) and λ ∈ F. The
eigenspace of T corresponding to λ,
denoted E(λ, T), is defined by
E(λ, T) = null(T − λI).
Eigenspaces
Definition: eigenspace, E(λ, T)
Suppose T ∈ L(V) and λ ∈ F. The
eigenspace of T corresponding to λ,
denoted E(λ, T), is defined by
E(λ, T) = null(T − λI).
v ∈ null(T − λI) ⇐⇒ Tv = λv.
Thus E(λ, T) is the set of eigenvectors of T
corresponding to λ, along with the 0 vector.
Eigenspaces
Definition: eigenspace, E(λ, T)
Suppose T ∈ L(V) and λ ∈ F. The
eigenspace of T corresponding to λ,
denoted E(λ, T), is defined by
E(λ, T) = null(T − λI).
v ∈ null(T − λI) ⇐⇒ Tv = λv.
Thus E(λ, T) is the set of eigenvectors of T
corresponding to λ, along with the 0 vector.
For T ∈ L(V) and λ ∈ F, the
eigenspace E(λ, T) is a subspace of V.
Eigenspaces
Definition: eigenspace, E(λ, T)
Suppose T ∈ L(V) and λ ∈ F. The
eigenspace of T corresponding to λ,
denoted E(λ, T), is defined by
E(λ, T) = null(T − λI).
v ∈ null(T − λI) ⇐⇒ Tv = λv.
Thus E(λ, T) is the set of eigenvectors of T
corresponding to λ, along with the 0 vector.
For T ∈ L(V) and λ ∈ F, the
eigenspace E(λ, T) is a subspace of V.
λ is an eigenvalue of T if
and only if E(λ, T) 6= {0}.
Eigenspaces
Definition: eigenspace, E(λ, T)
Example: Suppose the matrix of an
Suppose T ∈ L(V) and λ ∈ F. The operator T ∈ L(V) with respect to a
eigenspace of T corresponding to λ, basis v1 , v2 , v3 of V is the matrix
denoted E(λ, T), is defined by
8 0 0
E(λ, T) = null(T − λI). 0 5 0 .
0 0 5
v ∈ null(T − λI) ⇐⇒ Tv = λv.
Then
Thus E(λ, T) is the set of eigenvectors of T
corresponding to λ, along with the 0 vector. E(8, T) = span(v1 ), E(5, T) = span(v2 , v3 ).
For T ∈ L(V) and λ ∈ F, the
eigenspace E(λ, T) is a subspace of V.
λ is an eigenvalue of T if
and only if E(λ, T) 6= {0}.
Eigenspaces
Definition: eigenspace, E(λ, T)
Example: Suppose the matrix of an
Suppose T ∈ L(V) and λ ∈ F. The operator T ∈ L(V) with respect to a
eigenspace of T corresponding to λ, basis v1 , v2 , v3 of V is the matrix
denoted E(λ, T), is defined by
8 0 0
E(λ, T) = null(T − λI). 0 5 0 .
0 0 5
v ∈ null(T − λI) ⇐⇒ Tv = λv.
Then
Thus E(λ, T) is the set of eigenvectors of T
corresponding to λ, along with the 0 vector. E(8, T) = span(v1 ), E(5, T) = span(v2 , v3 ).
For T ∈ L(V) and λ ∈ F, the
eigenspace E(λ, T) is a subspace of V. If λ is an eigenvalue of an operator
T ∈ L(V), then T restricted to E(λ, T)
λ is an eigenvalue of T if is the operator of multiplication by λ.
and only if E(λ, T) 6= {0}.
Sum of Eigenspaces
Sum of eigenspaces is a direct sum
Suppose V is finite-dimensional and T ∈ L(V). Suppose also that
λ1 , . . . , λm are distinct eigenvalues of T. Then
E(λ1 , T) + · · · + E(λm , T)
is a direct sum. Furthermore,
dim E(λ1 , T) + · · · + dim E(λm , T) ≤ dim V.
Sum of Eigenspaces
Sum of eigenspaces is a direct sum
Suppose V is finite-dimensional and T ∈ L(V). Suppose also that
λ1 , . . . , λm are distinct eigenvalues of T. Then
E(λ1 , T) + · · · + E(λm , T)
is a direct sum. Furthermore,
dim E(λ1 , T) + · · · + dim E(λm , T) ≤ dim V.
Proof To show that E(λ1 , T)+· · ·+E(λm , T)
is a direct sum, suppose
u1 + · · · + um = 0,
where each uj is in E(λj , T).
Sum of Eigenspaces
Sum of eigenspaces is a direct sum
Suppose V is finite-dimensional and T ∈ L(V). Suppose also that
λ1 , . . . , λm are distinct eigenvalues of T. Then
E(λ1 , T) + · · · + E(λm , T)
is a direct sum. Furthermore,
dim E(λ1 , T) + · · · + dim E(λm , T) ≤ dim V.
Proof To show that E(λ1 , T)+· · ·+E(λm , T) Because eigenvectors corresponding to
is a direct sum, suppose distinct eigenvalues are linearly indepen-
u1 + · · · + um = 0, dent, this implies that each uj equals 0.
where each uj is in E(λj , T).
Sum of Eigenspaces
Sum of eigenspaces is a direct sum
Suppose V is finite-dimensional and T ∈ L(V). Suppose also that
λ1 , . . . , λm are distinct eigenvalues of T. Then
E(λ1 , T) + · · · + E(λm , T)
is a direct sum. Furthermore,
dim E(λ1 , T) + · · · + dim E(λm , T) ≤ dim V.
Proof To show that E(λ1 , T)+· · ·+E(λm , T) Because eigenvectors corresponding to
is a direct sum, suppose distinct eigenvalues are linearly indepen-
u1 + · · · + um = 0, dent, this implies that each uj equals 0.
This implies that E(λ1 , T) + · · · + E(λm , T)
where each uj is in E(λj , T). is a direct sum, as desired.
Diagonalizable Operators
Definition: diagonalizable
An operator T ∈ L(V) is called diagonalizable if the operator has a
diagonal matrix with respect to some basis of V.
Diagonalizable Operators
Definition: diagonalizable
An operator T ∈ L(V) is called diagonalizable if the operator has a
diagonal matrix with respect to some basis of V.
Example: Define T ∈ L(R2 ) by
T(x, y) = (41x + 7y, −20x + 74y).
Diagonalizable Operators
Definition: diagonalizable
An operator T ∈ L(V) is called diagonalizable if the operator has a
diagonal matrix with respect to some basis of V.
Example: Define T ∈ L(R2 ) by
T(x, y) = (41x + 7y, −20x + 74y).
The matrix of T with respect to the stan-
dard basis of R2 is
41 7
,
−20 74
which is not a diagonal matrix.
Diagonalizable Operators
Definition: diagonalizable
An operator T ∈ L(V) is called diagonalizable if the operator has a
diagonal matrix with respect to some basis of V.
Example: Define T ∈ L(R2 ) by However, T is diagonalizable, because
T(x, y) = (41x + 7y, −20x + 74y). the matrix of T with respect to the ba-
sis (1, 4), (7, 5) is
The matrix of T with respect to the stan-
69 0
dard basis of R2 is ,
0 46
41 7
,
−20 74 as you should verify.
which is not a diagonal matrix.
Diagonalizability
Conditions equivalent to diagonalizability
Suppose V is finite-dimensional and T ∈ L(V). Let λ1 , . . . , λm de-
note the distinct eigenvalues of T. Then the following are equivalent:
T is diagonalizable;
Diagonalizability
Conditions equivalent to diagonalizability
Suppose V is finite-dimensional and T ∈ L(V). Let λ1 , . . . , λm de-
note the distinct eigenvalues of T. Then the following are equivalent:
T is diagonalizable;
V has a basis consisting of eigenvectors of T;
Diagonalizability
Conditions equivalent to diagonalizability
Suppose V is finite-dimensional and T ∈ L(V). Let λ1 , . . . , λm de-
note the distinct eigenvalues of T. Then the following are equivalent:
T is diagonalizable;
V has a basis consisting of eigenvectors of T;
there exist 1-dimensional subspaces U1 , . . . , Un of V, each
invariant under T, such that
V = U1 ⊕ · · · ⊕ Un ;
Diagonalizability
Conditions equivalent to diagonalizability
Suppose V is finite-dimensional and T ∈ L(V). Let λ1 , . . . , λm de-
note the distinct eigenvalues of T. Then the following are equivalent:
T is diagonalizable;
V has a basis consisting of eigenvectors of T;
there exist 1-dimensional subspaces U1 , . . . , Un of V, each
invariant under T, such that
V = U1 ⊕ · · · ⊕ Un ;
V = E(λ1 , T) ⊕ · · · ⊕ E(λm , T);
Diagonalizability
Conditions equivalent to diagonalizability
Suppose V is finite-dimensional and T ∈ L(V). Let λ1 , . . . , λm de-
note the distinct eigenvalues of T. Then the following are equivalent:
T is diagonalizable;
V has a basis consisting of eigenvectors of T;
there exist 1-dimensional subspaces U1 , . . . , Un of V, each
invariant under T, such that
V = U1 ⊕ · · · ⊕ Un ;
V = E(λ1 , T) ⊕ · · · ⊕ E(λm , T);
dim V = dim E(λ1 , T) + · · · + dim E(λm , T).
Not Every Operator is Diagonalizable
Define T ∈ L(C2 ) defined by
T(w, z) = (z, 0).
Not Every Operator is Diagonalizable
Define T ∈ L(C2 ) defined by
T(w, z) = (z, 0).
Then 0 is the only eigenvalue of T and furthermore
E(0, T) = {(w, 0) ∈ C2 : w ∈ C}.
Not Every Operator is Diagonalizable
Define T ∈ L(C2 ) defined by
T(w, z) = (z, 0).
Then 0 is the only eigenvalue of T and furthermore
E(0, T) = {(w, 0) ∈ C2 : w ∈ C}.
Thus T is not diagonalizable.
Enough Eigenvalues Implies Diagonalizability
Enough eigenvalues implies diagonalizability
If T ∈ L(V) has dim V distinct eigenvalues, then T is diagonalizable.
Enough Eigenvalues Implies Diagonalizability
Enough eigenvalues implies diagonalizability
If T ∈ L(V) has dim V distinct eigenvalues, then T is diagonalizable.
Proof Suppose T ∈ L(V) has distinct
eigenvalues λ1 , . . . , λn , where
n = dim V.
Enough Eigenvalues Implies Diagonalizability
Enough eigenvalues implies diagonalizability
If T ∈ L(V) has dim V distinct eigenvalues, then T is diagonalizable.
Proof Suppose T ∈ L(V) has distinct
eigenvalues λ1 , . . . , λn , where
n = dim V.
For each j, let vj ∈ V be an
eigenvector corresponding to the
eigenvalue λj .
Enough Eigenvalues Implies Diagonalizability
Enough eigenvalues implies diagonalizability
If T ∈ L(V) has dim V distinct eigenvalues, then T is diagonalizable.
Proof Suppose T ∈ L(V) has distinct
eigenvalues λ1 , . . . , λn , where
n = dim V.
For each j, let vj ∈ V be an
eigenvector corresponding to the
eigenvalue λj .
Because eigenvectors corresponding
to distinct eigenvalues are linearly
independent, v1 , . . . , vn is linearly
independent.
Enough Eigenvalues Implies Diagonalizability
Enough eigenvalues implies diagonalizability
If T ∈ L(V) has dim V distinct eigenvalues, then T is diagonalizable.
Proof Suppose T ∈ L(V) has distinct A linearly independent list of n vectors
eigenvalues λ1 , . . . , λn , where in V is a basis of V; thus
n = dim V. v1 , . . . , vn
For each j, let vj ∈ V be an
eigenvector corresponding to the is a basis of V.
eigenvalue λj .
Because eigenvectors corresponding
to distinct eigenvalues are linearly
independent, v1 , . . . , vn is linearly
independent.
Enough Eigenvalues Implies Diagonalizability
Enough eigenvalues implies diagonalizability
If T ∈ L(V) has dim V distinct eigenvalues, then T is diagonalizable.
Proof Suppose T ∈ L(V) has distinct A linearly independent list of n vectors
eigenvalues λ1 , . . . , λn , where in V is a basis of V; thus
n = dim V. v1 , . . . , vn
For each j, let vj ∈ V be an
eigenvector corresponding to the is a basis of V.
eigenvalue λj . With respect to this basis consisting
Because eigenvectors corresponding of eigenvectors, T has a diagonal
to distinct eigenvalues are linearly matrix.
independent, v1 , . . . , vn is linearly
independent.
Fibonacci [Leonardo of Pisa] (1170 – 1250)
Linear Algebra Done Right, by Sheldon Axler