Linear Analysis

Lectured by B. J. Green (and B. Schlein)

Michaelmas Term 2010 (and 2008)

1 Normed Spaces 1
2 The Hahn-Banach Theorem 11
3 The Baire Category Theorem 16
4 The Space C(X) 23
5 Weak Topologies on Normed Spaces 33
6 Hilbert Space 39
7 Spectral Theory 48

Hahn-Banach handout (2011)

Examples Sheets

These notes combine the material lectured in 2008 and 2010.

Chapters 1, 3, 4, 6, 7 were lectured in 2010.
Chapters 2, 5 were lectured in 2008.

Last updated: Wed 29th May, 2019

These have not really been proof-read yet, and typing up analysis
is prone to mistakes. So please do let me know of any corrections:
[email protected]
LINEAR ANALYSIS (D) 24 lectures, Michaelmas term

Part IB Linear Algebra, Analysis II and Metric and Topological Spaces are essential.

Normed and Banach spaces. Linear mappings, continuity, boundedness, and norms. Finite-
dimensional normed spaces. [4]
The Baire category theorem. The principle of uniform boundedness, the closed graph theorem
and the inversion theorem; other applications. [5]
The normality of compact Hausdorff spaces. Urysohn’s lemma and Tiezte’s extension the-
orem. Spaces of continuous functions. The Stone-Weierstrass theorem and applications.
Equicontinuity: the Ascoli-Arzelà theorem. [5]
Inner product spaces and Hilbert spaces; examples and elementary properties. Orthonormal
systems, and the orthogonalization process. Bessel’s inequality, the Parseval equation, and
the Riesz-Fischer theorem. Duality; the self duality of Hilbert space. [5]
Bounded linear operations, invariant subspaces, eigenvectors; the spectrum and resolvent set.
Compact operators on Hilbert space; discreteness of spectrum. Spectral theorem for compact
Hermitian operators. [5]

Appropriate books

B. Bollobás Linear Analysis. 2nd Edition, Cambridge University Press 1999 (£22.99 paper-
back).
C. Goffman and G. Pedrick A First Course in Functional Analysis. 2nd Edition, Oxford
University Press 1999 (£21.00 Hardback)
W. Rudin Real and Complex Analysis. McGraw-Hill International Editions: Mathematics
Series (£31.00 Paperback)
1. Normed Spaces
In this course all vector spaces V will be over R or C. Usually it doesn’t matter which. They
won’t always (or even usually) be finite dimensional. Our vector spaces will have additional
analytic structure; this makes the theory much more interesting.

By a norm on V we mean a map k k : V → R>0 satisfying the following:

(i) kxk = 0 ⇐⇒ x = 0
(ii) kλxk = |λ| kxk for all λ ∈ R or C
(iii) kx + yk 6 kxk + kyk (‘subadditive’)

Think of kxk as the ‘length’ of x, but don’t rely on geometric intuition too much!

Examples

1. V = R, and kxk = |x| is absolute value.

p
2. V = Rn , with the Euclidean norm: kxk2 = |x1 |2 + . . . + |xn |2 , where x = (x1 , . . ., xn ).
Or with the ℓ1 -norm: kxk1 = |x1 | + . . . + |xn |
Or with the ℓ∞ -norm: kxk∞ = max |xi |
i=1,...,n

3. S = any set, V = B(S) = vector space of bounded real-valued functions on S.

‘sup norm’: kf k∞ = sup |f (s)|
s∈S

4. V = C[0, 1] = continuous functions on [0, 1].

Possible norm: kf k∞ = sup |f (t)|. (Actually max since f attains its bounds.)
t∈[0,1]

Also works in any metric (topological) space X in place of C[0, 1].

R1 
kf k1 = 0 |f (t)| dt 
R 1/2 proof that these are norms – exercise
1
kf k2 = 0 |f (t)|2 dt 

5. V = C k [a, b] = space of k-times continuously differentiable functions on [a, b].

kf k = sup |f (t)| + |f ′ (t)| + . . . + |f k (t)|
t∈[a,b]

6. V = Riemann-integrable integrable functions on [0, 1].

R1
kf k1 = 0 |f (t)| dt – only a seminorm, since one can have kf k1 = 0 with f 6= 0.

0 x ∈ (0, 1]
E.g., f (x) = .
1 x=0
R1
Make it a normed space by quotienting these out: f ∼ g if 0 |f (t) − g(t)| dt = 0.
P∞
7. ℓ1 = {(x1 , x2 , . . .) : i=1 |xi | < ∞}
P∞
ℓ2 = {(x1 , x2 , . . .) : i=1 |xi |2 < ∞} – Hilbert space
ℓ∞ = {(x1 , x2 , . . .) : supi |xi | < ∞}
8. ℓp -spaces. Let 1 6 p < ∞, V = Rn .
Pn 1/p
kxkp = ( i=1 |xi |p ) – this is called ℓnp . (Proof of triangle inequality: see below.)
P∞ P∞ 1/p
ℓp = {(x1 , x2 , x3 , . . .) : i=1 |xi |p < ∞} with kxkp = ( i=1 |xi |p ) .
This generalises example 7.

1
ℓp -norms, Hölder’s Inequality and Minkowski’s Inequality
1 1
Hölder’s Inequality. Suppose 1 < p < ∞. Define the conjugate index q by p + q = 1.
Let (ai )ni=1 and (bi )ni=1 be two sequences of complex numbers. Then
n
X X
n 1/p  X
n 1/q
|ai bi | 6 |ai |p |bi |q .
i=1 i=1 i=1

Remarks. Fancier statement is that |hx, yi| 6 kxkp kykq if x, y ∈ Cn .

In the special case p = q = 2, one recovers Cauchy-Schwarz.

The fancier statement is also true and easy with p = 1, q = ∞. By convention, p = 1

and q = ∞ are thought of as conjugate indices: ‘ 11 + ∞
1
= 1’.

We first need to prove:

The weighted AM-GM inequality. If a1 , . . ., an > 0, and if λ1 , . . ., λn ∈ [0, 1] satisfy

λ1 + . . . + λn = 1, then

aλ1 1 . . .aλnn 6 λ1 a1 + . . . + λn an .

This is a generalisation of the usual AM-GM, which is the case λi = 1/n.

Proof. We use the concavity of the function f (x) = log x, which follows from the fact that
f ′′ (x) = −1/x2 6 0. (Exercise: this really implies concavity.)

We have f (λ1 x1 + . . . + λn xn ) > λ1 f (x1 ) + . . . + λn f (xn ) – Jensen’s Inequality.

In the specific case of f (x) = log x we get

λ1 log x1 + . . . + λn log xn 6 log(λ1 x1 + . . . + λn xn ).

Exponentiating both sides gives weighted AM > GM. 2

ap bq
A particular example of this (in the case n = 2) is that ab 6 + (∗), whenever
p q
a, b ∈ [0, ∞) and p, q are conjugate indices.

Proof of Hölder’s Inequality. The inequality is homogeneous in the sense that if it is true
for ai , bi , then it is also true for λai and µbi for any λ, µ ∈ R \ {0}.
P P q
The inequality is manifestly true when either |ai |p or |bi | equals 0 (since then
eitherPall ai or all bP
i are 0). Otherwise, by an appropriate choice of λ, µ we may assume
n n
that i=1 |ai |p = i=1 |bi |q = 1.
Pn 1
Pn
But under these conditions, by (∗) we do indeed have i=1 |ai bi | 6 p i=1 |ai |p +
1 Pn q
q i=1 |bi | = 1. 2

Minkowski’s Inequality. Suppose ai , bi ∈ C and that 1 < p < ∞. Then

X
n 1/p X
n 1/p X
n 1/p
p p p
|ai + bi | 6 |ai | + |bi | .
i=1 i=1 i=1

2
In other words, the ℓp -norm on Rn (or Cn ) is indeed a norm, since we have the triangle
inequality.
n
X n
X n
X n
X
Proof. |ai + bi |p = |ai + bi |p−1 |ai + bi | 6 |ai + bi |p−1 |ai | + |ai + bi |p−1 |bi |.
i=1 i=1 i=1 i=1

By Hölder, this is at most

X n 1/q  X
n 1/p  X
n 1/q  X
n 1/p
(p−1)q p (p−1)q p
|ai + bi | |ai | + |ai + bi | |bi |
i=1 i=1 i=1 i=1

1 1
Now, (p − 1)q = p, since p + q = 1. Therefore this is
X
n 1/q X
n 1/p X
n 1/p !
p p p
|ai + bi | |ai | + |bi |
i=1 i=1 i=1

P
1/q
Divide though by |ai + bi |p .
X
n 1− q1 X
n 1/p X
n 1/p
|ai + bi |p 6 |ai |p + |bi |p .
i=1 i=1 i=1

1
But 1 − q = p1 , so done. 2

Banach Spaces
If V is a normed space with norm k k then it can be regarded as a metric space by defining
d(x, y) = kx − yk. (Easy exercise to see that this is a metric space.)

Definition. V is a Banach space if it is complete with respect to this metric. (I.e., Cauchy
sequences converge.)

Examples

• R is a Banach space.
• Rn with ℓ1 -norm: kxk1 = |x1 | + . . . + |xn |.
The metric induced on Rn is exactly the product metric on Rn , viewed as a product of
n copies of R.
Recall from Met&Top that if
X, Y are metric spaces then we can define a metric on
X × Y by d (x1 , y1 ), (x2 , y2 ) = dX (x1 , x2 ) + dY (y1 , y2 ).

∞
If X and Y are complete then so is X × Y . (Proof: exercise.) If (xn , yn ) n=1 is Cauchy
in X × Y then (needs justifying) (xn ) is Cauchy in X and (yn ) is Cauchy in Y . Let
xn → x∗ , yn → y ∗ , then (xn , yn ) → (x∗ , y ∗ ).

Proposition. ‘Any two norms on a finite-dimensional space are equivalent.’

Suppose k k and k k′ are two norms on Rn (or Cn ). Then they are equivalent in the
sense that there are c1 , c2 > 0 such that c1 kxk 6 kxk′ 6 c2 kxk for all x ∈ V .

(Exercise: this defines an equivalence relation.)

Proof. Since ‘equivalence’ is indeed an equivalence relation, it suffices to show that a given
norm k k is equivalent to the ℓ1 -norm, k k1 .

3
 We claim that x 7→ kxk is continuous in the metric
Pn induced by the ℓ1 -norm. To see
this, suppose that kv − wk1 6 δ, that is to say i=1 |vi − wi | 6 δ. Then
n
X
kv − wk = (vi − wi )ei with e1 , . . ., en the standard basis
i=1
n
X
6 |vi − wi | kei k using scalar multn and triangle ineq. for k k
i=1
n
X
6 M |vi − wi | where M = sup kei k
i=1 i

6 Mδ

So if kv−wk1 6 δ then kv−wk 6 M δ. By the triangle inequality, kvk−kwk 6 kv−wk.

So, in the ε-δ definition of continuity, we may take δ = ε/M . This establishes the claim.

Now, the unit sphere {x : kxk1 = 1} ⊂ Rn is closed and bounded, hence compact. So
the function x 7→ kxk, being continuous, is bounded and attains its bounds.

In particular, inf kxk = min kxk 6= 0, since the only vector with norm zero is 0.
kxk1 =1 kxk1 =1

So there are c1 , c2 > 0 such that c1 6 kxk 6 c2 for all x with kxk1 = 1. This implies
c1 kxk1 6 kvk 6 c2 kvk1 , for all v (take x = v/kvk1 ). 2

Corollary. Any finite-dimensional normed space is a Banach space.

Proof. Equivalent norms give rise to equivalent metrics: c1 d1 (x, y) 6 d2 (x, y) 6 c2 d1 (x, y).
And Cauchy sequences and convergence in one metric means the same as in the other.
(Exercise: check details.)

Norms and convex bodies

There is a correspondence between norms on Rn and closed, bounded, centrally symmetric

(i.e. x ∈ K ⇔ −x ∈ K) convex bodies K containing a ball Bε (0) about the origin.

For (⇒), it is easy to show that the triangle inequality for k k implies that {x : kxk 6 1} is
a convex body.

For (⇐), suppose K is such a convex body. We can define kxk to be the dilation factor of K
required to ‘hit‘ x, that is kxk = min{x ∈ λK}.
λ

Why is it a norm? First check it’s well-defined (using non-empty interior). The triangle
inequality follows since λK + µK ⊂ (λ + µ)K – a consequence of convexity.

✬✩
ℓ∞ ℓ2 ℓ1
❅
❅
❅
✫✪ ❅

4
Back to examples of Banach spaces

• C[0, 1] = continuous functions on [0, 1] with kf k = kf k∞ = supt |f (t)| is a Banach

space The completeness follows as a consequence of the fact that a uniformly convergent
sequence (fn )∞
n=1 of continuous functions converges to a continuous function f . (See
Analysis II.)
Note. (fn )∞ ∞
n=1 being uniformly convergent is the same thing as (fn )n=1 being a Cauchy
sequence in k k (or rather, in the metric space induced by this norm).
R1
• L1 = C[0, 1] = continuous functions on [0, 1], with kf k = kf k1 = 0 |f (t)| dt. This is
not a Banach space. We can have a Cauchy sequence of functions fn :

❅ ❆ ‘−→’
❅ ❆
1 1 1
2 1 2 1 2 1
 1

 0 if x > 2 + n1
1
E.g., fn (x) = 1 if x 6 2 − n1

 1
2 − 2 (x − 2 ) if 21 −
n 1 1
n 6x6 1 1
2 + n
2 2


This is Cauchy in L1 , since kfm − fn k 6 max m, n → 0 as m, n → ∞.
1


However, fn does not converge to any continuous function f . What would f 2 be?


Wlog, f 12 6 21 . Since f is continuous, f (x) 6 34 for x − 12 6 δ, for some δ.
1 1 δ
By contrast, fn (x) = 1 if 2 −δ 6x6 2 − 2 for n sufficiently large.
Hence, if n is big enough, kf − fn k1 > 41 2δ = 8δ . So fn 6→ f in L1 .

Exercise. On C[0, 1], the norms k k∞ and k k1 are inequivalent. (Easy.)

Remark. For essentially the same reason, C[0, 1] is not a Banach space in any Lp -norm,
1 6 p 6 ∞.

More examples
P∞ P
• ℓp = {(x1 , x2 , x3 , . . .) ∈ CN : i=1 |xi |p < ∞}, with the ℓp -norm, kxkp = ( |xi |p )1/p .
(This is a norm by Minkowski’s inequality – let n → ∞.)
This is complete and hence a Banach space.
Proof. Let (xn )∞n=1 be a Cauchy sequence in ℓp . Each xn is a sequence; write them as
x1 = (x11 , x12 , x13 , . . .), x2 = (x21 , x22 , x23 , . . .), . . .

We must prove that there exists x∗ ∈ ℓp such that kxn − x∗ kp → 0.

Observe that for any fixed i, |xni − xmi | 6 kxn − xm kp , and therefore (xni )∞
n=1 is
a Cauchy sequence. Since C is complete, there exists x∗i such that xni → x∗i .

Define x∗ = (x∗1 , x∗2 , . . .). Claim that xn → x∗ in ℓp .

Claim. Let m be such that kxm − xn kp 6 ε for all n > m. Then kxm − x∗ kp 6 ε.

Then we can conclude that kxm − x∗ kp → 0. And x∗ ∈ ℓp , since kx∗ kp 6 kxm kp +

kx∗ − xm kp 6 kxm0 kp + 1 (if m0 is large enough) < ∞.

It remains to prove the claim.

5
 Proof of claim. Suppose kxm − xn kp 6 ε for all n > m. Let R be arbitrary.
PR PR
Then i=1 |xmi − xni |p 6 εp . Let n → ∞, we get i=1 |xmi − x∗i |p 6 εp .
P∞
Let R → ∞, we get i=1 |xmi − x∗i |p 6 εp .

That is kxm − x∗ kp 6 ε, as required. 2

Note. In particular, ℓ2 = Hilbert space is complete.

Completion of Normed Spaces

This section is from the 2008 notes.

Let (V, k k) be a normed space. The construction of the completion of V is as follows. Let
CV be the set of all Cauchy sequences in V . This is a vector space: if x = (xn ) and y = (yn ),
then λx + µy = (λxn + µyn ).

Let NV = {(xn ) : xn → 0} be the subset of CV consisting of null sequences. This is in fact a

subspace of CV , and so V = CV /NV is a vector space.

We define a norm on V . For x ∈ CV , let p(x) = limn→∞ kxn k, which is well-defined as kxn k
is Cauchy in R. We have

(i) p(x) = 0 iff x ∈ NV

(ii) p(λx) = |λ| p(x)
(iii) p(x + y) 6 p(x) + p(y) for all x, y ∈ CV .

and p induces p on V , by p : V → R+ , [x] 7→ p(x).

This is well-defined, as x ∼ y ⇒ x − y ∈ NV ⇒ xn − yn → 0 ⇒ kxn k − kyn k → 0 ⇒

limn→∞ kxn k = limn→∞ kyn k ⇒ p(x) = p(y).

And p is a norm on V as p([x]) = 0 ⇔ p(x) = 0 ⇔ x ∈ NV ⇔ [x] = 0.

So (V , p) is a normed space.

We can regard (V, k k) as a dense subset of (V , p). Define ϕ : V → V , x 7→ [(x, x, . . .)]. Then
ϕ is linear and p(ϕ(x)) = kxk, so ϕ is isometric, and in particular injective.

Claim. ϕ(V ) = V , where ϕ(V ) is the closure of ϕ(V ) in V .

Proof. Let [x] ∈ V and ε
> 0. Then x = (xn ) is Cauchy, so kxm − xn k 6 ε for m, n > n0 .
Then p ϕ(xn0 ) − [x] = p(xn0 − x1 , xn0 − x2 , . . .) 6 ε. 2

If V is complete then V ∼
= V , with ϕ(V ) = V . For if [x] ∈ V , write x = (xn ) ∈ CV and
x∞ = limn→∞ xn . Then x∞ ∈ V by the completeness of V we have ϕ(x∞ ) = [x].

Claim. (V , p) is always complete.

Proof. Let ([xn ]) be a Cauchy sequence in V , so that (xn ) is Cauchy in V for all n. Since
ϕ(V ) = V , there exists zn ∈ V such that p(ϕ(zn ) − [xn ]) 6 21n , for each n ∈ N.

6


kzn − zm k = p ϕ(zn ) − ϕ(zm )




6 p ϕ(zn ) − [xn ] + p [xn ] − [xm ] + p [xm ] − ϕ(zm )
1
1
6 n
+ p [xn ] − [xm ] + m → 0 as n, m → ∞
2 2
So z = (zn ) is Cauchy in V .


p [z] − [xn ] = lim kzm − xnm k
m→∞
6 lim sup kzm − zn k + lim sup kzn − xnm k
m→∞ m→∞
1
6 → 0 as n → ∞
2n
So [xn ] → [z] as n → ∞ and thus (V , p) is complete. 2

Definition. Let V be a normed space. A completion of V is a triple (V , p, ϕ) such that

(V , p) is a complete normed space, and ϕ : V → V is linear, isometric, and ϕ(V ) = V .

Theorem. The completion of a normed space V exists and is uniquely determined up to an

isometric isomorphism.

Proof. Exercise.

Proposition. A normed space V is complete iff every absolutely convergent series converges.
P∞ PM
Proof. Suppose V is complete and n=1 kxn k < ∞. Then for M > N , we have n=N xn
PM
6 n=N kxn k → 0, as N → ∞. So the series is Cauchy, hence converges.

Conversely, suppose that every absolutely convergent series is convergent, and let (xn )
be a Cauchy sequence. It is sufficient to show that there exists subsequence xni con-
verges to some x ∈ V , as then: kxm − xk 6 kxm − xni k + kxni − xk → 0 as m, i → ∞.

To prove there is a convergent subsequence, choose n1 ∈ N such that kxn1 − xm k 6 21

for all m > n1 . Then choose n2 > n1 such that kxn2 − xm k 6 212 for all m > n2 . And
so on. Then xnk+1 = xn1 + (xn2 − xn1 ) + . . . + (xnk+1 − xnk ).
P∞ P∞ 1
But i=1 kxni+1 − xni k 6 i=1 2i = 1. So xni → x, for some x ∈ V . 2

Linear operators
Let X, Y be two normed spaces. A linear map T : X → Y is said to be continuous if T is
a continuous map in the metric spaces induced by the norms on X and Y .

That is, T must be continuous at every point x0 ∈ X, which means that for all ε > 0 there
exists δ > 0 such that kx − x0 kX 6 δ implies kT x − T x0 kY 6 ε.

Lemma. Let T : X → Y be a linear map between normed spaces. The following are
equivalent.

(1) T is continuous
(2) T is continuous at some x0
(3) T is bounded, meaning that T (BX (1)) ⊂ BY (R) for some R (where BX (r) =
{x ∈ X : kxk < r}). Equivalently, kT xkY 6 RkxkX .

7
As a result of this we always speak of bounded operators and never continuous ones.

Proof. (1) ⇒ (2). Trivial.

(2) ⇒ (3). If T is continuous at x0 then there exists δ > 0 such that kx − x0 kX < δ
implies kT x − T x0 kY 6 1. I.e., kT (x − x0 )k 6 1, by linearity of T .

Hence T (BX (δ)) ⊂ BY (1). By linearity of T again, this implies T (BX (1)) ⊂ BY (1/δ).

(3) ⇒ (1). Note (3) implies that kT xkY 6 RkxkX . Hence if ε > 0, take δ = ε/R. Then
if kx − x0 k 6 δ we have kT x − T x0 kY = kT (x − x0 )kY 6 Rkx − x0 kX 6 ε.

So T is continuous at x0 . 2

The infimum of all possible values of R for which kT xkY 6 RkxkX is called the operator
norm of T , written kT k. Equivalently, kT k = sup kT xkY .
kxkX =1

Examples

1. Linear maps between finite-dimensional spaces. These are always bounded.

Indeed, suppose X = Rm , Y = Rn , and that T : X → Y is given by a matrix A = (aij ).

P P P P P P
Let x = j λj ej , then T x = j λj T ej = j λj ( i aij ei ) = i ( j λj aij )ei . Hence
P P P
kT xk∞ = maxi | j λj aij | 6 maxi j |λj aij | 6 maxj |λj | · maxi j |aij | = kxk∞ · M ,
P
where M = maxi j |aij |, since kxk∞ = maxj |λj |.

Since any two norms on Rn are equivalent, T is bounded by any norm.

More simply. Let M = maxi kT ei k∞ . Then T is bounded as a map ℓm n

1 → ℓ∞ , since
P P P
kT xk∞ = k j λj T ej k∞ 6 j |λj | kT ej k∞ 6 M ( j |λj |) = M kxk1 .

So done since any two norms on a finite-dimensional vector space are equivalent.

P
2. Shift map. Let X = ℓ1 = {(x1 , x2 , . . .) : |xi | < ∞}.

Define T : X → X by T (x1 , x2 , . . .) = (0, x1 , x2 , . . .),

S : X → X by S(x1 , x2 , . . .) = (x2 , x3 , . . .).

Both are linear and kSk = kT k = 1. Note that S ◦ T = id, T ◦ S 6= id. (So existence of
a left inverse does not imply existence of a right inverse.)

3. Suppose X = C (1) [0, 1], the continuously differentiable

functions on [0, 1]. We previ-
ously mentioned the norm kf k = supt |f (t)| + |f ′ (t)| .

We also have the sup norm, kf k∞ = supt |f (t)|. Let Y = C[0, 1] with the sup norm,
and define T : X → Y by T (f ) = f ′ . This is certainly linear.

T is bounded if X is given the first norm, but not if X has the sup norm. (Take a
bounded function with unbounded derivative.)

8
Suppose X, Y are normed spaces. We write B(X, Y ) = {T : X → Y : kT k < ∞}.

When Y = R (or C) we write X ∗ = B(X, R) (or B(X, C)). This is the dual space of X.
Elements of X ∗ are called functionals.

B(X, Y ) is a vector space and in fact a normed space with the operator norm:

• If S, T ∈ B(X, Y ) we define (S + T )(x) = Sx + T x.

Then sup k(S + T )(x)k 6 sup kSxk + sup kT xk = kSk + kT k.
kxk=1 kxk=1 kxk=1

• Same with scalar multiplication: (λT )(x) = λT x. Exercise: kλT k = |λ| kT k.

The following is the most important fact about B(X, Y ).

Lemma. If Y is a Banach space (i.e. Y is complete) then so is B(X, Y ).

In particular, X ∗ is a Banach space for every X.

Proof. Suppose that (Tn )∞

n=1 is a Cauchy sequence in B(X, Y ), i.e. kTm − Tn k → 0 as
m, n → ∞. We need to find T ∈ B(X, Y ) such that Tn → T in the operator norm.

Let x ∈ X. The sequence (Tn x)∞

n=1 is Cauchy, because kTm x−Tn xkY 6 kTm −Tn k kxk,
which → 0 as m, n → ∞. But Y is a Banach space and so Tn x → x∗ .

Define T x = x∗ . It’s easy to check that T is linear: Tn (x + y) = Tn x + Tn y → x∗ + y ∗ ,

so (x + y)∗ = x∗ + y ∗ .

Also, Tn → T in the operator norm, as follows.

Let ε > 0 and suppose kTm − Tn k 6 ε for all n > m. Then kTm x − Tn xkY 6 εkxkX .
Letting n → ∞, we get kTm x − T xkY 6 εkxkX . For such an m, we have kTm − T k 6 ε.
Hence Tm → T in the operator norm.

Note kT k is bounded, since kT k 6 kTn k + kT − Tn k 6 kTn k + 1 for a suitable n. 2

(Compare the proof that ℓp is complete.)

Lemma. Suppose 1 < p < ∞. Let 1

p + 1
q = 1. Then ℓ∗p ∼
= ℓq (‘isometrically isomorphic’).

Notes. 1. (X, k kX ) and (Y, k kY ) are isometrically isomorphic if there is a linear

isomorphism θ : X → Y such that kθxkY = kxkX .

2. This gives us ‘another’ proof that ℓp is complete.

Proof. Let e1 = (1, 0, 0, . . .), e2 = (0, 1, 0, . . .), e3 = (0, 0, 1, 0, . . .), . . . in ℓp . (Exercise: this
is not a basis for ℓp as a vector space.)

Suppose that ϕ ∈ ℓ∗p . Let ϕ(ei ) = yi .

Clearly ϕ (x1 , . . ., xn , 0, 0, . . .) = ϕ(x1 e1 + . . . + xn en ) = x1 y1 + . . . + xn yn .

However, if x = (x1 , x2 , . . .) ∈ ℓp then we have (x1 , . . ., xn , 0, 0, . . .) → x in ℓp , because

P

p 1/p
kx − (x1 , . . ., xn , 0, 0, . . .)kp = i>n |xi | → 0. (Note: this is false in ℓ∞ .)

9

P∞

It follows that ϕ(x) = ϕ (x1 , x2 , . . .) = i=1 xi yi = limn→∞ ϕ (x1 , . . ., xn , 0, 0, . . .) .
However, the yi are not arbitrary.

Claim. ϕ ∈ ℓ∗p ⇐⇒ (y1 , y2 , . . .) ∈ ℓq . (This provides the isomorphism ℓ∗p ∼

= ℓq .)

Proof of claim. (⇐) is an immediate P consequence of Hölder’s inequality. Indeed, if

(y1 , y2 , . . .) ∈ ℓq then |ϕ(x)| = | xi yi | 6 kxkp kykq .

So ϕ is a bounded linear functional with kϕk 6 kykq . (In fact, equality occurs.)

P∞ P 1/p
∞
(⇒). Suppose conversely that i=1 xi yi 6 M i=1 |xi |p for all (xi ) ∈ ℓp .


|yi |q yi−1 if i 6 N
Let xi = , for fixed N .
0 if i > N

PN P 1/p
N
Then (xi ) ∈ ℓp , and so i=1 |yi |q 6 M i=1 |yi |(q−1)p .

P 1/q
N
Hence, since (q − 1)p = q and 1/p = 1 − 1/q, we have i=1 |yi |q 6 M.

But N was arbitrary, so we are done. 2

Corollary. ℓp is complete for 1 < p < ∞.

Proof. ℓp = ℓ∗q and every dual space is complete. 2

Note. Note. This doesn’t work for ℓ1 , because ℓ1 is much smaller than ℓ∗∞ . (In fact, ℓ∞ is
a pretty exotic object.)

ℓ∞ is the dual of ℓ1 . However, ℓ1 is the dual of c0 , the space of bounded sequences

(x1 , x2 , . . .) with xi → 0 as i → ∞. The proof is the same as for the above norm,
because the ‘convergence of the approximations’ now holds.

Adjoints

Suppose X, Y are normed spaces and T ∈ B(X, Y ).

Define the adjoint map T ∗ : Y ∗ → X ∗ by (T ∗ g)(x) = g(T x) for all g ∈ Y ∗ , x ∈ X.

Easy to show T ∗ is linear and bounded. Indeed: kT ∗ g(x)k = kg(T x)k 6 kgk kT xk 6
kgk kT k kxk, all g ∈ Y ∗ , x ∈ X.

Therefore, kT ∗ gk 6 kgk kT k, and therefore kT ∗ k 6 kT k.

(In fact, equality occurs: requires the Hahn-Banach theorem.)

10
2. The Hahn-Banach Theorem
Throughout this section, X ∗ is always a Banach space. But could X ∗ = {0}?

Definition. Let X be a real vector space. A function p : X → R is said be a convex

functional if:

(i) p(αx) = αp(x) for all α > 0, x ∈ X

(ii) p(tx + (1 − t)y) 6 tp(x) + (1 − t)p(y) for all t ∈ [0, 1], and all x, y ∈ X.

Note convexity follows from subadditivity, p(x + y) 6 p(x) + p(y), and positive homogeneity.

Lemma. Let X be a real vector space, p : X → R a convex functional on X, M ⊂ X a

linear subspace, and f : M → R a linear functional on M such that f (x) 6 p(x) for all
x ∈ M . Let x0 ∈ X \ M and write M ′ = M + hx0 i.

Then there exists F : M ′ → R such that F |M = f , and F (x) 6 p(x) for all x ∈ M ′

Proof. Any z ∈ M ′ can be written as z = y + tx0 , with y ∈ M , t ∈ R. We wish to define

F (x0 ) = c such that F (y + tx0 ) 6 p(y + tx0 ) for all y ∈ M , t ∈ R.

Now F (y + tx0 ) = F (y) + tF (x0 ) = f (y) + tc.

1


For t > 0, we require f (y) + tc 6 p(y + tx0 ), which is true iff c 6 t p(y + tx0 ) − f (y) =
p(y ′ + x0 ) − f (y ′ ), where y ′ = y/t.

For t < 0, let t = −s,
then we require f (y) − sc 6 p(y − sx0 ), which is true iff
c > 1s f (y) − p(y − sx0 ) = f (y ′′ ) − p(y ′′ − x0 ), where y ′′ = y/s.

Such a c exists if f (y ′′ ) − p(y ′′ − x0 ) 6 p(y ′ + x0 ) − f (y ′ ) for all y ′ , y ′′ ∈ M .

And indeed, f (y ′ ) + f (y ′′ ) = f (y ′ + y ′′ ) 6 p(y ′ + y ′′ ) 6 p(y ′ − x0 ) + p(y ′′ + x0 ). 2

The proof of Hahn-Banach will require Zorn’s Lemma.

Definitions. Let P be a set.

1. A partial ordering on P is a relation 6 such that, for all a, b, c ∈ P :

(i) a 6 a
(ii) if a 6 b and b 6 a, then a = b
(iii) if a 6 b and b 6 c, then a 6 c.
Then (P, 6) is a partially ordered set.
2. C ⊂ P is a total ordering if a 6 b or b 6 a for all a, b ∈ C.
3. m ∈ P is a maximal element if m 6 a implies m = a.
4. b ∈ P is an upper bound for S ⊂ P if a 6 b for all a ∈ S.

Zorn’s Lemma. Suppose that every totally ordered subset S of a non-empty, partially
ordered set P has an upper bound. Then P has a maximal element.

Proof. It can be proved that this is equivalent to the Axiom of Choice. 2

Theorem (Hahn-Banach theorem for real vector spaces). Let X be a real vector
space, p : X → R a convex functional on X, M ⊂ X a linear subspace and f : M → R
a linear functional on M such that f (x) 6 p(x) for all x ∈ M .

11
 Then there exists F : X → R, linear, such that F |M = f , and F (x) 6 p(x) for all
x ∈ X.
( )
f ⊂ X is a linear subspace such that M
M f ⊃ M, and
f e
Proof. Let F = (M , f ) : e .
f → R linear, fe|M = f, fe(x) 6 p(x) for all x ∈ X
f :M

f, fe) 6 (N
Define a partial ordering by (M e, e f⊂N
g ) if M e and e
g |M e
f = f.

F is non-empty since (M, f ) ∈ F . And if {(Mi , gi ) : i ∈ I} is a totally ordered subset,

S
f=
then it has an upper bound, given by M g(x) = gi (x) if x ∈ Mi .
i∈I Mi and e

By Zorn, F has a maximal element (N, F ). We must have N = X, as otherwise we

can pick some x0 ∈ X \ N and extend F to N + hx0 i, contradicting maximality. 2

Definition. Let X be a vector space over K = R or C. A map q : X → R is called a

seminorm if:
1. q(αx) = |α|q(x) for all α ∈ K, x ∈ X
2. q(x + y) 6 q(x) + q(y).

Theorem (Hahn-Banach, general case). Let X be a vector space over K = R or C,

q : X → R a seminorm on X, M ⊂ X a linear subspace and f : M → K such that
|f (x)| 6 q(x) for all x ∈ M .

Then there exists F : X → K, linear, with F |M = f , and |F (x)| 6 q(x) for all x ∈ X.

Proof. For K = R, apply Hahn-Banach for real vector spaces. Then there exists F : X → R
such that F |M = f , F (x) 6 q(x) and −F (x) = F (−x) 6 q(−x) = q(x) for all x ∈ X.
So |F (x)| 6 q(x) for all x ∈ X.

For K = C, consider M and X as real vector spaces. Then Re f : M → R is linear with

respect to R, and |Re f (x)| 6 |f (x)| 6 q(x) for all x ∈ M . So there exists g : X → R,
linear with respect to R, and g|M = Re f and |g(x)| 6 q(x) for all x ∈ X.

Let F (x) = g(x) − ig(ix). Then F (ix) = g(ix) − ig(−x) = g(ix) + ig(x) = iF (x), so F
is C-linear.

Then for all x ∈ M , we have Re F (x) = g(x) = Re f (x), and Im F (x) = −g(ix) =
−Re f (ix) = −Re if (x) = Im f (x).

So F (x) = f (x) for all x ∈ M and F |M = f .

Lastly, |F (x)| = | eiθ F (x) | = |F (eiθ x)| = |g(eiθ x)| 6 q(eiθ x) = q(x) for all x ∈ X. 2
| {z }
real

Consequences of Hahn-Banach
Throughout, K = R or C, and X is a normed space over K.

1. Let M be a linear subspace of X, and f ∈ M ∗ . Then there exists F ∈ X ∗ with F |M = f

and kF k = kf k.
Indeed if q(x) = kf k kxk, then q is a seminorm and by Hahn-Banach, there exists

12
 F : X → K linear with F |M = f and |F (x)| 6 kf k kxk for all x ∈ X. So kF k 6 kf k.
But since F |M = f , we have kF k = kf k.
2. Let x0 ∈ X \ {0}. Then there exists f ∈ X ∗ such that kf k = 1 and f (x0 ) = kx0 k.
Indeed, let M = hx0 i. Then the function fe(λx0 ) = λkxk for λ ∈ K has kfek = 1. By
Hahn-Banach, there is f ∈ X ∗ with kf k = 1 and f |M = fe. So f (x0 ) = fe(x0 ) = kx0 k.
3. Let Z be a linear subspace of X, and y ∈ X \ Z, and let d = dist(y, z) = inf z∈Z ky − zk.
Note d > 0. Then there exists F ∈ X ∗ such that kF k = 1, F |Z = 0, F (y) = d.
Indeed, let M = Z + hyi, and f : M → K be defined by f (z + ty) = td. So f is linear
and:
|f (z + ty)| |t|d d
kf k = sup = sup z 6 =1
z∈Z,t∈K kz + tyk z∈Z,t∈K |t| k t + yk d
So kf k = 1. By Hahn-Banach, there exists F : X → R such that kF k = 1 and F |M = f .
So F (z) = f (z) = 0 for all z ∈ Z, and F (y) = f (y) = d.
4. X ∗ separates the points of X. In other words, for all x, y ∈ X with x 6= y, there exists
f ∈ X ∗ such that f (x) 6= f (y).
Indeed if x 6= y, then w = x − y 6= 0. So by Hahn-Banach, there exists F ∈ X ∗ such
that kF k = 1 and F (x) − F (y) = F (x − y) = kx − yk > 0. So F (x) 6= F (y).

5. For all x ∈ X, kxk = sup |f (x)|.

f ∈X ∗ ,kf k61

Indeed, |f (x)| 6 kf k kxk 6 kxk for all f ∈ X ∗ with kf k 6 1. On the other hand, if
x ∈ X \ {0}, by Hahn-Banach there exists f ∈ X ∗ such that kf k = 1 and f (x) = kxk.
So |f (x)| = kxk and sup |f (x)| > kxk.
kf k61

6. Let T ∈ B(X, Y ), so that T ∗ : Y ∗ → X ∗ , f 7→ (T ∗ (f )), where (T ∗ (f ))(x) = f (T (x)).

Then T ∗ ∈ B(Y ∗ , X ∗ ) and kT ∗ k = kT k.
We already know kT ∗ k 6 kT k, and we want kT ∗k > kT k.
So, given ε > 0, find x0 ∈ X with kx0 k 6 1 and kT k 6 kT (x0 )k + ε. By Hahn-Banach,
there exists f ∈ Y ∗ such that kf k = 1 and f (T (x0 )) = kT (x0 )k.
Then kT k − ε 6 kT x0 k = f (T x0 ) = (T ∗ f )(x0 ), so kT ∗ f k > kT k − ε, and thus kT ∗ k >
kT k − ε. This is true for all ε, and so kT ∗ k > kT k.

Definition. A normed space X is called separable if it has a countable dense subset.

Example. ℓp (K) is separable for all 1 6 p < ∞, but ℓ∞ (K) is not separable.

Theorem. Let X be a normed space. If X ∗ is separable then X is separable.

Proof. Let {fk }k∈N be a dense subset of X ∗ . Then for all k ∈ N, there exists xk ∈ X with
kxk k = 1 such that |fk (xk )| > 12 kfk k. Let A be the set of finite linear combinations of
the {xk } with rational coefficients. Then A is countable.

Suppose that A 6= X. Then, by Hahn-Banach, there exists f ∈ X ∗ such that f 6= 0

and f |A = 0. So there is a subsequence kj → ∞ as j → ∞ such that fkj → f in X ∗ .

Then kfkj − f k → 0. But kfkj − f k > |fkj (xkj ) − f (xkj )| = |fkj (xkj )| > 21 kfkj k.

So fkj → 0 as j → ∞, which implies f = 0. /\/\ Therefore A = X. 2

13
Definition. Let X be a normed space. Define X ∗∗ = (X ∗ )∗ to be the second dual of X.
Then X ∗∗ is always a Banach space. Also define ϕX : X → X ∗∗ by x 7→ ϕX (x) where
ϕX (x) : X ∗ → K, f 7→ ϕX (x)(f ) = f (x).

Proposition. ϕX is linear and isometric.

Proof. Linearity is clear.

|ϕX (x)(f )|
|ϕX (x)(f )| = |f (x)| 6 kf k kxk, so sup 6 kxk, and so kϕX (x)k 6 kxk.
f 6=0 kf k

On the other hand, if x 6= 0 then by Hahn-Banach there exists fe ∈ X ∗ such that

kfek = 1 and kfe(x)k = kxk.

|ϕX (x)(fe)|
So |ϕX (x)(fe)| = |fe(x)| = kxk. So = kxk, and thus kϕX (x)k > kxk. 2
kfek

Definition. A normed space X is called reflexive if ϕX is surjective and thus defines an

isometric isomorphism between X and X ∗∗ .

Remark. Every reflexive space is complete. However, there are Banach spaces which are
not reflexive.

Theorem. Let X be a reflexive Banach space and M ⊂ X a closed linear subspace. Then
M is reflexive.

Proof. We have ϕX : X → X ∗∗ and ϕM : M → M ∗∗ , a map j : M ֒→ X such that

j(m) = m, and also j ∗ : M ∗ ֒→ X ∗ such that j ∗ (f )(m) = f (j(m)) = f (m).

In other words, for f ∈ M ∗ , we have j ∗ (f ) = f |M .

We also have j ∗∗ : M ∗∗ ֒→ X ∗∗ such that j ∗∗ (m∗∗ )(f ) = m∗∗ (j ∗ (f )) = m∗∗ (f |M ).

Let m∗∗ ∈ M ∗∗ . We want x ∈ M such that ϕM (x) = m∗∗ .

Since X is reflexive, there exists x ∈ X such that ϕX (x) = j ∗∗ (m∗∗ ). If x ∈

/ M then
by Hahn-Banach, there exists x∗ ∈ X ∗ such that x∗ |M = 0 and x∗ (x) = 1. (Note we
require here that M is closed.)

So 1 = x∗ (x) = ϕX (x)(x∗ ) = j ∗∗ (m∗∗ )(x∗ ) = m∗∗ (j ∗ (x∗ )) = m∗∗ (x∗ |M ) = 0. /\/\

Thus x ∈ M . We now need to show ϕM (x) = m∗∗ . To prove this, note that the
following diagram is commutative:
ϕX
X −→ X ∗∗
j↑ ↑ j ∗∗
ϕM
M −→ M ∗∗

For example, ϕX ◦ j = j ∗∗ ◦ ϕM .

For arbitrary m∗ ∈ M ∗ , let x∗ ∈ X ∗ be such that x∗ |M = m∗ . Then ϕM (x)(m∗ ) =

m∗ (x) = x∗ (x) = ϕX (x)(x∗ ) = j ∗∗ (m∗∗ )(x∗ ) = m∗∗ (j ∗ (x∗ )) = m∗∗ (x∗ |M ) = m∗∗ (m∗ ).

So ϕM (x) = m∗∗ . 2

14
Theorem. Let X be a Banach space. Then X is reflexive iff X ∗ is reflexive.

Proof. Suppose that X is reflexive. We must show that ϕX ∗ : X ∗ → X ∗∗∗ is surjective.

Let x∗∗∗ ∈ X ∗∗∗ , and define x∗ = x∗∗∗ ◦ ϕX . This is linear and continuous, so in X ∗ .

For any x∗∗ ∈ X ∗∗ , we have x∗∗ = ϕX (x) for some x ∈ X, since X is reflexive. Then:

ϕX ∗ (x∗ )(x∗∗ ) = x∗∗ (x∗ ) = ϕX (x)(x∗ ) = x∗ (x) = x∗∗∗ (ϕX (x)) = x∗∗∗ (x∗∗ ).

In other words, ϕX ∗ (x∗ ) = x∗∗∗ . Thus ϕX ∗ is surjective, as required.

Exercise: if X ∗ is reflexive then X is reflexive. 2

Examples. ℓp is reflexive if 1 < p < ∞. Note ℓ∗p = ℓq , so ℓ∗∗ ∗

p = ℓq = ℓp for 1 < p < ∞.

But ℓ1 and ℓ∞ are not reflexive. Indeed, if ℓ1 were reflexive, then ℓ1 =∼ ℓ∗∗ implies ℓ∗∗
1 1
∗ ∼
is separable and then ℓ1 = ℓ∞ is separable. /\/\ Thus ℓ∞ is also not reflexive.

15
3. Baire Category Theorem
Suppose X is a metric space. A subset A ⊂ X is said to be dense if it intersects every open
set U ⊂ X. Equivalently, A intersects every open ball Br (x0 ) = {x ∈ X : d(x, xo ) < r}.

E.g., Q is dense in R, since every open interval in R contains a point in Q.

∞
Theorem (Baire Category Theorem). Suppose T∞ (Gn )n=1 is a sequence of open dense
subsets of a complete metric space. Then n=1 Gn is dense.

The proof requires the following fairly standard lemma.

Lemma. Let X be a complete metric space and suppose F1 ⊃ TF2 ⊃ F3 ⊃ . . . are nested,
closed, non-empty subsets of X with diam(Fn ) → 0. Then ∞
n=1 Fn 6= ∅.

Recall diam(S) = sup d(x, y).

x,y∈S

Proof of lemma. Pick xn ∈ Fn for each i. Since d(xn , xm ) 6 diam(Fm ) → 0 as n > m →

∞, this is a Cauchy sequence.

Suppose xn → x∗ . Then, for each m, the sequence xm , xm+1 , xm+2 , . . . lies entirely in
Fm (as the sets are nested) and tends to x∗ .
T∞
Since Fm is closed, we have x∗ ∈ Fm , and so x∗ ∈ n=1 Fn . 2
T∞
Proof of BCT. Consider some ball Br0 (x0 ). We want a point of n=1 Gn inside here.

Since G1 is open and dense, Br0 (x0 ) meets G1 in an open set, and hence Br0 (x0 )
contains an open ball. In fact, it contains some closed ball Br1 (x1 ) with r1 > 0.
(Indeed, Bε/2 (t) ⊂ Bε (t) – that is, every open ball contains a non-empty closed ball.)

Now, Br1 (x1 ) meets G2 , so find a closed ball Br2 (x2 ) ⊂ Br1 (x1 ) ∩ G2 . Continue in this
fashion, finding ever smaller balls Br3 (x3 ) ⊂ Br2 (x2 ) ∩ G3 , etc. Do this in such a way
that rn → 0.

The closed balls Brn (xn ) are nested,

T∞ and hence by the lemma have some x∗ in their
intersection. By construction, x ∈ n=1 Gn . But also, x∗ ∈ Br0 (x0 ).
∗

T∞
Since x0 and r0 are arbitrary, we have that n=1 Gn is dense. 2

An easy corollary of this is the following.

Theorem (Baire Category Theorem II). Suppose X is aScomplete metric space, and
∞
that (Fn )∞
n=1 is a sequence of closed sets such that X = n=1 Fn .

Then at least one Fn has non-empty interior. That is, some Bε (x0 ), ε > 0, is in it.
T∞
Proof. Take Gn = X \ Fn . Apply Baire Category: since n=1 Gn = ∅, at least one of Gn is
not dense.

Thus some ball Bε (x0 ) doesn’t meet Gn , whence Bε (x0 ) ⊂ Fn . 2

Standard terminology: a set is meagre, or of first category, if it is a countable union of

16
closed sets with empty interior; it is of second category otherwise.

So Baire Category could be stated as: a complete metric space is of second category.

We will see various applications of the Baire Category Theorem.

Application: continuous, nowhere differentiable functions

Theorem. They exist. More precisely, there is a function f ∈ C[0, 1] (say) that is not
differentiable at any point x ∈ (0, 1).

Idea of the proof. Suppose

S∞not, i.e. every continuous function is differentiable somewhere.
Then write C[0, 1] = n=1 Sn as a union of countably many closed sets. Here, C[0, 1]
has the supremum norm – note this is a Banach space. Baire Category ⇒ one of the
Sn has non-empty interior. We’ll get a contradiction.

Proof. Define Sn to be the set of all f ∈ C[0, 1] such that there exists some x ∈ (0, 1) such
that ‘the slope of f is bounded near x by n’. That is to say |f (x)−f (y)|
|x−y| 6 n for all y
with 0 < |x − y| 6 n1 .

If f is differentiable at x then
S∞ (Mean Value Theorem) f lies in Sn for n big enough. If
the theorem is false, then n=1 Sn = C[0, 1]. So it is sufficient to show that each Sn is
(a) closed and (b) has empty interior.
(a) Suppose (fi )∞
i=1 is a sequence in Sn such that fi → f for some f ∈ C[0, 1]. We
aim to show that f ∈ Sn .

|fi (xi )−fi (y)| 1

For each i, there is xi such that |xi −y| 6 n whenever 0 < |xi − y| 6 n.

Since [0, 1] is compact, by Bolzano-Weierstrass there is a convergent subsequence

of the xi . To ease notation, wlog the sequence xi itself converges to some x.

Suppose that |x − y| < n1 . Then |xi − y| 6 1

n for i large enough. So, by continuity
and the fact that fi → f , we have
|f (x) − f (y)| |fi (xi ) − fi (y)|
= lim 6 n.
|x − y| i→∞ |xi − y|
(Exercise: show that fi (xi ) → f (x).)
1
Since f is continuous, the same holds for |x − y| = n too. Hence f ∈ Sn .

(b) We need to show Sn contains no ball Bε (f0 ). In other words, for any continuous
f0 ∈ C[0, 1] we want some f ∈ C[0, 1] with kf − f0 k∞ < ε but whose slope is not
bounded (in the sense of belonging to Sn ).

We’ll do this in two stages.

First, we’ll find a piecewise linear function f1 with kf1 − f0 k < 2ε . In fact, f1 will
be linear on each segment [ Mi , i+1
M ] for i = 0, 1, . . ., M − 1.

So define f1 ( Mi ) = f0 ( Mi ) and interpolate linearly elsewhere.

Since f0 is uniformly continuous, if M is sufficiently large then |f0 (x) − f0 (y)| < 4ε
1
whenever |x−y| 6 M . It is an easy exercise to show that this implies kf1 −f0 k < 2ε .

17
 We now define f = f1 + 2ε g, where g is a suitable function bounded by 1 everywhere,
so that kf − f0 k < ε.

Let M ′ ≫ M . For i = 0, . . ., M ′ − 1, define g( Mi ′ ) = (−1)i , and interpolate

linearly.

The slope of g at every point is at least 2M ′ , hence the slope of f at every point
is at least εM ′ − slope(f1 ). By taking M ′ large enough in terms of ε and slope(f1 )
(which is bounded since f1 is piecewise linear), we can make this > n. Thus
f∈ / Sn , as required. 2

Application: uniform boundedness principle

Theorem. Suppose that X is a complete metric space and that F is a collection of continuous
real-valued functions on X. Suppose that for each x ∈ X, sup |f (x)| < ∞.
f ∈F

Then there is a ball Bε (x0 ), ε > 0, such that sup sup |f (x)| < ∞.
f ∈F x∈Bε (x0 )

T
Proof. Define Sn = f ∈F {x ∈ X : |f (x)| 6 n}. Then Sn is an intersection of closed sets,
so is closed.
S∞
And n=1 Sn = X – indeed, x ∈ Sn whenever n > sup |f (x)|.
f ∈F

Since X is complete, by Baire Category, one of the Sn has non-empty interior. 2

Theorem (Banach-Steinhaus uniform boundedness principle). Suppose X is a Ba-

nach space and T is a family of bounded linear operators from X to some other normed
space Y . Suppose sup kT xk < ∞ for all x ∈ X.
T ∈T

Then sup kT k < ∞.

T ∈T

Proof. Apply the previous theorem to the functions x 7→ kT xk, for T ∈ T . This is a family
of continuous functions on a complete metric space X. Hence there is a ball Bε (x0 ),
ε > 0, such that sup sup kT xk < ∞. Call this M .
T ∈T x∈Bε (x0 )

ε
In particular, if z ∈ X has kzk < 2 then kT (x0 + z)k 6 M and kT (x0 − z)k 6 M .

Hence kT zk = k 12 T (x0 + z) − 21 T (x0 − z)k 6 M .

2M
Scaling up, we see that kT yk 6 for all y with kyk 6 1 and for all T .
ε
2M
Hence sup kT k 6 . 2
T ∈T ε

Application: divergence of Fourier series

Suppose f is a 2π-periodic continuous function – i.e. f (x) = f (x + 2π) for all x.
P
We would like to say that f can be expanded as a Fourier series: f (x) = k∈Z ak eikx .

18
 P
Heuristic computation: assume f (x) = ak eikx , and then
Z 2π Z 2π X X Z 2π
−imx i(k−m)x
f (x)e dx = ak e dx ∼ ak ei(k−m)x dx ∼ 2πam ,
0 0 k k 0


R 2π 2π if n = 0
as einx dx =
0 0 6 0
if n =
R 2π
Thus (we are tempted to say), ak = 1
2π 0 f (x)e−ikx dx = fb(k).
P b ikx
Question. How rigorous is this? Is f (x) = k f (k)e in any meaningful way?

Theorem (Kolmogorov). There exists a continuous function f , 2π-periodic, whose Fourier

series diverges at x = 0.
P
Proof. Let’s look at the partial sums of the Fourier series for f at 0, SN f (0) = |k|6N fb(k).

Note that f 7→ SN f (0) is a linear operator, which we’ll call ϕN . It is defined on the
space X of 2π-periodic continuous functions, which we’ll identify with {f ∈ C[0, 2π] :
f (0) = f (2π)}. Obviously, with the sup norm k k∞ this is a closed subspace of C[0, 2π],
and hence it is a Banach space.

We’ll show that each ϕN is bounded but that kϕN k → ∞ as N → ∞. The result
then follows from uniform boundedness: if supN kϕN f k < ∞ for all f ∈ X then
supN kϕN k < ∞, contradiction.
X Z 2π Z 2π
1 X 1
ϕN f = SN f (0) = fb(k) = f (x)e−ikx dx = f (x)DN (x)dx,
2π 0 2π 0
|k|6N |k|6N

P
where DN (x) = Dirichlet kernel = |k|6N eikx .
Z 2π
1
This makes it clear that ϕN is bounded: kϕN f k 6 sup |f (x)| · |DN (x)|dx.
x 2π 0
| {z }
this is certainly finite

Z 2π
1
Actually, kϕN k = |DN (x)|dx.
2π 0

Z 2π
1
To see this, take f (x) = e−i arg DN (x) , then kf k∞ = 1 and ϕN f = |DN (x)|dx.
2π 0

R 2π
All we have to show is that 0
|DN (x)|dx → ∞. We estimate DN (x) as follows.

e−iN x (e(2N +1)ix − 1)

DN (x) = e−iN x (1 + eix + . . . + e2N ix ) =
eix − 1
1 1 1
e(N + 2 )ix − e−(N + 2 )ix sin(N + 2 )x
= =
eix/2 − e−ix/2 sin 12 x

π(r+ 12 ) 1
This has peaks around x = N + 12
, say of width 10N .

π(r+ 12 )
Around such a peak, | sin(N + 21 )x| > 12 , but sin 21 x 6 12 x 6 2(N + 21 )
6 10r
N .

19
 1 N 1
Hence the contribution to the integral from the rth peak is at least 10N · 10r = 100r .

Z 2π N
X 1
But the harmonic series diverges, and |DN (x)|dx > > c log N .
0 r=1
100r

R 2π
(In fact, 0
|DN (x)|dx ∼ C log N .) 2
P
Remarks (non-examinable). What went wrong? We looked at SN f = |k|6N fb(k)eikx ,
and the cut-off at |k| = N is too sharp.
P |k| b
Look instead at Sf
Nf = |k|6N (1 − N )f (k)e
ikx
.
R 2π
Here, Sf
N f (0) = 0
f (x)KN (x)dx, where KN (x) is called the Fejér kernel.

 2
sin(N − 12 )x R 2π
KN (x) ∼ , and |KN (x)|dx < ∞ uniformly in N .
sin 12 x 0

The theory of Sf
N is rather nice.

Application: open mapping theorem

Theorem (Open Mapping Theorem). Suppose T : X → Y is a surjective bounded lin-
ear operator between two Banach spaces X, Y . Then T maps open sets to open sets.

Proof. It suffices to show (by scaling and linearity properties of T ) that T (BX (1)) is open,
where BX (1) is the open unit ball {x ∈ X : kxkX < 1}. It then follows easily that
T (BX (x0 , ε)) is open for any x0 ∈ X and any ε > 0.

Plan. (1) Apply the Baire Category Theorem to conclude that T (BX (1)) contains an
open ball BY (δ), δ > 0. Here we use the completeness of Y .
(2) Mess around a bit to show that in fact T (BX (1)) in fact contains BY (δ). Here
we use the completeness of X.

S∞
(1) T surjective implies that Y = n=1 T (BX (n)). But T (BX (n)) = nT (BX (1)) and
S∞ S∞
so Y = n=1 nT (BX (1)), and thus trivially Y = n=1 nT (BX (1)).

Applying Baire Category (2nd form), it follows that one of these sets nT (BX (1))
has non-empty interior. Hence T (BX (1)) has non-empty interior – say it contains
BY (y0 , δ) = {y ∈ Y : ky − y0 kY < δ}.

Note that T (BX (1)) is symmetric about the origin (i.e. if it contains y then it
contains −y) and is convex since BX (1) is.

If kzk < δ then T (BX (1)) contains both y0 + z and −y0 + z, and hence contains
1 1
2 (y0 + z) + 2 (−y0 + z) = z.

Thus T (BX (1)) ⊃ BY (δ), as required.

(2) For notational simplicity, assume δ = 1. (We can always replace T by Te = 1δ T ,

which doesn’t affect the open mapping property.)

20
 Assume, then, that T (BX (1)) ⊃ BY (1) (∗). We wish to conclude from this that
T (BX (1)) ⊃ BY (1).

Observe that (∗) implies that for any y ∈ Y and ε > 0 there is an x such that
kxkX 6 kykY and y = T x + y ′ with ky ′ kY 6 ε.

(Let r = kykY . By scaling BY (r) ⊂ T (BX (r)), so BY (r) ⊂ T (BX (r)), so y ∈

T (BX (r)), so given any ε > 0 there exists yb ∈ T (BX (r)) with ky − ybkY < ε, so
there exists x ∈ BX (r) with ky − T xkY < ε, and such an x has kxkX < r.)

Apply this repeatedly. Let y ∈ BY (1), and set y1 = y. Let ε1 , ε2 , . . . be a rapidly

decreasing sequence of positive reals to be chosen later.

Find x1 with kx1 k 6 ky1 k such that y1 = T x1 + y2 with ky2 k 6 ε1 .

Find x2 with kx2 k 6 ky2 k such that y2 = T x2 + y3 with ky3 k 6 ε2 . And so on.

Define x = x1 + x2 + x3 + . . .. If the εi are sufficiently rapidly decaying (e.g.

εi = 2−i ), this sum converges, so this makes sense. (Here, we have used the
completeness of X.)

Since T is bounded (and hence continuous), T x = limn→∞ T (x1 + . . . + xn ) =

limn→∞ (y − yn+1 ) = y (by construction).

Furthermore, kxk 6 kx1 k + kx2 k + . . . 6 ky1 k + ky2 k + . . . 6 kyk + ε1 + ε2 + . . . < 1,

if the εi are chosen sufficiently small (since y ∈ BY (1) means kyk < 1).

(Explicitly, if you like, if kyk = 1 − η, we can take εi = η i+1 .)

We have found x ∈ BX (1) such that T x = y. Since y was arbitrary, this implies
that T (BX (1)) ⊃ BY (1). 2

Corollary (Identity/Inversion Principle). Suppose X and Y are Banach spaces and

T : X → Y is bounded, linear and bijective. Then T −1 , which manifestly exists as a
function, is a bounded linear operator.

Proof. Only the boundedness requires proof. But if U ⊂ X is open then (T −1 )−1 U = T U
is open, by the open mapping theorem. 2

Theorem (Closed Graph Theorem). Suppose X, Y are Banach spaces, and T : X → Y

a linear operator.

Then T is bounded if and only if the graph Γ = {(x, T x) : x ∈ X} is closed in the

product topology on X × Y .

Remark. One direction is easy. If you have f : X → Y as a continuous function between

metric spaces X, Y then the graph Γ is closed. Indeed, if xn → x then f (xn ) → f (x).
(Hausdorff is okay too.)

The converse is not true in general.

Proof. X × Y is a normed space with norm k(x, y)kX×Y = kxkX + kykY . (Easy exercise:
the topology induced on X × Y is the product topology.)

21
So X × Y is a Banach space and Γ is a closed linear subspace of it. Hence Γ is also a
Banach space (with the same norm). Indeed, Γ is certainly a normed space. If (γn )∞
n=1
is a Cauchy sequence in Γ then it is certainly Cauchy in X × Y and hence converges to
some γ. But Γ closed ⇒ γ ∈ Γ.

Consider the map π : Γ → X defined by projection, namely π(x, y) = x. This is

obviously linear, bounded and bijective. Hence by the identity/inversion principle, the
inverse π −1 : X → Γ is a bounded linear operator. Thus for some constant C,

kxkX + kT xkY = k(x, T x)kX×Y = kπ −1 (x)kX×Y 6 CkxkX .

Hence kT xkY 6 (C − 1)kxkX . 2

22
4. The Space C(X)
Let X be a topological space. We’ll study the space C(X) of continuous R-valued functions
in some generality. (Much of the theory is easier when X is a metric space.)

Our spaces will be compact (aside: for much of the theory, locally compact will do – see, e.g.,
Rudin’s red book), and, importantly, Hausdorff.

Recall. X is Hausdorff if for every pair a, b of distinct points in X there are disjoint open
sets U, V with a ∈ U , b ∈ V .

Lemma. (Recall Met&Top.) Suppose X is a compact Hausdorff space and A ⊂ X. Then A

is compact iff A is closed.

Proof. A closed ⇒ A compact: easy. The converse is slightly trickier. Suppose A is compact
and that x ∈/ A. For each a ∈ A, the Hausdorff property gives open sets Ua ∋ a and
Va ∋ x which are disjoint. Since A is compact and the Ua s cover A, there is a finite
subcover: Ua1 ∪ . . . ∪ Uan , say.
Tn Sn
But then i=1 Vai is open, contains x and is disjoint from i=1 Uai and hence from A.
Since x was arbitrary, X \ A is open. 2

Definition. A topological space X is said to be normal if for every pair A, B ⊂ X of disjoint

closed sets, one can find disjoint open sets U, V with A ⊂ U and B ⊂ V .

Proposition. Suppose X is a compact Hausdorff space. Then X is normal.

Proof. Suppose A, B ⊂ X are disjoint closed sets. For any pair a ∈ A, b ∈ B, there are
disjoint open sets Ua,b and Va,b with a ∈ Ua,b and b ∈ Va,b .

Fix a. The sets Va,b , b ∈ B, form an open cover of B, and since B is compact (by the
lemma above), there is a finite subcover Va,b1 ∪ . . . ∪ Va,bn . Write Va for this set, and
write Ua for Ua,b1 ∩ . . . ∩ Ua,bn .

Then Ua , Va are open and disjoint, and a ∈ Ua whilst B ⊂ Va . But now the Ua form
an open cover of A, and since A is compact (again by the lemma), there is a finite
subcover Ua1 ∪ . . . ∪ Uam . Take U to be this set, and let V = Va1 ∩ . . . ∩ Vam .

Then U, V are open and disjoint, and A ⊂ U whilst B ⊂ V . 2

Lemma. Let X be a topological space. Then X is normal iff it has the ‘sandwiching prop-
erty’. This is: if A is closed, W is open, and A ⊂ W , then there is an open set U with
A ⊂ U ⊂ U ⊂ W.

Proof (sketch). Suppose X has SP. We want X to be normal. Let A, B ⊂ X be disjoint

closed sets. Take W = X \ B. Then W is open and A ⊂ W . Find the U guaranteed
by SP, and set V = X \ U . Then B ⊂ V .

The converse is essentially identical. 2

Theorem (Urysohn’s Lemma). Let X be a compact Hausdorff space, and let A, B ⊂ X

be disjoint closed sets. Then there is a continuous function f : X → [0, 1] with f |A = 0
and f |B = 1.

23
 dist(x, A)
Remark. If X is a metric space, take f (x) = .
dist(x, A) + dist(x, B)

Proof. We make an ‘onion decomposition’: f is constructed iteratively.

Set W = X \ B. This is an open set containing A. At the first step use the sandwiching
property twice to find open sets U0 , U1 such that

A ⊂ U0 ⊂ U0 ⊂ U1 ⊂ U1 ⊂ W.

Next, find an open set U 21 such that

A ⊂ U0 ⊂ U0 ⊂ U 21 ⊂ U 21 ⊂ U1 ⊂ U1 ⊂ W.

Next find open sets U 41 and U 43 such that

A ⊂ U0 ⊂ U0 ⊂ U 41 ⊂ U 41 ⊂ U 21 ⊂ U 21 ⊂ U 43 ⊂ U 43 ⊂ U1 ⊂ U1 ⊂ W.

Carry on doing this for all dyadic rationals (rationals with denominator a power of 2.)
We end up with open sets Ur for each dyadic rational r, with the crucial property that
if r < s then Ur ⊂ Us .

For x ∈ U1 , define f (x) = inf{r : x ∈ Ur }, and define f (x) = 1 otherwise. Certainly

f (x) = 0 on A (since A ⊂ Ur for all r) and f (x) = 1 on B (since B ⊂ X \ U1 ). We
need to show that f is continuous.

Claim. Suppose α ∈ (0, 1). Then f −1 ([0, α]) and f −1 ([α, 1]) are closed.
T
Proof. First, we show that f −1 ([0, α]) = r>α Ur .
T
If x ∈ LHS, then f (x) 6 α, so x ∈ Ur for all r > α, and thus x ∈ r>α Ur ⊂ RHS.
If x ∈ RHS, then x ∈ Ur for all r > α, and so x ∈ Us for all s > r > α, and hence,
since r was arbitrary, for all s > α. Therefore f (x) 6 α, and so x ∈ LHS.
S S
Second, we show f −1 ([α, 1]) = X \ r<α Ur . Equivalently, f −1 ([0, α)) = r<α Ur .
S
For: x ∈ f −1 ([0, α)) ⇔ f (x) < α ⇔ x ∈ Ur for some r < α ⇔ x ∈ r<α Ur .

This implies that f −1 (open interval) is open, whence f −1 (any open set) is open. So f
is continuous, as required. 2

Theorem (Tietze Extension Theorem). Suppose X is a compact Hausdorff space. Sup-

pose that S ⊂ X is closed and that f : S → [−1, 1] is a continuous function. Then
there is a continuous function fe : X → [−1, 1] such that fe|S = f .

Proof. We’ll apply Urysohn’s Lemma repeatedly to find continuous function on X which
approximate f better and better on S.

First (zeroth) step: let A0 = {x ∈ S : f (x) 6 − 13 } and B0 = {x ∈ S : f (x) > 13 }. Then

A0 and B0 are disjoint closed sets. Set f0 = f .

By (a rescaled version of) Urysohn’s Lemma, there is a continuous function g0 : X →

[− 31 , 13 ] such that g0 |A0 = − 31 and g0 |B0 = 13 .
2
Define f1 = f0 − g0 . Note that |f1 (x)| 6 3 for all x – i.e., f1 : S → [− 32 , 23 ].

24
 Now define A1 = {x ∈ S : f1 (x) 6 − 13 · 23 } and B1 = {x ∈ S : f1 (x) > 13 · 23 }.

As before, find a continuous function g1 : X → [− 13 · 32 , 13 · 23 ] such that g1 |A1 = − 31 · 23

and g1 |B1 = 31 · 23 .

n
Continue in this way, obtaining functions fn with |fn (x)| 6 23 for all x ∈ S, and

n
n
An = {x ∈ S : fn (x) 6 − 31 · 23 } and Bn = {x ∈ S : fn (x) > 13 · 23 }, and functions

n
n
n
n
gn : X → [− 31· 32 , 13· 23 ], continuous and with gn |An = − 31· 23 and gn |Bn = 13· 23 .


2 n
Define fe(x) = g0 (x) + g1 (x) + . . .. This converges uniformly on X, since kgn k 6 13 · 3 .
P∞

2 n


In fact, kfek∞ 6 n=0 kgn k∞ 6 1
3 1+ 2
3 + 3 + . . . = 1.

By construction, fe|S = f0 = f . 2

Stone-Weierstrass Theorem

We’ll prove a significant generalisation of the ‘well-known’ fact that any continuous function
on [a, b] can be uniformly approximated by polynomials.

Definition. Let X be a topological space. C(X) denotes real-valued continuous functions

on X together with the sup norm. Let A ⊂ C(X) be a set of functions. Then we say
that A is an algebra if whenever f, g ∈ A, we have f + g ∈ A, f g ∈ A, and λf ∈ A
for all λ ∈ R.

Examples. (a) X = [a, b] and A ⊂ C(X) is the set of real-coefficient polynomials p(x) =
c0 + c1 x + . . . + cn xn .

(b) X = [a1 , b1 ] × [a2 , b2 ] ⊂ R2 , with A the collection of all finite sums g1 (x)h1 (y) +
. . . + gn (x)hn (y), where gi : [a1 , b1 ] → R and hj : [a2 , b2 ] → R are continuous 1-variable
functions.

Theorem (Stone-Weierstrass). Let X be a compact Hausdorff topological space. Let

A ⊂ C(X) be an algebra of functions with the following separation-of-points prop-
erty: if x, y ∈ X are distinct and if s, t ∈ R, then there is some f ∈ A with f (x) = s
and f (y) = t.

Then A, the closure of A in C(X), is all of C(X). (I.e., A is dense in C(X).)

In other words, any continuous function on X can be uniformly approximated by functions

in A.

Remarks. The separation-of-points property is trivially satisfied in Examples 1 and 2. Thus

every continuous real-valued function f : [a, b] → R can be uniformly approximated by
polynomials.

Idea. The proof has two main steps, rather orthogonal to one another.
(1) Show that A is a lattice. That is, if f, g ∈ A then max(f, g) and min(f, g) ∈ A.
(2) Use the lattice property, compactness of X and separation-of-points property to
show A = C(X).

Proof. (1) It suffices to show that if f ∈ A then |f | ∈ A. For max(f, 0) = 12 (f + |f |) ∈ A,

25
 and so max(f, g) = max(f − g, 0) + g ∈ A, and similarly for min(f, g).

Rescaling if necessary, assume |f (x)| 6 1 for all x.

Claim. There is a sequence (Pn )∞

n=1 of polynomials with zero constant term such
that Pn (t) → |t| uniformly on [−1, 1] as n → ∞.

Then, Pn (f ) lies in A because A is an algebra. (If A is an algebra then so is A,

because if fn → f and gn → g, then fn + gn → f + g and fn gn → f g, so limits
are preserved.) But Pn (f ) → |f | uniformly, so since A is closed on C(X) (with
the sup norm, of course), we have |f | ∈ A.

Proof of claim.
√ It suffices to show that there are polynomials Qn such that
Qn (t) → t uniformly on [0, 1], since we may then take Pn (s) = Qn (s2 ).
√
Let δ > 0 be small. There is an analytic branch of z + δ in the ball B 1+δ ( 12 ),
√ 2
and hence there is a Taylor series: z + δ = c0 + c1 (z − 21 ) + c2 (z − 12 )2 + . . .,
uniformly convergent for |z − 21 | 6 21 , say.

Truncating this series at the ith term, and restricting to real values of z (which
certainly includes [0,√1]), we get a sequence of polynomials Ri,δ (t) such that
lim sup |Ri,δ (t) − t + δ| = 0.
i→∞ t∈[0,1]

√
Note that the constant term of Ri,δ (t) tends to δ as i → ∞ (put√t = 0). So
√
ei,δ (t) = Ri,δ (t) − Ri,δ (0), then lim sup |R
let R ei,δ (t) − t + δ + δ| = 0.
i→∞ t∈[0,1]

Now define Qn (t) = R ei(n), 1 (t), where i(n) is chosen large enough that
q q
n

ei(n), 1 (t) − t + 1 + 1 6 1 .
sup R n n n n
t∈[0,1]

√ √ q q
1
Then sup |Qn (t) − t| 6 n+ sup t + n1 − t + 1
n → 0 as n → ∞.
t∈[0,1] t∈[0,1]

(Exercise.)

This concludes the proof of (1).

(2) Consider any closed lattice A ⊂ C(X) with the separation-of-points property. Let
f ∈ C(X) be arbitrary. We need to show that f can be approximated arbitrarily
closely by functions from A. Let ε > 0 be arbitrary.

Given any x, y ∈ X, by separation-of-points, there is a function fx,y ∈ A such that

fx,y (x) = f (x) and fx,y (y) = f (y). Define Ux,y = {z ∈ X : fx,y (z) < f (z) + ε}.
These sets Ux,y are open, and by construction x, y ∈ Ux,y .

Fix x. Then the sets Ux,y form an open cover of X and so by compactness of X
there is a finite subcover Ux,y1 ∪ . . . ∪ Ux,yn . Define fx = min{fx,y1 , . . ., fx,yn }.
This fx ∈ A and fx (x) = f (x) and fx (z) < f (z) + ε for all z ∈ X, since z is in at
least one Ux,yi .

For each x, define Vx = {z ∈ X : fx (z) > f (z) − ε}. By construction, x ∈ Vx .

Hence the sets Vx form an open cover of X. By compactness we may pass to a

26
 finite subcover Vx1 ∪ . . . ∪ Vxn .

Define fe = max{fx1 , . . ., fxn }. Then by construction fe ∈ A and f (z) − ε < fe(z) <
f (z) + ε for all z ∈ X. Since ε was arbitrary, A is indeed dense in C(X), and
hence, since A is closed, A = C(X). 2

Corollary (Stone-Weierstrass, complex form). Let X be a compact Hausdorff space.

Suppose A is an algebra of complex-valued functions which satisfies the separation-of-
points property and the property that if f ∈ A then f ∈ A.

Then A = C C (X), the space of complex-valued continuous functions on X, with the

sup norm.

Proof. For every f ∈ A, the functions Re(f ) = 21 (f + f ) and Im(f ) = 2i1

(f − f ) both lie
R
in A. If follows that the real algebra A ∩ C (X) has the separation-of-points property.
Thus, by the real version of Stone-Weierstrass, A is dense in C R (X).

Since C C (X) = C R (X) + iC R (X), the result follows. 2

Corollary. Every continuous complex-valued 2π-periodic

PN function can be uniformly approx-
imated by finite sums of exponentials: g(θ) = n=−N an einθ .

(Compare with the example of a function with a divergent Fourier series.)

Proof. Apply the complex-valued Stone-Weierstrass theorem to R/2πZ = [0, 2π]/(0 ∼ 2π).

Easy to check that separation-of-points holds: consider functions ae−iθ + b + ceiθ . 2

Note. The an need not be the Fourier coefficients of f (if one is trying to approximate f ).

Next, a slightly more precise version of Stone-Weierstrass – more like the ‘standard’ version.

Theorem. Let X be a compact Hausdorff space. Then any proper closed subalgebra of
C(X) is contained in:
(i) the subalgebras Ax,y = {f ∈ C(X) : f (x) = f (y)}, x 6= y
(ii) the subalgebras Ax = {f ∈ C(X) : f (x) = 0}.

Proof. Fix x and y. Let A be an algebra and suppose A is proper. Define Vx,y =
{(f (x), f (y)) : f ∈ A} ⊂ R2 .

Assume that A is not contained in any Ax,y or Ax or Ay . Then the vector space Vx,y
contains points (a1 , b1 ), (a2 , b2 ), (a3 , b3 ) with a1 6= b1 , a2 6= 0, b3 6= 0.

The only way Vx,y could be a proper subspace of R2 yet still contain these three points
would be if Vx,y = h(t, u)i with t 6= u, t 6= 0, u 6= 0.

But Vx,y also contains (t2 , u2 ), and since (t, u) and (t2 , u2 ) span R2 , this is a contradic-
tion.

Hence Vx,y = R2 . That is to say, A has the separation-of-points property at x, y. Since

x, y were arbitrary, it follows from Stone-Weierstrass that A = C(X). 2

27
Corollary (typical formulation of Stone-Weierstrass). Let X be a compact Hausdorff
space, and A ⊂ C(X) a subalgebra which separates the points in the sense that for all
x, y ∈ X with x 6= y, there is f ∈ A with f (x) 6= f (y).

Then either A is dense in C(X), or there is x0 ∈ X such that f (x0 ) = 0 for all f ∈ A.

Proof. The assumption is that A 6⊂ Ax,y for any x, y. Hence A is everything, or else is
proper, and hence contained in Ax for some x. 2

What if we drop the assumption of compactness (e.g. X = R)?

There are certainly closed subalgebras of C(R) with the separation-of-points property which
are not dense in C(R). For example, C0 (R) = {f ∈ C(R) : f (x) → 0 as |x| → ∞}.

Similar examples exist in any locally compact space. (Recall a space is locally compact
 every x ∈ X lies in some neighbourhood with compact closure.) Indeed, define C0 (X) =
if
f ∈ C(X) : {x : |f (x)| > ε} is compact for all ε > 0 . This coincides with what we wrote
before when X = R. It is always a closed subalgebra of C(X). (Exercise: by Urysohn’s
Lemma, C0 (X) separates the points.)

Theorem (Stone-Weierstrass for locally compact spaces). Suppose X is a locally

compact Hausdorff space (e.g. R), and suppose A ⊂ C0 (X) is an algebra with the
separation-of-points property. Then A = C0 (X).

e of X. The underlying space is X ∪ {∞},

Proof. Consider the ‘one-point compactification’ X
where ∞ is some extra element. The open sets in X e are the open sets in X, together
with sets of the form (X \ K) ∪ {∞} for compact sets K ⊂ X. (E.g., if X = R then
{x > a} ∪ {∞} ∪ {x < −a} is an open set.)

One may check that X e becomes a compact Hausdorff space. ‘All of this is very easy’,
except note that the local compactness of X is required to check the Hausdorff property
e Indeed, if x ∈ X then there is an open set U with compact closure. Let
for X.
V = (X \ U ) ∪ {∞}. Then V is open, disjoint from U and contains ∞.

Another easy check: C0 (X) may be identified with the space {fe ∈ C(X)
e : fe(∞) = 0},
e e
where f ∈ C0 (X) is identified with the function f : X → R which equals f on X and
is 0 at ∞. It remains to check that this is continuous.

Under this identification, A is identified with an algebra Ae ⊂ C(X).

e Let B be the
e e
subalgebra of C(X) generated by A and the constant functions (to make sure that not
e
all functions vanish at ∞). Then B has the separation-of-points property in C(X).
Hence, by ordinary Stone-Weierstrass, B is dense.

g ∈ Ae and a constant c
Now suppose f ∈ C0 (X). Then for every ε > 0 there is some e
such that fe = ge + c +η, where kηk∞ < 2ε .
| {z }
∈B

Evaluating at ∞ we get (since f (∞) = ge(∞) = 0) that |c| < 2ε .

Thus kfe − gek∞ < ε, and hence kf − gk∞ < ε, where g ∈ A. 2

Example. Suppose f : [0, ∞) → R has f (x) → 0 as x → ∞. Then f can be uniformly

approximated using e−x .

28
** Non-examinable section **

A simpler proof of the Weierstrass approximation theorem

Theorem (Weierstass). Suppose f : [0, 1] → R is continuous. Then for all ε > 0 there is
a polynomial p ∈ R[x] with kf − pk∞ < ε on [0, 1].

Remark. Of course, this√ follows trivially from Stone-Weierstass. However, we need the
special case f (t) = t in the proof of that theorem.

Proof (Bernstein). Let n be an integer (to be chosen later), and define

n
X  
n i
pn (x) = bi,n (x)f (i/n), where bi,n (x) = x (1 − x)n−i .
i=0
i

We’ll show limn→∞ kpn − f k∞ = 0.

Of course, bi,n (x) is P(X1 + . . . + Xn = i), where the Xj are independent random
variables with mean x.
Pn
This implies i=0 bi,n (x) = 1. Also bi,n (x) is largest when i ∼ nx.

Thus it does at least look as if pn (x) ≈ f (x).

To prove this rigorously. . . Let M be a quantity (to be specified later) in terms of n.

Then
n
X
|pn (x) − f (x)| = bi,n (x)f (i/n) − f (x)
i=0
n
X
6 bi,n (x) |f (i/n) − f (x)|
i=0
X X
= ... + ...
i:|i−nx|6M i:|i−nx|>M
| {z } | {z }
S1 S2

Now |S1 | 6 sup |f (i/n) − f (x)|.

|i−nx|6M

m
Since f is uniformly continuous, this → 0 as long as n → 0 (∗)
X
Also, |S2 | 6 2kf k∞ bi,n (x) = 2kf k∞P(|X1 + . . . + Xn − nx| > M ).
i:|i−nx|>M

However, the mean of X1 + . . . + Xn − nx is 0 and its variance is 6 n, because it’s

the sum of n independent random variables Xi − x, each with variance 6 1. Hence
Chebyshev’s inequality gives
n n
P(|X1 + . . . + Xn − nx| > M ) 6 → 0 if → 0 (∗∗)
M2 M2
Choose M = n2/3 , and both (∗) and (∗∗) are satisfied. 2

** End of non-examinable section **

29
Arzelà-Ascoli Theorem
Suppose F ⊂ C(X). Say that F is precompact if its closure F is compact.

When is F precompact?

Exercise. {f ∈ C[0, 1] : kf k∞ 6 1} is not compact.

Say that F is uniformly bounded if sup kf k∞ < ∞.

f ∈F

Say that F is equicontinuous if for each x ∈ X and for all ε > 0 there is an open set
Ux ∋ x such that y ∈ Ux implies |f (y) − f (x)| < ε for all f ∈ F simultaneously. (I.e., in the
definition of continuity, the same open set works for all f ∈ F .)

Theorem (Arzelà-Ascoli). Let X be a compact Hausdorff topological space, and F ⊂

C(X). Then F is precompact iff F is uniformly bounded and equicontinuous.

Example. Lipschitz functions on [0, 1]. The set F of all f ∈ C[0, 1] with |f (x)| < c1
and |f (x) − f (y)| < c2 |x − y| (Lipschitz condition). Arzelà-Ascoli says that this is a
precompact set of functions. To see equicontinuity, note that |f (x)−f (y)| < ε whenever
y ∈ Bδ (x) with δ = cε2 . Actually, we could replace (x − y) with ψ(x − y) as long as
ψ(t) → 0 as t → 0. When ψ(t) = tα we talk of the ‘Hölder condition’ with exponent α.

Say that F is totally bounded if for every ε > 0 we can cover F by finitely many balls of
radius ε, say B(fi , ε) with fi ∈ F .

Note. Total boundedness of F and F are equivalent.

Proof. Suppose first that F has been covered by balls B(fi , 2ε ). If g ∈ F then there is a
sequence of functions in F with limit g. By the pigeonhole principle we can assume
these all lie in the same B(fi , 2ε ). Hence g ∈ B(fi , 2ε ) ⊂ B(fi , ε).

Conversely, suppose that F has been covered by balls B(fi , 2ε ). For each i, choose
gi ∈ F with d(fi , gi ) < ε2 . Then the balls B(gi , ε) cover F . 2

Proof of Arzelà-Ascoli. Recall that a metric space is compact iff it is complete and totally
bounded. Note that F is automatically complete, because it is a closed subset of the
complete metric space C(X) (with the sup norm). So all we need show is that:

Total Boundedness (TB) ⇐⇒ Uniform Boundedness (UB) + Equicontinuity (EQ).

TB ⇒ UB. For any ε > 0 there is some finite collection {f1 , . . ., fn } of functions such
that for all f ∈ F , there is i ∈ {1, . . ., n} such that kf − fi k∞ 6 ε.

In particular, sup kf k∞ 6 max kfi k∞ + ε. That is, F satisfies UB.

f ∈F i=1,...,n

TB ⇒ EQ. Note that each fi is continuous at x, and so there is an open set U con-
taining x such that if y ∈ U then |fi (x) − fi (y)| 6 3ε for i = 1, . . ., n. Now if f ∈ F
is arbitrary, let y ∈ U . Choose i such that kf − fi k∞ 6 3ε . Then:

ε ε ε
|f (y) − f (x)| 6 |f (y) − fi (y)| + |fi (y) − fi (x)| + |fi (x) − f (x)| 6 3 + 3 + 3 = ε.

Thus F is EQ.

30
 UB + EQ ⇒ TB. Let ε > 0. For each x ∈ X, EQ implies that there is a set Ux ∋ x
such that |f (y) − f (x)| 6 3ε whenever f ∈ F and y ∈ Ux . Since X is compact, we
may pass to a finite subcover Ux1 ∪ . . . ∪ Uxn .

Look at the vectors {(f (x1 ), . . ., f (xn )) : f ∈ F } ∈ Rn . Since F is UB these all lie
in some box [−m, m]n .

This means we can choose an 3ε -net from this set of vectors. I.e., there is a
collection of functions f1 , . . ., fk such that for each f ∈ F there is some i such that
|f (xj ) − fi (xj )| 6 3ε for j = 1, . . ., n.

We claim that kf − fi k∞ < ε. Indeed, if y ∈ Ux , then

|f (y) − fi (y)| 6 |f (y) − f (xj )| + |f (xj ) − fi (xj )| + |fi (xj ) − fi (y)|

ε ε ε
6 3 + 3 + 3 = ε.

Since the Uxj cover X, this is true for all y ∈ X.

We have shown that the balls of radius ε about f1 , . . ., fk in the sup norm cover
F . Since ε was arbitrary, the family F is indeed totally bounded. 2

Application: Peano’s existence theorem for ODEs

Theorem. Suppose Ψ : R2 → R is a continuous function. Then the ODE f ′ (x) = Ψ(x, f (x))
(or y ′ = Ψ(x, y)) has a solution f : (−η, η) → R for some η > 0, with f (0) = 0.

Proof. The idea is to construct approximate solutions to the equation in a suitable space of
functions V . We’ll ensure that V is compact, hence sequentially compact, and so our
sequence of approximate solutions will have a limit, which is a genuine solution.
 
1
Let M = max max Ψ(x, y), 1 , and let η = M.
x,y∈[−1,1]

Take V to be the subset of C[−η, η] with f (0) = 0 and Lipschitz |f (x) − f (x′ )| 6
M |x − x′ | for all x, x′ ∈ [−η, η].

By Arzelà-Ascoli, this V is compact.

Lemma. For each δ > 0 there is a δ-approximate solution to our ODE in V . That
means a function f ∈ V , differentiable except possibly at finitely many points,
and such that |f ′ (x) − Ψ(x, f (x))| 6 δ for x ∈ [−η, η] if f ′ (x) is defined.

Proof of lemma. Let n be a large positive integer, and xj = 2j n for j = −n, −(n − 1),
. . ., (n − 1), n. So −η = x−n < · · · < x−1 < x0 < x1 < · · · < xn = η.

Define f (0) = 0. Suppose that f has already been defined up to xi . Define f to be

piecewise linear on [xi , xi+1 ] with slope Ψ(xi , f (xi )). By induction on i, |f (x)| 6 1
for x ∈ [xi−1 , xi ] and it has slope 6 M on this interval. If this is known, then
by construction the slope on [xi , xi+1 ], namely Ψ(xi , f (xi )) is also at most M .
(Recall the definition of M .)

From this it follows that |f (x)| 6 M |x| 6 M η 6 1 for all x ∈ (xi , xi+1 ). It follows
that f ∈ V .

31
 Let us show that (if n is big enough), then f is a δ-approximate solution to the
ODE.

Suppose x ∈ (xi , xi+1 ). Then |f ′ (x) − Ψ(x, f (x))| = |Ψ(xi , f (xi )) − Ψ(x, f (x))|.
η 2M
However, |xi − x| 6 n and |f (xi ) − f (x)| 6 n .

So by uniform continuity of Ψ, this will be < δ, for n sufficiently large. 2

Now V is compact by Arzelà-Ascoli. Take a sequence of δ-approximate solutions with

δ → 0. Then there is a convergent subsequence. By passing to a further subsequence if
1
necessary, we get a convergent subsequence (fn )∞ n=1 with fn a n -approximate solution,
′ 1
that is |fn (x) − Ψ(x, fn (x))| 6 n for x ∈ (−η, η).

Suppose that fn → f in V . We claim that f is differentiable everywhere and that

f ′ (x) = Ψ(x, f (x)). Consider
Z x Z x
(since Ψ uniformly continuous
Ψ(t, f (t)) dt = lim Ψ(t, fn (t)) dt
0 n→∞ and fn → f uniformly)
Z0 x
(makes sense even though fn′ is
= lim fn′ (t) dt
n→∞ 0
undefined at finitely many points)

= lim fn (x) = f (x)

n→∞

Z x
Thus f (x) = Ψ(t, f (t)) dt, so f is differentiable and f ′ (x) = Ψ(x, f (x)). 2
0

32
5. Weak Topologies on Normed Spaces
On X, we always have the norm topology τk k , defined by: A ∈ τk k iff for all x ∈ A, there
exists ε > 0 such that Bε (x) ⊂ A. Alternatively, given a family F of functions f : X → Yf
for some topological spaces Yf , we can also define a topology as follows:
S = {f −1 (U ) : U ∈ τYf , f ∈ F }
\
τF = τ, over topologies τ on X with τ ⊃ S

Then τF is the smallest topology on X such that every f ∈ F is continuous. We can show
that τF is the set of arbitrary unions of finite intersections of sets of the form f −1 (U ), where
U ∈ τYf , f ∈ F .

Proposition. Let F be a family of maps f : X → Yf such that Yf is Hausdorff for all f ∈ F ,

and such that F separates the points of X. Then (X, τF ) is Hausdorff.

Proof. Let x, y ∈ X, x 6= y. Then there exists f ∈ F such that f (x) 6= f (y). So there are
W1 , W2 ∈ τYf such that f (x) ∈ W1 , f (y) ∈ W2 with W1 ∩ W2 = ∅.

Then x ∈ f −1 (W1 ) ∈ F , y ∈ f −1 (W2 ) ∈ F , and f −1 (W1 ) ∩ f −1 (W2 ) = ∅. 2

Lemma. Let f1 , . . ., fn be linear functionals on X and N = {x ∈ X : fiP (x) = 0 for all i}.
Then f (x) = 0 for all x ∈ N iff there exist α1 , . . ., αn such that f = αi fi .

Proof. Exercise. 2

Recall that a locally convex space is a topological vector space with a basis consisting of
convex sets.

Theorem. Let X be a vector space and F a vector space of linear functionals on X, sepa-
rating the points of X. Then (X, τF ) is a locally convex space, and (X, τF )∗ = F .

Proof. Since F separates points of X, we know (X, τ F ) is Hausdorff. We first check for the
existence of a neighbourhood basis consisting of convex sets. Indeed, there is a basis of
the neighbourhoodTof 0 consisting of the sets of the form V = {x ∈ X : |fi (x)| < εi for
all i = 1, . . ., n} = fi−1 (Bε (0)) for some fi ∈ F and εi > 0.

V is convex, as if x, y ∈ V and α ∈ [0, 1] then |fi (αx+(1−α)y)| = |αfi (x)+(1−α)fi (y)|

< α|fi (x)| + (1 − α)|fi (y)| < εi . So αx + (1 − α)y ∈ V .

Next we check the continuity of the sum and scalar multiplication. Note that if V is a
0-neighbourhood as before, then 12 V + 21 V = V and (x + 12 V ) + (y + 12 V ) = x + y + V .
So + : X × X → X. If z = x + y and z ∈ U ∈ τF , then there is a V as above such that
z + V ⊂ U , and so +−1 (x + y + V ) ⊃ (x + 12 V ) × (y + 21 V ), and thus + is continuous,
as preimages of open sets are open.

Multiplication is similar. Therefore (X, τF ) is a locally convex space.

(X, τF )∗ ⊃ F is clear. To show (X, τF )∗ ⊂ F , take f ∈ (X, τF )∗ . Then f −1 ({α :

|α| < 1}) ⊃ V , where V = {x ∈ X : |fi (x)| < 1 for all i = 1, . . ., n} for appropiate
f1 , . . ., fn ∈ F .

Suppose x0 ∈ X such that fi (x0 ) = 0 for all i. Then x0 ∈ V and moreover αx0 ∈ V
P α. So |f (αx0 )| < 1 for all α, and thus f (x0 ) = 0. By the previous lemma,
for all
f = αi fi ∈ F . 2

33
Definition. Let X be a normed space, and X ∗ be the dual space. The topology on X
generated by F = X ∗ is known as the weak topology on X and is denoted by τw .

Thus U ⊂ X is open in the weak topology (‘w-open’) iff for every x ∈ U there are
f1 , . . ., fn ∈ X ∗ and ε1 , . . ., εn > 0 such that {y ∈ X : |fi (x) − fi (y)| < εi ∀ i} ⊂ U .
(By replacing fi by ε1i fi , we may assume each εi = 1.)

On X ∗ , we may define the topology generated by F = {ϕX (x) : x ∈ X} ⊂ X ∗∗ , known

as the weak-star topology and denoted by τw∗ , where ϕX is the natural map from X
to X ∗∗ , given by ϕX (x)(f ) = f (x).

Thus F ⊂ X ∗ is open in the weak-star topology (‘w∗ -open’) iff for every f ∈ F there
are x1 , . . ., xn ∈ X and ε1 , . . ., εn > 0 such that {g ∈ X ∗ : |ϕX (xi )(f ) − ϕX (xi )(g)| <
εi ∀ i} = {g ∈ X ∗ : |f (xi ) − g(xi )| < εi ∀ i} ⊂ F . (As before, we may assume each
εi = 1.)

Note that on X, τw ⊂ τk k but (X, τw )∗ = (X, τk k )∗ . And on X ∗ , τw∗ ⊂ τw ⊂ τk k .

Remark. If X is reflexive, then τw∗ = τw on X ∗ .

Lemma. Let X be a normed space. Then τw = τk k iff dim X < ∞.

Proof. (⇐). Wlog, X = Rn or Cn with norm kxk = maxi |xi |, where x = (x1 , . . ., xn )t . Let
fi (x) = xi . Then fi ∈ X ∗ for all i = 1, . . ., n. For every ε > 0, Bε (0) = {x ∈ X :
|fi (x)| < ε for all i} ∈ τw . So τk k ⊂ τw , and therefore τk k = τw .

(⇒). Suppose W is a w-open neighbourhood of 0. Then there exist f1 , . . ., fn ∈ X ∗

such that W ⊃ {x ∈ X : |fi (x)| < 1 ∀ i}. So N = {x ∈ X : fi (x) = 0 ∀ i} ⊂ W . If
dim X = ∞ then dim N = ∞. So every w-open neighbourhood of 0 contains an infinite
dimensional linear space. Then Bε (0) = {x ∈ X : kxk < ε} ∈
/ τw . /\/\ 2

Definition. Let X be a normed space. A sequence (xn ) in X is said to converge weakly

(‘w-convergent’) to x ∈ X, written xn ⇀ x as n → ∞, if for all w-open neighbourhoods
V of 0, there exists n0 > 0 such that xn ∈ x + V for all n > n0 .

Lemma. Let (xn ) be a sequence in a normed space X. Then xn ⇀ x ⇐⇒ f (xn ) → f (x)

for all f ∈ X ∗ .

Proof. (⇒). Assume xn ⇀ x, f ∈ X ∗ . Then for all ε > 0, V = {x ∈ X : |f (x)| < ε} ∈ τw .

So there exists n0 > 0 such that xn ∈ x + V for all n > n0 . Then |f (xn − x)| < ε for
all n > n0 and |f (xn ) − f (x)| < ε for all n > n0 .

(⇐). It is enough to show that xn ∈ x + V for all n sufficiently large, where V =

{x ∈ X : |fi (x)| < 1 for i = 1, . . ., k} for some f1 , . . ., fk ∈ X ∗ . Then for every i, there
exists n0 (i) such that |fi (xn ) − fi (x)| < 1 for all n > n0 (i). If n > max n0 (i) then
|fi (xn − x)| < 1 for all i, and so xn ∈ x + V for all n > max n0 (i). 2

Remark. We proved that for f ∈ X ∗ , xn ⇀ x implies f (xn ) → f (x). This is not the
definition of continuity. Only in metric spaces is continuity equivalent to sequential
continuity. In general, continuity implies sequential continuity.

Lemma. Let xn ⇀ x. Then xn is bounded and kxk 6 lim inf n→∞ kxn k.

Proof. For each f ∈ X ∗ , f (xn ) → f (x), and so there exists cf such that cf > |f (xn )| =

34
 |ϕX (xn )(f )| for all n > 1. Let F = {ϕX (xn ) : n > 1} be a family of linear functionals
on X ∗ such that for all f ∈ X ∗ , there exists cf > 0 with |T (f )| 6 cf for all T ∈ F . By
Banach-Steinhaus, there exists c > 0 such that kT k < c for all T ∈ F .

So kxn k = kϕX (xn )k 6 c for all n > 1 and so xn is bounded. By Hahn-Banach, there
exists f ∈ X ∗ such that kf k = 1 and f (x) = kxk. Then kxk = f (x) = limn→∞ f (xn ) 6
lim inf n→∞ kf k kxn k = lim inf n→∞ kxn k. 2

Definition. A subset S ⊂ X ∗ is a fundamental subset of X ∗ if the closure of the span of

S is X ∗ . In other words, hSi = X ∗ .

Lemma. Let (xn ) be a bounded sequence in X, and S ⊂ X ∗ a fundamental subset. Then

xn ⇀ x iff f (xn ) → f (x) for all f ∈ S.

Proof. As S ⊂ X ∗ , that xn ⇀ x implies f (xn ) → f (x) for all f ∈ S follows by a previous

lemma.

∈ X ∗ and ε > 0. Then there are f1 , . . ., fm ∈ S and scalars α1 , . . ., αm such that

Let f P
kf − αi fi k < ε. Then
P P P
|f (xn ) − f (x)| 6 |f (xn ) − αi fi (xn )| + | αi fi (xn ) − αi fi (x)|
P
+ | αi fi (x) − f (x)|
P P
6 kf − αi fi k · (kxn k + kxk) + |αi | |fi (xn ) − fi (x)|
6 cε for all n sufficiently large, for some c. 2
1 1
Example. Let X = ℓp , 1 < p < ∞, so that X ∗ = ℓq , where p + q = 1. Then xn ⇀ x iff xn
(i) (i)
is bounded and xn →x for each i ∈ N.

For example, xn = (0, . . ., 0, 1, 0, . . .) ⇀ 0 as n → ∞, but kxn k = 1 so xn 6→ 0.

Indeed, take S = {fi = (0, . . ., 0, 1, 0, , , , ) : i ∈ N}, a fundamental subset of ℓq . Then

(i)
xn ⇀ x iff xn is bounded and fi (xn ) → fi (x) for all i > 1. But fi (xn ) = xn , so
(i)
fi (x) = x .

For X = ℓ1 , we have xn ⇀ x iff xn → x. Although in ℓ1 , strong and weak convergence

are equivalent, τw ( τk k , as k k is continuous with respect to τk k but only sequentially
continuous with respect to τw .

Definition. A system F of subsets of a set S is said to be of finite character if for A ⊂ S,

A ∈ F iff every finite subset of A is in F .

Lemma (Tukey’s Lemma). Let F be a set system of finite character and F ∈ F . Then
F has a maximal element containing F .

Proof. Let F0 =
S{A ∈ F : F ⊂ A}. If C ⊂ F0 is totally ordered with respect to inclusion,
then D = C∈C C ∈ F , because its finite subsets are in F . Also D ⊃ F , so D ∈ F0
and D is an upper bound for C. By Zorn’s Lemma, F0 has a maximal element. This is
also a maximal element for F and contains F .

Definition. A system F of subsets of Tna given set has the finite intersection property if
for all F1 , . . ., Fn ∈ F , we have i=1 Fi 6= ∅.

Proposition. Let X be a topological space. Then X compact iff for every system F of

35
 T
closed subsets of X with the finite intersection property we have A∈F A 6= ∅.
S
Proof. X compact ⇐⇒ for all collections of open sets (Uα )α∈ΛS with X = Uα ,
there exist α1 , . . ., αn such that X = Uαi
T
⇐⇒ for all collections of closed sets (Vα )α∈Λ
T with ∅ = V α,
there exist α1 , . . ., αn such that ∅ = Vαi , by taking Vα = Uαc
⇐⇒ if F = (Vα )α∈Λ is a systemTof closed sets with the finite
intersection property then Vα 6= ∅. 2

T compact iff every collection F of subsets of X with the finite

It follows easily that X is
intersection property has A∈F A 6= ∅.

Theorem (Tychonoff ’s Theorem). The product of a collection of compact spaces is com-

pact with respect to the product topology.
Q
Proof. Let {Xγ : γ ∈ Γ} be a collection of compact spaces and let X = Xγ . Let A
be a system of subsets of X with
T the finite intersection property. To show that X is
compact, we want to show that A∈A A 6= ∅.

Let F be the collections of all systems of subsets of X which have the finite intersection
property. By definition, F is of finite character, so by Tukey’s Lemma there exists a
maximal systemT B of subsets
T of X having the finite intersection
T property such that
B ⊃ A. Since B∈B B ⊂ A∈A A, it is enough to show that B∈B B 6= ∅.

Because B is maximal, we know that if B1 , . . ., Bn ∈ B then B1 ∩ . . . ∩ Bn ∈ B. (∗)

Similarly, if C ⊂ X such is that C ∩ B 6= ∅ for all B ∈ B, then C ∈ B. (∗∗)

We may write x ∈ X as (xγ )γ∈Γ with xγ ∈ Xγ . For each γ ∈ Γ, let pγ : X → Xγ

be the projection (xγ ′ )γ ′ ∈Γ 7→ xγ . Then the system {pγ (B) : B ∈ B} of subsets of Xγ
has the finite intersection property: for if B1 , . . ., Bn ∈ B then pγ (B1 ) ∩ . . . ∩ pγ (Bn ) ⊃
pγ (B1 ∩ . . . ∩ Bn ) 6= ∅, by (∗).
T
Then Xγ being compact implies B∈B pγ (B) 6= ∅. So there exists xγ ∈ Xγ such that
T
xγ ∈ B∈B pγ (B). This means that for each open neighbourhood Uγ of xγ and all
B ∈ B, we have Uγ ∩ pγ (B) 6= ∅, and hence pγ−1 (Uγ ) ∩ B 6= ∅. So, by (∗∗), p−1
γ (Uγ ) ∈ B.

Let x = (xγ )γ∈Γ , and let U be an open neighbourhood of x. By definition of the

product topology on X, thereTexist γ1 , . . .γn ∈ Γ and Uγi an open neighbourhood of
n
xγi , 1 6 i 6 n, such that U ⊃ i=1 p−1
γi (Uγi ).

Tn Tn
However, each p−1 (Uγi ) ∈ B, and so i=1 pγ−1
i
(Uγi ) ∈ B, so i=1 p−1
γi (Uγi ) ∩ B 6= ∅.

T
So, for all B ∈ B, we have U ∩ B 6= ∅, so x ∈ B, and thus x ∈ B∈B B, as required. 2

Theorem (Banach-Alaoglu Theorem). Let X be a normed space. Then the closed unit
ball B(X ∗ ) = {f ∈ X ∗ : kf k 6 1} is compact in the τw∗ topology (‘w∗ -compact’).

(Note this is sometimes just called Alaoglu’s Theorem.)

Q
Proof. For x ∈ X, let Dx = {ξ ∈ K : |ξ| 6 kxk} ⊂ K. Let D = x∈X Dx , equipped with the
product topology. Each Dx is compact, so D is compact with respect to the product
topology, by Tychonoff’s theorem. Note that we may write ξ ∈ D as (ξx )x∈X .

36
 Define ϕ : B(X ∗ ) → D, f 7→ (f (x))x∈X .

ϕ is well-defined: |f (x)| 6 kf k kxk 6 kxk, so f (x) ∈ Dx for all x, so (f (x))x∈X ∈ D.

ϕ is injective: if ϕ(f ) = ϕ(g) then f (x) = g(x) for all x ∈ X, so f = g.

ϕ is a homeomorphism. We show ϕ is continuous – the proof for ϕ−1 is similar.

Suppose U ⊂ D is an open neighbourhood of ϕ(f0 ) = (f0 (x))x∈X . Then there exist

x1 , . . ., xn ∈ X and ε1 , . . ., εn > 0 such that U ⊃ {ξ ∈ D : |ξxi − f0 (xi )| < εi ∀ i}. Then:


ϕ−1 {ξ ∈ D : |ξxi − f0 (xi )| < εi ∀ i} = {f ∈ B(X ∗ ) : |f (xi ) − f0 (xi )| < εi ∀ i} ∈ τw∗

I.e., ϕ is continuous.

To prove ϕ(B(X ∗ )) is compact, it is enough to show that it is closed, since D is compact.

Suppose ξ = (ξx )x∈X ∈ ϕ(B(X ∗ )). Define f : X → K, x 7→ ξx . We claim that f is

linear, i.e. that f ∈ X ∗ .

Let x, y ∈ X and α, β ∈ K. Since ξ ∈ ϕ(B(X ∗ )), there exists a sequence fn ∈ B(X ∗ )

such that ϕ(fn ) → ξ, i.e. fn → f , as n → ∞, for all x ∈ X fixed. Then:

|f (αx + βy) − αf (x) − βf (y)| 6 |f (αx + βy) − fn (αx + βy)|

+ |fn (αx + βy) − αfn (x) − βfn (y)|
+ |αfn (x) − αf (x)| + |βfn (y) − βf (y)|
→ 0 as n → ∞

So f is linear. And since |f (x)| = |ξx | 6 kxk for all x ∈ X, f is continuous and kf k 6 1,
i.e. f ∈ B(X ∗ ). And by definition of f , we have ξ = ϕ(f ) ∈ ϕ(B(X ∗ )), which is thus
closed. 2

Theorem. Let X be a Banach space. Then ϕX (B(X)) is w∗ -dense in B(X ∗∗ ), where ϕX is

the usual map X → X ∗∗ , ϕX (x)(f ) = f (x), and B(X) = {x ∈ X : kxk 6 1}.

Proof. Let U be w∗ -open in B(X ∗∗ ) and let θ ∈ U . Recall that this means there are
f1 , . . ., fn ∈ X ∗ and ε1 , . . ., εn > 0 such that {ψ ∈ X ∗∗ : |θ(fi ) − ψ(fi )| < εi ∀ i} ⊂ U .

To show that ϕX (B(X)) is dense, we want to show that U contains a ψ of the form
ϕX (x) for some x ∈ B(X). So it is enough
P to show that for every
P f1 , . . ., fn and ε > 0,
there exists x ∈ B(X) such that ε > |θ(fi ) − ϕX (x)(fi )| = |θ(fi ) − fi (x)|.

Wlog, the fi are linearly independent. Define J : X → Kn by x 7→ (f1 (x), . . ., fn (x)).

Then J is linear, continuous and surjective but in general not injective. So define an
equivalence relation x ∼ y iff Jx = Jy. Let [x] = {y ∈ X : x ∼ y}, and define the
b = X/ ker J = {[x] : x ∈ X}, with k[x]k = inf y∼x kyk.
quotient X

Now define Jb : X b → Kn by J([x])

b = J(x). Then Jb is a linear bijection so there exists
a unique [z] ∈ X b such that J([z])
b b this equals
= (θ(f1 ), . . ., θ(fn )). By definition of J,
J(z) = (f1 (z), . . ., fn (z)).

By Hahn-Banach, there exists fb ∈ X

b ∗ such that kfbk = 1 and fb([z]) = k[z]k.

37
 Let b = fb ◦ Jb−1 ∈ (Kn )∗ , say b = (b1 , . . ., bn ). Then
P P
θ(b ◦ J) = θ ( b
bi fi ) = bi θ(fi ) = (b ◦ J)([z]) = fb([z]) = k[z]k

b
Since |(b ◦ J)(x)| = |(b ◦ J)([x])| = |fb([x])| 6 kfbk k[x]k = k[x]k 6 kxk, we know
kb ◦ Jk 6 1 and thus k[z]k 6 kθk kb ◦ Jk 6 kθk 6 1.

P since k[z]k = inf x∼z kxk, there existsx x ∈ X such that x ∼ z and
So, P kxk 6 1 + ε. Then
|θ(fi )−fi (x)| = 0, and taking x0 = 1+ε we have kx0 k 6 1 and |θ(fi )−fi (x0 )| 6 cε,
for some c. 2

Corollary. Let X be a Banach space. Then X is reflexive iff B(X) is w-compact.

Proof. Use the fact that ϕX : (X, τW ) → (X ∗∗ , τW

∗
) is continuous, and if X is reflexive then
−1 ∗∗ ∗
ϕX : (X , τW ) → (X, τW ) is also continuous.

So if X is reflexive, then X ∗ is reflexive, and so on X ∗∗ , τW = τW

∗
. By Banach-Alaoglu,
∗∗ −1 ∗∗
B(X ) is w-compact and B(X) = ϕX (B(X )) is also w-compact.

Conversely, suppose B(X) is w-compact. Then ϕX (B(X)) is w∗ -compact and thus

w∗ -closed in B(X ∗∗ ). By the previous result, ϕX (B(X)) is w∗ -dense in B(X ∗∗ ). Hence
ϕX (B(X)) = B(X ∗∗ ), and by linearity ϕX (X) = X ∗∗ .

Definition. A sequence (xn ) in X is said to be weakly Cauchy (w-Cauchy), if for every

w-open neighbourhood V of 0, we have xn − xm ∈ V for all m, n large enough.

Theorem. A reflexive Banach space is w-complete, i.e. every w-Cauchy sequence in X is

w-convergent.

Proof. Let (xn ) be a w-Cauchy sequence in X, and f ∈ X ∗ . Then f (xn ) is Cauchy in K

and thus f (xn ) converges in K. So f (xn ) is bounded in K and |ϕX (xn )(f )| 6 cf for all
n. By Banach-Steinhaus, there exists c > 0 such that kxn k = kϕX (xn )k 6 c for all n.

Define T : X ∗ → K, f 7→ limn→∞ ϕX (xn )(f ). Then T is linear and bounded, since

|T (f )| 6 lim supn→∞ kxn k kf k 6 ckf k. So T ∈ X ∗∗ . As X is reflexive, T = ϕX (x) for
some x ∈ X, and f (xn ) → f (x) for all f ∈ X ∗ . Then xn ⇀ x. 2

Theorem. Let X be a reflexive Banach space. Then B(X) is sequentially compact.

Proof. Let (xn ) be a sequence in B(X). Define X1 = h(xn )i. Then X1 is separable and
reflexive. So X1∗ is separable. Let (fn ) be a dense sequence in X1∗ .

(f1 (xn )) is a bounded sequence in K, so there exists a subsequence x1,j of xn such that
f1 (x1,j ) converges.

(f2 (xn )) is a bounded sequence in K, so there exists a subsequence x2,j of xn such that
f2 (x2,j ) converges.

Continue to find sequences xj,k for all j, k ∈ N. Let vk = xk,k . It follows that fn (vk )
converges for all n. So f (vk ) converges for all f ∈ X1∗ , so (vk ) is a w-Cauchy sequence
and therefore w-convergent. 2

Corollary. Let X be a reflexive Banach space. Then every bounded sequence in X has a
w-convergent subsequence.

38
6. Hilbert Spaces
Let V be a vector space over R. Then an inner product on V is a map h , i : V × V → R
satisfying:

(i) h , i is bilinear, that is: hx, y + zi = hx, yi + hx, zi

hx + y, zi = hx, zi + hy, zi (redundant, given (ii))
hx, λyi = λhx, yi

(ii) h , i is symmetric, that is: hx, yi = hy, xi

(iii) positive definiteness: hx, xi > 0 with equality iff x = 0

Now let V be a vector space over C. Then a Hermitian inner product on V is a map
h , i : V × V → R satisfying:

(i) h , i is sesquilinear (‘one-and-a-half’), we have: hx, y + zi = hx, yi + hx, zi

hx + y, zi = hx, zi + hy, zi
hλx, yi = λhx, yi
hx, λyi = λhx, yi

(ii) h , i is Hermitian, that is: hx, yi = hy, xi

(iii) positive definiteness: hx, xi > 0 with equality iff x = 0
(Note hx, xi is automatically real.)

Examples

1. V = Cn and hx, yi = xT Ay, where A is a positive definite Hermitian matrix (i.e.

AT = A), in the standard basis.
The most basic example is hx, yi = x1 y1 + . . . + xn yn , the standard inner product (with
A = I).
P∞
2. V = ℓx = {(x1 , x2 , . . .) ∈ CN : i=1 |xi |2 < ∞}.
P∞
Take hx, yi = i=1 xi yi . (This does converge, since xi yi 6 21 (|xi |2 + |yi |2 ).
Rb
3. V = C C [a, b] and define hf, gi = a f (x)g(x)dx.
(Note, hf, f i = 0 iff f = 0, since f is continuous.)

p
Inner products (over R or C) may be used to define norms: define kxk = hx, xi.

Why is this a norm? The triangle inequality needs proof, using:

Proposition (Cauchy-Schwarz). Suppose x, y ∈ V , a vector space over C, and that h , i

is a Hermitian inner product on V . Then |hx, yi| 6 kxk kyk.

Proof. It suffices to prove it with y replaced by y ′ = λy, where |λ| = 1, since |hx, y ′ i| =
|hx, yi| and ky ′ k = kyk. By choosing λ appropriately, we may assume that hx, yi ∈ R.

Now let t ∈ R and note that hx + ty, x + tyi > 0, i.e. kxk2 + 2thx, yi + t2 kyk2 > 0. (Note
hx, yi = hy, xi = hy, xi.) As a quadratic in t, this has a non-positive discriminant, that
is 4hx, yi2 − 4kxk2 kyk2 6 0. Rearranging gives the result. 2

39
It follows that k k satisfies the triangle inequality:
kx + yk2 = hx + y, x + yi = kxk2 + kyk2 + hy, xi + hx, yi 6 (kxk + kyk)2 .

Definition. If V with the norm k k induced from a (Hermitian) inner product is complete
(i.e. if V is a Banach space), then V is called a Hilbert space.

Examples 1 and 2 above are Hilbert spaces: 1 because every finite-dimensional normed space
is a Banach space, and 2 because we proved that ℓp is complete for all p > 1.

But 3 is not complete. We showed before (see page 5) that C[a, b] is not complete with the
Rb Rb
L1 -norm kf k = a |f (t)|dt. Here our norm is the L2 -norm kf k = ( a |f (t)|2 dt)1/2 , but the
same example (functions converging to a step function) works.

Completeness

e This is constructed as equivalence

Given any metric space X, one can define its completion X.
∞
classes of Cauchy sequences (xn )n=1 where (xn ) ∼ (yn ) if xn − yn → 0 as n → ∞. (Think of
making R out of Q.)

Can turn Xe into a complete metric space – it is the unique (up to isomorphism) smallest
complete metric space containing X.

This construction works well with respect to inner products and norms. In particular if
V comes equipped a Hermitian inner product h , i (which induces a norm k k) then the
completion Ve of V as a metric space is a Hilbert space (i.e. can be given the structure of a
Hilbert space).

The completion of C C [a, b] with respect to the L2 -norm is a very important space, called
L2 [a, b]. Remarkable fact: there is a reasonably explicit description of L2 [a, b] as the space
Rb
of Lebesgue measurable functions with a |f |2 < ∞.
p
A few facts about inner product norms, kxk = hx, xi.

Polarisation identity
• Over R, hx, yi = 12 (kx + yk2 − kxk2 − kyk2 ). (Trivial proof.)

So knowing the norm tells us the inner product (if there is one).
Z 2π
1
• Over C, hx, yi = kx + eiθ yk2 eiθ dθ.
2π 0

Z 2π
1
Expanding out, gives: kxk2 eiθ + hy, xie2iθ + kyk2 eiθ + hx, yi dθ.
2π 0

Z 2π 
2π if n = 0
But einθ dθ =
0 0 if n ∈ Z \ {0}

So this will give hx, yi.

• Alternatively, use fourth roots of unity:

hx, yi = 1
4 kx + yk2 + ikx + iyk2 − kx − yk2 − ikx − iyk2 .

40
Parallelogram identity. kx − yk2 + kx + yk2 = 2kxk2 + 2kyk2. (This is not satisfied by all
norms, only inner product norms.)

Pythagoras’ theorem. Suppose x1 , . . ., xn are mutually orthogonal, i.e. hxi , xj i = 0 if

i 6= j. Then kx1 + . . . + xn k2 = kx1 k2 + . . . + kxn k2 . (Proof is trivial.)

Another name for a (Hermitian) inner product space is a Euclidean space. So a Hilbert
space is a complete Euclidean space.

Suppose X is some Euclidean space and x ∈ X. Define the orthogonal complement to be

x⊥ = {y ∈ X : hx, yi = 0}.

This is a closed linear subspace (and hence a Hilbert space), as follows. Obviously, x⊥
is closed under addition and scalar multiplication. It is closed because if yn → y then
|hx, y − yn i| 6 kxk ky − yn k → 0 as n → ∞. So |hx, yi − hx, yn i| → 0. But yn ∈ X ⊥ and thus
hx, yi = 0.
T
More generally, if S ⊂ X is any set at all, then we define S ⊥ = x∈S x⊥ . This is also a
closed linear subspace (as indeed is any intersection of closed linear subspaces).

Suppose that V = lin S = closed linear space of S, the closure of all linear combinations of
elements of S. Then S ⊥ = V ⊥ . It’s clear that V ⊥ ⊂ S ⊥ . (In fact, if A ⊂ B then B ⊥ ⊂ A⊥ .)

Conversely, if x ∈ S ⊥ then x is orthogonal to all linear combinations of elements of S and

hence to the closure of this set by the same argument as before. Finally note that (S ⊥ )⊥ ⊃ S.
In fact (S ⊥ )⊥ = (V ⊥ )⊥ ⊂ V = lin S.

Theorem. Let X be a Euclidean space and suppose that Y ⊂ X is a complete subspace.

Then every x ∈ X can be written as y + y ∗ where y ∈ Y and y ∗ ∈ Y ⊥ .

This decomposition is unique and in fact it gives a direct sum decomposition X =

Y ⊕ Y ⊥ . Finally, the projection operators PY , PY ⊥ are bounded, self-adjoint linear
operators with norm 6 1.

Proof. Idea is to take y to be the ‘nearest’ point to x. Need: (1) to understand what this
means, and (2) check that x − y ⊥ Y .
1
(1) Let d = inf y∈Y kx − yk. Take a sequence (yn )∞ 2 2
n=1 ⊂ Y with kx − yn k 6 d + n .

We claim that this is a Cauchy sequence. Recall the parallelogram law, kv + wk2 +
kv − wk2 = 2kvk2 + 2kwk2 . Apply this with v = x − yn , w = x − ym . We obtain

kyn − ym k2 = 2kx − yn k2 + 2kx − ym k2 − k2x − yn − ym k2

6 2 d2 + n1 + 2 d2 + m1
− 4d2 (∗)


= 2 n1 + m
1
→ 0 as m, n → ∞

(For (∗), note that 21 (yn + ym ) ∈ Y , hence kx − 12 (yn + ym )k > inf y∈Y kx − yk = d.)

It follows that yn tends to some limit y. We obviously have kx−yk = limn→∞ kx−
yn k = d.

(2) We need to show that x − y ⊥ Y . Suppose not. Let x − y = y ∗ . Then there is

some z ∈ Y such that hy ∗ , zi 6= 0. Multiplying z through by an appropriate scalar

41
 λ ∈ C, we can assume that hy ∗ , zi ∈ R and is > 0.

Consider vectors y ′ = y + εz for small postive ε ∈ R. Clearly y ′ ∈ Y . Also,

kx − y ′ k2 = hx − (y + εz), x − (y + εz)i
= hy ∗ − εz, y ∗ − εzi
= ky ∗ k2 − 2εhy ∗ , zi + ε2 kzk2
= kx − yk2 − 2εhy ∗ , zi + ε2 kzk2

Taking ε sufficiently small, we see that kx − y ′ k < kx − yk, contrary to the

minimality of kx − yk.

Uniqueness. Suppose x = y1 + y1∗ = y2 + y2∗ . Then 0 = y + y ∗ with y = y1 − y2

and y ∗ = y1∗ − y2∗ . Taking inner products with y, we obtain kyk2 = hy, yi = 0, and
so y = 0. Thus y1 = y2 and y1∗ = y2∗ .

The fact that PY : x 7→ y is linear is straightforward. Indeed, if x1 = y1 + y1∗

and x2 = y2 + y2∗ , then x1 + x2 = (y1 + y2 ) + (y1∗ + y2∗ ) and this is the unique
decomposition of x1 + x2 in Y ⊕ Y ⊥ .

Note PY ⊥ = id − PY , so this is also a linear operator.

The boundedness of PY and PY ⊥ is immediate from Pythagoras’ theorem: kxk2 =

kPY xk2 + kPY ⊥ xk2 and so kPY xk, kPY ⊥ xk 6 kxk.

To see self-adjointness, note hx, PY x′ i = hPY x, PY x′ i = hPY x, x′ i.

Again, since PY ⊥ = id − PY , this is self-adjoint as well. 2

Remark. The theorem is most often applied when X is a Hilbert space, in which case it
suffices that Y be a closed subspace (as it’s then automatically complete).

Corollary. Suppose X is a Hilbert space and Y ⊂ X. Then (Y ⊥ )⊥ = Y .

Proof. Apply the theorem twice to get X = Y ⊕ Y ⊥ and X = (Y ⊥ )⊥ ⊕ Y ⊥ . Since Y ⊂

(Y ⊥ )⊥ , it follows that indeed Y = Y ⊥⊥ . 2

Corollary/theorem (Riesz representation theorem). Suppose X is a Hilbert space

and that ϕ : X → C is a bounded linear functional. Then there is some x0 ∈ X
such that ϕ(x) = hx, x0 i.

Furthermore, anything of this type is a bounded linear functional and the identification
f : x0 → ϕ gives a bijection X → X ∗ . This is an anti-isometry, in that f (λx + µx′ ) =
λf (x) + µf (x′ ). Furthermore, kϕk = kx0 k, where kϕk means the norm of ϕ considered
as an operator, i.e. supkxk=1 |ϕ(x)|.

This gives X ∗ with the dual norm the structure of a Hilbert space.

Proof. Let Y = ker(ϕ). This is a closed linear subspace of X. Hence X = Y ⊕ Y ⊥ by the

result from last time.

It’s easy to see that dim Y ⊥ 6 1, since ϕ|Y ⊥ has trivial kernel. If ϕ = 0 then we’re
done, otherwise dim Y ⊥ = 1.

42
 Suppose Y ⊥ = hzi for some z ∈ X. Let x = y + λz be an arbitrary element of X. Then
ϕ(x) = λϕ(z). Also, hx, zi = hλz, zi = λkzk2 .

z ϕ(z)
Hence, defining x0 = ϕ(z), we get hx, x0 i = hx, zi = λϕ(z) = ϕ(x).
kzk2 kzk2

The rest of the proof is straightforward.

We have kϕk = sup |ϕ(x)| = sup hx, x0 i 6 sup kxk kx0 k = kx0 k.
kxk=1 kxk=1 kxk=1

 
x0 hx0 , x0 i
On the other hand, ϕ = = kx0 k, so kϕk > kx0 k as well. 2
kx0 k kx0 k

A remark on adjoints

Recall that if T : X → Y is a map between two normed spaces then we defined T ∗ : Y ∗ → X ∗

by (T ∗ f )(x) = f (T x). Suppose now that X = Y = Hilbert space, and that X and X ∗ have
been identified using the Riesz representation theorem.

Then hT x, x0 i = (T ∗ x0 )(x) = hx, T ∗ x0 i. That is, hT x, yi = hx, T ∗ yi, as expected.

Note that T ∗∗ = T . We showed before, for general normed spaces, that kT ∗k 6 kT k. Hence
kT k = kT ∗∗k 6 kT ∗k, thus kT k = kT ∗ k.

** Non-examinable section **

Von Neumann’s L2 -ergodic theorem

What is an ergodic theorem? Let X be a nice space and suppose Ψ : X → X is a nice map.
Take a point x0 ∈ X and look at iterates x0 , Ψ(x0 ), Ψ2 (x0 ), . . .

‘Typically’, we’d expect this ‘orbit’ to become equidistributed on X. Ergodic theory studies
this phenomenon. When are time averages and space averages the same?

Von Neumann’s ergodic theorem. Let H be a Hilbert space, and suppose that T : H →
H has norm at most 1. Let Y be the closed subspace of H consisting of T -invariant
vectors, i.e. T y = y. Let π : H → Y be the orthogonal projection.

1
Write SN x = N (x + T x + . . . + T N −1 x) – ‘time average’. Then SN : x 7→ π(x).

Suppose X is a compact metric space, and take H = L2 (X), the completion of the space
C(X) of continuous functions on X. E.g., X = [0, 1], C(X) has L2 -norm. Let Ψ : X → X
be a map.

This will induce a map T : H → H by defining T f (x) = f (Ψ(x)). Von Neumann’s ergodic
theorem, in this context, say that either

(i) time averages converge to space averages:

1

R
SN f = N f (x) + f (Ψ(x)) + . . . + f (Ψn−1 (x)) → constant (= X f )
(ii) there is an invariant vector, i.e. T f = f , or in other words, f (x) = f (Ψ(x)) for all x.
Taking level sets of f , i.e. sets of the form {x : f (x) > t}, one gets a set A with
Ψ(A) = A.

43
Proof. First of all note that kT ∗k 6 1 as well. Next suppose that x is T -invariant, i.e.
T x = x. Claim that T ∗ x = x as well.

To see this, write kx − T ∗ xk2 = hx − T x, xi + hx, x − T xi − kxk2 + kT ∗ xk2 = −kxk2 +

kT ∗ xk2 6 0.

Hence indeed x = T ∗ x.

Main idea. Identify Y ⊥ as the closed subspace M spanned by cocycles ∂g = g − T g.

Proof of idea. First of all, suppose f ∈ Y . Claim hf, ∂gi = 0 for all g ∈ H. Indeed,
hf, ∂gi = hf, g − T gi = hf, gi − hT ∗ f, gi = 0, since f is T -invariant. Hence if
f ∈ Y then f is orthogonal to M . Conversely suppose f is orthogonal to M , then
hf, ∂f i = 0.

But kf − T f k2 = hf − T f i + hf − T f, f i − kf k2 + kT f k2 6 0. Therefore f = T f
and hence f ∈ Y .

So Y ⊥ is the closed subspace spanned by cocycles g − T g = ∂g.

Let x ∈ X and ε > 0. Since X = Y + Y ⊥ , we can decompose x = π(x) + ∂g + h, where

khk 6 2ε . Look at each part separately.

π(x) is T -invariant, so SN (π(x)) = π(x).

1

1
Furthermore, SN (∂g) = N (g−T g)+(T g−T 2g)+. . .+(T N −1 g−T N g) = N
N (g−T g),

2 ε
and hence kSN (∂g)k 6 N kgk 6 2 for N sufficiently large.

Finally, kSN hk 6 khk 6 2ε .

ε ε
It follows that if N is big enough then kSN x − π(x)k 6 2 + 2 = ε.

** End of non-examinable section **

Orthonormal systems
Let X be a Euclidean space. A collection of elements (ϕi )i∈I is said to be orthonormal if
kϕi k = 1 for all i, and if hϕi , ϕj i = 0 when i 6= j.

Let X be a Hilbert space, and let (ϕi )i∈I be an orthonormal space. We say that the system
is complete if we can’t add another element ϕ and still have an orthonormal system.

Lemma. An orthonormal system (ϕi )i∈I is complete iff its closed linear span is X.

Proof. Suppose (ϕi )i∈I has closed linear span X. Then ((ϕi )i∈I )⊥ = X ⊥ = {0}.

Conversely, suppose (ϕI )i∈I is complete, and let Y be its closed linear span. Then
Y ⊥ = {0}. But X = Y ⊕ Y ⊥ and so X = Y . 2

Gram-Schmidt orthonormalisation. Suppose x1 , . . ., xn ∈ X are linearly independent.

Then there are orthonormal vectors y1 , . . ., yn with the linear spans hx1 , . . ., xi i =
hy1 , . . ., yi i for 1 6 i 6 n.

44
Proof. Proceed by induction on i. Suppose y1 , . . ., yi are defined.

Then define yei+1 = xi+1 − Pi xi+1 , where Pi is the orthogonal projection on to the
closed subspace hx1 , . . ., xi i = hy1 , . . ., yi i.

yei+1
Set yi+1 = . (Note, ke
yi+1 k 6= 0 since xi+1 ∈
/ hx1 , . . ., xi i.)
ke
yi+1 k

Clearly yei+1 and hence yi+1 are orthogonal to hx1 , . . ., xi i = hy1 , . . ., yi i. Also clearly,
hy1 , . . ., yi+1 i = hy1 , . . ., yi , yei+1 i = hx1 , . . ., xi+1 i. 2

Explicitly, the orthogonal projection of X onto hy1 , . . ., yi i is π(x) = hx, y1 iy1 + . . . + hx, yi iyi .
Pi
For π(x) ∈ hy1 , . . ., yi i and hx−π(x), yj i = hx, yj i− k=1 hx, yk ihyk , yi i = hx, yj i−hx, yj i = 0.

Corollary. Let X be a separable Hilbert space (i.e. X has a countable subset with dense
linear span).

Equivalently, X has a countable dense subset (taking rational linear combinations of

the above countable subset).

Then X has a complete orthonormal system (ϕn )∞

n=1 (possibly finite).

Proof. Take a countable spanning set (xn )∞ ∞

n=1 , that is a countable set with h(xn )n=1 i − X.

Thin it out if necessary so that the xi are linearly independent. Then apply Gram-
Schmidt. 2

Examples

(1) ℓ2 is a separable Hilbert space, and the orthonormal system ei = (0, . . ., 0, 1, 0, 0, . . .) is

complete.
(2∗ ) L2 (π) = L2 ([0, 2π]/(0 ∼ 2π)) = completion of L2 -norm of continuous 2π-periodic
functions on the circle.
The system (ϕn )n∈Z defined by ϕn (x) = einx is orthonormal.

1
R 2π 1 if m = n
Recall hf, gi = 2π f (x)g(x)dx. Easy check that hϕn , ϕm i =
0 0 if m 6= n
Much less obviously, (ϕn ) is complete. We saw, using Stone-Weierstrass, that every
∞
continuous
P periodic function can be uniformly approximated (in L or sup) by finite
sums an ϕn .
 R 1/2
1 2π 2
However, kf − gk2 = 2π 0
|f (t) − g(t)| dt 6 kf − gk∞ , and so the same is true
2
in L2 . That is, h(ϕn )∞n=1 i ⊃ C[0, 2π] and hence, since C[0, 2π] = L [0, 2π], we have
2
h(ϕn )∞
n=1 i = L [0, 2π] = X.
Note that the map en ↔ ϕn gives an isomorphism ℓ2 ↔ L2 [0, 2π]. We’ll see soon that
there is only one separable Hilbert space up to isomorphism, namely ℓ2 .

Reisz-Fischer theorem. Suppose (ϕn )∞ n=1 is an orthonormal system in a Hilbert space X.

Let (an )∞
n=1 be a sequence of complex numbers.
P∞ P P P
1/2
Then n=1 an ϕn ∈ X iff |an |2 < ∞ and k an ϕn k = |an |2 .

45
 PN 2
PN 2
Proof. Write xN = n=1 an ϕn . Then PkxN −x2
M k for N > M is n=M+1 |an | by Pythago-
ras. So indeed xN converges iff |an | does.
PN
Note (Pythagoras) that kxN k2 = n=1 |an |2 . 2
P∞
Further remarks. If x = n=1 cn ϕn then the cn are called the Fourier coefficients of x
with respect to (ϕn )∞
n=1 .

Pm
We have cn = hx, ϕn i. To see this, write xm = i=1 ci ϕi . Then xm → x.
Pm
Also, hxm , ϕn i = i=1 ci hϕi , ϕn i = cn if m > n. So limm→∞ hxm , ϕn i = cn = hx, ϕn i.

Bessel’s inequality and Parseval’s identity (Proposition). Let X be a Hilbert space,

and let (ϕn )∞
n=1 be an orthonormal system. Let V be the closed subspace spanned by
the ϕ. (Thus if (ϕn ) is complete, V = X.)
P∞
Then n=1 hx, ϕn iϕn = ρV x, where ρV is orthogonal projection onto V . (∗).
∞
X
In particular: |hx, ϕn i|2 = kρV xk2 6 kxk2 – Bessel’s inequality.
n=1

P∞
If x, y ∈ X, then n=1 hx, ϕn ihy, ϕn i = hρV x, ρV yi = hρV x, yi = hx, ρV yi, by self-
adjointness. (∗∗)

So if (ϕn )∞
n=1 is complete then ρV = I and so:

∞
X ∞
X
|hx, ϕn i|2 = kxk2 and hx, ϕn ihy, ϕn i = hx, yi – Parseval’s identities.
n=1 n=1

Consequently, if X admits a complete orthogonal system (ϕn ), then X is isometric to

ℓ2 , the isometry being given by the Fourier transform: x 7→ (hx, ϕn i)∞
n=1 .

(Recall, isometric means isomorphic, with the isomorphism preserving the inner product.)

The statements (∗) and (∗∗) require proof.

P∞
Proof of (∗). Let’s first note that n=1 hx, ϕn iϕn exists.
Pm
To see this, write xm = n=1 hx, ϕn iϕn . It’s easy to check that x − xm is orthogonal to
ϕ1 , . . ., ϕm . Hence x = xm + (x − xm ) is a decomposition of x into orthogonal vectors,
and so by Pythagoras kxm k 6 kxk.
Pm P∞ P∞
But kxm k2 = n=1 |hx, ϕn i|2 . Hence n=1 |hx, ϕn i|2 < ∞, and so n=1 hx, ϕn iϕn
exists.
P∞
We’ve essentially shown that x − n=1 hx, ϕn iϕn = limm→∞ (x − xm ).
P∞
Since x−xm is orthogonal to ϕ1 , . . ., ϕm , we have that x− n=1 hx, ϕn iϕn is orthogonal
to all of the ϕn
P P∞
Since ∞ ∞
n=1 hx, ϕn iϕn is in the closed linear span of (ϕn )n=1 , we have n=1 hx, ϕn iϕn =
ρV x, as claimed. 2

46
Proof of (∗∗). This is straightforward.
Pm
hρV x, ρV yi = limm→∞ hxm , ym i, where xm = n=1 hx, ϕn iϕn and similarly for ym .
Pm P∞
So hρV x, ρV yi = limm→∞ n=1 hx, ϕn ihy, ϕn i = n=1 hx, ϕn ihy, ϕn i.

This is also hx, ρV yi = hρV x, yi by the self-adjointness of ρV . 2

** Non-examinable section **

It follows from this that

P Fourier series converge in L2 . Indeed, if f is a continuous 2π-periodic
e
function and SN f = |n|6N f (n)einx , then kSN f − f k2 → 0.

Proof. Work in L2 [0, 2π], the Hilbert space obtained by completing the continuous functions
C[0, 2π] in the L2 -norm. The exponentials ϕn (x) = einx , n ∈ Z, are a complete
orthonormal system (by Stone-Weierstrass).
Z 2π
1
f (n) = f (x)e−inx dx = hf, ϕn i.
2π 0
X
Thus kSN f − f k2 = |hf, ϕn i|2 → 0 as N → ∞.
|n|>N

Easier proof: approximate

P f by trig. polynomials using Stone-Weierstass. If p is a trig.
inx
polynomial (i.e. |n|6N an a ) then SN ′ p = 0 for N ′ > N . 2

This doesn’t contradict the earlier counterexample, because kSN f − f k2 → 0 does not imply
SN f (x) → f (x) pointwise.

Famous (hard) theorem by Carleson: if f ∈ L2 then SN f → f pointwise for almost all x.

** End of non-examinable section **

47
7. Spectral Theory
Let T : X → X be a bounded linear operator on a Banach space.

Can we talk about eigenvectors and eigenvalues? Can we diagonalise?

Definition. If T x = λx, x 6= 0, then x is an eigenvector of eigenvalue λ.

Examples. Let X = ℓ2 .


1. Right shift: T (x1 , x2 , . . .) = (0, x1 , x2 , . . .).
This has no eigenvalues. Suppose (0, x1 , x2 , . . .) = λ(x1 , x2 , . . .). Then λ 6= 0 we have
inductively that x1 = x2 = . . . = 0. And if λ = 0 then we also have x1 = x2 = . . . = 0.


2. Left shift: T (x1 , x2 , . . .) = (x2 , x3 , . . .).
Every λ ∈ C with |λ| < 1 is an eigenvalue, with eigenvector (1, λ, λ2 , λ3 , . . .).

It is no longer the case that λ is an eigenvalue ⇐⇒ T − λI is not invertible.

(⇒) is still true, since if T x = λx then x ∈ ker(T − λI).

(⇐) is not true: for the right shift, T itself is not invertible, but 0 is not an eigenvalue.

Definition. The spectrum of T is σ(T ) = {λ ∈ C : T − λI is not invertible}.

If λ is an eigenvalue (i.e. there exists x 6= 0 for which T x = λx) then we say that λ lies in
the point spectrum, σp (T ).

Clearly, σp (T ) ⊂ σ(T ), because if T − λI is invertible then (T − λI)x = 0 ⇒ x = 0, and so

λ∈/ σP (T ). However, we can have σp (T ) 6= σ(T ), as with the right shift.

Remark. If X is finite dimensional then σp (T ) = σ(T ), since a linear map T : X → X is an

isomorphism iff ker T = {0}.

General properties of the spectrum

Proposition. σ(T ) is a compact subset of {z ∈ C : |z| 6 kT k}.

Proof. All of this is a consequence of a simple observation of Neumann:

Claim. If kT k < 1, then I − T is invertible, with inverse I + T + T 2 + T 3 + . . .

Proof of claim. Indeed, the operator on the right hand side is well-defined since

kT m+1 + . . . + T n k 6 kT m+1 k + . . . + kT n k 6 kT km+1 + . . . + kT kn ,

since kABk 6 kAk kBk.

And this → 0 as m, n → ∞ since 1 + x + x2 + . . . converges with |x| < 1.

Thus (1 + T + . . . + T n )∞ n=1 is a Cauchy sequence in the operator norm, and so

I + T + T 2 + . . . is well-defined since the space B(X) of bounded linear operators
is a Banach space..

48
 Also,

(I − T )(I + T + T 2 + . . .) = lim (I − T )(I + T + . . . + T n )

n→∞
= lim I − T n+1
n→∞
= I

where, of course, limits are in the operator norm.

The same works on the left, so I + T + T 2 + . . . is an inverse for I − T , as claimed.

It follows immediately that σ(T ) ⊂ {z ∈ C : |z| 6 kT k}. Indeed, if λ > kT k then

I − λ1 T is invertible as above, and hence so is T − λI.

To show that σ(T ) is compact, it suffices to show it’s closed, for which is suffices to
show that the invertible operators are an open subset of B(X).


So let T ∈ B(X) be invertible. If S is another operator, write S = T I + T −1 (S − T ) .

But if kS − T k < kT −1 k−1 then kT −1(S − T )k 6 kT −1 k kS − T k < 1, and so by Neu-

mann’s observation, S is the product of two invertible operators and hence is invertible
itself. 2

Proposition. Suppose T : X → X is a bounded linear operator from a Banach space X to

itself, and suppose λ ∈ σ(T ).

Then either T − λI fails to have dense image, or else λ is an approximate eigenvalue

of T , meaning there is a sequence (xn )∞
n=1 of unit vectors in X with (T − λI)xn → 0.

Proof. Replacing T with T − λI, we may assume λ = 0. Suppose 0 is not an approximate

eigenvalue for T , so we can’t choose unit vectors xn with T xn → 0. Thus T is bounded
below, i.e. kT xk > εkxk, some ε > 0 (∗).

In particular, T is injective (as ker T = {0}). Assume also that Y = im T is dense.

Take an arbitrary z ∈ X, and let y1 , y2 , . . . be a sequence of elements in Y with yn → z.

Since T is bounded below, T −1 |Y is bounded, with norm at most 1ε , by (∗).

Hence T −1 yn converges to some x. But then T x = lim T (T −1 yn ) = lim yn = z.

n→∞ n→∞

It follows that im T = X, so therefore T is invertible and kT −1k < 1ε .

Hence 0 ∈
/ σ(T ), and the result follows. 2

Remark. Conversely, it’s clear that if T − λI doesn’t have dense image, or if λ is an ap-
proximate eigenvalue of T , then λ ∈ σ(T ).

From now on, we start specialising, considering in turn the following classes of operators:

• X is a Hilbert space, T : X → X is compact and self-adjoint

• X is a Banach space, T is compact
• X is a Banach space, T is bounded

49
Compact Operators
Let X, Y be Banach spaces, and T : X → Y a bounded linear operator.

Definition. T is compact if T (BX (1)) is a (pre)compact subset of Y (that is to say, it has

compact closure) where here BX (1) = {x ∈ X : kxk 6 1}.

(Actually, if follows from the open mapping theorem that T (BX (1)) is automatically
closed, so the ‘pre’ is superfluous.)

Note. BX (1) will always, if X is infinite dimensional, be non-compact. So this is a strong

condition on T .

Example 1. Finite-rank operators: im T is finite dimensional.

Example 2. Interval operators.

Let X = C[0, 1], say, and let K : [0, 1]2 → C be a continuous function ‘kernel’. Define
R1
T : X → X by T f (x) = 0 f (y)K(x, y)dy.

Claim. T is a compact operator on X.

Proof. We need to show that the set {T f : kf k∞ 6 1} is precompact in X. By the Arzelà-

Ascoli theorem, it suffices to show that it’s uniformly bounded and equicontinuous.
Z 1
UB. kT f k∞ 6 kf k∞ sup |K(x, y)|dy 6 C for all f with kf k∞ 6 1,
x 0

Z 1
where C = sup |K(x, y)|dy 6 kKk∞ , which is finite since [0, 1]2 is compact.
x 0

Z 1
EQ. |T f (x1 ) − T f (x2 )| = f (y) (K(x1 , y) − K(x2 , y)) dy
0
Z 1
6 kf k∞ |K(x1 , y) − K(x2 , y)| dy
0

But K is uniformly continuous, so in particular if δ is small enough, then we have

|K(x1 , y) − K(x2 , y)| 6 kfεk∞ whenever |x1 − x2 | 6 δ and for all y.

So if |x1 − x2 | 6 δ then |T f (x1 ) − T f (x2 )| 6 ε for all f with kf k∞ 6 1. 2

Example 3. Hilbert-Schmidt operators.

Here, let X = Y = ℓ2 .

Suppose (ei )∞ P∞orthonormal system). If T : X → X

i=1 is an orthonormal basis (i.e. a complete
is a bounded linear operator then define kT kHS = i=1 kT ei k2 , if this is finite – in which
case we say that T is Hilbert-Schmidt.

We note first of all that kT kHS doesn’t depend on the (ei )∞

i=1 .

Proof. Let (e′i )∞

i=1 be another orthonormal basis.
P P
Then by Parseval’s identity, kT ei k2 = j |hT ei , e′j i|2 = j |hei , T ∗ e′j i|2 .

50
 P P P P
Hence i kT ei k2 = j i |hei , T ∗ e′j i|2 = j kT ∗ e′j k2 , by Parseval again.
P P
ei )∞
Similarly, if (e i=1 is another orthonormal system then i ei k2 =
kT e j kT ∗ e′j k2 .
P P
Hence kT eik2 = ei k2 , as required.
kT e 2

We claim that Hilbert-Schmidt operators are compact. Fix a basis (ei )∞

i=1 . Define the matrix
coefficients aij = hT ej , ei i, just like in finite dimensions.
P P 
Then T x = i j a ij xj ei , just like in finite dimensions.

P P
(To prove this rigorously, consider x(n) = (x1 , . . ., xn , 0, . . .), note T x(n) = i j6n aij xj ei ,
and let n → ∞.)
P
Hence if x ∈ BX (1), that is if kxk 6 1, then (T x)i = j aij xj , and so by Cauchy-Schwarz
P 1/2
2
|(T x)i | 6 j |a ij | = bi .

But if T is Hilbert-Schmidt then

X XX XX X
|bi |2 = |aij |2 = |hT ej , ei i|2 = kT ej k2 = kT kHS < ∞.
i i j i j j

P
So in fact T is Hilbert-Schmidt if and only if i,j |aij |2 < ∞.

Hilbert cube, {y = (y1 , y2 , . . .) ∈ ℓ2 : |yi | 6 bi },

It follows that T (BX (1)) is contained in the P
where (bi ) is a fixed sequence of reals with |bi |2 < ∞.
Q∞
(Note, ‘the’ Hilbert cube is the special case bi = 1i , that is the set 1 1
i=1 [− i , i ] ⊂ ℓ2 .)

We’ll show that any Hilbert cube is compact.

∗∗ Proof 1 (non-examinable). The Hilbert cube is homeomorphic to a countably infinite

direct product of closed intervals, with the product topology. And Tychonoff’s theorem
(equivalent to the Axiom of Choice) states that any product of compact spaces is
compact. 2∗∗

Proof 2. Diagonalisation argument. Given a sequence (x(i) )∞

i=1 in the Hilbert cube, one can
pass to a convergent subsequence as follows.

(i)
Look at the first coordinate x1 , i = 1, 2, . . .. Since these live in a closed interval, pass
to a convergent subsequence. With this new sequence, take a subsequence for which
the second coordinates converge, and so on.

Now take the sequence consisting of the nth element of the nth sequence, and then all
coordinates converge. But any such sequence in a Hilbert cube converges. 2

Some general facts about compact operators

Proposition. Let X, Y be Banach spaces. Then B0 (X, Y ), the space of compact operators
from X to Y , is a closed linear subspace of B(X, Y ), the space of all bounded linear
operators from X to Y .

51
Proof. A compact operator is bounded, since T (BX (1)) is bounded.

We need that if S, T ∈ B0 (X, Y ) then S + T, λT ∈ B0 (X, Y ) for λ ∈ C. Let (xn )∞

n=1 be
a sequence of unit vectors in X. We’ll show that there is a subsequence of the xn such
that (S + T )xn converges. Then (S + T )(BX (1)) is sequentially compact.

To do this, first use compactness of S to pass to a subsequence x′n for which Sx′n
converges. Now use compactness of T to pass to a further subsequence x′′n for which
T x′′n also converges. Then (S + T )x′′n converges.

The fact that T ∈ B0 (X, Y ) implies λT ∈ B0 (X, Y ) is clear.

Now let (Tn )∞

n=1 be a sequence of compact operators with Tn → T in the operator
norm. We must show that T is compact. We’ll show that T (BX (1)) is totally bounded.

Let ε > 0 and let n be such that kTn − T k 6 2ε . Since Tn is compact, Tn (BX (1)) is
totally bounded and so there are y1 , . . ., ym ∈ Y such that for all x ∈ BX (1) there is i
such that kTn x − yi k 6 2ε .
ε ε ε
But then kT x − yi k 6 kT x − Tn xk + kTn x − yi k 6 kT − Tn k kxk + 2 6 2 + 2 = ε.

Since ε > 0 was arbitrary, T (BX (1)) is totally bounded. 2

Example 4 (of compact operators). Multiplication operators on Hilbert space.

Let (αi )∞i=1 be
a sequence of complex numbers tending to 0. Define T : ℓ2 → ℓ2 by

T (x1 , x2 , . . .) = (α1 x1 , α2 x2 , . . .). Fairly clearly, T is bounded and linear, with kT | =
max |αi |.

Furthermore, T is compact, being
the limit in the operator norm of the finite-rank operators
Tn defined by Tn (x1 , x2 , . . .) = (α1 x1 , . . ., αn xn , 0, 0, . . .). Each Tn is compact, as are all
finite-rank operators. And


k(T − Tn ) (x1 , x2 , . . .) k = k(0, . . ., 0, αn+1 xn+1 , αn+2 xn+2 , . . .)k
6 sup |αm | k(0, . . ., 0, xn+1 , xn+2 , . . .)k
m>n
6 sup |αm | kxk.
m>n

Thus kT − Tn k 6 supm>n |αm | → 0 as m → ∞.

It is extremely difficult to find a compact operator T : X → Y between Banach spaces which

is not a limit of finite-rank operators. The first example was found by Per Enfio in 1973.

However, there is no such example in Hilbert space.

Proposition. Every compact operator on Hilbert space is a limit of finite-rank operators.

Proof. Let T : X → X be compact, and let ε > 0. Since T (BX (1)) is totally bounded,
there are points y1 , . . ., ym ∈ X such that for each x ∈ BX (1) there is i such that
kT x − yi k 6 2ε .

Let Y = hy1 , . . ., ym i be the linear span of y1 , . . ., ym , and consider π : X → Y , the

orthogonal projection. Set Te = π ◦ T . Since Y is finite dimensional, Te is finite rank.

52
 We’ll show kT − Tek 6 ε. Let x ∈ BX (1) and choose yi so that kT x − yi k 6 2ε .

Then kT x − Texk 6 kT x − yi k + kTex − yi k 6 kT x − yi k + kT x − yi k 6 ε

2 + ε
2 = ε.

(In fact, kz − yi k > kπ(z) − yi k for any z, by Pythagoras.)

So kT − Tek 6 ε. Since ε was arbitrary, T is indeed a limit of finite-rank operators. 2

Proposition. Let X be any Banach space, and suppose λ 6= 0 is an approximate eigenvalue

for some compact operators T : X → X. (Recall this means that there is a sequence
of unit vectors (xn )∞
n=1 with (T − λI)xn → 0.)

Then λ is actually an eigenvalue of T .

Proof. Since T (BX (1)) is compact, there is a convergent subsequence of the (xn ). Assume,
relabelling if necessary, that T xn → z. But then, xn → λ1 z. Let z ∗ = λ1 z.

But then (T − λI)z ∗ = limn→∞ (T − λI)xn = 0, and kz ∗ k = 1.

Hence z ∗ is an eigenvector of T with eigenvalue λ. 2

Spectral theorem for compact, self-adjoint operators on Hilbert space

Revision (from Linear Algebra). If T : CnP → Cn is a self-adjoint (Hermitian) linear
map, i.e. hT x, yi = hx, T yi where hx, yi = ni=1 xi yi .

Then the eigenvalues of T are all real, eigenvectors corresponding to distint eigenvalues
are orthogonal, and T is diagonalisable, meaning that there is an orthonormal basis for
Cn consisting of eigenvectors of T . (Apply Gram-Schmidt within each eigenspace.)

Aim. To prove an analogous result in Hilbert space.

Theorem (Spectral theorem). Let T : X → X be such an operator. Then all eigenvalues

are real, and eigenspaces corresponding to distinct eigenvalues are orthogonal.

Furthermore, these eigenspaces are finite-dimensional, and there are only finitely many
eigenvalues λ with |λ| > ε for any ε > 0.

There is an orthonormal
P basis for X consisting of eigenvectors of T . More precisely,
we may write T = λ ρV λ, where ρV is the orthogonal projection onto the (finite-
dimensional) eigenspace Eλ corresponding to λ.

Conversely, any operator of this form is compact and self-adjoint.

We start with an observation about the spectrum of bounded (not necessarily compact)
self-adjoint operators on Hilbert space.

Lemma. Let T : X → X be such an operator. Then if λ ∈ σ(T ), then λ is an approxi-

mate eigenvalue for T . That is, σ(T ) = σap (T ), the approximate point spectrum, i.e.
approximate eigenvalues of T .

Proof. We saw before that if λ ∈ σ(T ) then either λ is an approximate eigenvalue or

im (T − λI) is not dense in X. Wlog take λ = 0, else replace T with T − λI.

53
 Suppose then that 0 is not an approximate eigenvalue. Then T is bounded below, say
kT xk > εkxk.

But then (im T )⊥ = ker T because T is self-adjoint: if hx, T yi = 0 for all y, then
hT x, yi = 0 for all y, so T x = 0.

Since T is bounded below, ker T = {0}, so (im T )⊥ = {0}, so im T is dense in X. /\/\ 2

Corollary. Suppose now that T is compact and self-adjoint. Then σ(T ) \ {0} = σp (T ) \ {0}.

That is, for λ ∈ σ(T ), if λ 6= 0 then λ is an eigenvalue of T .

Proof. Combine the preceding lemma with the result from earlier – that non-zero approxi-
mate eigenvalues are actual eigenvalues (for compact operators).

Proposition. Suppose T is compact and self-adjoint. Then T has at least one eigenvalue.
More specifically, either kT k or −kT k is an eigenvalue.

In particular, a non-zero compact self-adjoint operator has a non-zero eigenvalue.

Proof. We’ll show that T 2 − kT k2 I is not invertible. Then it follows that at least one of
T − kT kI and T + kT kI is not invertible. That is, σ(T ) contains one of ±kT k, and
from the corollary above it then follows that one of ±kT k lies in σp (T ). (The proof is
trivial if T = 0, so we may assume not.)

We’ll show that T 2 − kT k2I is not invertible by exhibiting an approximate eigenvector.

Take a sequence (xn )∞n=1 of unit vectors with kT xn k → kT k. Then

2
T 2 xn − kT k2xn = kT 2 xn k2 + kT k4 − kT k2hxn , T 2 xn i − kT k2 hT 2 xn , xn i
= kT 2 xn k2 + kT k4 − 2kT k2kT xn k2
6 2kT k4 − 2kT k2kT xn k2 since kT 2xn k 6 kT 2 k 6 kT k2
→ 0 by the choice of the xn

That is, (T 2 − kT k2I)xn → 0, which is exactly what we wanted. 2

Proof of the spectral theorem. We now prove the theorem.

• That all eigenvalues are real is true in exactly the same way as in the finite-
dimensional case. That is, λkxk2 = hT x, xi = hx, T xi = λkxk2 , whence λ = λ.

• Ditto the fact that eigenvectors corresponding to distinct eigenvalues are orthog-
onal: if T x = λx and T y = µy, then λhx, yi = hT x, yi = hx, T yi = µhx, yi, so if
λ 6= µ then hx, yi = 0.

• We turn now to the finite dimensionality of the eigenspaces Eλ , λ 6= 0, as well as

the finiteness of the set {λ : λ an eigenvalue of T, |λ| > ε} for every ε > 0.

If either of these statements fails, then there is some ε > 0, together with an
infinite orthonormal sequence (xn )∞n=1 with T xn = λn xn , |λn | > ε for all n.

(If the first statement fails, use Gram-Schmidt on an eigenspace Eλ with dim Eλ =
∞. If the second fails, simply pick one unit vector from each eigenspace.)

54
 But then kT xn − T xm k2 = kλn xn − λm xm k2 = λ2n + λ2m > 2ε2 .

So then (T xn )∞n=1 does not have a convergent subsequence, contrary to the com-
pactness of T .

• Finally, let V be the closed linear span of the eigenspaces Eλ , λ 6= 0. We may

write X = V ⊕ V ⊥ .

We claim that T preserves V ⊥ . Indeed, if x ⊥ Eλ then hx, yi = 0 for all y ∈ Eλ ,

and therefore hT x, yi = hx, T yi = λhx, yi = 0. Thus T x ⊥ Eλ as well.

Note that T |V ⊥ cannot have any non-zero eigenvalues, since V ⊥ is orthogonal to

all eigenspaces of T . It follows from the proposition that T |V ⊥ = 0.

To get an orthonormal basis for X consisting of eigenvectors of T , pick orthonormal

bases for the Eλ s, together with any orthonormal basis for V ⊥ .

Suppose this basis consists of eigenvectors e1 , e2 , . . . with T ei = λi ei , together

with a basis (fi )∞ ⊥
i=1 for V .

Then T (x1 e1 + . . . + xn en + . . . + y1 f1 + . . .) = λ1 x1 + . . . + λn xn + . . . + 0 + 0 + . . ..

This completes the proof of the spectral theorem. 2

55
The Hahn-Banach Theorem BJG October 2011

Conspicuous by its absence from this course (Cambridge Mathematical Tripos Part
II, Linear Analysis) is the Hahn-Banach theorem. A simple version of it is as follows.
Theorem 1 (Hahn-Banach). Let V, Ṽ be normed spaces with V ⊆ Ṽ . Let φ : V → R
be a bounded linear functional. Then there is a bounded linear functional φ̃ : Ṽ → R
which extends φ in the sense that φ̃|V = φ, and for which kφ̃k = kφk.
Proof. Roughly speaking, the idea is to extend φ to φ̃ “one dimension at a time”.
Suppose, then, that Ṽ = V + hwi, where w ∈ / V . By rescaling we may assume, without
loss of generality, that kφk = 1. We are forced to define

φ̃(v + w) = φ(v) + tλ

for all v ∈ V and t ∈ R, where λ must not depend on v or on t. A map φ̃ defined in this
way will always be a linear functional, but our task is to show that by judicious choice
of λ we may ensure that kφ̃k 6 1. For this we require that

|φ(v) + tλ| 6 kv + twk

for all v ∈ V and t ∈ R. By replacing v by v/t and using linearity of φ, it suffices to

establish this in the case t = 1; that is to say, we must show that there is λ ∈ R such
that
|φ(v) + λ| 6 kv + wk
for all v ∈ V . Equivalently,

−kv + wk − φ(v) 6 λ 6 kv + wk − φ(v).

There will be such a λ if, and only if,

−kv + wk − φ(v) 6 kv 0 + wk − φ(v 0 )

for all v, v 0 ∈ V . Indeed, we could then take any λ ∈ [m, M ] with

m := sup −kv + wk − φ(v), M := inf

0
kv 0 + wk − φ(v 0 ).
v∈V v ∈V

Rearranging, the inequality we are required to prove is

φ(v 0 ) − φ(v) 6 kv 0 + wk + kv + wk

for all v, v 0 ∈ V . However the left-hand side is φ(v 0 − v) which, since kφk = 1, has
magnitude at most kv 0 − vk. The result is now a consequence of the triangle inequality.
We have proved the Hahn-Banach theorem when Ṽ is obtained from V by the addition
of one vector. This is already enough to prove the whole theorem when Ṽ is finite-
dimensional (by incrementing the dimension of V one step at a time). Essentially the
1
2

same argument works in the infinite-dimensional case, too, although Zorn’s lemma is
needed to make this rigorous.
Consider the set of all extensions of φ, that is to say pairs (V 0 , φ0 ) where V ⊆ V 0 ⊆ Ṽ ,
φ0 |V = φ, and kφ0 k 6 kφk. There is an obvious partial order on this set: namely, say
that (V1 , φ1 )  (V2 , φ2 ) if and only if V1 ⊆ V2 and φ2 |V1 = φ1 . Every chain in this partial
order has an upper bound. Indeed if (Vi , φi )i∈I is a chain, then an upper bound for it is
(V 0 , φ0 ), where V 0 = i∈I Vi and φ0 equals φi on Vi , for all i. By Zorn’s lemma, there is
S

a maximal element (V0 , φ0 ). However by the special case of the theorem proved above,
we could extend φ0 to V0 + hwi for any w ∈ / V0 . The only possible conclusion is that
there is no w ∈ / V0 , or in other words V0 = Ṽ and φ0 is defined on all of Ṽ .

Remark. Zorn’s lemma is equivalent to the axiom of choice, and so we have used
the axiom of choice in proving Hahn-Banach. It is known that Hahn-Banach is strictly
weaker than the axiom of choice, but cannot be proven in ZF.

Let us derive some consequences of the theorem. The main point is that, without it,
we are essentially powerless to construct a good supply of bounded linear functionals
on a typical normed space X. With it, however, we immediately see that X ∗ is quite
rich; indeed for any x ∈ X there is some φ ∈ X ∗ such that φ(x) 6= 0. More specifically,
there is some φ ∈ X ∗ with kφk = 1 such that φ(x) = kxk. To see these facts, simply
take V to be the subspace spanned by hxi and Ṽ := X, and extend the linear functional
φ0 : V → R defined by φ0 (tx) = tkxk.
One may think of this geometrically in terms of convex bodies admitting supporting
hyperplanes. Consider the unit ball B := {x ∈ X : kxk 6 1} (which is the most
general form of a convex set) and let x0 ∈ B have norm 1. As just remarked, there is
a linear functional φ : X → R with φ(x0 ) = 1 and kφk 6 1. Consider the hyperplane
H := {x ∈ X : φ(x) = 1}. Then H meets B at x0 (and possibly at other points).
However if x ∈ B then φ(x) 6 kφkkxk 6 1, and so all of B lies in the half-space
{x ∈ X : φ(x) 6 1}. H is called a supporting hyperplane for B.

The following interesting fact is little more than a rephrasing of the above.

Theorem 2. Let X be a normed space. Then the natural map from X to X ∗∗ is an

isometry.

Proof. The natural map in question associates to x ∈ X the functional x̂ on X ∗ defined

by x̂(φ) := φ(x). It is easy to see that kx̂k 6 kxk. To get an inequality in the other
direction, choose φ as described above. Then |x̂(φ)| = |φ(x)| = kxk = kxkkφk, and so
indeed kx̂k > kxk.

A slightly more complicated observation in the same vein is the following.

 3

Theorem 3. Let X, Y be normed spaces, and suppose that T : X → Y is a bounded

linear map. Let T ∗ : Y ∗ → X ∗ be its dual. Then kT ∗ k = kT k.

Proof. We remark that it is very easy to see that kT ∗ k 6 kT k; indeed we already

remarked on this in the main part of the course. Let ε > 0 be arbitrary. By definition
of kT k, there is some x ∈ X, x 6= 0, with kT xk > (kT k − ε)kxk. Using the remark
above, choose a linear functional φ ∈ X ∗ with kφk = 1 and φ(T x) = kT xk. Then

T ∗ φ(x) = φ(T x) = kT xk > (kT k − ε)kxk,

which certainly means that

kT ∗ φk > kT k − ε = (kT k − ε)kφk.

It follows that
kT ∗ k > kT k − ε.
Since ε > 0 was arbitrary, the result follows.

Here is another fact we were unable to establish in the course proper.

Theorem 4. The dual of `∞ is strictly bigger than `1 .

Proof. Certainly the dual (`∞ )∗ contains `1 , since if (bi )i∈N ∈ `1 then the map
P
(ai )i∈N 7→ i ai bi is a bounded linear functional. Note that any functional of this
type is determined by its values on `∞ 0 , the closed subspace of `
∞
consisting of se-
quences which tend to zero. This is obviously a proper subspace of `∞ , and so the
quotient space `∞ /`∞ 0 is a nontrivial normed space. By Hahn-Banach we may find a
nontrivial bounded linear functional ψ on it. This pulls back under the quotient map
π : `∞ → `∞ /`∞ ∞ ∗
0 to give a nontrivial functional φ ∈ (` ) defined by φ(x) := ψ(π(x)).
Since φ is trivial on `∞ 1
0 , it does not come from ` .

The applications we have given so far are perhaps not very “surprising”. The next
one is rather more so.

Theorem 5 (Finitely additive measure on Z). Write P(Z) for the set of all subsets of
Z. Then there is a “measure” µ : P(Z) → [0, 1] which is normalised so that µ(Z) = 1,
is shift-invariant in the sense that µ(A + 1) = µ(A), and is finitely-additive in the sense
that µ(A1 ∪ · · · ∪ Ak ) = µ(A1 ) + µ(A2 ) + · · · + µ(Ak ) whenever A1 , . . . , Ak are disjoint.

Proof. For the purposes of this proof write `∞ for the Banach space of bounded
sequences (xn )n∈Z indexed by Z. We will in fact construct a linear functional φ ∈ (`∞ )∗
which is shift-invariant in the sense that φ((xn )n∈Z ) = φ((xn+1 )n∈Z ), positive in the
sense that φ((xn )n∈Z ) > 0 whenever xn > 0 for all n, and normalised so that φ(1) = 1,
where 1 is the constant sequence of 1s.
4

Clearly such a φ gives rise to a finitely additive measure on Z by defining µ(A) :=

φ(xA ), where (xA )n = 1 if n ∈ A and 0 otherwise.
Let V be the subspace of `∞ spanned by the constant sequence 1 and the space V0
of sequences of the form (xn+1 − xn )n∈Z . It is trivial to check that 1 is not in V0 , so we
may unambiguously define

φ0 ((xn+1 − xn + c)n∈Z ) := c

on V . For any ε > 0 there must be some n such that xn+1 − xn > −ε, and so if c > 0
we certainly have k(xn+1 − xn − c)n∈Z k = supn |xn+1 − xn + c| > |c| − ε. Since ε was
arbitrary, we actually have k(xn+1 − xn + c)n∈Z k > |c|. The same conclusion holds if
c 6 0, and therefore kφ0 k 6 1. By the Hahn-Banach theorem there is an extension of
φ0 to a linear functional φ on all of `∞ such that kφk 6 1. This is obviously normalised
so that φ(1) = 1, and φ is pretty clearly shift-invariant since (xn )n∈Z and the shifted
sequence (xn+1 )n∈Z differ by an element of V0 , on which φ0 is defined and equal to zero.
It remains to confirm that φ is positive, and for this we may suppose without loss of
generality that kxk = 1, so that 0 6 xn 6 1 for all n. Then k1 − xk 6 1, and so

1 − φ(x) = φ(1 − x) 6 k1 − xk 6 1,

which obviously means that φ(x) > 0 as required.

Linear Analysis Example Sheet 1 BJG, Michaelmas 2010

In this sheet all normed spaces are over R.

1. Prove that if V is a normed space then the addition map P : V ×V → V

defined by P (v1 , v2 ) = v1 +v2 and the scalar multiplication map S : R×V →
V defined by S(λ, v) = λv are both continuous.
2. When does equality occur in Hölder’s inequality?
3. Let X, Y and Z be normed spaces and let T : Y → Z and S : X → Y
be bounded linear maps. Show that the composition T S is bounded and
that kT Sk 6 kT kkSk. Show, by providing an example, that equality need
not occur.
4. If x ∈ Rn , define |x|1/2 = (|x1 |1/2 + · · · + |xn |1/2 )2 . Is this a norm?
5. Let X be the vector space of continuously differentiable functions on
[0, 1]. Show that X is incomplete in the uniform norm k · k∞ , but complete
in the norm kf k = kf k∞ + kf 0 k∞ .
6. Find, in terms of n, the best constants C, C 0 such that
Ckxk2 6 kxk4 6 C 0 kxk2
for all x ∈ Rn .
7*. Let X = R[0, 1]/ ∼ be the space of Riemann integrable functions
modulo functions 1
R 1 with Riemann integral zero. Is X complete in the L -
norm kf k := 0 |f (t)| dt?
8. Suppose that V is a normed space. If x ∈ V \ {0} write π(x) = x/kxk.
Is it true that kπ(x) − π(y)k 6 kx − yk whenever kxk, kyk > 1?
9. Suppose that Y and Z are dense subspaces of a normed space X. Must
Y ∩ Z be dense in X?
10. A Banach space X is said to be separable if it contains a countable
dense set. Show that `1 is separable, but `∞ is not.
11. Does there exist a Banach space with countably infinite dimension?
12. Is it true that every linear operator on a Banach space is bounded?
13. Give an example of a normed space X and an unbounded linear map
T : X → R which vanishes on a dense subspace of X.
Linear Analysis Example Sheet 2 BJG, Michaelmas 2010

1. Let c0 be the vector space of all sequences (x1 , x2 , . . . ) with xi → 0

as i → ∞. Show that c0 is a Banach space with the sup norm. Show that
c∗0 = `1 .
2. Suppose that T is an unbounded linear functional on a normed space
X. Show that ker T is dense in X.
3. Suppose that T : `∞ → R is a linear functional with the property that
T (x1 , x2 , . . . ) > 0 whenever x1 , x2 , . . . > 0. Show that T is bounded.
4. Let X be a Banach Space. Show that the unit ball of X is compact if
and only if X is finite-dimensional.
5. A metric space is said to be separable if it has a countable dense subset.
Prove that every compact metric space X is separable. Show furthermore
that the space C(X) of continuous functions on X, with the metric induced
by the sup norm, is separable.
6. Let X be a normed space. The aim of this exercise is to show that X
is separable if X ∗ is. Suppose that (φn )∞n=1 is a countable dense subset of
X ∗ , and take a sequence of unit vectors xn ∈ X such that |φn (xn )| > 21 kφn k.
Suppose that V , the closed subspace of X spanned by the xn , is not all of
X. A consequence of the Hahn-Banach theorem is that there is φ ∈ X ∗ with
φ|V = 0 but kφk = 1. Show that this leads to a contradiction.
7. Show that `∞ is not separable. Hence, or otherwise, show that X ∗∗
need not be isometrically isomorphic to X when X is a Banach space.
8. Let f : (0, ∞) → R be a continuous function such that for every x > 0
we have f (nx) → 0 as n → ∞. Show that f (x) → 0 as x → ∞.
9. A subset E of a metric space is said to be perfect if it is closed and
every point of E is a limit point of E. Show that a perfect subset of R
is uncountable. Hence, or otherwise, show that R cannot be written as a
countable disjoint union of closed intervals.
10. Suppose that k · k is a complete norm on C[0, 1] (i.e. C[0, 1] is a
Banach space with this norm) with the property that each evaluation map
f → f (x), x ∈ [0, 1], is continuous. Show that the topology induced on
C[0, 1] by this norm is the same as that induced by the sup norm.
11. Suppose that fn : [0, 1] → R is a sequence of continuous functions
tending pointwise to 0. Is there necessarily an interval on which fn → 0
uniformly?
12*. Let f : R → R be an infinitely differentiable function such that for
every x ∈ R there is an n with f (m) (x) = 0 for all m > n. Prove that f is a
polynomial.
Linear Analysis Example Sheet 3 BJG, Michaelmas 2010

1. Check carefully that if X is a locally compact Hausdorff space then

the one-point compactification X̃ is a compact Hausdorff space.
2. Show that if X is an infinite compact Hausdorff space then C(X) is
not compact.
3. Let f : [0, 2π] → R be a continuous function with f (0) = f (2π), and
R 2π
suppose that fˆ(n) = 0 for all n ∈ Z, where fˆ(n) = 0 f (x)e−inx dx. Show
that f = 0.
4. Let X be a locally compact Hausdorff space. Show that Cc (X), the
space of continuous real-valued functions on X with compact support (that
is, {x : f (x) = 0} is compact) is dense in C0 (X).
5. Let A be the collection of all functions f ∈ C([0, 1]) which are bounded
pointwise by 1 and such that |f (x)−f (y)| 6 |x−y| for all x, y. Show directly
that A is sequentially compact (hint: given a sequence of functions (fi ), first
pass to a subsequence which converges at every rational).
6. Show that the set of all x = (x1 , x2 , . . . ) ∈ `2 with |xn | 6 1/n for all n
is compact.
7. Does there exist a real Hilbert space with countably infinite dimension
as a vector space over R?
8. Let {en : n ∈ Z} be an orthonormal basis for `2 , and consider the set
{fn : n ∈ Z}, where fn := en − en−1 . Show that the linear span of {fn }
is dense in `2 . Is the same true if a single element is removed from {fn }?
What if two elements are removed?
9. Suppose that S is a countably infinite subset of `2 with the property
that the linear span of S 0 is dense in `2 whenever S \ S 0 is finite. Show that
there is some S 0 whose linear span is dense in `2 and for which S \ S 0 is
infinite.
10. Show√that there exists an infinite set S of unit vectors in `2 such that
kx − yk > 2 whenever x, y are distinct elements of S. Can the constant
√
2 be improved?
11. Show that the norm on `p is not induced from an inner product unless
p = 2.
12. Construct an inner produce space X and a proper closed subspace
Y ( X for which Y ⊥ = {0}.
1
13 ++ . By considering the function f (x) = 1+25x 2 , or otherwise, demon-

strate the Runge phenomenon: there is a continuous function f : [0, 1] →

R whose minimal degree polynomial interpolants at {0, 1/n, 2/n, . . . , (n −
1)/n, 1} do not tend uniformly to f as n → ∞.
Linear Analysis Example Sheet 4 BJG, Michaelmas 2010

1. Let p : C → C is a polynomial. Suppose that T : X → X is a bounded

linear operator on a Hilbert space. Show that the spectrum of p(T ) is
{p(λ) : λ ∈ σ(T )}. Extend this result to rational functions (quotients of two
polynomials). How are the spectra of T and T ∗ related?
2. Suppose that U : X → X is a unitary map on a Hilbert space X: that
is to say, hU x, U yi = hx, yi for all x, y ∈ X. Show that σ(U ) is contained in
the unit circle of the complex plane. By considering i(U + I)(U − I)−1 , or
otherwise, show that σ(U ) 6= ∅.
3. Let T : X → X be a bounded linear operator on a Hilbert space.
Show that if one of the maps T, T ∗ , T T ∗ , T ∗ T is compact then they all are.
4. Let S, T : X → X be bounded linear operators on a Hilbert space.
Show that if one of them is compact then ST is compact.
5. Let T : X → X be a compact self-adjoint operator on a Hilbert space.
Show that there is a unique compact self-adjoint operator T + : X → X
such that kT + xk = kT xk for all x and T + is positive in the sense that
hT + x, xi > 0 for all x ∈ X.
6. Let (en )n∈Z be a complete orthonormal system in a Hilbert space X,
and consider the bounded linear map T : X → X which maps en to en+1 .
Describe σ(T ).
7. A bounded linear operator T : X → X on a Hilbert space is said to
be normal if T commutes with T ∗ . Show that if T is normal then every
element of σ(T ) is an approximate eigenvalue.
8. Is every compact operator on Hilbert space Hilbert-Schmidt?
9. Suppose that Tn are bounded operators on a Hilbert space X such
that Tn x → 0 for all x ∈ X. Is it true that Tn∗ x → 0?
10. Is there a bounded self-adjoint operator on a Hilbert space with no
eigenvalues?
11. Show that any compact set K ⊆ C is the spectrum of some bounded
operator on `2 .
12. Construct a linear operator T on Hilbert space with empty spectrum.