Diagonalization of Normal Operators
April 18, 2007
Here is a proof of the diagonalizability of normal operators which does
not use Schur’s theorem.
Lemma
Let
T
:
V
→
W
be a linear operator. Assume that its adjoint
T
*
exists. Then
Ker
(
T
) =
Im
(
T
*
)
⊥
. If
V
is ﬁnite dimensional, then
Ker
(
T
)
⊥
=
Im
(
T
*
).
Proof:
Recall that an element
w
of
W
is zero if and only if (
w

w
0
) = 0
for all
w
0
∈
W
. Thus
x
∈
Ker
(
T
) if and only if (
Tx

y
) = 0 for all
y
∈
W
,
i.e.
, if and only if (
x

T
*
y
) = 0 for all
y
∈
W
,
i.e.
, if and only if
x
is
orthogonal to the image of
T
*
. This proves the ﬁrst statement. It follows
that (
Ker
(
T
))
⊥
= (
Im
(
T
*
)
⊥
)
⊥
, and the latter is
Im
(
T
*
) when
V
is ﬁnite
dimensional.
Lemma
Let
A
and
B
be linear operators on
V
. If
A
and
B
commute,
then
Ker
(
A
) and
Im
(
A
) are
B
invariant.
Proof:
If
x
∈
Ker
(
A
), then
AB
(
x
) =
BA
(
x
) = 0, so
B
(
x
)
∈
Ker
(
A
).
Thus
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
This is the end of the preview. Sign up
to
access the rest of the document.
 Spring '08
 GUREVITCH
 Linear Algebra, Ker, Normal Operators

Click to edit the document details