1.

Banach Spaces
Denition 1.1. A (real) complex normed space is a (real) complex vector
space X together with a map : X R, called the norm and denoted | |,
such that
(i) |x| 0, for all x X, and |x| = 0 if and only if x = 0.
(ii) |x| = [[|x|, for all x X and all C (or R).
(iii) |x +y| |x| +|y|, for all x, y X.
Remark 1.2. If in (i) we only require that |x| 0, for all x X, then
| | is called a seminorm.
Remark 1.3. If X is a normed space with norm | |, it is readily checked
that the formula d(x, y) = |x y|, for x, y X, denes a metric d on
X. Thus a normed space is naturally a metric space and all metric space
concepts are meaningful. For example, convergence of sequences in X means
convergence with respect to the above metric.
Denition 1.4. A complete normed space is called a Banach space.
Thus, a normed space X is a Banach space if every Cauchy sequence in
X converges (where X is given the metric space structure as outlined above).
One may consider real or complex Banach spaces depending, of course, on
whether X is a real or complex linear space.
Examples 1.5.
1. If R is equipped with the norm || = [[, R, then it becomes a real
normed space. More generally, for x = (x
1
, x
2
, . . . , x
n
) R
n
dene
|x| =
_
n

i=1
[x
i
[
2
_
1/2
Then R
n
becomes a real Banach space (with the obvious component-wise
linear structure).
1
1.2 Functional Analysis Gently Done Mathematics Department
In a similar way, one sees that C
n
, equipped with the similar norm, is a
(complex) Banach space.
2. Equip (([0, 1]), the linear space of continuous complex-valued functions
on the interval [0, 1], with the norm
|f| = sup[f(x)[ : x [0, 1].
Then (([0, 1]) becomes a Banach space. This norm is called the supremum
(or uniform) norm and is often denoted | |

. Notice that convergence with

respect to this norm is precisely that of uniform convergence of the functions
on [0, 1].
Suppose that we now equip (([0, 1]) with the norm
|f|
1
=
_
1
0
[f(x)[ dx.
One can check that this is indeed a norm but (([0, 1]) is no longer complete
(so is not a Banach space). In fact, if h
n
is the function given by
h
n
(x) =
_

_
0, 0 x
1
2
n(x
1
2
),
1
2
< x
1
2
+
1
n
1,
1
2
+
1
n
< x 1
then one sees that (h
n
) is a Cauchy sequence with respect to the norm | |
1
.
Suppose that h
n
h in ((([0, 1]), | |
1
) as n . Then
_
1/2
0
[h(x)[ dx =
_
1/2
0
[h(x) h
n
(x)[ dx |h h
n
|
1
0
and so we see that h vanishes on the interval [0,
1
2
]. Similarly, for any 0 <
<
1
2
, we have
_
1
1
2
+
[h(x) 1[ dx |h h
n
|
1
0
as n . Therefore h is equal to 1 on any interval of the form [
1
2
+ , 1],
0 < <
1
2
0. This means that h is equal to 1 on the interval [
1
2
, 1]. But such
a function h is not continuous, so we conclude that (([0, 1]) is not complete
with respect to the norm | |
1
.
3. Let S be any (non-empty) set and let X denote the set of bounded
complex-valued functions on S. Then X is a Banach space when equipped
Kings College London Banach Spaces 1.3
with the supremum norm |f| = sup[f(s)[ : s S (and the usual pointwise
linear structure ).
In particular, if we take S = N, then X is the linear space of bounded
complex sequences. This Banach space is denoted

(or sometimes

(N)).
With S = Z, the resulting Banach space is denoted

(Z).
4. The set of complex sequences, x = (x
n
), satisfying
|x|
1
=

n=1
[x
n
[ <
is a linear space under componentwise operations (and | |
1
is a norm).
Moreover, one can check that the resulting normed space is complete. This
Banach space is denoted
1
.
5. The Banach space
2
is the linear space of complex sequences, x = (x
n
),
satisfying
|x|
2
= (

n=1
[x
n
[
2
)
1/2
< .
In fact,
2
has the inner product
< x, y >=

n=1
x
n
y
n
and so is a (complex) Hilbert space.
The following suggests that there is not a great deal of excitement to be
got from nite-dimensional normed spaces.
Suppose that X is a nite-dimensional normed vector space with basis
e
1
, . . . , e
n
. Dene a map T : C
n
R by
T(
1
, . . . ,
n
) = |
1
e
1
+ +
n
e
n
| .
Then the inequality

|
1
e
1
+ +
n
e
n
||
1
e
1
+ +
n
e
n
|

|(
1

1
)e
1
+ +(
n

n
)e
n
|
shows that T is continuous on C
n
. Now, T(
1
, . . . ,
n
) = 0 only if every

i
= 0. In particular, T does not vanish on the unit sphere, z : |z| = 1, in
C
n
. By compactness, T attains its bounds on the unit sphere and is therefore
strictly positive on this sphere. Hence there is m > 0 and M > 0 such that
m
_

n
k=1
[
k
[
2
|
1
e
1
+ +
n
e
n
| M
_

n
k=1
[
k
[
2
.
1.4 Functional Analysis Gently Done Mathematics Department
We have shown that the norm | | on X is equivalent to the usual Euclidean
norm on X determined by any particular basis. Consequently, any nite-
dimensional linear space can be given a norm, and, moreover, all norms on a
nite-dimensional linear space X are equivalent: for any pair of norms | |
1
and | |
2
there are positive constants ,

such that
|x|
1
|x|
2

|x|
1
for every x X. (We say that two norms | | and [[[ [[[ on X are equivalent if
there are strictly positive constants m and M such that m|x| [[[x[[[ M|x|
for every x X. This means that the norms induce the same open sets in
X and equivalently (to anticipate results from the next section) that the
identity map is continuous both from (X, | |) to (X, [[[ [[[) and also from
(X, [[[ [[[) to (X, | |).
Proposition 1.6. Suppose that X is a normed space. Then X is complete
if and only if the series

n=1
x
n
converges, where (x
n
) is any sequence in X
satisfying

n=1
|x
n
| < .
Proof. Suppose that X is complete, and let (x
n
) be any sequence in X such
that

n=1
|x
n
| < . Let > 0 be given and put y
n
=

n
k=1
x
k
. Then, for
n > m,
|y
m
y
n
| =
_
_
_
_
n

k=m+1
x
k
_
_
_
_

k=m+1
|x
k
|
<
for all suciently large m and n, since

n=1
|x
n
| < . Hence (y
n
) is a
Cauchy sequence and so converges since X is complete, by hypothesis.
Conversely, suppose

n=1
x
n
converges in X whenever

n=1
|x
n
| < .
Let (y
n
) be any Cauchy sequence in X. We must show that (y
n
) converges.
Now, since (y
n
) is Cauchy, there is n
1
N such that |y
n
1
y
m
| <
1
2
whenever
m > n
1
. Furthermore, there is n
2
> n
1
such that |y
n
2
y
m
| <
1
4
whenever
m > n
2
. Continuing in this way, we see that there is n
1
< n
2
< n
3
< . . .
such that |y
n
k
y
m
| <
1
2
k
whenever m > n
k
. In particular, we have
|y
n
k+1
y
n
k
| <
1
2
k
Kings College London Banach Spaces 1.5
for k N. Set x
k
= y
n
k+1
y
n
k
. Then
n

k=1
|x
k
| =
n

k=1
|y
n
k+1
y
n
k
|
<
n

k=1
1
2
k
.
It follows that

k=1
|x
k
| < . By hypothesis, there is x X such that

m
k=1
x
k
x as m , that is,
m

k=1
x
k
=
m

k=1
(y
n
k+1
y
n
k
)
= y
n
m+1
y
n
1
x.
Hence y
n
m
x +y
n
1
in X as m . Thus the Cauchy sequence (y
n
) has
a convergent subsequence and so must itself converge.
We shall apply this result to quotient spaces, to which we now turn. Let
X be a vector space, and let M be a vector subspace of X. We dene an
equivalence relation on X by x y if and only if x y M. It is
straightforward to check that this really is an equivalence relation on X. For
x X, let [x] denote the equivalence class containing the element x and
denote the set of equivalence classes by X/M. The denitions [x] + [y] =
[x + y] and [x] = [x], for C and x, y X, make X/M into a linear
space. (These denitions are meaningful since M is a linear subspace of
X. For example, if x x

and y y

, then x + y x

+ y

, so that the
denition is independent of the particular representatives taken from the
various equivalence classes.) We consider the possibility of dening a norm
on the quotient space X/M. Set
|[x]| = inf|y| : y [x].
Note that if y [x], then y x so that y x M; that is, y = x + m for
some m M. Hence
|[x]| = inf|y| : y [x] = inf|x +m| : m M
= inf|x m| : m M , since M is a subspace,
and this is the distance between x and M in the usual metric space sense.
The zero element of X/M is [0] = M, and so |[x]| is the distance between x
1.6 Functional Analysis Gently Done Mathematics Department
and the zero in X/M. Now, in a normed space it is certainly true that the
norm of an element is just the distance between it and zero; |z| = |z 0|.
So the denition of |[x]| above is perhaps a reasonable choice.
To see whether this does give a norm or not we shall consider the various
requirements. First, suppose that ,= 0, and consider
|[x]| = |[x]|
= inf
mM
|x +m|
= inf
mM
|x +m| , since ,= 0,
= [[ inf
mM
|x +m|
= [[ |[x]|.
If = 0, this equality remains valid because [0] = M and inf
mM
|m| = 0.
Next, we consider the triangle inequality;
|[x] + [y]| = |[x +y]|
= inf
mM
|x +y +m|
= inf
m,m

M
|x +m+y +m

|
inf
m,m

M
(|x +m| +|y +m

|)
= |x| +|y|.
Clearly, |[x]| 0 and, as noted already, |[0]| = 0, so | | is a seminorm
on the quotient space X/M. To see whether or not it is a norm, all that
remains is the investigation of the implication of the equality |[x]| = 0.
Does this imply that [x] = 0 in X/M? We will see that the answer is no, in
general, but yes if M is closed, as the following argument shows.
Proposition 1.7. Suppose that M is a closed linear subspace of the
normed space X. Then | | as dened above is a norm on the quotient space
X/M called the quotient norm.
Proof. According to the discussion above, all that we need to show is that
if x X satises |[x]| = 0, then [x] = 0 in X/M; that is, x M.
So suppose that x X and that |[x]| = 0. Then inf
mM
|x + m| = 0,
and hence, for each n N, there is z
n
M such that |x + z
n
| <
1
n
. This
means that z
n
x in X as n . Since M is a closed subspace, it
follows that x M and hence [x] = 0 in X/M, as required.
Kings College London Banach Spaces 1.7
Proposition 1.8. Let M be a closed linear subspace of a normed space X
and let : X X/M be the canonical map (x) = [x], x X. Then is
continuous.
Proof. Suppose that x
n
x in X. Then
|(x
n
) (x)| = | [x
n
] [x] |
= | [x
n
x] |
= inf
mM
|x
n
x +m|
|x
n
x| , since 0 M,
0 as n .
Proposition 1.9. For any closed linear subspace M of a Banach space X,
the quotient space X/M is a Banach space under the quotient norm.
Proof. We know that X/M is a normed space, so all that remains is to
show that it is complete. We use the criterion established above. Suppose,
then, that ([x
n
]) is any sequence in X/M such that

n
|[x
n
]| < . We
show that there is [y] X/M such that

k
n=1
[x
n
] [y] as k .
For each n, |[x
n
]| = inf
mM
|x
n
+ m|, and therefore there is m
n
M
such that
|x
n
+m
n
| |[x
n
]| +
1
2
n
,
by denition of the inmum. Hence

n
|x
n
+m
n
|

n
_
|[x
n
]| +
1
2
n
_
< .
But (x
n
+ m
n
) is a sequence in the Banach space X, and so the limit
lim
k

k
n=1
(x
n
+ m
n
) exists in X. Denote this limit by y. Then we
have
1.8 Functional Analysis Gently Done Mathematics Department
|
k

n=1
[x
n
] [y]| = |
k

n=1
[x
n
y]|
= inf
mM
|
k

n=1
x
n
y +m|
|
k

n=1
x
n
y +
k

n=1
m
n
|
= |
k

n=1
(x
n
+m
n
) y|
0 , as k .
Hence

k
n=1
[x
n
] [y] as k and we conclude that X/M is complete.
Example 1.10. Let X be the linear space (([0, 1]) and let M be the subset
of X consisting of those functions which vanish at the point 0 in [0, 1]. Then
M is a linear subspace of X and so X/M is a vector space.
Dene the map : X/M C by setting ([f]) = f(0), for [f] X/M.
Clearly, is well-dened (if f g then f(0) = g(0)) and we have
([f] +[g]) = ([f +g])
= f(0) +g(0)
= ([f]) +([g])
for any , C, and f, g X. Hence : X/M C is linear. Furthermore,
([f]) = ([g]) f(0) = g(0)
f g
[f] = [g]
and so we see that is one-one.
Given any C, there is f X with f(0) = and so ([f]) = .
Thus is onto. Hence is a vector space isomorphism between X/M and
C; i.e. X/M

= C as vector spaces.
Now, it is easily seen that M is closed in X with respect to the | |

norm and so X/M is a Banach space when given the quotient norm. We
have
|[f]| = inf|g|

: g [f]
= inf|g|

: g(0) = f(0)
= [f(0)[ (take g(s) = f(0) for all s [0, 1]).
Kings College London Banach Spaces 1.9
That is, |[f]| = [([f])[, for [f] X/M, and so preserves the norm. Hence
X/M

= C as Banach spaces.
Now consider X equipped with the norm | |
1
. Then M is no longer
closed in X. We can see this by considering, for example, the sequence (g
n
)
given by
g
n
(s) =
_
ns, 0 s 1/n
1, 1/n < s 1.
Then g
n
M, for each n N, and g
n
1 with respect to | |
1
, but 1 / M.
The quotient norm is not a norm in this case. Indeed, inf|g|
1
: g
[f] = 0 for all [f] X/M. To see this, let f X, and, for n N, set
h
n
(s) = f(0)(1 g
n
(s)), with g
n
dened as above. Then h
n
(0) = f(0) and
|h
n
|
1
= [f(0)[/2n. Hence
inf|g|
1
: g(0) = f(0) |h
n
|
1
[f(0)[/2n
which implies that
inf|g|
1
: g [f] = 0.
The quotient norm on X/M assigns norm zero to all vectors.
2. Linear Operators
Denition 2.1. A linear operator T between normed spaces X and Y is a
map T : X Y such that
T(x +x

) = Tx +Tx

for all , C (or R in the real case) and all x, x

X.
Denition 2.2. The linear operator T : X Y is said to be bounded if
there is some k > 0 such that
|Tx| k|x|
for all x X. If T is bounded, we dene |T| to be
|T| = infk : |Tx| k|x|, x X .
We will see shortly that | | really is a norm on the set of bounded linear
operators from X into Y . The following result is a direct consequence of the
denitions.
Proposition 2.3. Suppose that T : X Y is a bounded linear operator.
Then we have
|T| = sup|Tx| : |x| 1
= sup|Tx| : |x| = 1
= sup
|Tx|
|x|
: x ,= 0 .
Proof. This follows using the fact that if |x| 1 (and x ,= 0), then we
have |Tx| |Tx|/|x| = |Tx/|x||.
1
2.2 Functional Analysis Gently Done Mathematics Department
Note that if T is bounded, then, by the very denition of |T|, we have
|Tx| |T||x|, for any x X. Thus, a bounded linear operator maps any
bounded set in X into a bounded set in Y . In particular, the unit ball in X
is mapped into (but not necessarily onto) the ball of radius |T| in Y . This
will be used repeatedly without further comment.
The next result is a basic consequence of the linear structure.
Theorem 2.4. Given a linear operator T : X Y , for normed spaces X
and Y , the following three statements are equivalent.
(i) T is continuous at some point in X.
(ii) T is continuous at every point in X.
(iii) T is bounded on X.
Proof. Clearly (ii) implies (i). We shall show that (i) implies (ii). Suppose
that T is continuous at x
0
X. Let x X and let > 0 be given. Then
there is > 0 such that if |z x
0
| < then |Tz Tx
0
| < . But then we
have
|Tx

Tx| = |Tx

Tx +Tx
0
Tx
0
|
= |T(x

x +x
0
) Tx
0
| , by the linearity of T,
<
whenever |(x

x + x
0
) x
0
| < , i.e., |x

x| < . This shows that T is

continuous at any x X.
(Alternatively, one could argue as follows. Suppose that (x
n
) is a sequence in
X such that x
n
x. Then x
n
x +x
0
x
0
and so T(x
n
x +x
0
) Tx
0
,
since T is assumed to be continuous at x
0
. Thus Tx
n
Tx + Tx
0
Tx
0
,
since T is linear. In other words Tx
n
Tx 0, or Tx
n
Tx.)
Next we show that (iii) implies (ii). Let > 0 be given. By (iii), there is
k > 0 such that
|Tx

Tx| = |T(x

x)| k|x

x|
for any x, x

X. Putting = /k, we conclude that if |x

x| < , then
|Tx

Tx| < . Thus T is continuous at x X. In fact, this estimate

shows that T is uniformly continuous on X, and that therefore, continuity
and uniform continuity are equivalent in this context. In other words, the
notion of uniform continuity can play no special role in the theory of linear
operators, as it does, for example, in classical real analysis.
Finally, we show that (ii) implies (iii). If T is assumed to be continuous at
every point of X, then, in particular, it is continuous at 0. Hence, for given
Kings College London Linear Operators 2.3
> 0, there is > 0 such that |Tx| < whenever |x 0| = |x| < . Now,
for any x ,= 0, z = x/2|x| has norm equal to /2 < . Hence |Tz| < .
But |Tz| = |Tx|/2|x| and so we get the inequality
|Tx| <
2|x|

which is valid for any x X with x ,= 0. Therefore

|Tx|
2|x|

holds for any x X, and we conclude that T is bounded.

(Alternatively, suppose that T is continuous at 0, but is not bounded. Then
for each n N there is x
n
X such that |Tx
n
| > n|x
n
|. Evidently x
n
,= 0.
Put z
n
= x
n
/n|x
n
|. Then |z
n
| = 1/n 0, and so z
n
0 in X. However,
|Tz
n
| = |(Tx
n
/n|x
n
|)| = |Tx
n
|/n|x
n
| > 1 for all n N, and so (Tz
n
)
does not converge to 0. This contradicts the assumed continuity of T at 0.)
Remark 2.5. To establish the continuity of a given linear operator, it is
enough to show continuity at 0. However, it is often (marginally) easier to
check boundedness than continuity.
Denition 2.6. The set of bounded linear operators from a normed space
X into a normed space Y is denoted B(X, Y ). If X = Y , one simply writes
B(X) for B(X, X).
Proposition 2.7. The space B(X, Y ) is a normed space when equipped
with its natural linear structure and the norm | |.
Proof. For S, T B(X, Y ) and any , C, the linear operator S +T
is dened by (S + T)x = Sx + Tx for x X. Furthermore, for any
x X,
|Sx +Tx| |Sx| +|Tx| (|S| +|T|)|x|
and we see that B(X, Y ) is a linear space. To see that | | is a norm on
B(X, Y ), note rst that |T| 0 and |T| = 0 if T = 0. On the other hand,
if |T| = 0, then
0 = |T| = sup
|Tx|
|x|
: x ,= 0
which implies that |Tx| = 0 for every x X (including, trivially, x = 0).
That is, T = 0.
2.4 Functional Analysis Gently Done Mathematics Department
Now let C and T B(X, Y ). Then
|T| = sup|Tx| : |x| 1
= sup[[ |Tx| : |x| 1
= [[ sup|Tx| : |x| 1
= [[ |T| .
Finally, we see from the above, that for any S, T B(X, Y ),
|S +T| = sup|Sx +Tx| : |x| 1
sup(|S| +|T|)|x| : |x| 1
= |S| +|T|
and the proof is complete.
Proposition 2.8. Suppose that X is a normed space and Y is a Banach
space. Then B(X, Y ) is a Banach space.
Proof. All that needs to be shown is that B(X, Y ) is complete. To this end,
let (A
n
) be a Cauchy sequence in B(X, Y ); then
|A
n
A
m
| = sup
_
|A
n
x A
m
x|
|x|
: x ,= 0
_
0 ,
as n, m . It follows that for any given x X, |A
n
x A
m
x| 0, as
n, m , i.e., (A
n
x) is a Cauchy sequence in the Banach space Y . Hence
there is some y Y such that A
n
x y in Y . Set Ax = y. We have
A(x +x

) = lim
n
A
n
(x +x

)
= lim
n
(A
n
x +A
n
x

)
= lim
n
A
n
x + lim
n
A
n
x

= Ax +Ax

,
for any x, x

X and C. It follows that A : X Y is a linear operator.

Next we shall check that A is bounded. To see this, we observe rst that for
suciently large m, n N, and any x X,
|A
n
x A
m
x| |A
n
A
m
||x| |x|
Kings College London Linear Operators 2.5
Taking the limit n gives the inequality |AxA
m
x| |x|. Hence, for
any suciently large m,
|Ax| |Ax A
m
x| +|A
m
x|
|x| +|A
m
||x|
and we deduce that |A| 1 +|A
m
|. Thus A B(X, Y ).
We must now show that, indeed, A
n
A with respect to the norm in
B(X, Y ). Let > 0 be given. Then there is N N such that
|A
n
x A
m
x| |A
n
A
m
||x| |x|
for any m, n > N and for any x X. Taking the limit n , as before,
we obtain
|Ax A
m
x| |x|
for any m > N and any x X. Taking the supremum over x X with
|x| 1 yields |A A
m
| for all m > N. In other words, A
m
A in
B(X, Y ) and the proof is complete.
Remark 2.9. Note that
[ |A| |A
m
| [ |AA
m
| 0
so that (|A
m
|) converges to |A|.
Remark 2.10. If S and T belong to B(X), then ST : X X is dened
by STx = S(Tx), for any x X. Clearly ST is a linear operator. Also,
|STx| |S||Tx| |S||T||x|, which implies that ST is bounded and
|ST| |S||T|. Thus B(X) is an example of an algebra with unit (the
unit is the bounded linear operator 1lx = x, x X). If X is complete, then
so is B(X). In this case B(X) is an example of a Banach algebra.
Examples 2.11.
1. Let A = (a
ij
) be any n n complex matrix. Then the map x Ax,
x C
n
, is a linear operator on C
n
. Clearly, this map is continuous (where C
n
is equipped with the usual Euclidean norm), and so therefore it is bounded.
By slight abuse of notation, let us also denote by A this map, x Ax.
To nd |A|, we note that the matrix A

A is self adjoint and positive,

and so there exists a unitary matrix V such that V A

AV
1
is diagonal:
V A

AV
1
=
_
_
_
_

1
0 . . . 0
0
2
. . . 0
.
.
.
.
.
.
.
.
.
.
.
.
0 . . .
n
_
_
_
_
2.6 Functional Analysis Gently Done Mathematics Department
where each
i
0, and we may assume that
1

2
. . .
n
. Now, we
have
|A|
2
= sup|Ax| : |x| = 1
2
= sup|Ax|
2
: |x| = 1
= sup(A

Ax, x) : |x| = 1
= sup(V A

AV
1
x, x) : |x| = 1
= sup
_
n
k=1

k
[x
k
[
2
:

n
k=1
[x
k
[
2
= 1
_
=
1
.
It follows that |A| =
1
, the largest eigenvalue of A

A.
2. Let K : [0, 1] [0, 1] C be a given continuous function on the unit
square. For f (([0, 1]), set
(Tf)(s) =
_
1
0
K(s, t)f(t) dt .
Evidently, T is a linear operator T : (([0, 1]) (([0, 1]). Setting M =
sup[K(s, t)[ : (s, t) [0, 1] [0, 1], we see that
[Tf(s)[
_
1
0
[K(s, t)[ [f(t)[ dt
M
_
1
0
[f(t)[ dt .
Thus, |Tf|
1
M|f|
1
, so that T is a bounded linear operator on the space
((([0, 1]), | |
1
).
3. With T dened as in example 2, above, it is straightforward to check that
|Tf|

M|f|
1
and that
|Tf|
1
M|f|

so we conclude that T is a bounded linear operator from ((([0, 1]), | |

1
) to
((([0, 1]), | |

) and also from ((([0, 1]), | |

) to ((([0, 1]), | |
1
).
4. Take X =
1
, and, for any x = (x
n
) X, dene Tx to be the sequence
Tx = (x
2
, x
3
, x
4
, . . . ). Then Tx X and satises |Tx|
1
|x|
1
. Thus T
is a bounded linear operator from
1

1
, with |T| 1. In fact, |T| = 1
(take x = (0, 1, 0, 0, . . . )). T is called the left shift on
1
.
Kings College London Linear Operators 2.7
Similarly, one sees that T :

is a bounded linear operator, with

|T| = 1.
5. Take X =
1
, and, for any x = (x
n
) X, dene Sx to be the sequence
Sx = (0, x
1
, x
2
, x
3
, . . . ). Clearly, |Sx|
1
= |x|
1
, and so S is a bounded linear
operator from
1

1
, with |S| = 1. S is called the right shift on
1
.
As above, S also denes a bounded linear operator from

to

, with
norm 1.
Theorem 2.12. Suppose that X is a normed space and Y is a Banach
space, and suppose that T : X Y is a linear operator dened on some
dense linear subset D(T) of X. Then if T is bounded (as a linear operator
from the normed space D(T) to Y ) it has a unique extension to a bounded
linear operator from all of X into Y . Moreover, this extension has the same
norm as T.
Proof. By hypothesis, |Tx| |T||x|, for all x D(T), where |T| is the
norm of T as a map D(T) Y . Let x X. Since D(T) is dense in X, there
is a sequence (
n
) in D(T) such that
n
x, in X, as n . In particular,
(
n
) is a Cauchy sequence in X. But
|T
n
T
m
| = |T(
n

m
)| |T| |
n

m
| ,
and so we see that (T
n
) is a Cauchy sequence in Y . Since Y is complete,
there exists y Y such that T
n
y in Y . We would like to construct
an extension

T of T by dening

Tx to be this limit, y. However, to be
able to do this, we must show that the element y does not depend on the
particular sequence (
n
) in D(T) converging to x. To see this, suppose that
(
n
) is any sequence in D(T) such that
n
x in X. Then, as before,
we deduce that there is y

, say, in Y , such that T

n
y

. Now consider
the combined sequence
1
,
1
,
2
,
2
, . . . in D(T). Clearly, this sequence also
converges to x and so once again, as above, we deduce that the sequence
(T
1
, T
1
, T
2
, T
2
, . . . ) converges to some z, say, in Y . But this sequence
has the two convergent subsequences (T
k
) and (T
m
), with limits y and y

,
respectively. It follows that z = y = y

. Therefore we may unambiguously

dene the map

T : X Y by the prescription

Tx = y, where y is given as
above.
Note that if x D(T), then we can take
n
D(T) above to be
n
= x
for every n N. This shows that

Tx = Tx, and hence that

T is an extension
of T. We show that

T is a bounded linear operator from X to Y .
2.8 Functional Analysis Gently Done Mathematics Department
Let x, x

X and let C be given. Then there are sequences (

n
) and
(

n
) in D(T) such that
n
x and
n

in X. Hence
n
+
n

x+x

,
and by the construction of

T, we see that

T(x +x

) = lim
n
T(
n
+
n

) , using the linearity of D(T),

= lim
n
T
n
+T
n

Tx +

Tx

.
It follows that

T is a linear map. To show that

T is bounded and has the same
norm as T, we rst observe that if x X and if (
n
) is a sequence in D(T)
such that
n
x, then, by construction, T
n


Tx, and so |T
n
| |

Tx|.
Hence, the inequalities |T
n
| |T| |
n
|, for n N, imply (by taking
the limit) that |

Tx| |T| |x|. Therefore |

T| |T|. But since

T is an
extension of T we have that
|

T| = sup|

Tx| : x X, |x| 1
sup|

Tx| : x D(T), |x| 1

= sup|Tx| : x D(T), |x| 1
= |T| .
The equality |

T| = |T| follows.
The uniqueness is immediate; if S is also a bounded linear extension of T
to the whole of X, then S

T is a bounded (equivalently, continuous) map
on X which vanishes on the dense subset D(T). Thus S

T must vanish on
the whole of X, i.e., S =

T.
Remark 2.13. This process of extending a densely-dened bounded linear
operator to one on the whole of X is often referred to as extension by
continuity. If T is densely-dened, as above, but is not bounded on (D(T))
there is no obvious way of extending T to the whole of X. Indeed, such
a gaol may not even be desirable, as we will see laterfor example, the
Hellinger-Toeplitz theorem.
3. Baire Category Theorem and all that
We shall begin this section with a result concerning fullness of complete
metric spaces.
Denition 3.1. A subset of a metric space is said to be nowhere dense if
its closure has empty interior.
Example 3.2. Consider the metric space R with the usual metric, and let
S be the set S = 1,
1
2
,
1
3
,
1
4
, . . . . Then S has closure S = 0, 1,
1
2
,
1
3
, . . .
which has empty interior.
We shall denote the open ball of radius r around the point a in a metric
space by B(a; r). The statement that a set S is nowhere dense is equivalent
to the statement that the closure, S of S, contains no open ball B(a; r) of
positive radius.
The next theorem, the Baire Category theorem, tells us that countable
unions of nowhere dense sets cannot amount to much.
Theorem 3.3. (Baire Category theorem) The complement of any countable
union of nowhere dense subsets of a complete metric space X is dense in X.
Proof. Suppose that A
n
, n N, is a countable collection of nowhere dense
sets in the complete metric space X. Set A
0
= X

nN
A
n
. We wish to
show that A
0
is dense in X. Now, X

nN
A
n
X

nN
A
n
, and a
set is nowhere dense if and only if its closure is. Hence, by taking closures
if necessary, we may assume that each A
n
, n = 1, 2, . . . is closed. Suppose
then, by way of contradiction, that A
0
is not dense in X. Then X A
0
,= .
Now, X A
0
is open, and non-empty, so there is x
0
X A
0
and r
0
> 0
such that B(x
0
; r
0
) X A
0
, that is, B(x
0
; r
0
) A
0
= . The idea of the
proof is to construct a sequence of points in X with a limit which does not
lie in any of the sets A
0
, A
1
, . . . . This will be a contradiction, since X is the
union of the A
n
s.
1
3.2 Functional Analysis Gently Done Mathematics Department
We start by noticing that since A
1
is nowhere dense, the open ball
B(x
0
; r
0
) is not contained in A
1
. This means that there is some point
x
1
B(x
0
; r
0
) A
1
. Furthermore, since B(x
0
; r
0
) A
1
is open, there is
0 < r
1
< 1 such that B(x
1
; r
1
) B(x
0
; r
0
) and also B(x
1
; r
1
) A
1
= .
Now, since A
2
is nowhere dense, the open ball B(x
1
; r
1
) is not contained in
A
2
. Thus, there is some x
2
B(x
1
; r
1
)A
2
. Since B(x
1
; r
1
)A
2
is open, there
is 0 < r
2
<
1
2
such that B(x
2
; r
2
) B(x
1
; r
1
) and also B(x
2
; r
2
) A
2
= .
Similarly, we argue that there is some point x
3
and 0 < r
3
<
1
3
such that
B(x
3
; r
3
) B(x
2
; r
2
) and also B(x
3
; r
3
) A
3
= .
Recursively, we obtain a sequence x
0
, x
1
, x
2
, . . . in X and positive real
numbers r
0
, r
1
, r
2
, . . . satisfying 0 < r
n
<
1
n
, for n N, such that
B(x
n
; r
n
) B(x
n1
; r
n1
)
and B(x
n
; r
n
) A
n
= .
For any m, n > N, both x
m
and x
n
belong to B(x
N
; r
N
), and so
d(x
m
, x
n
) d(x
m
, x
N
) +d(x
n
.x
n
)
<
1
N
+
1
N
.
Hence (x
n
) is a Cauchy sequence in X and therefore there is some x X such
that x
n
x. Since x
n
B(x
n
; r
n
) B(x
N
; r
N
), for all n > N, it follows
that x B(x
N
; r
N
). But by construction, B(x
N
; r
N
) B(x
N1
; r
N1
) and
B(x
N1
; r
N1
) A
N1
= . Hence x / A
N1
for any N. This is our
required contradiction and the result follows.
Remark 3.4. The theorem implies, in particular, that a complete metric
space cannot be given as a countable union of nowhere dense sets. In other
words, if a complete metric space is equal to a countable union of sets, then
not all of these can be nowhere dense; that is, at least one of them has a
closure with non-empty interior. Another corollary to the theorem is that if
a metric space can be expressed as a countable union of nowhere dense sets,
then it is not complete.
As a rst application of the Baire Category Theorem, we will consider the
Banach-Steinhaus theorem, also called the Principle of Uniform Boundedness
for obvious reasons.
Kings College London Baire Category Theorem and all that 3.3
Theorem 3.5. (Banach-Steinhaus) Let X be a Banach space and let
T be a family of bounded linear operators from X into a normed space Y
such that for each x X the set |Tx| : T T is bounded. Then the set
of norms |T| : T T is bounded.
Proof. For each n N, let A
n
= x : |Tx| n for all T T. Then each
A
n
is a closed subset of X. Indeed,
A
n
=

TF
x : |Tx| n =

TF
T
1
(y : |y| n)
and T
1
(y : |y| n) is closed because y : |y| n is closed in Y and
every T in T is continuous. Moreover, by hypothesis, each x X lies in
some A
n
. Thus, we may write
X =

_
n=1
A
n
.
By the Baire Category Theorem, together with the fact that each A
n
is
closed, it follows that there is some m N such that A
m
has non-empty
interior. Suppose, then, that x
0
is an interior point of A
m
, that is, there is
r > 0 such that x : |x x
0
| < r A
m
. By the denition of A
m
, we may
say that if x is such that |x x
0
| < r then |Tx| m, for every T T.
But then for any x X with |x| < r, we have |x +x
0
x
0
| < r and so
|Tx| = |T(x +x
0
) Tx
0
|
|T(x +x
0
)| +|Tx
0
|
m+m
for every T T. Hence, for any x X, with x ,= 0, we see that rx/2|x|
has norm r/2 < r and so |T(rx/2|x|)| 2m for any T T. It follows
that |Tx| 4m|x|/r, for any x X and any T T. This implies that
|T| 4m/r for any T T, i.e., |T| : T T is bounded.
The following is an application of the Uniform Boundedness Principle to
the study of the joint continuity of bilinear maps.
3.4 Functional Analysis Gently Done Mathematics Department
Proposition 3.6. Suppose that X and Y are normed spaces and that
B(, ) : X Y C is a separately continuous bilinear mapping. Then
B(, ) is jointly continuous. (That is, if B(x, ) : Y C is continuous for
each xed x X, and if B(, y) : X C is continuous for each xed y Y
then B(, ) is jointly continuous.)
Proof. For each x X, dene T
x
: Y C by T
x
(y) = B(x, y), for y Y .
By hypothesis, each T
x
is a bounded linear operator from Y into C. Since
C is complete, we may extend T
x
, by continuity, to a bounded linear map
on Y , the completion of Y . Thus we have extended B(, ) to X Y whilst
retaining the separate continuity. In other words, we may assume, without
loss of generality, that Y is complete, that is, we may assume that Y is a
Banach space.
Now, for each xed y Y , x B(x, y) is a bounded linear operator from
X into C. Hence
[B(x, y)[ C
y
|x| for all x X
for some constant C
y
0. In terms of T
x
, this becomes
[T
x
(y)[ C
y
|x| for all x X.
It follows that the family [T
x
(y)[ : x X, |x| = 1 is bounded, for each
xed y Y . By the Uniform Boundedness Principle, the set |T
x
| : x
X, |x| = 1 is bounded. That is, there is some K 0 such that
|T
x
| K for all x X with |x| = 1.
But T
x
is linear in x (T
x
= T
x
for C) and so we deduce that
|T
x
| K|x| for all x X.
Hence, for any given (x
0
, y
0
) X Y and > 0,
[B(x, y) B(x
0
, y
0
)[ = [B(x, y y
0
) +B(x, y
0
) B(x
0
, y
0
)[
[B(x, y y
0
)[ +[B(x x
0
, y
0
)[
= [T
x
(y y
0
)[ +[B(x x
0
, y
0
)[
K|x| |y y
0
| +C
y
0
|x x
0
|
<
provided |x x
0
| and |y y
0
| are suciently small.
Kings College London Baire Category Theorem and all that 3.5
The next result we will discuss is the Open Mapping Theorem, but rst
a few preliminary remarks will help to clarify things.
Let A be a subset of a normed space X. For x X, we use the notation
x +A to denote the set
x +A = z X : z = x +a, a A
and for C, A denotes the set
A = z X : z = a, a A .
Then one readily checks that
B(a; r) = a +B(0; r)
and that for any C, (with ,= 0)
B(a; r) = a +B(0; r) = B(a; [[r) .
Now suppose that X and Y are normed spaces and that T : X Y is a
linear operator. Then
T(B(a; r)) = Ta +TB(0; r) .
Furthermore, if now A X is any set, then T(A) = T(A). Also, if
x T(A), then there is a sequence (a
n
) in A such that T(a
n
) x.
Hence Ta
n
x and so x T(A). Conversely, if x T(A), there is a
sequence (a
n
) in A such that Ta
n
x/. Hence T(a
n
) x and we see
that x T(A). It follows that, for any C,
T(A) = T(A) .
We say that a subset A X is symmetric if a A implies that a A.
We say that the subset A is convex if a +(1 )a

A, for any 0 1,
whenever a and a

belong to A. If A is symmetric or convex, then the same

is true of the sets T(A) and T(A).
3.6 Functional Analysis Gently Done Mathematics Department
Proposition 3.7. Suppose that T : X Y is a bounded linear operator,
where X is a Banach space and Y is a normed space. Suppose that for some
> 0 and R > 0
B(0; ) T(B(0; R)) .
Then B(0; ) T(B(0; R)).
Proof. Let > 0 be given, and let y B(0; ). Then y T(B(0; R)), by
hypothesis, and so there is y
1
T(B(0; R)) such that |y y
1
| < . That
is, there is x
1
B(0; R) such that y
1
= Tx
1
satises |y Tx
1
| < . In
other words, y Tx
1
B(0; ). But B(0; ) T(B(0; R)) implies that
B(0; ) = B(0; ) T(B(0; R)) = T(B(0; R)) ,
i.e., y Tx
1
B(0; ) T(B(0; R)).
Hence, as before, there is a point x
2
B(0; R) such that
|(y Tx
1
) Tx
2
| <
2
.
That is, y Tx
1
Tx
2
B(0;
2
) T(B(0;
2
R)).
Continuing in this way, (i.e., by recursion) we obtain a sequence of points
x
1
, x
2
, x
3
, . . . in X such that x
n
B(0;
n1
R) and such that
y Tx
1
Tx
2
Tx
n
B(0;
n
) T(B(0;
n
R)) .
It follows that y =

n=1
Tx
n
. However, |x
n
| <
n1
R which means that

n
|x
n
| < . Since X is complete, it follows that

x
n
converges, i.e.,
there is some x X such that

n
k=1
x
k
x, as n . But T is continuous
and therefore

n
k=1
Tx
k
= T(

n
k=1
x
k
) Tx, as n . It follows that
y = Tx. Furthermore,
|
n

k=1
x
k
|
n

k=1
|x
k
|
<
n

k=1

k1
R
< R/(1 )
so that |x| = lim
n
|

n
k=1
x
k
| R/(1 ) < R/(1 2) (suppose < 1/2).
Hence x B(0; R/(1 2)), and so y = Tx T(B(0; R/(1 2)). We have
proved, so far, that
B(0; ) T
_
B
_
0;
R
1 2
__
Kings College London Baire Category Theorem and all that 3.7
for any 0 < < 1/2.
Let y B(0; ). Then |y| < . Let d > 0 be such that |y| < d < .
Then
y B(0; d) =
d

B(0; )

T
_
B
_
0;
R
1 2
__
= T
_
B
_
0; R
d
(1 2)
__
.
Since d/ < 1, we can choose suciently small that Rd/(1 2) < R. It
follows that y T(B(0; R)). Hence B(0; ) T(B(0; R)).
The Open Mapping Theorem is a simple consequence of this last result
and the Baire Category theorem.
Theorem 3.8. (Open Mapping Theorem) Suppose that both X and
Y are Banach spaces and that T : X Y is a bounded linear operator
mapping X onto Y . Then T is an open map, i.e., T maps open sets in X
into open sets in Y .
Proof. Let G be an open set in X. We wish to show that T(G) is open
in Y . If G = , then T(G) = T() = and there is nothing to prove.
So suppose that G is non-empty. Let y T(G). Then there is x G such
that y = Tx. Since G is open, there is r > 0 such that B(x; r) G. Hence
T(B(x; r)) T(G). To show that T(G) is open it is certainly enough to
show that T(B(x; r)) contains an open ball of the form B(y; r

), for some
r

> 0. Now, B(y; r

) = y +B(0; r

) and
T(B(x; r)) = T(x +B(0; r)) = Tx +T(B(0; r)) = y +T(B(0; r))
in Y . Hence, the statement that B(y; r

) T(B(x; r)), for some r

> 0, is
equivalent to the statement that y+B(0; r

) Tx+T(B(0; r)), for some r

>
0, which, in turn, is equivalent to the statement that B(0; r

) T(B(0; r)),
for some r

> 0. We shall prove this last inclusion.

Any x X lies in the open ball B(0; n) whenever n > |x|, and so the
collection B(0; n) : n N covers X. Since T : X Y maps X onto Y , we
deduce that
Y =

_
n=1
T(B(0; n)) .
3.8 Functional Analysis Gently Done Mathematics Department
By the Baire Category Theorem, not all the sets T(B(0; n)) can be nowhere
dense, that is, there is some N N such that T(B(0; N)) has non-empty
interior. Thus there is some y Y and > 0 such that
B(y; ) T(B(0; N)) .
Since T(B(0; N)) is symmetric, it follows that also
B(y; ) T(B(0; N)) .
Furthermore, T(B(0; N)) is convex, so if |w| < , we have that w + y
B(y; ) and w y B(y; ) and so
w =
1
2
(w +y) +
1
2
(w y) T(B(0; N)) .
In other words,
B(0; ) T(B(0; N)) .
By the proposition, we deduce that B(0; ) T(B(0; N)). Hence
B
_
0;
r
N
_

r
N
T(B(0; N))
= T(B(0; r)) .
Taking r

=
r
N
completes the proof.
As a corollary to the Open Mapping Theorem, we have the following
theorem.
Theorem 3.9. (Inverse Mapping Theorem) Any one-one and onto
bounded linear mapping between Banach spaces has a bounded inverse.
Proof. Suppose that T : X Y is a bounded linear mapping between the
Banach spaces X and Y , and suppose that T is both injective and surjective.
Then it is straightforward to check that T is invertible and that its inverse,
T
1
: Y X, is a linear mapping. We must show that T
1
is bounded.
To see this, note that by the Open Mapping Theorem, the image under
T of the open unit ball in X is an open set in Y and contains 0. Hence there
is some r > 0 such that
B(0; r) T(B(0; 1)) .
Kings College London Baire Category Theorem and all that 3.9
Thus, for any y Y with |y| < r, there is x X with |x| < 1 such that
y = Tx. That is, if |y| < r, then |T
1
y| < 1. It follows that T
1
is
bounded and that |T
1
|
1
r
.
(Alternatively, we can simply remark that for any open set G in X, its
image, T(G), under T is open, by the Open Mapping Theorem. But T(G)
is precisely the pre-image of G under the inverse T
1
. It follows that T
1
is
a continuous mapping.)
Suppose that X and Y are normed spaces and that T : X Y is a linear
operator dened on a dense linear subspace D(T) of X. The linearity of
the domain of denition D(T) of T is of course necessary to even state the
linearity of T. The point is that we do not assume, for the moment, that
D(T) = X, or that T is bounded. We simply think of T as a linear operator
from the normed space D(T) into Y .
Denition 3.10. Let X and Y be normed spaces and let T : X Y be
a linear operator with dense linear domain D(T). The graph of T, denoted
(T), is the subset of the direct sum X Y given by
(T) = x y X Y : x D(T), y = Tx .
Thus, (T) = x Tx : x D(T).
It is readily seen that (T) is a linear subspace of X Y . The space
X Y is equipped with the norm
|x y| = |x| +|y|
for x y X Y . X Y is complete with respect to this norm if and only
if both X and Y are complete.
Theorem 3.11. (Closed Graph Theorem) Suppose that X and Y
are Banach spaces and that T : X Y is a linear operator with domain
D(T) = X. Then T is bounded if and only if the graph of T is closed in
X Y .
Proof. Suppose rst that T is bounded, and suppose that ((x
n
y
n
)) is
a sequence in (T) such that (x
n
, y
n
) (x, y) in X Y . It follows that
x
n
x and y
n
y and therefore, in particular, Tx
n
Tx in Y . However,
y
n
= Tx
n
, and so Tx
n
y. We conclude that y = Tx and that (x, y)
(T). Thus (T) is closed in X Y .
3.10 Functional Analysis Gently Done Mathematics Department
Conversely, suppose that (T) is closed in the Banach space X Y .
Then (T) is itself a Banach space (with respect to the norm inherited from
X Y ). Dene maps
1
: (T) X,
2
: (T) Y by the assignments

1
: x Tx x and
2
: x Tx Tx. Evidently, both
1
and
2
are
norm decreasing and so are bounded linear operators. Moreover, it is clear
that
1
is both injective and surjective. It follows, by the Inverse Mapping
Theorem, that
1
1
: X (T) is bounded. But then T : X Y is given
by T =
2

1
1
,
x

1
1
(x, Tx)

2
Tx,
which is the composition of two bounded linear maps and therefore T is
bounded.
Remark 3.12. The closed graph theorem can be a great help in establishing
the boundedness of linear operators between Banach spaces. Indeed, in order
to show that a linear operator T : X Y is bounded, one must establish
essentially two things; rstly, that if x
n
x in X, then (Tx
n
) converges in
Y and, secondly, that this limit is Tx. The closed graph theorem says that
to prove that T is bounded it is enough to prove that its graph is closed
(provided, of course, that X and Y are Banach spaces). This means that
we may assume that x
n
x and Tx
n
y, for some y Y , and then need
only show that y = Tx. In other words, thanks to the closed graph theorem,
the convergence of (Tx
n
) can be taken as part of the hypothesis rather than
forming part of the proof itself.
Example 3.13. Let X = (([0, 1]) equipped with the supremum norm,
| |

. Dene an operator T on X by setting D(T) = (

1
([0, 1]), the linear
subspace of continuously dierentiable functions on [0, 1], and, for x D(T),
put
Tx(t) =
dx
dt
(t) , 0 t 1 .
Note that D(T) is a dense linear subspace of (([0, 1]) (by the Weierstrass
approximation theorem, for example) and T is a linear operator. We shall
see that the graph of T is closed, but that T is unbounded. To see that T
is unbounded, we observe that if g
n
denotes the function g
n
(t) = t
n
, n N,
t [0, 1], then g
n
D(T) and Tg
n
= ng
n1
for n > 1. But |g
n
|

= 1 and
|Tg
n
|

= n, so it is clear that T is unbounded.

To show that (T) is closed, suppose that x
n
x in X with x
n
D(T),
and suppose that Tx
n
y in X. We must show that y D(T) and that
Kings College London Baire Category Theorem and all that 3.11
y = Tx. For ant t [0, 1], we have
_
t
0
y(s) ds =
_
t
0
lim
n
dx
n
ds
ds
= lim
n
_
t
0
dx
n
ds
ds , since convergence is uniform,
= lim
n
(x
n
(t) x
n
(0))
= x(t) x(0).
Thus
x(t) = x(0) +
_
t
0
y(s) ds , for 0 t 1,
with y (([0, 1]). Hence x (
1
([0, 1]) and
dx
ds
= y on [0, 1]. That is,
x D(T) and Tx = y. We conclude that x y (T) and therefore (T)
is closed.
Notice that T is not dened on the whole of X and so there is no conict
with the closed graph theorem. In fact, we can deduce from the closed graph
theorem, that there is no way in which we can extend the denition of T to
include every element of X in its domain of denition without spoiling either
linearity or the closedness of its graph.
Denition 3.14. A linear operator T : X Y with dense linear domain
D(T) is said to be closed if its graph is a closed subset of X Y .
The concept of closed linear operator plays a very important role in the
theory of unbounded operators in a Hilbert space.
Theorem 3.15. (Hellinger-Toeplitz) Let A : H H be a linear
operator on the Hilbert space H with D(A) = H and suppose that
< x, Ay > = < Ax, y >
for every x, y H. Then A is bounded.
Proof. We show that the graph, (A), of A is closed. Suppose, then, that
x
n
x, and that Ax
n
y. For any z H, we have
< z, y > = lim
n
< z, Ax
n
>
= lim
n
< Az, x
n
>
= < Az, x >
= < z, Ax > .
3.12 Functional Analysis Gently Done Mathematics Department
Hence < z, y Ax >= 0 for all z H. Taking z = y Ax, it follows that
y Ax = 0, that is y = Ax. Thus x y (A), and we conclude that (A)
is closed. By the closed graph theorem, it follows that A is bounded.
Remark 3.16. This thorem says that an everywhere-dened symmetric
linear operator on a Hilbert space is necessarily bounded. Thus, unbounded
symmetric operators cannot be everywhere dened. The domain of deni-
tion is a central issue in the theory of unbounded operators. It should be
emphasized that unbounded operators should not be considered as some-
what pathological and of no particular interest. Indeed, the example above,
the operator of dierentiation, could not really be thought of as especially
pathological. It turns out that many examples of operators in applications,
for example, in the mathematical theory of quantum mechanics, are un-
bounded. Indeed, one can show that operators P, Q satisfying the famous
Heisenberg commutation relation, PQQP = i, cannot both be bounded.
4. The Hahn-Banach Theorem
We now turn to a discussion of the Hahn-Banach theoremthis is con-
cerned with the extension of a continuous linear functional on a subspace
of a normed space to the whole of the space. First we must discuss Zorns
lemma.
Denition 4.1. A partially ordered set is a non-empty set P on which is
dened a relation _ (a partial ordering) satisfying:
(a) x _ x, for all x P;
(b) if x _ y and y _ x, then x = y;
(c) if x _ y and y _ z, then x _ z.
Note that it can happen that a particular pair of elements of P are not
comparable, that is, neither x _ y nor y _ x need hold.
Examples 4.2. 1. Let P be the set of all subsets of a given set, and let _
be given by set inclusion .
2. Set P = R, and let _ be the usual ordering on R.
3. Set P = R
2
, and dene _ according to the prescription (x

, y

) _ (x

, y

)
provided that both x

and y

in R.
4. Any subset of a partially ordered set inherits the partial ordering and so
is itself a partially ordered set.
Denition 4.3. An element m in a partially ordered set (P, _) is said to
be maximal if m _ x implies that x = m. Thus, a maximal element cannot
be majorized by any other element.
Example 4.4. Let P be the half-plane in R
2
given by P = (x, y) : x+y
0, equipped with the partial ordering as in example 3, above. Then one
sees that each point on the line x+y = 0 is a maximal element. Thus P has
many maximal elements. Note that P has no largestelement, i.e., there is
no element z P satisfying x _ z, for all x P.
1
4.2 Functional Analysis Gently Done Mathematics Department
Denition 4.5. An upper bound for a subset A in a partially ordered set
(P, _) is an element x P such that a _ x for all a A.
A partially ordered set (P, _) is called a chain (or totally ordered, or linearly
ordered) if for any pair x, y P, either x _ y or y _ x holds. In other words,
P is totally ordered if every pair of points in P are comparable.
A subset C of a partially ordered set (P, _) is said to be totally ordered (or
a chain in P) if for any pair of points c

, c

C, either c

_ c

or c

_ c

;
that is, C is totally ordered if any pair of points in C are comparable.
We now have sucient terminology to state Zorns lemma which we shall
take as an axiom.
Zorns lemma. Let P be a partially ordered set. If each totally ordered
subset of P has an upper bound, then P possesses at least one maximal
element.
Remark 4.6. As stated, the intuition behind the statement is perhaps not
evident. The idea can be roughly outlined as follows. Suppose that a is
any element in P. If a is not itself maximal then there is some x P with
a _ x. Again, if x is not a maximal element, then there is some y P such
that x _ y. Furthermore, the three elements a, x, y form a totally ordered
subset of P. If y is not maximal, add in some greater element, and so on. In
this way, one can imagine having obtained a totally ordered subset of P. By
hypothesis, this set has an upper bound, , say. (This means that we rule
out situations such as having arrived at, say, the natural numbers 1, 2, 3, . . . ,
(with their usual ordering) which one could think of having got by starting
with 1, then adding in 2, then 3 and so on.) Now if is not a maximal
element, we add in an element greater than and proceed as before. Zorns
lemma can be thought of as stating that this process eventually must end
with a maximal element.
Zorns lemma can be shown to be logically equivalent to the axiom of
choice (and, indeed, to the Well-Ordering Principle). We recall the axiom of
choice.
Axiom of Choice. Let A

: J be a family of non-empty sets,

indexed by the non-empty set J. Then there is a mapping : J
_

such that () A

for each J.
Thus, the axiom says that we can choose a family a

with a

,
for each J, namely, the range of . As a consequence, this axiom gives
substance to the cartesian product

.
Kings College London The Hahn-Banach Theorem 4.3
After these few preliminaries, we return to consideration of the Hahn-
Banach theorem whose proof rests on an application of Zorns lemma.
Denition 4.7. A linear map : X C, from a linear space X into C is
called a linear functional. A real-linear functional on a real linear space is a
real-linear map : X R.
Example 4.8. Let X be a (complex) linear space, and : X C a linear
functional on X. Dene : X R by (x) = Re (x) for x X. Then is a
real-linear functional on X if we view X as a real linear space. Substituting
ix for x, it is straightforward to check that
(x) = (x) i(ix)
for any x X. On the other hand, suppose that u : X R is a real linear
functional on the complex linear space X (viewed as a real linear space). Set
(x) = u(x) iu(ix)
for x X. Then one sees that : X C is (complex) linear. Further-
more, u = Re , and so we obtain a natural correspondence between real
and complex linear functionals on a complex linear space X via the above
relations.
It is more natural to consider the Hahn-Banach theorem for real normed
spaces; the complex case will be treated separately as a corollary.
Theorem 4.9. (Hahn-Banach extension theorem) Suppose that M
is a (real-) linear subspace of a real normed space X and that : M R
is a bounded real-linear functional on M. Then there is a bounded linear
functional on X which extends and with || = ||; that is, : X R,
M = , and
|| = sup
_
[(x)[
|x|
: x X, x ,= 0
_
= || = sup
_
[(x)[
|x|
: x M, x ,= 0
_
.
Proof. One idea would be to extend to the subspace of X obtained by
enlarging M by one extra dimension and then to keep doing this. However,
one must then give a convincing argument that eventually one does, indeed,
exhaust the whole of X in this way. To circumvent this problem, the idea is
to use Zorns lemma and to show that any maximal extension of must, in
fact, already be dened on the whole of X.
4.4 Functional Analysis Gently Done Mathematics Department
Let x
0
be a non-zero element of X with x
0
/ M and let M
1
be the (real)
linear space spanned by M and x
0
(if M = X, there is nothing to prove).
Then any element of M
1
has the form x + x
0
, where x M and R.
If x + x
0
= x

x
0
, then x x

+ (

)x
0
= 0 and so it follows that
x = x

and =

. In other words, this representation of each element of

M
1
is unique. We dene
1
: M
1
R by

1
(x +x
0
) = (x) +a
where a is any xed real number. It is evident that
1
is real-linear and that

1
M = . Now, extending any bounded linear operator cannot decrease
its norm, and so
1
has the same norm as provided we can nd a so that
() [(x) +a[ || |x +x
0
|
for all x M and R. This clearly holds for = 0 and so we may suppose
that ,= 0. Then we may replace x by x, and divide through by [[, to
obtain the requirement
[(x) a[ || |x x
0
|
for all x M. We rewrite this as the requirement that
(x) || |x x
0
|
. .
A
x
a (x) +|| |x x
0
|
. .
B
x
for all x M. It is possible to nd such a real number a if and only if
all the closed intervals [A
x
, B
x
] have a common point. This is equivalent to
A
x
B
y
for all x, y M. To see that this holds, we note that (since is
real-valued)
(x y) || |x y|
for all x, y M. Thus
(x) (y) || |x y|
|| |x x
0
| +|| |x
0
y| .
and hence we obtain A
x
B
y
for all x, y M, as required. Thus there
is some a R such that the inequality () holds, and we conclude that
|
1
| = ||. We shall now set things up so that Zorns lemma becomes
applicable. Let E be the collection of extensions e : M
e
R of which
Kings College London The Hahn-Banach Theorem 4.5
satisfy |e| = ||. E is partially ordered by declaring e
1
_ e
2
if e
2
is an
extension of e
1
(that is, if M
e
1
M
e
2
and e
2
M
e
1
= e
1
). Let C be
a totally ordered subset of E. Then it follows that

eC
M
e
is a linear
subspace of X. Dene e

: M

R, where M

eC
M
e
, by e

(x) = e(x),
for x M

, where e is such that x M

e
M

. It is clear that e

is well-
dened and is an extension of satisfying [e

(x)[ || |x| for all x M

.
Hence e

, dened on M

, is an element of E and is an upper bound for C. By

Zorns lemma, it follows that E possesses a maximal element, , say, dened
on some linear subspace M

, say, of X. Now, if M

were a proper subspace

of X, we could repeat our earlier argument to obtain an extension of , with
the same norm, which would therefore be an element of E and contradict the
maximality of . We conclude that is dened on the whole of X, and the
proof is complete.
Corollary 4.10. (Hahn-Banach extension theorem for complex
vector spaces) Suppose that M is a linear subspace of a complex normed
space X and that : M C is a bounded linear functional on M. Then
can be extended to a bounded linear functional on X with || = ||.
Proof. Consider X as a real linear space, and let (x) = Re (x), for x M.
Then
[(x)[ = [ Re (x)[
|| |x| , x M ,
so is bounded with || ||. For x M, write (x) = e
i
, where
= [(x)[ = 0. Then (e
i
x) = , and so (e
i
x) = Re (e
i
x) = =
[(x)[. It follows that [(x)[ : x M [(x)[ : x M and so || = ||.
By the theorem, has a real-linear extension L to X with |L| = ||. For
x X, set
(x) = L(x) iL(ix) .
Then : X C is complex-linear and, since (x) = (x) i(ix), x M,
we see that extends . We must show that [(x)[ || |x|, for x X,
thus giving || = ||. To see this, we note that for x X there is C
with [[ = 1 such that (x) = [(x)[. Then
[(x)[ = (x)
= (x) R since the left hand side is real
4.6 Functional Analysis Gently Done Mathematics Department
= L(x)
|L| |x|
= || [[ |x|
= || |x|
as required.
It is an immediate corollary that any (non-zero) normed space has a
non-zero continuous linear functional. In fact, it has many as we shall now
show.
Theorem 4.11. Let X be a normed space, and let x
0
X, with x
0
,= 0.
Then there is a continuous linear functional on X with || = 1 such that
(x
0
) = |x
0
|.
Proof. Let M be the subspace M = x
0
: C, and dene
0
: M C
by
0
(x
0
) = |x
0
|, C. Evidently,
0
is linear, and
[
0
(x
0
)[ = [[ |x
0
| = |x
0
|
for all C, which implies that |
0
| = 1 (as a map from M into C).
By the Hahn-Banach theorem,
0
has an extension to X with the required
properties.
Remark 4.12. The result above implies that the set of bounded linear
functionals on a normed space X separates the points of X, i.e., if x
1
and
x
2
are any two points of X with x
1
, = x
2
, then there is a bounded linear
functional : X C such that (x
1
) ,= (x
2
). One simply applies the
above to the non-zero element x
0
= x
1
x
2
. Put another way, this result
says that if x X is such that (x) = 0 for every bounded linear functional
on X, then x = 0. The following theorem says that the set of bounded
linear functionals on X also separates points and closed subspaces.
Kings College London The Hahn-Banach Theorem 4.7
Theorem 4.13. Let M be a proper closed linear subspace of a normed
space X and suppose that x
0
/ M. Then there is a bounded linear functional
on X such that (x) = 0 for all x M, but (x
0
) ,= 0.
Proof. Let M

be the subspace of X generated by M and x

0
. Then any
element x, say, of M

can be written uniquely as x = z + x

0
, for z M
and some C (if also x = z

x
0
, then subtracting, we obtain that
z z

= (

)x
0
, and this implies that

= (since x
0
/ M) and
therefore z = z

). Dene : M

C by (z +x
0
) = , for z M. Clearly
: M

C is linear, (x
0
) = 1, and is zero on M. To see that is
bounded, we note that since x
0
/ M and M is closed, there is r > 0 such
that B(x
0
; r) M = ; that is, |z x
0
| > r for all z M. Hence, for ,= 0,
and z M,
|z +x
0
| = [[ |
z

+x
0
|
> [[ r
since z/ M. Thus
[(z +x
0
)[ = [[ <
1
r
|z +x
0
| .
It follows that || 1/r and so : M

C is bounded. By the Hahn-

Banach theorem, we may extend to the whole of X, and the result follows.
Remark 4.14. Thus, if M is a closed subspace of X and x
0
X is such
that all bounded linear functionals which vanish on M also vanish on x
0
,
then x
0
M.
Denition 4.15. Suppose that : X C is a linear functional on the
linear space X. The kernel of is its null space; ker = x X : (x) = 0.
Proposition 4.16. Let be a linear functional on the linear space X.
Then ker is a linear subspace of X of codimension one.
Proof. It is clear that ker is a linear subspace of X. If is not zero, there
is some x
0
X such that (x
0
) ,= 0. Then, for any x X,
x =
(x)
(x
0
)
x
0
+x
(x)
(x
0
)
x
0
.
Clearly, x
(x)
(x
0
)
x
0
ker and hence X/ ker is one-dimensional.
4.8 Functional Analysis Gently Done Mathematics Department
As a corollary, we see that a linear functional is determined by its kernel,
up to a constant of proportionality.
Corollary 4.17. Linear functionals
1
and
2
on the linear space X have
the same kernel if and only if they are proportional.
Proof. Suppose that ker
1
= ker
2
and let x
0
/ ker
1
. (If no such x
0
exists, then both
1
and
2
are zero on X.) Without loss of generality, we
may suppose that
1
(x
0
) = 1. Then for any x X, we have x =
1
(x) x
0
+z,
with z ker
1
= ker
2
. Hence

2
(x) =
1
(x)
2
(x
0
)
which shows that
1
and
2
are proportional.
The converse is clear.
Theorem 4.18. Suppose that : X C is a linear functional on the
normed space X. Then is bounded if and only if ker is closed.
Proof. It is clear that if is bounded (equivalently, continuous), then ker
is closed.
Conversely, suppose that ker is closed. If ker = X, it follows that is
zero and so is certainly bounded. Suppose, then, that ker ,= X. Then there
is x
0
X such that x
0
/ ker . By hypothesis, ker is closed, and so there is
r > 0 such that B(x
0
; r) ker = . By replacing x
0
by x
0
/(x
0
), we may
assume that (x
0
) = 1.
Suppose that x X, x / ker . Then (x) ,= 0, and

x
(x)
+x
0
ker .
It follows that
x
(x)
+x
0
/ B(x
0
; r); that is,
_
_
_
_
_

x
(x)
+x
0
_
x
0
_
_
_
_
r ,
that is,
|x|
[(x)[
r ,
that is,
[(x)[
1
r
|x|
for all x / ker . But this inequality still holds even if x ker , and we
conclude that is bounded.
Kings College London The Hahn-Banach Theorem 4.9
Proposition 4.19. Suppose that : X C is a linear functional on the
normed space X. Then is unbounded if and only if ker is a proper dense
subset of X.
Proof. Suppose that ker is dense in X. If is bounded, then it follows
that must vanish on the whole of X. So ker ,= X demands that be
unbounded.
Conversely, suppose that ker is not dense in X. Then there is some
x
0
/ ker and some r > 0 such that B(x
0
; r) ker = . We now argue as
before to deduce that is bounded. It follows that if is unbounded, then
ker is a dense subset of X. Furthermore, ker must be a proper subset of
X since cannot vanish on the whole of X otherwise it would clearly be
bounded.
5. Hamel Bases
Denition 5.1. A nite set of elements x
1
, . . . , x
n
in a complex (real)
vector space is said to be linearly independent if and only if

1
x
1
+ +
n
x
n
= 0
with
1
, . . . ,
n
C (or R) implies that
1
= =
n
= 0. A subset A
in a vector space is said to be linearly independent if and only if each nite
subset of A is.
Denition 5.2. A linearly independent subset A in a vector space X is
called a Hamel basis of X if and only if any non-zero element x X can be
written as
x =
1
u
1
+ +
m
u
m
for some m N, non-zero
1
, . . . ,
m
C (or R) and distinct elements
u
1
, . . . , u
m
A.
In other words, A is a Hamel basis of X if it is linearly independent and
if any element of X can be written as a nite linear combination of elements
of A.
Note that if A is a linearly independent subset of X and if x X can
be written as x =
1
u
1
+ +
m
u
m
, as above, then this representation is
unique. To see this, suppose that we also have that x =
1
v
1
+ +
k
v
k
,
for non-zero
1
, . . . ,
k
C and distinct elements v
1
, . . . , v
k
A. Taking the
dierence, we have that
0 =
1
u
1
+ +
m
u
m

1
v
1

k
v
k
.
Suppose that m k. Now v
1
is not equal to any of the other v
j
s and so, by
independence, cannot also be dierent from all the u
i
s. In other words, v
1
is equal to one of the u
i
s. Similarly, we argue that every v
j
is equal to some
1
5.2 Functional Analysis Gently Done Mathematics Department
u
i
and therefore we must have m = k and v
1
, . . . , v
m
is just a permutation
of u
1
, . . . , u
m
. But then, again by independence,
1
, . . . ,
m
is the same
permutation of
1
, . . . ,
m
. The uniqueness of the representation of x as a
nite linear combination of elements of A follows.
We will use Zorns lemma to prove the existence of a Hamel basis.
Theorem 5.3. Every vector space X possesses a Hamel basis.
Proof. Let S denote the collection of linearly independent subsets of X,
partially ordered by inclusion. Let S

: J be a totally ordered subset

of S. Put S =

. We claim that S is linearly independent. To see this,

suppose that x
1
, . . . , x
m
are distinct elements of S and suppose that

1
x
1
+ +
m
x
m
= 0
for non-zero
1
, . . . ,
m
C (or R). Then x
1
S

1
, . . . , x
m
S

m
for some

1
, . . . ,
m
J. Since S

is totally ordered, there is some

J such
that S

1
S

, . . . , S

m
S

. Hence x
1
, . . . , x
m
S

. But S

is linearly
independent and so we must have that
1
= =
m
= 0. We conclude
that S is linearly independent, as claimed.
It follows that S is an upper bound for S

in S. Thus every totally

ordered subset in S has an upper bound and so, by Zorns lemma, S possesses
a maximal element, M, say. We claim that M is a Hamel basis.
To see this, let x X, x ,= 0, and suppose that an equality of the form
x =
1
u
1
+ +
k
u
k
is impossible for any k N, distinct elements u
1
, . . . , u
k
M and non-zero

1
, . . . ,
k
C (or R). Then, for any distinct u
1
, . . . , u
k
M, an equality of
the form
x +
1
u
1
+ +
k
u
k
= 0
must entail = 0. But then this means that
1
u
1
= =
k
u
k
= 0, by
independence. Hence x, u
1
, . . . , u
k
are linearly independent. It follows that
Mx is linearly independent, which contradicts the maximality of M. We
conclude that x can be written as
x =
1
u
1
+ +
m
u
m
for suitable m N, u
1
, . . . , u
m
M, and non-zero
1
, . . . ,
m
C (or R);
that is, M is a Hamel basis of X.
Kings College London Hamel Bases 5.3
The next result is a corollary of the preceding method of proof.
Theorem 5.4. Let A be a linearly independent subset of a linear space
X. Then there is a Hamel basis of X containing A; that is, any linearly
independent subset of a linear space can be extended to a Hamel basis.
Proof. Let S denote the collection of linearly independent subsets of X
which contain A. Then S is partially ordered by set-theoretic inclusion. As
above, we apply Zorns lemma to obtain a maximal element of S, which is a
Hamel basis of X and contains A.
The existence of a Hamel basis proves useful in the construction of various
pathological examples, as we shall see. We rst consider the existence of
unbounded linear functionals. It is easy to give examples on a normed space.
For example, let X be the linear space of those complex sequences which are
eventually zero thus (a
n
) X if and only if a
n
= 0 for all suciently
large n (depending on the particular sequence). Equip X with the norm
|(a
n
)| = sup [a
n
[, and dene : X C by (a
n
) ((a
n
)) =

n
a
n
.
Evidently is an unbounded linear functional on X. Another example is
furnished by the functional f f(0) on the normed space (([0, 1]) equipped
with the norm |f| =
_
1
0
[f(s)[ ds.
It is not quite so easy, however, to nd examples of unbounded linear
functionals or everywhere-dened unbounded linear operators on Banach
spaces. To do this, we shall use a Hamel basis. Indeed, let X be any innite-
dimensional normed space and let M be a Hamel basis. We dene a linear
operator T : X X via its action on M as follows. Let u
1
, u
2
, . . . be any
sequence of distinct elements of M, and set
Tu
k
= ku
k
, k = 1, 2, . . .
and
Tv = 0, for v M, v ,= u
k
for any k N.
Then if x X, with x = w
1
+ +
m
w
m
, w
j
M,
1
, . . . ,
m
C, we
put
Tx =
1
Tw
1
+ +
m
Tw
m
.
It is clear that T is a linear operator dened on the whole of X. Moreover,
|Tu
k
| = k|u
k
|, for any k N, and it follows that T is unbounded.
We can rene this a little. Dene Tu
k
= ku
k
as above, but now let
Tv = v for any v M with v ,= u
k
, any k. Then T is everywhere dened
and unbounded. Furthermore, it is easy to see that T : X X is one-one
5.4 Functional Analysis Gently Done Mathematics Department
and onto. If X is a Banach space, the inverse mapping theorem implies that
T
1
must also be unbounded.
Now let : M R be the map w (w) = |w|, w M. By linearity,
we can extend to a linear map on X. Then the map : X C given by
x (Tx) is an everywhere-dened unbounded linear functional on X.
We can use the concept of Hamel basis to give an example of a space which
is a Banach space with respect to two inequivalent norms. It is not dicult
to give examples of linear spaces with inequivalent norms. For example,
([0, 1] equipped with the | |

and | |
1
norms is such an example. It is
a little harder to nd examples where the space is complete with respect to
each of the two inequivalent norms. To give such an example, we will use
the fact that any Hamel basis for an innite dimensional, separable linear
space has cardinality 2

0
. In fact, all we need to know is that if X and Y
are separable, innite dimensional spaces with Hamel bases M
X
and M
Y
,
respectively, then M
X
and M
Y
are isomorphic as sets.
Example 5.5. Set X =
1
and Y =
2
and, for k N, let e
k
be the
element e
k
= (
km
)
mN
of
1
and let f
k
denote the corresponding element of

2
. For n N, let a
n
=

n
k=1
1
n
e
k

1
and let b
n
=

n
k=1
1
n
f
k

2
. Then
|a
n
|
1
= 1, for all n N, whereas |b
n
|
2
= 1/

n. Let A = a
n
: n N and
B = b
n
: n N. Then A and B are linearly independent subsets of
1
and

2
, respectively. Hence they may be extended to Hamel bases M
X
and M
Y
of
X and Y . Since both
1
and
2
are separable, M
X
and M
Y
are isomorphic.
It follows that the map dened by (a
n
) = b
n
, for n N, extends to an
isomorphism mapping M
X
onto M
Y
. By linearity, this map extends to an
isomorphism, which we denote also by , from
1
onto
2
.
We dene a new norm [[[ [[[ on
1
by setting
[[[x[[[ = |(x)|
2
for x X =
1
. This is a norm because is linear and injective. To see
that X is complete with respect to this norm, suppose that (x
n
) is a Cauchy
sequence with respect to [[[ [[[. Then ((x
n
)) is a Cauchy sequence in
2
.
Since
2
is complete, there is some y
2
such that |(x
n
) y|
2
0. Now,
is surjective and so we may write y as y = (x) for some x
1
. We have
|(x
n
) y|
2
= |(x
n
) (x)|
2
= [[[x
n
x[[[
and it follows that [[[x
n
x[[[ 0 as n . In other words,
1
is complete
with respect to the norm [[[ [[[.
Kings College London Hamel Bases 5.5
We claim that the norms | |
1
and [[[ [[[ are not equivalent norms on

1
. Indeed, we have that (a
n
) = b
n
and so [[[a
n
[[[ = |b
n
|
2
= 1/

n 0 as
n . However, |a
n
|
1
= 1 for all n.
6. Projections
Let X be a linear space and suppose that V and W are subspaces of X
such that V W = 0 and X = spanV, W; then any x X can be written
uniquely as x = v +w with v V and w W. In other words, X = V W.
Dene a map P : X V by Px = v, where x = v +w X, with v V and
w W, as above. Evidently, P is a well-dened linear operator satisfying
P
2
= P. P is called the projection onto V along W. We see that ran P = V
(since Pv = v for all v V ), and also ker P = W (since if x = v + w and
Px = 0 then we have 0 = Px = v and so x = w W).
Conversely, suppose that P : X X is a linear operator such that
P
2
= P, that is, P is an idempotent. Set V = ran P and W = ker P.
Evidently, W is a linear subspace of X. Furthermore, for any given v V ,
there is x X such that Px = v. Hence
Pv = P
2
x = Px = v
and we see that (1l P)v = 0. Hence V = ker(1l P) and it follows that
V is also a linear subspace of X. Now, any x X can be written as x =
Px + (1l P)x with Px V = ran P and (1l P)x W = ker P. We have
seen above that for any v V , we have v = Pv. If also v W = ker P
then Pv = 0, so that v = Pv = 0. It follows that V W = 0 and so
X = V W.
Now suppose that X is a normed space and that P : X X is a bounded
linear operator such that P
2
= P. Then both V = ran P = ker(1l P) and
W = ker P are closed subspaces of X and X = V W.
Conversely, suppose that X is a Banach space and X = V W, where
V and W are closed linear subspaces of X. Dene P : X V as above so
that P
2
= P and V = ran P = ker(1l P) and W = ker P. We wish to show
that P is bounded. To see this we will show that P is closed and then appeal
to the closed-graph theorem. Suppose, then, that x
n
x and Px
n
y.
1
6.2 Functional Analysis Gently Done Mathematics Department
Now, Px
n
V for each n and V is closed, by hypothesis. It follows that
y V and so Py = y. Furthermore, (1l P)x
n
= x
n
Px
n
x y and
(1l P)x
n
W for each n and W is closed, by hypothesis. Hence xy W
and so P(x y) = 0, that is, Px = Py. Hence we have Px = Py = y and
we conclude that P is closed. Thus P is a closed linear operator from the
Banach space X onto the Banach space V . By the closed-graph theorem, it
follows that P is bounded. Note that |P| 1 unless ran P = 0. Indeed,
|P| = sup|Px| : |x| = 1
sup|Pv| : |v| = 1, v V = 1 .
We have therefore proved the following theorem.
Theorem 6.1. Suppose that V is a closed subspace of a Banach space X.
Then there is a closed subspace W such that X = V W if and only if there
exists a bounded idempotent P with ran P = V .
Denition 6.2. We say that a closed subspace V in a normed space is
complemented if there is a closed subspace W such that X = V W.
Theorem 6.3. Suppose that V is a nite-dimensional subspace of a normed
space X. Then V is closed and complemented.
Proof. Let v
1
, . . . , v
m
be linearly independent elements of X which span
V . Using this basis, we may identify V with C
m
. Moreover, the norm on
V is equivalent to the usual Euclidean norm on C
m
. In particular, there
is some constant K such that if v V is given by v =

m
i=1

i
v
i
then
_

m
i=1
[
i
[
2
K|v|. Dene
i
: V C by linear extension of the rule

i
(v
j
) =
ij
for 1 i, j m. Then, with the notation above, for any v V ,
[
i
(v)[ = [
i
[ K|v|
and so we see that each
i
is a bounded linear functional on V . By the
Hahn-Banach theorem, we may extend these to bounded linear functionals
on X, which we will also denote by
i
. Then if v V is given by v =

1
v
1
+ +
m
v
m
we have
i
(v) =
i
and so
v =
1
(v)v
1
+ +
m
(v)v
m
.
Dene P : X X by
Px =
1
(x)v
1
+ +
m
(x)v
m
.
It is clear that P is a bounded linear operator on X with range equal to V .
Also we see that P
2
= P. Hence V = ker(1l P) is closed, since (1l P)
is bounded, and W = ker P is a closed complementary subspace for V . We
note that W =

m
i=1
ker
i
.
7. The Dual Space
Let X be a normed space. The space of all bounded linear functionals
on X, B(X, C), is denoted by X

and called the dual space of X. Since C is

complete, X

is a Banach space.
The Hahn-Banach theorem assures us that X

is non-trivial; indeed, X

separates the points of X. Now, X

is a normed space in its own right, so

we may consider its dual, X

; this is called the bidual or double dual of X.

Let x X, and consider the mapping
X

: (x) .
Evidently, this is a linear map : X

C. Moreover,
[(x)[ || |x| , for every X

so we see that this is a bounded linear map from X

into C, that is, it denes

an element of X

. In fact, this leads to an isometric embedding of X into

X

, as we now show.
Theorem 7.1. Let X be a normed space, and for x X, let
x
: X

C
be the evaluation map
x
() = (x), X

. Then x
x
is an isometric
linear map of X into X

.
Proof. We have seen that
x
X

for each x X. It is easy to see that

x
x
is linear;

x+y
() = (x +y) = (x) +(y)
=
x
() +
y
()
for all x, y X, and C. Also
[
x
()[ = [(x)[ || |x| for all X

1
7.2 Functional Analysis Gently Done Mathematics Department
shows that |
x
| |x|. However, by the Hahn-Banach theorem, for any
given x X, x ,= 0, there is

such that |

| = 1 and

(x) = |x|. For

this particular

, we then have
[
x
(

)[ = [

(x)[ = |x| = |x| |

| .
We conclude that |
x
| = |x|, and the proof is complete.
Thus, we may consider X as a subspace of X

via the linear isometric

embedding x
x
.
Denition 7.2. A Banach space X is called reexive if X = X

via the
above embedding.
Note that X

is a Banach space, so X must be a Banach space if we are

to be able to identify X with X

.
Theorem 7.3. A Banach space X is reexive if and only if X

is reexive.
Proof. If X = X

, then X

= X

. This can be seen as follows. To say

that X = X

means that each element of X

has the form

x
for some
x X. Now let

be the corresponding association of X

into X

(z) = z() for X

and z X

.
We have X

via

. Let X

. Any z X

has the form

x
, x X, and so (z) = (
x
).
Dene : X C by (x) = (
x
). Then
[(x)[ = [(
x
)[
|| |
x
|
= || |x| .
It follows that X

. Moreover,

(
x
) =
x
()
= (x) = (
x
)
and so

= ; i.e., X

= X

via .
Kings College London The Dual Space 7.3
Now suppose that X

= X

and suppose that X ,= X

. Then there is
X

such that ,= 0 but vanishes on X in X

; i.e., (
x
) = 0 for all
x X.
But then can be written as =

for some X

, since X

= X

and so
(
x
) =

(
x
)
=
x
() = (x)
which gives 0 = (
x
) = (x) for all x X, i.e., = 0 in X

.
It follows that

= 0 in X

and so = 0. This is a contradiction.

Hence X = X

.
Corollary 7.4. Suppose that the Banach space X is not reexive. Then
the natural inclusions X X

. . . and X

. . . are all
strict.
Proof. If, for sake of argument X

= X

, then this says that X

is
reexive. But this implies that X

is reexive, which in turn means that

X

is. Finally we conclude that X is reexive.

We shall now compute the duals of some of the classical Banach spaces.
For any p R with 1 p < , the space
p
is the space of complex sequences
x = (x
n
) such that
|x|
p
=
_

n=1
[x
n
[
p
_1
p
< .
We shall consider the cases 1 < p < and show that for such p, | |
p
is a
norm and that
p
is a Banach space with respect to this norm. We will also
show that the dual of
p
is
q
, where q is given by the formula
1
p
+
1
q
= 1. It
therefore follows that these spaces are reexive. At this stage, it is not even
clear that
p
is a linear space, never mind whether or not | |
p
is a norm.
We need some classical inequalities.
Proposition 7.5. Let a, b 0 and , > 0 with + = 1. Then
a

a +b
with equality if and only if a = b.
Proof. We note that the function t e
t
is strictly convex; for any x, y R,
e
(x+y)
< e
x
+e
y
. Putting a = e
x
, b = e
y
gives the required result.
7.4 Functional Analysis Gently Done Mathematics Department
The next result we shall need is Holders inequality.
Theorem 7.6. Let p > 1 and let q be such that
1
p
+
1
q
= 1, (q is called the
exponent conjugate to p.) Then, for any x = (x
n
)
p
and y = (y
n
)
q
,

n=1
[x
n
y
n
[ |x|
p
|y|
q
.
If p = 1, the above is valid if we set q = .
Proof. The case for p = 1 and q = is easy to see. So suppose that p > 1.
Without loss of generality, we may suppose that |x|
p
= |y|
q
= 1. Then let
=
1
p
, =
1
q
, a = [x
n
[
p
, b = [y
n
[
q
and use the previous proposition.
Proposition 7.7. For any x = (x
n
)
p
, with p > 1,
|x|
p
= sup

n=1
x
n
y
n

: |y|
q
= 1 .
The equality holds for p = 1 and q = , and also for the pair p = , q = 1.
Proof. Holders inequality implies that the right hand side is not greater
than the left hand side. For the converse, consider y = (y
n
) with y
n
=
sgn x
n
[x
n
[
p/q
/|x|
p/q
p
if 1 < p < , with y
n
= sgn x
n
if p = 1, and with
y
n
=
nm
, m N, if p = .
As an immediate corollary, we obtain Minkowskis inequality.
Corollary 7.8. For any x, y
p
, p 1, we have x +y
p
and
|x +y|
p
|x|
p
+|y|
p
.
Proof. This follows directly from the triangle inequality and the preceding
proposition.
Kings College London The Dual Space 7.5
Theorem 7.9. For any 1 p ,
p
is a Banach space. Moreover, if
1 p < , the dual of
p
is
q
, where q is the exponent conjugate to p.
Furthermore, for each 1 < p < , the space
p
is reexive.
Proof. We have already discussed these spaces for p = 1 and p = . For
the rest, it follows from the preceding results that
p
is a linear space and
that | |
p
is a norm on
p
. The completeness of
p
, for 1 < p < , follows
in much the same way as that of the proof for p = 1.
To show that
p

=
q
, we use the pairing as in Holders inequality.
Indeed, for any y = (y
n
)
q
, dene
y
on
p
by
y
: x = (x
n
)

n
x
n
y
n
.
Then Holders inequality implies that
y
is a bounded linear functional on

p
and the subsequent proposition (with the roles of p and q interchanged)
shows that |
y
| = |y|
q
.
To show that every bounded linear functional on
p
has the above form,
for some y
q
, let
p

, where 1 p < . Let y

n
= (e
n
), where
e
n
= (
nm
)
mN

p
. Then for any x = (x
n
)
p
,
(x) =
_

n
x
n
e
n
_
=
_

n
x
n
(e
n
)
_
=

n
x
n
y
n
.
Hence, replacing x
n
by sgn(x
n
y
n
) x
n
, we see that

n
[x
n
y
n
[ || |x|
p
.
For any N N, denote by y

the truncated sequence (y

1
, y
2
, . . . , y
N
, 0, 0, . . . ).
Then

n
[x
n
y

n
[ || |x|
p
and, taking the supremum over x with |x|
p
= 1, we obtain the estimate
|y

|
q
|| .
It follows that y
q
( and that |y|
q
||). But then, by denition,

y
= , and we deduce that y
y
is an isometric mapping onto
p

. Thus,
the association y
y
is an isometric linear isomorphism between
q
and

.
Finally, we note that the above discussion shows that
p
is reexive, for
all 1 < p < .
7.6 Functional Analysis Gently Done Mathematics Department
We shall now consider c
0
, the linear space of all complex sequences which
converge to 0, equipped with the supremum norm
|x|

= sup[x
n
[ : n N , for x = (x
n
) c
0
.
One checks that c
0
is a Banach space ( a closed subspace of

). We shall
show that the dual of c
0
is
1
, that is, there is an isometric isomorphism
between c

0
and
1
. To see this, suppose rst that z = (z
n
)
1
. Dene

z
: c
0
C to be : x

n
z
n
x
n
, x = (x
n
) c
0
. It is clear that
z
is
well-dened for any z
1
and that
[
z
(x)[

n
[z
n
[ [x
n
[ |z|
1
|x|

.
Thus we see that
z
is a bounded linear functional with norm |
z
| |z|
1
.
By taking x to be the element of c
0
whose rst m terms are equal to 1, and
whose remaining terms are 0, we see that |
z
|

m
n=1
[z
n
[. It follows that
|
z
| = |z|
1
, and therefore z
z
is an isometric mapping of
1
into c

0
.
We shall show that every element of c

0
is of this form and hence z
z
is
onto.
To see this, let c

0
, and, for n N, let z
n
= (e
n
), where e
n
c
0
is the sequence all of whose terms are zero except for the n
th
term which is
equal to 1, i.e., e
n
is the sequence (
nm
)
mN
. For any given N N, let
v =
N

k=1
sgn z
k
e
k
.
Then v c
0
and |v|

= 1 and
[(v)[ =
N

k=1
[z
k
[ || |v|

= || .
It follows that z = (z
n
)
1
and that |z|
1
||. Furthermore, for any
element x = (x
n
) c
0
,
(x) =
_

n=1
x
n
e
n
_
=
_

n=1
x
n
(e
n
)
_
=
z
(x) .
Hence =
z
, and the proof is complete.
Kings College London The Dual Space 7.7
Remark 7.10. Exactly as above, we see that the map z
z
is a linear
isometric mapping of
1
into the dual of

. Furthermore, if is any element

of the dual of

, then, in particular, it denes a bounded linear functional

on c
0
. Thus, the restriction of to c
0
is of the form
z
for some z
1
. It
does not follow, however, that has this form on the whole of

. Indeed,
c
0
is a closed linear subspace of

, and, for example, the element y = (y

n
),
where y
n
= 1 for all n N, is an element of

which is not an element of

c
0
. Then we know, from the Hahn-Banach theorem, that there is a bounded
linear functional , say, such that (x) = 0 for all x c
0
and such that
(y) = 1. Thus, is an element of the dual of

which is clearly not

determined by an element of
1
. The dual of

is strictly larger than

1
.
Theorem 7.11. Suppose that X is a Banach space and that X

is
separable. Then X is separable.
Proof. Let
n
: n = 1, 2, . . . be a countable dense subset of X

. For each
n N, let x
n
X be such that |x
n
| = 1 and [
n
(x
n
)[
1
2
|
n
|. Let S be the
set of nite linear combinations of the x
n
s with rational complex coecients.
Then S is countable. We claim that S is dense in X. To see this, suppose
the contrary; that is, suppose that S is a proper closed linear subspace of X.
Then there exists a non-zero bounded linear functional X

such that
vanishes on S. Since X

and
n
: n N is dense in X

, there is some
subsequence (
n
k
) such that
n
k
as k , that is,
|
n
k
| 0
as n . However,
|
n
k
| [(
n
k
)(x
n
k
)[, since |x
n
k
| = 1,
= [
n
k
(x
n
k
)[, since vanishes on S,

1
2
|
n
k
|
and so it follows that |
n
k
| 0, as k . But
n
k
implies that
|
n
k
| || and therefore || = 0. This forces = 0, which is a con-
tradiction. We conclude that S is dense in X and that, consequently, X is
separable.
7.8 Functional Analysis Gently Done Mathematics Department
Theorem 7.12. For 1 p < the space
p
is separable, but the space

is non-separable.
Proof. Let S denote the set of sequences of complex numbers (z
n
) such that
(z
n
) is eventually zero (i.e., z
n
= 0 for all suciently large n, depending on
the sequence) and such that z
n
has rational real and imaginary parts, for
all n. Then S is a countable set and it is straightforward to verify that S is
dense in each
p
, for 1 p < .
Note that S is also a subset of

, but it is not a dense subset. Indeed,

if x denotes that element of

all of whose terms are equal to 1, then

|x |

1 for any S.
To show that

is not separable, consider the subset A of elements whose

components consist of the numbers 0, 1, . . . , 9. Then A is uncountable and
the distance between any two distinct elements of A is at least 1. It follows
that the balls x : |x a|

< 1/2 : a A are pairwise disjoint. Now,

if B is any dense subset of

, each ball will contain an element of B, and

these will all be distinct. It follows that B must be uncountable.
Remark 7.13. The example of
1
shows that a separable Banach space
need not have a separable dualwe have seen that the dual of
1
is

,
which is not separable. This also shows that
1
is not reexive. Indeed, this
would require that
1
be isometrically isomorphic to the dual of

. Since

1
is separable, an application of the earlier theorem would lead to the false
conclusion that

is separable.
8. Topological Spaces
Denition 8.1. Let X be a nonempty set and suppose that T is a collection
of subsets of X. T is called a topology on X if the following hold;
(i) T, and X T;
(ii) if U

: J is an arbitrary collection of elements of T, labelled by

J, then

J
U

T;
(iii) if, for any k N, U
1
, U
2
, . . . , U
k
T, then

k
i=1
U
i
T.
The elements of T are called open sets, or T-open sets. The pair (X, T) is
called a topological space.
Examples 8.2.
1. T = , X called the indiscrete topology.
2. T is the set of all subsets of X discrete topology.
3. X = 0, 1, 2 and T = , X, 0, 1, 2 .
4. X any metric space, T the set of open sets in the usual metric space sense.
Thus a topological space is a generalization of a metric space.
Denition 8.3. For any non-empty subset A of a topological space (X, T),
the induced topology, T
A
, on A is dened to be that given by the collection
A T = A U : U T of subsets of A. (It is readily veried that T
A
is a
topology on A.)
Denition 8.4. A topological space (X, T) is said to be metrizable if there
is a metric on X such that T is as in example 4 above.
Remark 8.5. Not every topology is metrizable. For example, example 1,
above (provided X consists of more than one point).
Many of the usual concepts in metric space theory appear in the theory
of topological spaces but suitably rephrased in terms of open sets.
1
8.2 Functional Analysis Gently Done Mathematics Department
Denition 8.6. A subset F of a topological space X is said to be closed if
and only if its complement X F is open, i.e., belongs to T.
A point a X is an interior point of a set A X if there is U T such
that a U and U A. (Thus, a set G is open if and only if each of its
points is an interior point of G.)
The set of interior points of the set A is denoted by

A.
The point x is a limit point (accumulation point) of the set A if and only
if for every open set U, with x U, it is true that U A contains some point
distinct from x, i.e., the set A U x ,= .
The point a A is said to be an isolated point of A if there is an open
set U such that a U but U A a = .
The closure of the set A, written A, is the union of A and its set of limit
points,
A = A limit points of A .
Proposition 8.7. The closure A of A is the smallest closed set containing
A, that is,
A =

F : F is closed and F A .
Proof. We shall rst show that A is closed. Let y X A. Then y / A,
and y is not a limit point of A. Hence there is an open set U such that y U
and U A = . But then no point of U can be a limit point of A so we
deduce that U A = . It follows that X A is open and so, by denition,
A is closed.
Now suppose that F is closed and that A F. Then X F is open. Let
z X F. Then there is some open set U such that z U and U X F.
In particular, U A = , and so z / A and z is not a limit point of A. Hence
X F X A and we see that A F. The result follows.
Denition 8.8. A family of open sets U

: J is said to be an open
cover of a set B X if B

.
Denition 8.9. A subset K in a topological space is said to be compact if
every open cover of K contains a nite subcover.
By taking complements, open sets become closed sets, unions are replaced
by intersections and the notion of compactness can be rephrased as follows.
Kings College London Topological Spaces 8.3
Theorem 8.10. Let K be a subset of a topological space (X, T). The
following statements are equivalent.
(i) K is compact.
(ii) If F

J
is any family of closed sets in X such that K

J
F

= ,
then K

I
F

= for some nite subset I J.

(iii) If F

J
is any family of closed sets in X such that K

I
F

,= ,
for every nite subset I J, then K

J
F

,= .
Remark 8.11. The statements (ii) and (iii) are contrapositives. The prop-
erty in statement (iii) is called the nite intersection property (of the family
F

J
of closed sets).
Denition 8.12. A set N is a neighbourhood of a point x in a topological
space (X, T) if and only if there is U T such that x U and U N.
Thus, a set U belongs to T if and only if U is a neighbourhood of each
of its points. Note that N need not itself be open. For example, in any
metric space (X, d), the closed sets z X : d(a, z) r, for r > 0, are
neighbourhoods of the point a.
Denition 8.13. A topological space (X, T) is said to be a Hausdor topo-
logical space if and only if for any pair of distinct points x, y X, (x ,= y),
there exist sets U, V T such that x U, y V and U V = .
We can paraphrase the Hausdor property by saying that any pair of
distinct points can be separated by disjoint open sets. Example 3 above is
an example of a non-Hausdor topological space.
Proposition 8.14. A non-empty subset A of the topological space (X, T)
is compact if and only if A is compact with respect to the induced topology,
that is, if and only if (A, T
A
) is compact.
If (X, T) is Hausdor then so is (A, T
A
).
Proof. Suppose rst that A is compact in (X, T), and let G

be an open
cover of A in (A, T
A
). Then each G

has the form G

= A U

for some
U

T. It follows that U

is an open cover of A in (X, T). By hypothesis,

there is a nite subcover, U
1
, . . . , U
n
, say. But then G
1
, . . . , G
n
is an open
cover of A in (A, T
A
); that is, (A, T
A
) is compact.
Conversely, suppose that (A, T
A
) is compact. Let U

be an open cover
of A in (X, T). Set G

= AU

. Then G

is an open cover of (A, T

A
). By
hypothesis, there is a nite subcover, say, G
1
, . . . , G
m
. Clearly, U
1
, . . . , U
m
is an open cover for A in (X, T). That is, A is compact in (X, T).
8.4 Functional Analysis Gently Done Mathematics Department
Suppose that (X, T) is Hausdor, and let a
1
, a
2
be any two distinct points
of A. Then there is a pair of disjoint open sets U, V in X such that a
1
U
and a
2
V . Evidently, G
1
= AU and G
2
= AV are open in (A, T
A
), are
disjoint, and a
1
G
1
and a
2
G
2
. Hence (A, T
A
) is Hausdor, as claimed.
Remark 8.15. Note that it is quite possible for (A, T
A
) to be Hausdor
whilst (X, T) is not. A simple example is provided by example 3 above with
A given by A = 0, 1. In this case, the induced topology on A coincides
with the discrete topology on A.
Proposition 8.16. Let (X, T) be a Hausdor topological space and let
K X be compact. Then K is closed.
Proof. Let z XK. Then for each x K, there are open sets U
x
, V
x
such
that x U
x
, z V
x
and U
x
V
x
= . Evidently, U
x
: x K is an open
cover of K and therefore there is a nite number of points x
1
, x
2
, . . . , x
n
K
such that K U
x
1
U
x
n
.
Put V = V
x
1
V
x
n
. Then V is open, and z V . Furthermore,
V V
x
i
for each i implies that V U
x
i
= for 1 i n. Hence V K = ,
and therefore z V and V X K. Thus, X K is open, and K is closed.
Example 8.17. Let X = 0, 1, 2 and T = , X, 0, 1, 2 . Then the
set K = 2 is compact, but X K = 0, 1 is not an element of T. Thus,
K is not closed. As we have already noted, (X, T) is not Hausdor.
Proposition 8.18. Let (X, T) be a topological space and let K be compact.
Suppose that F is closed and F K. Then F is compact. (In other words,
closed subsets of compact sets are compact.)
Proof. Let U

: J be any given open cover of F. We augment this

collection by the open set X F. This gives an open cover of K;
K (X F)
_
_
J
U

_
.
Since K is compact, there are elements
1
,
2
, . . . ,
m
in J such that
K (X F) U

1
U

2
U

m
.
But then we see that
F U

1
U

2
U

m
.
We conclude that F is compact.
Kings College London Topological Spaces 8.5
We now consider continuity of mappings between topological spaces. The
denition is the obvious rewriting of the standard result from metric space
theory.
Denition 8.19. Let (X, T) and (Y, S) be topological spaces and suppose
that f : X Y is a given mapping. We say that f is continuous if and only
if f
1
(V ) T for any V S.
Many of the standard results concerning continuity in metric spaces have
analogues in this more general setting.
Theorem 8.20. Let (X, T) and (Y, S) be topological spaces with (X, T)
compact. Suppose that f : X Y is a continuous surjection. Then (Y, S) is
compact. (In other words, the image of a compact space under a continuous
mapping is compact.)
Proof. Let V

be any given open cover of Y . Then f

1
(V

) is an open
cover of X. Hence there are indices
1
,
2
, . . . ,
m
such that
X = f
1
(V

1
) f
1
(V

m
) .
Since f is onto, it follows that
Y = V

1
V

m
and so Y is compact.
Theorem 8.21. Let X be a compact topological space, Y a Hausdor
topological space and f : X Y a continuous injective surjection. Then
f
1
: Y X exists and is continuous (and so X is also Hausdor).
Proof. Clearly f
1
exists as a mapping from Y onto X. Let F be a closed
subset of X. To show that f
1
is continuous, it is enough to show that f(F)
is closed in Y . Now, F is compact in X and, as above, it follows that f(F)
is compact in Y . But Y is Hausdor and so f(F) is closed in Y .
Denition 8.22. A continuous bijection f : X Y , between topological
spaces (X, T) and (Y, S), with a continuous inverse is called a homeomor-
phism.
A homeomorphism f : X Y sets up a one-one correspondence between
the open sets in X and those in Y , via U f(U). The previous theorem
says that a continuous bijection from a compact space onto a Hausdor space
is a homeomorphism. It follows that both spaces are compact and Hausdor.
8.6 Functional Analysis Gently Done Mathematics Department
Denition 8.23. Let T
1
, T
2
be topologies on a set X. We say that T
1
is
weaker (or coarser or smaller) than T
2
if T
1
T
2
(alternatively, we say
that T
2
is stronger (or ner or larger) than T
1
).
The stronger (or ner) a topology the more open sets there are. It is
immediately clear that if f : (X, T) (Y, S) is continuous, then f is also
continuous with respect to any topology T

on X which is stronger than

T, or any topology S

on Y which is weaker than S. In particular, if X

has the discrete topology or Y has the indiscrete topology, then every map
f : X Y is continuous.
Let X be a given (non-empty) set, let (Y, S) a topological space and let
f : X Y be a given map. We wish to investigate topologies on X which
make f continuous. Now, if f is to be continuous, then f
1
(V ) should be
open in X for all V open in Y . Let T =

, where the intersection is over all

topologies T

on X which contain all the sets f

1
(V ), for V S. (The discrete
topology on X is one such.) Then T is a topology on X (any intersection of
topologies is also a topology). Moreover, T is evidently the weakest topology
on X with respect to which f is continuous. We can generalise this to
an arbitrary collection of functions. Suppose that (Y

, S

) : I is a
collection of topological spaces, indexed by I, and that F = f

: X Y
is a family of maps from X into the topological spaces (Y

, S

). Let T be the
intersection of all those topologies on X which contain all sets of the form
f
1

(V

), for f

F and V

. Then T is a toplogy on X and it is the

weakest topology on X with respect to which every f

F is continuous.
T is called the (X, F)-topology on X.
Theorem 8.24. Suppose that each (Y

0
, S

0
) is Hausdor and that F
separates points of X, i.e., if a, b X with a ,= b, then there is some f

F
such that f

(a) ,= f

(b). Then the (X, F)-topology is Hausdor.

Proof. Suppose that a, b X, with a ,= b. Then, by hypothesis, there is
some I such that f

(a) ,= f

(b). Since (Y , S

) is Hausdor, there exist

elements U, V S

such that f

(a) U, f

(b) V and U V = . But

then f
1

(U) and f
1

(V ) are open with respect to the (X, F)-topology and

a f
1

(U), b f
1

(V ) and f
1

(U) f
1

(V ) = .
To describe the (X, F)-topology somewhat more explicitly, it is conve-
nient to introduce some terminology.
Denition 8.25. A collection B of open sets is said to be a base for the
topology T on a space X if and only if each element of T can be written as
a union of elements of B.
Kings College London Topological Spaces 8.7
Examples 8.26.
1. The open sets x : d(x, a) < r, a X, r Q, r > 0, are a base for the
usual topology in a metric space (X, d).
2. The rectangles (x, y) R
2
: [x a[ <
1
n
, [y b[ <
1
m
, with (a, b) R
2
and n, m N, form a base for the usual Euclidean topology on R
2
.
3. The singleton sets x, x X, form a base for the discrete topology on
any non-empty set X.
Proposition 8.27. The collection of open sets B is a base for the topology
T on a space X if and only if for each non-empty set G T and x G there
is some B B such that x B and B G.
Proof. Suppose that B is a base for the topology T and suppose that G T
is non-empty. Then G can be written as a union of elements of B. In
particular, for any x G, there is some B B such that x B and B G.
Conversely, suppose that for any non-empty set G T and for any x G,
there is some B
x
Bsuch that x B
x
and B
x
G. Then G

x GB
x

G, which shows that B is a base for T.
Denition 8.28. A collection S of subsets of a topology T on X is said
to be a sub-base for T if and only if the collection of intersections of nite
families of members of S is a base for T.
Example 8.29. The collection of subsets of R consisting of those intervals
of the form (a, ) or (, b), a, b R, is a sub-base for the usual topology
on R.
Proposition 8.30. Let X be any non-empty set and let S be any collection
of subsets of X which covers X, i.e., for any x X, there is some A S
such that x A. Let B be the collection of intersections of nite families of
elements of S. Then the collection T of subsets of X consisting of together
with arbitrary unions of elements of members of B is a topology on X, and
is the weakest topology on X containing the collection of sets S. Moreover,
S is a sub-base for T, and B is a base for T.
Proof. Clearly, T and X T, and any union of elements of T is also a
member of T. It remains to show that any nite intersection of elements of T
is also an element of T. It is enough to show that if A, B T, then AB T.
If A or B is the empty set, there is nothing more to prove, so suppose that
8.8 Functional Analysis Gently Done Mathematics Department
A ,= and B ,= . Then we have that A =

and B =

for
families of elements A

and B

belonging to B. Thus
A B =
_

=
_
,
(A

) .
Now, each A

is an intersection of a nite number of elements of S, and the

same is true of B

. It follows that the same is true of every A

, and so
we see that AB T, which completes the proof that T is a topology on X.
If T

is any topology on X which contains the collection S, then certainly

T

must also contain B. But then T

must contain arbitrary unions of families

of subsets of B, that is, T

must contain T. It follows that T is the weakest

topology on X containing S. From the denitions, it is clear that S is a
sub-base for T and that B is a base for T.
Remark 8.31. The topology T can also be described as follows: a non-
empty set G belongs to T if and only if for any x G there is some S B
such that x S and S G. Indeed, if G has this property then it is clearly
a union of members of S. Conversely, if G is a union of elements S

, say, of
S and x G, then certainly there is some
0
such that x S

0
. Evidently,
we also have that S

0
G.
Remark 8.32. We can therefore describe the (X, F)-topology on X deter-
mined by the family of maps f

: I, discussed earlier, as the topology

with sub-base given by the collection f
1

(V ) : I, V S

.
9. Weak and weak

-topologies
Consider a normed space X with dual space X

. In particular, these are

both topological spaces with respect to the topologies induced by the norms.
We wish to consider topologies dierent from these norm topologies. First
we will see how we can use X

to dene a topology on X. The dual, X

, of X
is a collection of (all bounded linear) maps from X into the same topological
space, C. Thus, we can consider the (X, X

)-topology on Xcalled the

weak topology on the normed space X. It is Hausdor since X

separates
the points of X.
The weak topology on X is therefore the weakest topology on X making
every member of the dual, X

, continuous. A non-empty set G in X is open

with respect to the weak topology if and only if for each point a G there
is B such that a B and B G, and where B has the form
B =
1
1
(U
1
)
1
n
(U
n
)
for some n N,
1
, . . . ,
n
X

and U
1
, . . . , U
n
open sets in C. (This is just
the statement that the sets of the form (U), with X

and U open in
C form a sub-base.) Now, a
1
j
(U
j
) is equivalent to
j
(a) U
j
, and if
U
j
is open in C, there is some
j
> 0 such that the open ball B(
j
(a);
j
) is
contained in U
j
. It follows that the set G is open with respect to the weak
topology if and only if there is B as above but of the form
B =
1
1
(B(
1
(a);
1
))
1
n
(B(
n
(a);
n
))
By taking = min
1
, . . . ,
n
, this is equivalent to the existence of some B
as above of the form
B =
1
1
(B(
1
(a); ))
1
n
(B(
n
(a); )) .
However,

1
j
(B(
j
(a); )) = x X : [
j
(x)
j
(a)[ <
1
9.2 Functional Analysis Gently Done Mathematics Department
so that

1
1
(B(
1
(a); ))
1
n
(B(
n
(a); ))
= x X : [
j
(x)
j
(a)[ < , 1 j n .
Finally we arrive at the following characterisation of the weak topology on
the normed space X. A non-empty subset G of X is open with respect to the
weak topology if and only if for any a G there is n N,
1
, . . . ,
n
X

and > 0 such that

x X : [
j
(x)
j
(a)[ < , 1 j n G.
Equivalently, we can say that a set is open with respect to the weak topology
if and only if it is a union of sets of the above form. We introduce the following
notation; for given
1
, . . . ,
n
X

and > 0, we write

N(a;
1
, . . . ,
n
, ) = x X : [
j
(x)
j
(a)[ < 1 j n .
Then each N(a;
1
, . . . ,
n
, ) contains a and is (X, X

) open. A non-
empty subset G in X is weakly open if and only if for any a G we have
N(a;
1
, . . . ,
n
, ) G for some n N,
1
, . . . ,
n
X

and > 0. The sets

N(a;
1
, . . . ,
n
, ) play a role analogous to that of the open balls in a metric
space.
Denition 9.1. Let x be a point in a topological space (X, T). A collection
^
x
of neighbourhoods of x is said to be a neighbourhood base at x if and
only if for each open set G with x G there is some N ^
x
such that
x N and N G. (Note that the members of ^
x
need not themselves be
open sets.)
Thus, we see that the family
^
a
= N(a;
1
, . . . ,
n
, ) : n N, > 0,
1
, . . . ,
n
X
is an open neighbourhood base at a for the weak topology.
In order to discuss the relationship between the norm and the weak
topologies on X, we shall need the following result.
Kings College London Weak and weak

-topologies 9.3
Proposition 9.2. Let
1
, . . . ,
n
be bounded linear functionals on an
innite dimensional normed space X. Then there exists x X such that
x ,= 0 and
1
(x) = =
n
(x) = 0.
Proof. First we note that X

is innite dimensional. In fact, if X

were
nite dimensional, then its dual, X

, would also be nite dimensional. But

we have seen that X is isometrically isomorphic to a subspace of X

. Since
X is innite dimensional the same must be true of X

and hence also of

X

.
Let
1
, . . . ,
m
be linearly independent elements spanning the nite di-
mensional subspace generated by
1
, . . . ,
n
. Then m n. Since X

is
innite dimensional, there is X

independent of
1
, . . . ,
m
. Dene
T : X C
m+1
by
Tx = ((x),
1
(x), . . . ,
m
(x)) .
It is clear that T is a linear operator and so its range ran T is a linear subspace
of C
m+1
. If ran T ,= C
m+1
, there would exist a vector = (
0
,
1
, . . . ,
m
)
in C
m+1
orthogonal to the range of T, that is,

0
(x) +
1

1
(x) + +
m

m
(x) = 0
for all x X. But this is precisely the statement that

0
+
1

1
+ +
m

m
= 0
in X

. This contradicts the assumed linear independence of ,

1
, . . . ,
m
.
It follows that ran T = C
m+1
. In particular, there is x X such that
Tx = (1, 0, 0, . . . , 0). That is, there is x X such that (x) = 1 and

j
(x) = 0 for all 1 j m. Hence x ,= 0 and also
i
(x) = 0 for 1 i n
and the proof is complete.
Theorem 9.3. The weak topology on a normed space is weaker than the
norm topology. If X is innite dimensional, then the weak topology on X is
strictly weaker than the norm topology on X.
Proof. Every element of X

is continuous when X is equipped with the

norm topology. The weak topology is the weakest topology on X with this
property, so it is immediately clear that the weak topology is weaker than
the norm topology.
9.4 Functional Analysis Gently Done Mathematics Department
Now suppose that X is innite-dimensional. We shall exhibit a set which
is open with respect to the norm topology but not with respect to the weak
topology. We consider the open unit ball
G = x X : |x| < 1 .
Clearly G is open with respect to the norm topology on X. We claim that
G is not weakly open. If, on the contrary, G were weakly open, then, since
0 G, there would be n N,
1
, . . . ,
n
X

and > 0 such that

N(0;
1
, . . . ,
n
, ) G.
By the previous proposition, there is x X with x ,= 0 such that
i
(x) = 0
for all 1 i n. Put y = 2x/|x|. Then |y| = 2 and
i
(y) = 0 for
all 1 i n. It follows that y N(0;
1
, . . . ,
n
, ) but y / G. Hence
N(0;
1
, . . . ,
n
, ) cannot be contained in G, which is a contradiction. We
conclude that G is not weakly open and therefore the weak topology on X
is strictly weaker than the norm topology.
Remark 9.4. If we put y = kx/|x| in the argument above, we see that
y N(0;
1
, . . . ,
n
, ) and |y| = k. It follows that no weak neighbourhood
N(0;
1
, . . . ,
n
, ) can be norm bounded. Moreover, for any a X, we have
N(a;
1
, . . . ,
n
, ) = a + N(0;
1
, . . . ,
n
, ) and so none of the non-empty
weakly open sets can be norm bounded.
We now turn to a discussion of a topology on X

, the dual of the normed

space X. The idea is to consider X as a family of maps : X

C given by
x : (x), for x X and X

this is the map

x
dened earlier.
Denition 9.5. The weak

-topology on X

, the dual of the normed space

X, is the (X

, X)-topology, where X is considered as a collection of maps

from X

C as above.
Remark 9.6. The weak

-topology is also called the w

-topology on X

.
Since X separates points of X

, the w

-topology is Hausdor. In view of the

identication of X as a subset of X

, we see that the w

-topology is weaker
than the (X

, X

)-topology on X

. Of course, we have equality if X is

reexive. The converse is also true.
Theorem 9.7. The normed space X is reexive if and only if the weak
and weak

topologies on X

coincide.
Proof. We will not do this here.
Kings College London Weak and weak

-topologies 9.5
By repeating our earlier analysis, we see that a non-empty subset G X

is w

-open if and only if for each G there is m N, elements x

1
, . . . , x
m
in X and > 0 such that
N(; x
1
, . . . , x
m
, )

= X

: [(x
i
) (x
i
)[ < , 1 ilem G.
An open neighbourhood base at 0 for the w

-topology on X

is given by
^
0
= N(0; x
1
, . . . , x
m
, ) : m N, x
1
, . . . , x
m
X, > 0 .
An open neigbourhood base at X

is given by
^

= N + : N ^
0
.
One can show the following.
Theorem 9.8. Let X be a normed space and Y a family of bounded linear
functionals on X which separates points of X. Then the bounded linear
functional on X is (X, Y )-continuous if and only if Y . In particular,
the only w

-continuous linear functionals on X

are the elements of X.

A subset K of a metric space is compact if and only if any sequence in
K has a subsequence which converges to an element of K, but this need no
longer be true in a topological space. We will see an example of this later.
We have seen that in a general topological space nets can be used rather
than sequences, so the natural question is whether there is a sensible notion
of that of subnet of a net, generalising that of subsequence of a sequence.
Now, a subsequence of a sequence is obtained simply by leaving out various
termsthe sequence is labelled by the natural numbers and the subsequence
is labelled by a subset of these. The notion of a subnet is somewhat more
subtle than this.
Denition 9.9. A map F : J I between directed sets I and J is said to
be conal if for any I there is some

J such that J() _ whenever

_

. In other words, F is eventually greater than any given I.

Suppose that (x

)
I
is a net indexed by I and that F : J I is a conal
map from the directed set J into I. The net (y

)
J
= (x
F()
)
J
is said to
be a subnet of the net (x

)
I
.
It is important to notice that there is no requirement that the index set
for the subnet be the same as that of the original net.
9.6 Functional Analysis Gently Done Mathematics Department
Example 9.10. If we set I = J = N, equipped with the usual ordering,
and let F : J I be any increasing map, then the subnet (y
n
) = (x
F(n)
) is
a subsequence of the sequence (x
n
).
Example 9.11. Let I = N with the usual order, and let J = N equipped
with the usual ordering on the even and odd elements separately but where
any even number is declared to be greater than any odd number. Thus I
and J are directed sets. Dene F : J I by F() = 3. Let I
be given. Set

= 2 so that if _

in J, we must have that is

even and greater than

in the usual sense. Hence F() = 3

= 2 in I and so F is conal. Let (x

n
)
nI
be any sequence of
real numbers, say. Then (x
F(m)
)
mJ
= (x
3m
)
mJ
is a subnet of (x
n
)
nI
.
It is not a subsequence because the ordering of the index set is not the
usual one. Suppose that x
2k
= 0 and x
2k1
= 2k 1 for k I = N.
Then (x
n
) is the sequence (1, 0, 3, 0, 5, 0, 7, 0 . . . ). The subsequence (x
3m
)
mN
is (3, 0, 9, 0, 15, 0, . . . ) which clearly does not converge in R. However, the
subnet (x
3m
)
mJ
does converge, to 0. Indeed, for m _ 2 in J, we have
x
3m
= 0.
Proposition 9.12. Let (x

)
I
be a net in the space X and let A be a family
of subsets of X such that
(i) (x

)
I
is frequently in each member of A;
(ii) for any A, B A there is C A such that C A B.
Then there is a subnet (x
F()
)
J
of the net (x

)
I
such that (x
F()
)
J
is even-
tually in each member of A.
Proof. Equip A with the ordering given by reverse inclusion, that is, we
dene A _ B to mean B A for A, B A. For any A, B A, there is
C A with C A B, by (ii). This means that C _ A and C _ B and we
see that A is directed with respect to this partial ordering.
Let E denote the collection of pairs (, A) I A such that x

A;
E = (, A) : I, A A, x

A .
Dene (

, A

) _ (

, A

) to mean that

in I and A

_ A

in A.
Then _ is a partial order on E. Furthermore, for given (

, A

), (

, A

)
in E, there is I with _

and _

, and there is A A such

that A _ A

and A _ A

. But (x

) is frequently in A, by (i), and therefore

there is _ I such that x

A. Thus (, A) E and (, A) _ (, A

),
(, A) _ (, A

) and it follows that E is directed. E will be the index set for

the subnet.
Kings College London Weak and weak

-topologies 9.7
Next, we must construct a conal map from E to I. Dene F : E I
by F((, A)) = . To show that F is conal, let
0
I be given. For any
A A there is _
0
such that x

A (since (x

) is frequently in each
A A. Hence (, A) E and F((, A)) = _
0
. So if (

, A

) _ (, A) in
E, then we have
F((

, A

)) =

_ _
0
.
This shows that F is conal and therefore (x
F((,A))
)
E
is a subnet of (x

)
I
.
It remains to show that this subnet is eventually in every member of A.
Let A A be given. Then there is I such that x

A and so (, A) E.
For any (

, A

) E with (

, A

) _ (, A), we have
x
F((

,A

))
= x

A.
Thus (x
F((,A))
)
E
is eventually in A.
Theorem 9.13. A point x in a topological space X is a cluster point of
the net (x

)
I
if and only if some subnet converges to x.
Proof. Suppose that x is a cluster point of the net (x

)
I
and let N be the
family of neighbourhoods of x. Then if A, B N, we have A B N, and
also (x

) is frequently in each member of N. By the preceding proposition,

there is a subnet (y

)
J
eventually in each member of N, that is, the subnet
(y

) converges to x.
Conversely, suppose that (y

)
J
= (x
F()
)
J
is a subnet of (x

)
I
con-
verging to x. We must show that x is a cluster point of (x

)
I
. Let N be any
neighbourhood of x. Then there is
0
J such that x
F()
N whenever
_
0
. Since F is conal, for any given

I there is

J such that
F() _

whenever _

. Let _
0
and _

. Then F() _

and
y

= x
F()
N. Hence (x

)
I
is frequently in N and we conclude that x is
a cluster point of the net (x

)
I
, as claimed.
In a metric space, compactness is equivalent to sequential compactness
( the Bolzano-Weierstrass property). In a general topological space, this
need no longer be the case. However, there is an analogue in terms of nets.
9.8 Functional Analysis Gently Done Mathematics Department
Theorem 9.14. A topological space (X, T) is compact if and only if every
net in X has a convergent subnet.
Proof. Suppose that every net has a convergent subnet. Let G

I
be an
open cover of X with no nite subcover. Let F be the collection of nite
subfamilies of the open cover, ordered by set-theoretic inclusion. For each
F = G

1
, . . . , G

m
F, let x
F
be any point in X such that x
F
/

m
j=1
G

j
.
Note that such x
F
exists since G

has no nite subcover. By hypothesis,

the net (x
F
)
FF
has a convergent subnet or, equivalently, by the previous
theorem, a cluster point x, say. Now, since G is a cover of X, there is
some

such that x G

. But then, by denition of cluster point, (x

F
)
F
is frequently in G

. Thus, for any F

F, there is F _ F

F such that
x
F
G

. In particular, if we take F

= G

, we deduce that there is

F = G

1
, . . . , G

k
such that F _ G

, that is, G

F, and such
that x
F
G

. Hence G

= G

i
for some 1 i k, and
x
F
G

k
_
j=1
G

j
.
But x
F
/

k
j=1
G

j
, by construction. This contradiction implies that every
open cover has a nite subcover, and so (X, T) is compact.
For the converse, suppose that (X, T) is compact and let (x

)
I
be a net
in X. Suppose that (x

)
I
has no cluster points. Then, for any x X, there
is an open neighbourhood U
x
of x and
x
I such that x

/ U
x
whenever
_
x
. The family U
x
: x X is an open cover of X and so there
exists x
1
, . . . , x
n
X such that

n
i=1
U
x
i
= X. Since I is directed there is
_
i
for each i = 1, . . . , n. But then x

/ U
x
i
for all i = 1, . . . , n, which
is impossible since the U
x
i
s cover X. We conclude that (x

)
I
has a cluster
point, or, equivalently, a convergent subnet.
Denition 9.15. A universal net in a topological space (X, T) is a net with
the property that, for any subset A of X, it is either eventually in A or
eventually in X A, the complement of A.
The concept of a universal net leads to substantial simplication of the
proofs of various results, as we will see.
Kings College London Weak and weak

-topologies 9.9
Proposition 9.16. If a universal net has a cluster point, then it converges
(to the cluster point). In particular, a universal net in a Hausdor space can
have at most one cluster point.
Proof. Suppose that x is a cluster point of the universal net (x

)
I
. Then
for each neighbourhood N of x, (x

) is frequently in N. However, (x

) is
either eventually in N or eventually in X N. Evidently, the former must
be the case and we conclude that (x

) converges to x. The last part follows

because in a Hausdor space a net can converge to at most one point.
At this point, it is not at all clear that universal nets exist!
Examples 9.17.
1. It is clear that any eventually constant net is a universal net. In particular,
any net with nite index set is a universal net. Indeed, if (x

)
I
is a net in X
with nite index set I, then I has a maximum element,

, say. The net is

therefore eventually equal to x

. For any subset A X, we have that (x

)
I
is eventually in A or eventually in X A depending on whether x

belongs
to A or not.
2. No sequence can be a universal net, unless it is eventually constant. To
see this, suppose that (x
n
)
nN
is a sequence which is not eventually constant.
Then the set S = x
n
: n N is an innite set. Let A be any innite subset
of S such that S A also innite. Then (x
n
) cannot be eventually in either
of A or its complement. That is, the sequence (x
n
)
N
cannot be universal.
We shall show that every net has a universal subnet. First we need the
following lemma.
Lemma 9.18. Let (x

)
I
be a net in a topological space X. Then there is
a family ( of subsets of X such that
(i) (x

) is frequently in each member of (;

(ii) if A, B ( then A B (;
(iii) for any A X, either A ( or X A (.
Proof. Let denote the collection of families of subsets of X satisfying the
conditions (i) and (ii):
= T : T satises (i) and (ii).
Evidently X so ,= . The collection is partially ordered by set
inclusion:
T
1
_ T
2
if and only if T
1
T
2
, for T
1
, T
2
.
9.10 Functional Analysis Gently Done Mathematics Department
Let T

be a totally ordered family in , and put

T =

. We shall
show that

T . Indeed, if A

T, then there is some such that A T

,
and so (x

) is frequently in A and condition (i) holds.

Now, for any A, B

T, there is
1
and
2
such that A T

1
, and B
T

2
. Suppose, without loss of generality, that T

1
_ T

2
. Then A, B T

2
and therefore A B T

2


T, and we see that condition (ii) is satised.
Thus

T as claimed.
By Zorns lemma, we conclude that has a maximal element, (, say. We
shall show that ( also satises condition (iii).
To see this, let A X be given. Suppose, rst, that it is true that (x

)
is frequently in A B for all B (. Dene T

by
T

= C X : A B C, for some B ( .
Then C T

implies that A B C for some B in ( and so (x

) is
frequently in C. Also, if C
1
, C
2
T

, then there is B
1
and B
2
in ( such that
A B
1
C
1
and A B
2
C
2
. It follows that A (B
1
B
2
) C
1
C
2
.
Since B
1
B
2
(, we deduce that C
1
C
2
T

. Thus T

.
However, it is clear that A T

and also that if B ( then B T

.
But ( is maximal in , and so T

= ( and we conclude that A (, and (iii)

holds.
Now suppose that it is false that (x

) is frequently in every A B, for

B (. Then there is some B
0
( such that (x

) is not frequently in AB
0
.
Thus there is
0
such that x

X (A B
0
) for all _
0
. That is, (x

)
is eventually in X (A B
0
)

A, say. It follows that (x

) is frequently in

AB for every B (. Thus, as above, we deduce that

A (. Furthermore,
for any B (, B B
0
( and so

A B B
0
(. But

A B B
0
= (X (A B
0
)) (B B
0
)
= ((X A) (X B
0
)) B B
0
= (X A) B B
0
(X B
0
) B B
0

. .
=
= (X A) B B
0
and so we see that (x

) is frequently in (X A) B B
0
and hence is
frequently in (X A) B for any B (. Again, by the above argument, we
deduce that X A (. This proves the claim and completes the proof of
the lemma.
Kings College London Weak and weak

-topologies 9.11
Theorem 9.19. Every net has a universal subnet.
Proof. To prove the theorem, let (x

)
I
be any net in X, and let ( be a
family of subsets as given by the lemma. Then, in particular, the conditions
of Proposition 9.12 hold, and we deduce that (x

)
I
has a subnet (y

)
J
such
that (y

)
J
is eventually in each member of (. But, for any A X, either
A ( or X A (, hence the subnet (y

)
J
is either eventually in A or
eventually in X A; that is, (y

)
J
is universal.
Theorem 9.20. A topological space is compact if and only if every
universal net converges.
Proof. Suppose that (X, T) is a compact topological space and that (x

)
is a universal net in X. Since X is compact, (x

) has a convergent subnet,

with limit x X, say. But then x is a cluster point of the universal net (x

)
and therefore the net (x

) itself converges to x.
Conversely, suppose that every universal net in X converges. Let (x

) be
any net in X. Then (x

) has a subnet which is universal and must therefore

converge. In other words, we have argued that (x

) has a convergent subnet

and therefore X is compact.
Corollary 9.21. A non-empty subset K of a topological space is compact
if and only if every universal net in K converges in K.
Proof. The subset K of the topological space (X, T) is compact if and only
if it is compact with respect to the induced topology T
K
on K. The result
now follows by applying the theorem to (K, T
K
).
10. Product Spaces
Suppose that (X
1
, T
1
) and (X
2
, T
2
) are topological spaces and consider
the cartesian product
Y = X
1
X
2
= (x
1
, x
2
) : x
1
X
1
, x
2
X
2
.
We would like to give Y a topology using those on X
1
and X
2
. Dene T to
be the collection of subsets of Y give as follows: T and the non-empty
set G belongs to T if and only if for any point (a
1
, a
2
) G there are open sets
U T
1
and V T
2
such that (a
1
, a
2
) U V . In other words, G belongs
to T if and only if each of its points is contained in an open rectangle
U V which itself lies in G. One checks that T is indeed a topology
called the product topology on the cartesian product X
1
X
2
given by T
1
and T
2
. A set is open (with respect to T) if and only if it is a union of sets
of the form U V ; that is, the sets of the form U V form a base for the
product topology. Since U V = U X
2
X
1
V , we see that the sets
U X
2
, X
1
V : U T
1
, V T
2
constitute a sub-base for the topology
T.
Example 10.1. Let (X
1
, T
1
) = (X
2
, T
2
) = (R, T
Euclidean
). Then the carte-
sian product X
1
X
2
is just R
2
, and we see that a non-empty set G R
2
is
open with respect to the product topology if and only if each of its points is
contained in an open rectangle also lying in G. Thus, the product topology
is precisely the usual Euclidean topology on R
2
. Rectangles are as good as
discs.
The projection maps, p
1
and p
2
, on the cartesian product X
1
X
2
, are
dened by
p
1
: X
1
X
2
X
1
, (x
1
, x
2
) x
1
p
2
: X
1
X
2
X
1
, (x
1
, x
2
) x
2
.
1
10.2 Functional Analysis Gently Done Mathematics Department
For any open set U X
1
(that is, U T
1
), we have p
1
1
(U) = UX
2
T and
so it follows that p
1
: X
1
X
2
X
1
is continuous. Similarly, p
2
: X
1
X
2

X
2
is continuous. This property characterises the product topology as we
now show.
Proposition 10.2. The product topology T is the weakest topology on the
cartesian product X
1
X
2
such that both p
1
and p
2
are continuous.
Proof. Suppose that S is a topology on Y = X
1
X
2
with respect to which
both p
1
and p
2
are continuous. Then for any U T
1
, p
1
1
(U) S. But
p
1
1
(U) = U X
2
and so U X
2
S for all U T
1
. Similarly, X
1
V S
for all V T
2
. Since these sets form a sub-base for T we deduce that T S,
as required.
We would like to generalise this to an arbitrary cartesian product of
topological spaces. Let (X

, T

) : I be a collection of topological
spaces indexed by the set I. We recall that X =

, the cartesian
product of the X

s, is dened to be the collection of maps from I into the

union

satisfying () X

for each I. We can think of the value

() as the
th
coordinate of the point in X. The idea is to construct
a topology on X =

built from the individual topologies T

. Two
possibilities suggest themselves. The rst is to construct a topology on X
such that it is the weakest topology with respect to which all the projection
maps p

are continuous. The second is to construct the topology on

X whose open sets are unions of super rectangles, that is, sets of the form

, where U

for every I.
In general, these two topologies are not the same, as we will see. Consider
the rst construction. We wish to dene a topology T on X making every
projection map p

continuous. This means that T must contain all sets of

the form p
1

(U

), for U

, and also nite intersections of such sets,

and also arbitrary unions of such nite intersections. So we dene T to be
the topology on X with sub-base given by the sets p
1

(U

), for U

.
For reasons we will discuss later, this topology turns out to be the more
appropriate and is taken as the denition of the product topology.
Denition 10.3. The product topology on the cartesian product of the
topological spaces (X

, T

) : I is that with sub-base given by the sets

p
1

(U

), for U

.
Clearly, this agrees with the case discussed earlier for the product of
just two spaces. Moreover, this denition is precisely the statement that
T is the (

, F)topology, where F is the family of projection maps

Kings College London Product Spaces 10.3
p

: I. In other words, T is the weakest topology on

with
respect to which every projection map p

is continuous.
Remark 10.4. Let G be a non-empty open set in X, equipped with the
product topology, and let G. Then, by denition of the topology, there
exist
1
, . . . ,
n
I and open sets U

i
in X

i
, 1 i n, such that
p
1

1
(U

1
) p
1

1
(U

1
) G.
Hence there are open sets S

, I, such that

G and where all

but a nite number of the S

are equal to the whole space X

. This means
that G can dier from X in at most a nite number of components.
Now let us consider the second candidate for a topology on X. Let S be
the topology on X with base given by the sets of the form

, where
V

for I. Thus, a non-empty set G in X belongs to S if and only

if for any point x in G there exist V

such that
x

G.
Here there is no requirement that all but a nite number of the V

are equal
to X

.
Denition 10.5. The topology on the cartesian product

constructed
in this way is called the box-topology on X.
Evidently, in general, S is strictly ner than the product topology T.
We shall use the notation T
product
and T
box
for the product and box
topologies, respectively.
Proposition 10.6. A net (x

) converges in (

, T
product
) if and only
if (p

(x

)) converges in (X

, T

) for each I.
Proof. Suppose that x

x in (

, T
product
). Then p

(x

) p

(x)
for each , since p

is continuous.
Converesely, suppose that p

(x

) z

in (X

, T

) for each I. We
shall show that x

z in (

, T
product
) where z is given by z() = z

.
Indeed, let G be any neighbourhood of z in (

, T
product
). Then there
are
1
, . . . ,
n
and open sets U

1
, . . . , U

n
such that
z p
1

1
(U

1
) p
1

n
(U

n
) G.
Now, p

(x

) z

, for each , so there is

j
such that p

j
(x

) U

j
whenever _
j
, 1 j n. Let

_
j
for all 1 j n. Then if _

,
10.4 Functional Analysis Gently Done Mathematics Department
we have p

j
(x

) U

j
and so x

p
1

j
(U

j
) for all 1 j n. In other
words, for _

,
x

p
1

1
(U

1
) p
1

n
(U

n
) G.
Hence x

z in (

, T
product
) as required.
Example 10.7. Let I = N, let X

be the open interval (2, 2) for each

N, and let T

be the usual (Euclidean) topology on X

. Let x
n

k
X
k
be the element x
n
= (
1
n
,
1
n
,
1
n
, . . . ); that is, p
k
(x
n
) =
1
n
for all k I = N.
Clearly p
k
(x
n
) 0 as n , for each k and so the sequence (x
n
) converges
to z in (

k
X
k
, T
product
) where z is given by p
k
(z) = 0 for all k.
However, (x
n
) does not converge to z with respect to the box-topology.
Indeed, to see this, let G =

k
A
k
where A
k
is the open set A
k
= (
1
k
,
1
k
)
T
k
. Then G is open with respect to the box-topology and is a neighbourhood
of z but x
n
/ G for any n N. It follows that, in fact, (x
n
) does not converge
at all with respect to the box-topology (if it did, then the limit would have
to be the same as that for the product topology, namely z).
Remark 10.8. This is a rst indication that the box-topology may not be
very useful (apart from being a possible source of counter-examples).
Suppose that each (X

, T

), I, is compact. What can be said about

the product space

with respect to the product and the box topologies?

Example 10.9. Let I = N and let X
k
= [0, 1] for each k I = N and
equip each X
k
with the Euclidean topology. Then each (X
k
, T
k
) is compact.
However, the product space

k
X
k
is not compact with respect to the box-
topology. To see this, let I
k
(t) be the open disk in X
k
with centre t and
radius
1
k
;
I
k
(t) = [0, 1]
_
t
1
k
, t +
1
k
_
[0, 1] .
Evidently, the diameter of I
k
(t) is at most
2
k
. For each x

k
X
k
let G
x
be
the set
G
x
=

k
I
k
(x(k))
so G
x
is the product of the open sets I
k
(x(k)), each centred on the k
th
component of x and with diameter at most
2
k
. The set G
x
is open with respect
to the box-topology and can be pictured as an ever narrowing tube centred
on x = (x(k)).
Kings College London Product Spaces 10.5
Clearly, G
x
: x

k
X
k
is an open cover of

k
X
k
(for the box-
topology). We shall argue that this cover has no nite subcover this
because the tails of the G
x
s become too narrow. Indeed, for any points
x
1
, . . . , x
n
in

k
X
k
, and any m N, we have
p
m
(G
x
1
G
x
n
) = I
m
(x
1
(m)) I
m
(x
n
(m)) .
Each of the n intervals I
m
(x
j
(m)) has diameter not greater than
2
m
, so any
interval covered by their union cannot have length greater than
2n
m
. If we
choose m > 3n, then this union cannot cover any interval of length greater
than
2
3
, and in particular, it cannot cover X
m
. It follows that G
x
1
, . . . , G
x
n
is
not a cover for

k
X
k
and, consequently,

k
X
k
is not compact with respect
to the box-topology.
That this behaviour cannot occur with the product topology this being
the content of Tychonovs theorem which shall now discuss. It is convenient
to rst prove a result on the existence of a certain family of sets satisfying
the nite intersection property (p).
Proposition 10.10. Suppose that F is any collection of subsets of a given
set X satisfying the p. Then there is a maximal collection D containing F
and satisfying the p, i.e., if F F

and if F

satises the p, then F

D.
Furthermore;
(i) if A
1
, . . . , A
n
D, then A
1
A
n
D;
(ii) if A is any subset of X such that A D ,= for all D D, then A D.
Proof. As might be expected, we shall use Zorns lemma. Let ( denote
the collection of those families of subsets of X which contain F and satisfy
the p. Then F (, so ( is not empty. Evidently, ( is ordered by set-
theoretic inclusion. Suppose that is a totally ordered set of families in (.
Let A =
_
S
S. Then F A, since F S, for all S . We shall show
that A satises the p. To see this, let S
1
, . . . , S
n
A. Then each S
i
is an
element of some family S
i
that belongs to . But is totally ordered and so
there is i
0
such that S
i
S
i
0
for all 1 i n. Hence S
1
, . . . , S
n
S
i
0
and
so S
1
S
n
,= since S
i
0
satises the p. It follows that A is an upper
bound for in (. Hence, by Zorns lemma, ( contains a maximal element,
D, say.
(i) Now suppose that A
1
, . . . , A
n
D and let B = A
1
A
n
. Let
D

= D B. Then any nite intersection of members of D

is equal to
a nite intersection of members of D. Thus D

satises the p. Clearly,

F D

, and so, by maximality, we deduce that D

= D. Thus B D.
10.6 Functional Analysis Gently Done Mathematics Department
(ii) Suppose that A X and that A D ,= for every D D. Let
D

= D A, and let D
1
, . . . , D
m
D

. If D
i
D, for all 1 i m, then
D
1
D
m
,= since Dsatises the p. If some D
i
= A and some D
j
,= A,
then D
1
D
m
has the form D
1
D
k
A with D
1
, . . . , D
k
D.
By (i), D
1
D
k
D and so, by hypothesis, A (D
1
. . . D
k
) ,= .
Hence D

satises the p and, again by maximality, we have D

= D and
thus A D.
We are now ready to prove Tychonovs theorem. In fact we shall present
three proofs. The rst is based on the previous proposition, the second uses
the idea of partial cluster points together with Zorns lemma, and the third
uses universal nets.
Theorem 10.11. (Tychonovs Theorem) Let (X

, T

) : I be any
given collection of compact topological spaces. Then the Cartesian product
(

, T
product
) is compact.
Proof. (1st proof) Let F be any family of closed subsets of

satisfying
the p. We must show that

FF
,= . By the previous proposition, there
is a maximal family D of subsets of

satisfying the p and with F D.

(Note that the mebers of D need not all be closed sets.)
For each I, consider the family p

(D) : D D. Then this family

satises the p because D does. Hence p

(D) : D D satises the p.

But this is a collection of closed sets in the compact space (X

, T

), and so

DD
p

(D) ,= .
That is, there is some x

such that x

(D) for every D D.

Let x

be given by p

(x) = x

, i.e., the
th
coordinate of x is
x

. Now, for any I, and for any D D, x

(D) implies that for any

neighbourhood U

of x

we have U

(D) ,= . Hence p
1

(U

) D ,=
for every D D. By the previous proposition, it follows that p
1

(U

)
D. Hence, again by the previous proposition, for any
1
, . . . ,
n
I and
neighbourhoods U

1
, . . . , U

n
of x

1
, . . . x

n
, respectively,
p
1

1
(U

1
) p
1

n
(U

n
) D.
Furthermore, since D has the p, we have that
p
1

1
(U

1
) p
1

n
(U

n
) D ,=
Kings College London Product Spaces 10.7
for every D D, every nite family
1
, . . . ,
n
I and neighbourhoods
U

1
, . . . , U

n
of x

1
, . . . x

n
, respectively.
We shall show that x D for every D D. To see this, let G be any
neighbourhood of x. Then, by denition of the product topology, there is a
nite family
1
, . . . ,
m
I and open sets U

1
, . . . , U

m
such that
x p
1

1
(U

1
) p
1

m
(U

m
) G.
But we have shown that for any D D,
D p
1

1
(U

1
) p
1

m
(U

m
) ,=
and therefore D G ,= . We deduce that x D, the closure of D, for any
D D. In particular, x F = F for every F F. Thus

FF
F ,=
it contains x. The result follows.
Proof. (2nd proof) Let (

)
A
be any given net in X =

iI
X
i
. We shall
show that (

) has a cluster point. For each i I, (

(i)) is a net in the

compact space X
i
and therefore has a cluster point z
i
, say, in X
i
. However,
the element X given by (i) = z
i
need not be a cluster point of (

).
(For example, let I = 1, 2, X
1
= X
2
= [1, 1] with the usual topology and
let (
n
) be the sequence ((x
n
, y
n
)) =
_
((1)
n
, (1)
n+1
)
_
in X
1
X
2
. Then
1 is a cluster point of both (x
n
) and (y
n
) but (1, 1) is not a cluster point
of the sequence ((x
n
, y
n
)).) The idea of the proof is to consider the set of
partial cluster points, that is, cluster points of the net (

) with respect to
some subset of components. These are naturally partially ordered, and an
appeal to Zorns lemma assures the existence of a maximal such element.
One shows that this is truly a cluster point of (

) in the usual sense.

For given X and J I, J ,= , let J denote the element
of the partial cartesian product

jJ
X
j
whose j
th
component is given by
J(j) = (j), for j J. In other words, J is obtained from by
simply ignoring the components in each X
j
for j / J. Let g

jJ
X
j
.
We shall say that g is a partial cluster point of (

) if g is a cluster point
of the net (

J)
A
in the topological space

jJ
X
j
. Let P denote the
collection of partial cluster points of (

). Now, for any j I, X

j
is compact,
by hypothesis. Hence, (

(j))
A
has a cluster point, x
j
, say, in X
j
. Set
10.8 Functional Analysis Gently Done Mathematics Department
J = j and dene g

i{j}
X
i
= X
j
by g(j) = x
j
. Then g is a partial
cluster point of (

), and therefore P is not empty.

The collection P is partially ordered by extension: that is, if g
1
and g
2
are
elements of P, where g
1

jJ
1
X
j
and g
2

jJ
2
X
j
, we say that g
1
_ g
2
if J
1
J
2
and g
1
(j) = g
2
(j) for all j J
1
. Let g

jJ

X
j
: be
any totally ordered family in P. Set J =

and dene g

jJ
X
j
by setting g(j) = g

(j), j J, where is such that j J

. Then g is
well-dened because g

: is totally ordered. It is clear that g _ g

for each . We claim that g is a partial cluster point of (

). Indeed, let
G be any neighbourhood of g in X
J
=

jJ
X
j
. Then there is a nite set
F in J and open sets U
j
X
j
, for j F, such that g
jF
p
1
j
(U
j
) G.
By denition of the partial order on P, it follows that there is some
such that F J

, and therefore g(j) = g

(j), for j F. Now, g

belongs to
P and so is a cluster point of the net (

)
A
. It follows that for any
A there is

_ such that p
j
(

) U
j
for every j F. Thus

G,
and we deduce that g is a cluster point of (

J)

. Hence g is a partial
cluster point of (

) and so belongs to P.
We have shown that any totally ordered family in P has an upper bound
and hence, by Zorns lemma, P possesses a maximal element, , say. We
shall show that the maximality of implies that it is, in fact, not just a
partial cluster point but a cluster point of the net (

). To see this, suppose

that

jJ
X
j
, with J I, so that is a cluster point of (

J)
A
.
We shall show that J = I. By way of contradiction, suppose that J ,= I
and let k I J. Since is a cluster point of (

J)
A
in

jJ
X
j
, it
is the limit of some subnet (
()
J)
B
, say. Now, (
()
(k))
B
is a net
in the compact space X
k
and therefore has a cluster point, X
k
, say. Let
J

= J k and dene

jJ

X
j
by

(j) =
_
(j) j J
j = k
.
We shall show that

is a cluster point of (

)
A
. Let F be any nite
subset in J and, for j F, let U
j
be any open neighbourhood of

(j) in X
j
,
and let V be any open neighbourhood of

(k) = in X
k
. Since (
()
)
B
converges to in

jJ
X
j
, there is
1
B such that
()
(j)
B
U
j
for
each j F for all _
1
. Furthermore, (
()
(k)
B
is frequently in V . Let

0
A be given. There is
0
B such that if _
0
then () _
0
. Let

2
B be such that
2
_
0
and
2
_
1
. Since (
()
(k)
B
is frequently
in V , there is _
2
such that
()
(k) V . Set = () A. Then
_
0
,

(k) V and, for j F,

(j) =
()
(j) U
j
. It follows that
Kings College London Product Spaces 10.9

is a cluster point of the net (

)
A
, as required. This means that

P. However, it is clear that _

and that ,=

. This contradicts
the maximality of in P and we conclude that, in fact, J = I and therefore
is a cluster point of (

)
A
.
We have seen that any net in X has a cluster point and therefore it follows
that X is compact.
Finally, we will consider a proof using universal nets.
Proof. (3rd proof) Let (

)
A
be any universal net in X =

iI
X
i
. For
any i I, let S
i
be any given subset of X
i
and let S be the subset of X given
by
S = X : (i) S
i
.
Then (

) is either eventually in S or eventually in X S. Hence we have

that either (

(i)) is either eventually in S

i
or eventually in X
i
S
i
. In
other words, (

(i))
A
is a universal net in X
i
. Since X
i
is compact, by
hypothesis, (

(i)) converges;

(i) x
i
, say, for i I. Let X be given
by (i) = x
i
, i I. Then we have that p
i
(

) =

(i) x
i
= (i) for each
i I and therefore

in X. Thus every universal net in X converges,

and we conclude that X is compact.
We will use Tychonovs theorem to show that the unit ball in the dual
space of a normed space is compact in the w

-topology. To do this, it is
necessary to consider the unit ball of the dual space as a suitable cartesian
product. By way of a preamble, let us discuss the dual space X

of the
normed space X as a cartesian product. Each element in X

is a (linear)
function on X. The collection of values (x), as x runs over X, can be
thought of as an element of a cartesian product with components given by
the (x). Specically, for each x X, let Y
x
be a copy of C, equipped with
its usual topology. Let Y =

xX
Y
x
=

xX
C, equipped with the product
topology. To each element X

, we associate the element

Y given by

(x) = (x), i.e., the x

th
-coordinate of

is (x) C = Y
x
.
If
1
,
2
X

, and if

1
=

2
, then

1
and

2
have the same coordinates
so that
1
(x) =

1
(x) =

2
(x) =
2
(x) for all x X. In other words,

1
=
2
, and so the correspondence

of X

Y is oneone. Thus X

can be thought of as a subset of Y =

xX
C.
Suppose now that

is a net in X

such that

in X

with respect
to the w

-topology. This is equivalent to the statement that

(x) (x)
for each x X. But then

(x) = p
x
(

) (x) = p
x
(

) for all x X,
10.10 Functional Analysis Gently Done Mathematics Department
which, in turn, is equivalent to the statement that

with respect to
the product topology on

xX
C.
We see, then, that the correspondence

respects the convergence

of nets when X

is equipped with the w

-topology and Y with the product

topology. It will not come as a surprise that this also respects compactness.
Consider now X

1
, the unit ball in the dual of the normed space X. For
any x X and X

1
, we have that [(x)[ |x|. Let B
x
denote the ball
in C given by
B
x
= C : [[ |x| .
Then the above remark is just the observation that (x) B
x
for every
X

1
. We equip B
x
with its usual metric topology, so that it is compact. Let
Y =

xX
B
x
equipped with the product topology. Then, by Tychonovs
theorem, Y is compact.
Let X

1
. Then, as above, determines an element

of Y by setting

(x) = p
x
(

) = (x) B
x
.
The mapping

is oneone. Let

Y denote the image of X

1
under this
map ;

Y = Y : =

some X

1
.
Proposition 10.12.

Y is closed in Y .
Proof. Let (

) be a net in

Y such that

in Y . Then p
x
(

) p
x
()
in B
x
, for each x X. Each

is of the form

for some

1
. Hence
p
x
(

) =

(x) (x)
for each x X

1
. It follows that for any a C and elements x
1
, x
2
X
(ax
1
+x
2
) = lim

(ax
1
+x
2
)
= lima

(x
1
) +

(x
2
)
= a(x
1
) +(x
2
) .
That is, the map x (x) is linear on X. Furthermore, (x) = p
x
() B
x
,
i.e., [(x)[ |x|, for x X. We conclude that the mapping x (x)
denes an element of X

1
. In other words, if we set (x) = (x), x X, then
X

1
and

= . That is,

Y and so

Y is closed, as required.
Kings College London Product Spaces 10.11
Remark 10.13. We know that Y is compact and since

Y is closed in Y ,
we conclude that

Y is also compact.
Theorem 10.14. (Banach-Alaoglu) Let X be a normed space and let
X

1
denote the unit ball in X

, the dual of X;
X

1
= X

: || 1 .
Then X

1
is a compact subset of X

with respect to the w

-topology.
Proof. First let us note that X

1
is closed in X

with respect to the w

-
topology. To see this, let (

) be a net in X

1
such that

in X

in the w

-topology. This means that

(x) (x) for each x X. But

[

(x)[ |x| for each and so we also have that [(x)[ |x|. Thus X

1
and therefore X

1
is w

-closed, as claimed.
To show compactness, we will exploit the above identication with

Y .
Suppose F

is a family of closed sets in X

1
satisfying the nite intersection
property. The proof is complete if we can show that the whole family has a
non-empty intersection.
Let A

denote the image of F

in Y so A

Y . We claim that A

is
closed in Y . To see this, suppose that (

) is a net in A

such that

in Y . Each

has the form

for some

. Now, A

Y and

Y is closed and so we have that

Y ; that is, =

for some X

1
. But

in Y implies that

in X

1
with respect to the w

-topology.
Since

and F

is w

-closed we deduce that F

. Hence

and therefore A

is closed.
Now, A

also has the nite intersection property and since A

is closed,
for each , and

Y is compact, we deduce that

,= and therefore

,= . It follows that X

1
is w

-compact.
Example 10.15. Let X =

and for each n N let

m
: X C be the
map
m
(x) = x
m
, where x = (x
n
)

. Thus
m
is simply the evaluation
of the m
th
coordinate on

.
We have [
m
(x)[ = [x
m
[ |x|

and so we see that

m
X

1
for each
m N. We claim that the sequence (
m
)
mN
in X

1
has no w

-convergent
subsequence, despite the fact that X

1
is w

-compact. Indeed, let (

m
k
)
kN
be
any subsequence. Then
m
k
in the w

-topology if and only if

m
k
(x)
(x) in C for every x X =

. Let z be the particular element of X =

10.12 Functional Analysis Gently Done Mathematics Department

given by z = (z
n
) where
z
n
=
_
1 , n = m
2j
, j N,
1 , otherwise.
Then
m
k
(z) = 1 if k is even and is equal to 1 if k is odd. So the subsequence
(
m
k
(z)) cannot converge in C.