Summary in Preparation for Test in Topology

Jacob Shapiro
September 4, 2013

Abstract
A summary (largely of Munkres’ topology book, and hopefully more
examples and exercises) in preparation for a test with Prof. Damien
Calaque at ETH in the summer of 2013.

Contents
0.1 Open Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
0.2 Misc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

I Topological Spaces 4
1 Chapter 2 Topological Spaces and Continuous Functions 4
1.1 §12 Topological Spaces . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 §13 Basis for a Topology . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 §14 The Order Topology . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 §17 Closed Sets and Limit Points . . . . . . . . . . . . . . . . . . 9
1.5 §18 Continuous Functions . . . . . . . . . . . . . . . . . . . . . . 16
1.6 §16 The Subspace Topology . . . . . . . . . . . . . . . . . . . . . 22
1.7 §19 The Product Topology . . . . . . . . . . . . . . . . . . . . . . 25
1.8 §15 The Product Topology on X × Y . . . . . . . . . . . . . . . . 27
1.9 §22 The Quotient Topology . . . . . . . . . . . . . . . . . . . . . 28
1.10 §20, §21 The Metric Topology . . . . . . . . . . . . . . . . . . . . 35

2 Chapter 3 Connectedness and Compactness 41

2.1 §23 Connected Spaces . . . . . . . . . . . . . . . . . . . . . . . . 41
2.2 §24 Connected Subspaces of the Real Line (and path connectedness) 43
2.3 §25 Components and Local Connectedness . . . . . . . . . . . . . 45
2.4 §26 Compact Spaces . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.5 §27 Compact Subspaces of the Real Line . . . . . . . . . . . . . . 51
2.6 §28 Limit Point Compactness . . . . . . . . . . . . . . . . . . . . 53
2.7 §29 Local Compactness (note: not in material for test–utility for
measure theory) . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

1
3 Chapter 4 Countability and Separation Axioms (not part of the
curriculum for the test–for competeness reasons) 57
3.1 §31 The Separation Axioms . . . . . . . . . . . . . . . . . . . . . 57
3.2 §35 Imbeddings of Manifolds . . . . . . . . . . . . . . . . . . . . 57

4 Chapter 7 Complete Metric Spaces and Function Spaces 58

4.1 §43 Complete Metric Spaces . . . . . . . . . . . . . . . . . . . . . 58
4.2 §44 Theorem 44.1–The Peano Space-Filling Curve . . . . . . . . 61
4.3 §45 Compactness in Metric Spaces . . . . . . . . . . . . . . . . . 61
4.4 §46 Pointwise and Compact Convergence . . . . . . . . . . . . . . 63

II Algebraic Topology 63
5 Chapter 9 The Fundamental Group 64
5.1 §51 Homotopy of Paths . . . . . . . . . . . . . . . . . . . . . . . 64
5.2 §52 The Fundamental Group . . . . . . . . . . . . . . . . . . . . 67
5.3 §53 Covering Spaces . . . . . . . . . . . . . . . . . . . . . . . . . 69
5.4 §54 The Fundamental Group of the Circle . . . . . . . . . . . . . 71
5.5 §55 Retractions and Fixed Points . . . . . . . . . . . . . . . . . . 73
5.6 §58 Deformation Retracts and Homotopy Type . . . . . . . . . . 75
5.7 §59 The Fundamental Group of S n . . . . . . . . . . . . . . . . . 76
5.8 §60 Fundamental Groups of Some Surfaces . . . . . . . . . . . . . 77

6 Chapter 11–The Seifert-van Kampen Theorem 79

6.1 §67 Direct Sums of Abelian Groups . . . . . . . . . . . . . . . . . 79
6.2 §68 Free Products of Groups . . . . . . . . . . . . . . . . . . . . 79
6.3 §69 Free Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
6.4 §70 The Seifert-van Kampen Theorem . . . . . . . . . . . . . . . 81
6.5 §71 The Fundamental Group of a Wedge of Circles . . . . . . . . 81

7 Chapter 13–Classification of Surfaces 82

7.1 §79 Equivalence of Covering Spaces . . . . . . . . . . . . . . . . . 82
7.2 §80 The Universal Covering Space . . . . . . . . . . . . . . . . . 82
7.3 §81 Covering Transformations (Deck Transformations) . . . . . . 83

0.1 Open Questions

1. HWS04 “Open and closed maps”, question 2.
2. HWS05 “ultrametrics” question 3.
3. Lemma 23.1 proof.
4. Example 7 in chapter 23.

5. HWS06 “application to analysis” (why is A ∈ Open (Ω)?)

2
 6. HWS09E01, the end of the argument.
7. HWS10E01 – walk through again.
8. HWS11E05 – walk through again.

0.2 Misc
1. Matrix multiplication on M atn (R) is continuous.
2. Matrix inversion on GL (n, R) is continuous.

3. A topological space which is also a group, with composition and inverse

functions being continuous, is called a topological group.
4. Torus
   0    0  
2 x x p 2 x x
(a) T1 := R / ∼ where ∼ 0 ⇐⇒ ∃ ∈ Z : 0 = +
  y y q y y
p
2π .
q
 

[0, 2π)×[0, 2π) θ
(b) Let r, R ∈ R : R > 2r. Define p ∈ R3 by p :=
  ϕ
[R + r cos (ϕ)] cos (θ)
 [R + r cos (ϕ)] sin (θ) .
r sin (ϕ)
Define T2 := p ([0, 2π) × [0, 2π)).
(c) Observe that T1 ' T2 ' S 1 × S 1 . This is the torus. It is connected
and compact. It is not homeomorphic to [0, 2π) × [0, 2π).
   0    0 
x x x
(d) T1 is a topologiacl group, where m , := .
y ∼ y0 ∼ y0 ∼
      
x −x 0
i := , and eT1 := .
y ∼ −y ∼ 0 ∼


5. If f ∈ C S 1 , R then ∃a ∈ S 1 : f (a) = f (−a).
6. Dini’s theore: theorem 7.14 in Rudin’s PMA.
7. The least upper bound property:

• Any non-empty set of that has an upper bound necessarily has a least
upper bound.

8. (HWS09E01) If X is an ordered set (equipped with the order topology) in

which every closed interval is compact, then X has the least upper bound
property.

3
Part I
Topological Spaces
1 Chapter 2 Topological Spaces and Continuous
Functions
1.1 §12 Topological Spaces
1.1.1 Definition of a Topology
A topology on a set X, is a set, T ⊆ P (X), such that:
1. ∅, X ∈ T
some set
(finite, countable, or uncountable) and if Uα ∈ T ∀ α ∈ A,
2. If A is S
then α∈A Uα ∈ T .
Tn
3. If n ∈ N and Ui ∈ T ∀ i ∈ {1, 2, . . . , n} then ( i=1 Ui ) ∈ T .
If U ∈ T then U is called an open set. Thus T is also denoted as Open (X).
If (X\U ) ∈ T then U is called closed set. All the closed sets are denoted by
Closed (X).

• The discrete topology is defined as P (X).

• The indiscrete / trivial topology is defined as {X, ∅}.
• The cofinite topology is defined as Tf := {∅} ∪ {A ⊆ X | |X \ A| < ∞}.

– Apply de Morgan’s laws for the arbitrary union.

• The cocountable topology is defined as Tc := {∅}∪{A ⊆ X | |X \ A| = |N|}.

1.1.2 Example 1 for a Topology

X := {a, b, c}, Open (X) := {X, ∅, {a, b} , {b} , {b, c}}.
• But S1 := {X, ∅, {a} , {b}} (no union)

• and S2 := {X, ∅, {a, b} , {b, c}} (no intersection)

are not topologies on X!

1.1.3 Example for a Topology

• All subsets of X which are supersets of a given set, plus the empty set.

4
1.1.4 Counter Example for a Topology
• All subsets of X which intersect a given set, plus the empty set. Then de-
pending on the nature of the given set, this may or may not be a topology
on X.
• X = R, S ∈ 2R and O := {(−∞, b) : b ∈ S ∪ {∞}}. Then this is a
topology only if for any collection of open sets, the supremum of b of each
of those sets is in S.

1.1.5 Coarser vs. Finer

If T 0 ⊃ T then T 0 is finer than T , or equivalently, T is coarser than T 0 . If one
set is finer or coarser than another, the two are comparable.
• For any set, Tdiscrete ⊃ Tcocountable ⊃ Tcof inite ⊃ Ttrivial .

1.1.5.1 Counterexample If X := {a, b, c}, the topology {X, ∅, {a, b} , {b, c} , {b}}
is not comparable to the topology {X, ∅, {a, b}, {a}}, because one doesn’t con-
tain the other and vice versa.

1.1.5.2 Counterexample The lower-limit topology is not comparable with

the K-topology.

1.2 §13 Basis for a Topology

1.2.1 Definition of a Basis for a Topology
If X is a set, a “basis for a topology on X” is a set B ⊂ P (X), whose elements
are called basis elements, such that:
1. ∀ x ∈ X ∃ B ∈ B : x ∈ B.
2. IfB1 , B2 ∈ B and x ∈ (B1 ∩ B2 ) then ∃ B3 ∈ B such that x ∈ B3 ⊂
(B1 ∩ B2 ).

1.2.2 The Topology Generated by a Basis

If B is a basis for a topology, the topology that it generates is defined as TB =
{U ∈ P (X) : ∀ x0 ∈ U ∃ B0 ∈ B : x0 ∈ B0 ⊂ U }.
Thus each basis element is itself an open set. V
The closed sets in a topology generated by B are ClosedS
(X) = {C ∈ P (X) : ∀x0 ∈ / C ∃ B0 ∈ B : x0 ∈ B0 B
Furthermore, it can be shown (lemma 13.1) that TB = α∈A αB : B α ∈ B ∀ α ∈ A .
But this representation is not unique (we can find many different ways to write
unions of basis elements for a given open set).

1.2.2.1 Example The set of all circles in R2 is a basis for the standard
topology on R2 .
The set of all rectangles in R2 is a basis for the standard topology R2 .

5
 S
1.2.2.2 Example If X is any set, x∈X {{x}} is a basis for the discrete
topology.

1.2.3 Example
{(a, b) : a, b ∈ Q} is a countable basis for the standard topology on R.

1.2.4 Example
{[a, b) : a, b ∈ Q} is a basis for a topology generating a topology different than
the lower limit topology on R.

1.2.5 Example
If X = R and B := {(−∞, b) : b ∈ R} ∪ {(a, ∞) : a ∈ R} then B is not a basis
for a topology because (a, ∞)∩(−∞, a + 2) does not contain any basis element.

1.2.6 Example
If S 6= ∅ is some set, define X := {{bn } : bn ∈ S∀n ∈ N}. If ω is some finite
sequence, then let Bω be the set of all sequences that begin with ω. Then
B := {Bω : ω is a finite sequence} is a basis for a topology on X.

1.2.7 Lemma 13.2 Going from Topology to Basis

Let X be a set and let T be a topology on it.
Let C be a subset of T such that ∀U ∈ T , ∀x ∈ U ∃C ∈ C : x ∈ C ⊂ U .
Then C is a basis for T .

1.2.8 Lemma 13.3–Comparing Bases

Let B and B 0 be bases for topologies T and T 0 respectively. Then T 0 is finer
than T if and only if ∀ x ∈ X, ∀ (B ∈ B : x ∈ B) ∃ B 0 ∈ B 0 : x ∈ B 0 ⊂ B.

1.2.9 Some topologies on R

1. The Standard Topology is given by the basis {(a, b) | a, b ∈ R ∧ a < b}.

2. The Lower Limit Topology is given by the basis {[a, b) | a, b ∈ R ∧ a < b}.
Rl denotes the induced topological space.

 is given by the basis {(a, b) | a, b ∈ R ∧ a < b}∪{(a, b) \K | a, b ∈ R ∧ a < b}

3. The K-Topology
where K := n1 | n ∈ N . RK denotes the induced topological space.

1.2.9.1 Lemma 13.4 Open (Rl ) ) Open (R) and Open (RK ) ) Open (R)
but Open (Rl ) \Open (RK ) 6= ∅ and Open (RK ) \Open (Rl ) 6= ∅.

6
1.2.10 Definition of a Subbasis for a Topology
S

A subbasis for a topology is a set S ⊂ P (X)Tsuch that S∈S S = X.
n
The basis generated by S is given by BS = { i=1 Si : Si ∈ S, n ∈ N}.
Note: The topology generated by BS is the coarsest topology to contain S.
Note: The topology generated by some subbasis may not have as a basis that
given subbasis.

1.3 §14 The Order Topology

Let X be a set with a total order <. That is, if a, b, c ∈ X then:
1. Antisymmetry: ((a < b) ∧ (b < a)) ⇒ a = b.
2. Transitivity: ((a < b) ∧ (b < c)) ⇒ a < c.
3. Totality: ((a < b) ∨ (b < a)) (this implies reflexivity).

1.3.1 Definition of the Order Topology

Let X be a set such that |X| > 1.
The order topology is given by the basis
{(a, b) | a, b ∈ X ∧ a < b}∪{[a0 , b) | b ∈ X, a0 < x ∀x ∈ X}∪{(a, b0 ] | a ∈ X, x < b0 ∀x ∈ X}
Note that if X has no smallest (largest) element then there are no elements in
the second (third) set, which is still perfectly fine.

1.3.2 Example 1
The standard topology on R is the same as the order topology on R with the
relation ≤ on numbers.

1.3.3 Example 2—The Dictionary Order on R2 (lexicographical or-

der)
The dictionary order on R2 , 2, is defined as:
   0
x x
2 0 ⇐⇒ (x < x0 ) ∨ ((x = x0 ) ∧ (y < y 0 ))
y y
Under 2, R2 has no smallest
 or largest
     element,
  sothe order topology
 on R2
a c a c a c
is given by the basis , | 2 ∧ , ∈ R2 .
b d b d b d

1.3.4 Example 3—The Discrete Topology on N

N has a smallest element under the order ≤, namely, the number 1. Thus the
order topology on N with ≤ is the discrete topology. {1} = [1, 2), which is a
basis element, so {1} is open. If n ∈ N\ {1} then {n} = (n − 1, n + 1) which is
also a basis element.

7
1.3.5 Example 4—The Dictionary Order on {1, 2} × N
 
2
The order topology in the dictionary order of this set is not discrete:
1
 
2
is not open. Proof: Assume otherwise, thus ∃B2,1 ∈ B : ∈ B2,1 ⊂
  1
2
, where B is the set of basis elements in the order topology, that is, B =
1              
i i 1 2 1 i
, | i ∈ {1, 2} ∧ n < m ∪ , ∪ [ , ) | (i = 1 ∧ m > 1) ∨ i = 2 .
n m n m 1  m
2
So it is clear that the only basis element that contains contains also
1
   
1 2
elements of the form which is a contradiction to B2,1 ⊂ . Note
n 1
 
1
that any other one-point set is open, in particular, is open because
      1
1 1 1
=[ , ) is a basis element.
1 1 2

1.3.6 Rays
(HWS02E01)
If X is an totally ordered set, and a is any element in X:
The following sets are open in the order topology, and are called open rays:
(a, ∞) , (−∞, a).
Proof: If X has a largest element, called b0 then (a, ∞) = (a, b0 ] which is a
basis element. S
If X has no largest element, then (a, ∞) = x∈X, x>a (a, x), and (a, x) is a
basis element. Similarly for (−∞, a).
The following sets are closed in the order topology, and are called closed
rays: [a, ∞), (−∞, a].
As it turns out, C := {(a, ∞) | a ∈ X} ∪ {(−∞, a) | a ∈ X} is a subbasis for
a topology on X.
Claim: C generates the order topology. Tn
Proof: The basis for the topology generated by C is given by BC = { i=1 Ci : Ci ∈ C, n ∈ N}.
Thus using Lemma 13.3 we can compare BC with the basis that generates the
order topology:
⊂ BC generates a topology finer than the order topology.
Take any x ∈ X and any basis element B in the order topology containing
x.
Case 1: B is of the form (a, b).
Then (a, b) = (a, ∞) ∩ (−∞, b) is a basis element in BC .
Case 2: B is of the form [a0 , b), then B is itself an open ray, namely, (−∞, b).
Case 3: B is of the form (a, b0 ]. Similar to case 2.
⊃ The order topology is finer than the topology generated by BC .

8
 Take any x ∈ X and any basis element B ∈ BC containing x. Since B is
composed of finite intersections of open sets (we proved the Ci ’s–the open rays,
are open in the order topology), it is itself open in the order topology. Thus
by the definition of a basis for a topology ∃B 0 in the basis generating the order
topology such that x ∈ B 0 ⊂ B.
QED

1.4 §17 Closed Sets and Limit Points

1.4.1 Definition of Closed Sets
A set A ∈ P (X) is said to be closed if X\A ∈ Open (X).

1.4.2 Example—Closed Intervals in R

[a, b] ∈ Closed (R) because R\ [a, b] = (−∞, a) ∪ (b, ∞) ∈ Open (R).

1.4.3 Example—Closed Rays in R

[a, ∞) ∈ Closed (R) because R\[a, ∞) = (−∞, a) ∈ Open (R).

1.4.4 Example—Neither closed nor open intervals in R

[a, b) ∈
/ Open (R) ∧ [a, b) ∈
/ Closed (R).

1.4.5 Example—Closed in R2
  
x
| x ≥ 0 ∧ y ≥ 0 = [0, ∞)×[0, ∞) is closed in R2 , because R2 \ ([0, ∞) × [0, ∞)) =
y
(R × (−∞, 0)) ∪ ((−∞, 0) × R) is open.

1.4.6 Example—The Co-Finite Topology

In the co-finite topology on a set X, Closed (X) = {X}∪{A ∈ P (X) | |A| < ∞}.

1.4.7 Example—The Discrete Topology

In the discrete topology on a set X, Open (X) = P (X). But ∀A ∈ P (X),
(X\A) ∈ P (X). Thus (X\A) ∈ Open (X). Thus A ∈ Closed (X). So
Closed (X) = P (X).

1.4.8 Example—Clopen sets in the Subspace Topology

Consider Y := [0, 1] ∪ (2, 3) endowed with the subspace topology from R
(open sets are intersections of open sets in the ambient space with the space).
Thus [0, 1] ∈ Open (Y ) because (−0.0001, 1.0001) ∈ Open (R) and [0, 1] =
(−0.0001, 1.0001) ∩ Y . (2, 3) ∈ Open (Y ) as well because (2, 3) ∈ Open (R) and
(2, 3) = (2, 3) ∩ Y .

9
 But [0, 1] = Y \ (2, 3), so [0, 1] ∈ Closed (Y ).
Similarly, (2, 3) = Y \ [0, 1], so (2, 3) ∈ Closed (Y ).

1.4.9 Theorem 17.1—Parallel Universe of Structures of Closed Sets

(HWS01E06)
Let X be a topological space. Then the following conditions hold:
1. ∅, X ∈ Closed (X).
2. If A is some
countable, or uncountable) and if Fα ∈ Closed (X) ∀ α ∈
T set (finite,
A, then α∈A F α ∈ Closed (X).
Sn
3. If n ∈ N and Fi ∈ Closed (X) ∀ i ∈ {1, 2, . . . , n} then ( i=1 Fi ) ∈
Closed (X).
Proof using de Morgan’s laws.
Note: Kuratowski’s Axioms (HWS02E02)

1.4.10 Theorem 17.2—Criterion for Closed Sets in Subspace Topol-

ogy
Let Y be a subspace of X. Then a set A ∈ Closed (Y ) iff ∃F ∈ Closed (X) such
that A = F ∩ Y .

1.4.11 Theorem 17.3—(another) Criterion for Closed Sets in Sub-

space Topology
Let Y be a subspace of X. If A ∈ Closed (Y ) and Y ∈ Closed (X) then
A ∈ Closed (X).

1.4.12 Interior and Closure

The interior of a set A ⊂ X is defined as
[
int (A) := V
V ⊂A, V ∈Open(X)

The closure of a set A ⊂ X is defined as

\
closure (A) := C
C⊃A, C∈Closed(X)

Some properties of interior and closure:

• closure (A) ∈ Closed (X)
• int (A) ∈ Open (X)
• int (A) ⊆ A ⊆ closure (A)
• A ∈ Open (X) ⇒ A = int (A)
• A ∈ Closed (X) ⇒ A = closure (A)

10
1.4.13 Boundary
(HWS02E02)
The boundary of A is defined as ∂A := closureX (A) ∪ closureX (X\A).
Another possibility: ∂A = closureX (A) \intX (A)..
Note:
• A ∈ Open (X) ⇐⇒ A = closureX (A) \∂A.

• A ∈ Open (X) ∩ Closed (X) ⇐⇒ ∂A = ∅.

• X\ (intX (A)) = closureX (X\A)
• ∂∂A 6= ∂A. For example, with A = Q =⇒ ∂A = R =⇒ ∂∂A = ∂R = ∅.
• closureX (intX (closureX (intX (A)))) = closureX (intX (A)).

• intX (closureX (intX (closureX (A)))) = intX (closureX (A)).

• ∂ (closureX (A)) 6= ∂A. For example, A = Q, ∂A = R, ∂ (closureX (A)) =
∂R = ∅.

1.4.14 Theorem 17.4—Closure in Subspace or Ambient Space

Let Y be a subspace of X. Let A ∈ P (Y ). Then:

closureY (A) = closureX (A) ∩ Y

1.4.15 [I] Theorem 17.5—Another Characterization of Closure

Let X be a topological space and A ∈ P (X).
1. closure (A) = {x ∈ X | Ux ∩ A 6= ∅ ∀ Ux ∈ Open (X) s.t. x ∈ Ux }.

2. If a basis for a topology on X, B, generates Open (X), then closure (A) =

{x ∈ X | Bx ∩ A 6= ∅ ∀ Bx ∈ B s.t. x ∈ Bx }.
Proof of 1. by contradiction in both directions of containment.

1.4.16 Neighborhoods
If X is a topological space and x ∈ X, a neighborhood of x, U , is a subset of
X such that ∃ V ∈ Open (X) : x ∈ V, V ⊂ U . This is equivalent to saying that
x ∈ int (V ).
Some authors (Munkres for instance) define a neighborhood already to be
open, others only require it to contain an open set containing x.
From this point onwards in this text we shall follow Munkres’ convention for
neighborhoods.
In this terminology, x ∈ closure (A) iff every neighborhood of x intersects
A.

11
1.4.17 Example—Closure of Subsets in R
• closureR ((0, 1]) = [0, 1] because every neighborhood of 0 intersects (0, 1],
and every point outside of [0, 1] has a neighborhood that does not intersect
(0, 1].
• If B = n1 | n ∈ N then closureR (B) = B ∪ {0}.


• If C = {0} ∪ (1, 2) then closureR (C) = [1, 2] ∪ {0}.

• closureR (Q) = R
• closureR (N) = N
• closureR ({x ∈ R | x > 0}) = {x ∈ R | x ≥ 0}
• If Y := (0, 1] ⊂ R then A := 0, 12 ∈ P (Y ). closureR (A) = 0, 12 . Thus

 

by theorem 17.4, closureY (A) = 0, 12 ∩ Y = (0, 12 ].

1.4.18 Limit Points Definition

If X is a topological space and and A ∈ P (X) and if x ∈ X, we say that x
is a limit point, or a cluster point, or a point of accumulation of A if every
neighborhood of x intersects A in some point other than x itself (that is, if
∀U ∈ Open (X) s.t. x ∈ U, A ∩ (U \ {x}) 6= ∅). Said differently, x is a limit
point of A if x ∈ closureX (A\ {x}).

1.4.19 Example—Limit Points of Subsets in R

• The set of limit points of (0, 1] in R is [0, 1].

• B = n1 | n ∈ N has 0 as its single limit point.



• C = {0} ∪ (1, 2) has [1, 2] as the set of its limit points.

• Q has R as the set of its limit points.
• N has no limit points in R.

• {x ∈ R | x > 0} has {x ∈ R | x ≥ 0} as the set of its limit points.

1.4.20 Theorem 17.6—Closure and the Set of Limit Points of a Set

Let X be a topological space and let A ∈ 2X .
Define A0 := {x ∈ X | x ∈ closureX (A\ {x})} (which is the set of all limit
points of A in X).
Claim: closureX (A) = A ∪ A0
Proof by inclusion of both sides.

12
1.4.21 Corollary 17.7—A ∈ Closed (X) ⇔ A0 ⊂ A
Proof: A ∈ Closed (X) ⇔ A = closureX (A), but by Theorem 17.6, closureX (A) =
A ∪ A0 .
So we get A ∈ Closed (X) ⇔ A = A ∪ A0 ⇔ A0 ⊂ A.

1.4.22 Hausdorff Spaces

1.4.23 Definition of Convergence in Topological Spaces
Let X be a topological space. Let f : N → X be a map. We can think of f (n)
as a sequence.
The sequence of points f (N) ⊂ X converges to the point a ∈ X iff ∀ U ∈
Open (X) such that a ∈ U , ∃ maU ∈ N such that f ({n ∈ N | n ≥ maU }) ⊂ U .
If we consider N with the co-finite topology, then that translates to:
What is the topology on N so that f → a iff f is continuous at ∞?
(23:53) (+zeno) You actually need the one-point compactification of N, call
it N~ which is N with the point oo adjoined, and a set U in N~ is open iff either
oo is not in it, or else oo is in it and N\U is finite. (23:53) (+zeno) THen the
sequence f(1), f(2), ... of points of X converges the point x = f(oo) iff f : N~ ->
X is continuous at oo.
• (HWS03E05) (b) If a sequence does not converge in a given topology, it
will also not converge in a finer topology.

1.4.23.1 Example with the Cocountable Topology (HWS02E06) (c)

A sequence {xn } converges to x in (R, Tcocountable ) ⇐⇒ ∃m ∈ N such that
∀n ≥ m, xn = x.

1.4.24 One-point sets are closed in R or R2

{x0 } is closed in either one of R or R2 becaues (according to 1.4.15) every
point different from x has a neighborhood that does not intersect {x0 }, so
closure ({x0 }) = {x0 }, and so {x0 } would be closed in either R or R2 .
In R or R2 , a sequence cannot convergence to more than point.

1.4.25 Example—One-point sets are not always closed, and conver-

gence is strange
Take X := {a, b, c} with the topology Open (X) := {X, ∅, {a, b} , {b} , {b, c}}.
Claim: {b} ∈
/ Closed (X) Proof: X\ {b} = {a, c} ∈ / Open (X).
Claim: The sequence {xi } where xi := b ∀i ∈ N convergences to the point b
and to the points a and to c. Proof: The open sets that contain a are exactly
X and {a, b}. Since xi = b is in both of these ∀i ∈ N, xi → a. Same goes for c.

13
1.4.26 Definition of Hausdorff Spaces
A topological space X is called Hausdorff iff ∀x1 , x2 ∈ X such that x1 6= x2 ,
∃Ui ∈ Open (X) : xi ∈ Ui ∀i ∈ {1, 2} and U1 ∩ U2 = ∅.

1.4.27 Theorem 17.8—|A| < ∞ ⇒ A ∈ Closed (X) for Hausdorff X

Proof:
Case 1: A = {x0 }. Take any y ∈ X\ {x0 }. Since X is Hausdorff, we know
that there exist two disjoint neighborhoods of x0 and y respectively. But that
means that every point y 6= {x0 } has a neighborhood that does not intersect
{x0 }, so that closureX ({x0 }) = {x0 }.
Case 2: A = {x1 , x2 , . . . , xn } for some n ∈ N. Since A is the finite
union of closed sets (by case 1), by 1.4.9 (parallel universe for closed sets)
A ∈ Closed (X).

1.4.28 Example—non-Hausdorff Space with point-sets being closed

Claim: R with the co-finite topology is not a Hausdorff space.
Proof: Take any x, y ∈ R such that x 6= y. Any neighborhood of x is of the
form R\ {x1 , x2 , x3 , . . . , xn } where n ∈ N and the same for any neighborhood
of y. The intersection of any two such neighborhoods will thus certainly not be
disjoint.
Claim: Every finite set in R with the co-finite topology is closed.
Proof: The complement of every such set, by definition, is open.

1.4.29 T1 axiom
Every finite set is closed.

1.4.30 [I] Theorem 17.9—Characterization of limit points in T1 spaces

Let X be a T1 space, let A ∈ 2X , and let x ∈ X.
Claim: x ∈ A0 ⇔ ∀U ∈ Open (X) s.t. x ∈ U, |U ∩ A| = ∞.
Proof: ⇐ This way works always (we don’t have to assume T1 ).
If every neighborhood of x intersects A in infinitely many points, then it
must intersect A\ {x}. Thus x ∈ A0 .
⇒ Assume that x ∈ A0 , and assume that the theorem does not hold, that is,
∃U0 ∈ Open (X) such that x ∈ U0 and |U0 ∩ A| < ∞. Thus |U0 ∩ (A\ {x})| <
∞ Since |U0 ∩ (A\ {x})| < ∞ and we are in a T1 space, (U0 ∩ (A\ {x})) ∈
Closed (X). So (X\ (U0 ∩ (A\ {x}))) ∈ Open (X).
Observe that x ∈ (X\ (U0 ∩ (A\ {x}))).
Thus (X\ (U0 ∩ (A\ {x}))) ∩ U0 is a neighborhood of x. But observe that
((X\ (U0 ∩ (A\ {x}))) ∩ U0 ) ∩ (A\ {x}) = ∅, which contradicts our assumption
that x ∈ A0 .

14
1.4.31 Theorem 17.10—Sequences in a Hausdorff Space Converge
to at Most One Point
Proof: Let X be a Hausdorff space, with a point x ∈ X, and let f : N → X be
a map.
Furthermore assume that limn→∞ f (n) = x. Take any point y ∈ X\ {x}.
Since X is Hausdorff, ∃ Nx , Ny two neighborhoods of x and y such that
Nx ∩ Ny = ∅. 

But limn→∞ f (n) = x. So in particular, ∃ mxUx ∈ N such that f n ∈ N | n ≥ mxUx ⊂
Ux . 

Thus necessarily f n ∈ N | n ≥ mxUx ∩Uy = ∅ which excludes limn→∞ (n) =
y.

1.4.32 Theorem 17.11—Characterization of Hausdorff Spaces

Claim:

1. If X is a set with a total order <, then the order topology on X is Haus-
dorff.
2. If X and Y are Hausdorff, then X × Y with the product topology is
Hausdorff.

3. If X is a Hausdorff space and A ∈ 2X , then A with the subspace topology

is Hausdorff.
Proof:
1.
Take any x ∈ X and any y ∈ X\ {x}. WLOG x < y.
Case 1: @z ∈ X such that x < z < y.
In this case, if there is a smallest element in X, a0 , take a neighborhood of
x to be Nx := [a0 , y). If there is no smallest element in X, take any element
smaller than x, call it a and define Nx := (a, y).
If there is a largest element in X, b0 , take a neighborhood of y to be Ny :=
(x, b0 ]. Otherwise, take any element larger than y, call it b, and define Ny :=
(x, b). Then Nx ∩ Ny = ∅ are two disjoint neighborhoods of x and y.
Case 2: If such a z ∈ X exists, define the right end point of Nx to be z
instead and the left end point of Ny to be z.

2.
Let r1 ∈ X ×  Y and let r2 ∈ (X × Y ) \ {r1 }.
x
Write ri = i ∀i ∈ {1, 2}. Since r1 6= r2 , we must have either x1 6= x2
yi
or y1 6= y2 (or both). WLOG assume that x1 6= x2 . Since X is Hausdorff,
∃Nx1 , Nx2 disjoint neighborhoods of x1 , x2 respectively in X. Take any neigh-
borhoods of y1 , y2 (since we may have y1 = y2 , we will not assume that these

15
neighborhoods are disjoint): Ny1 , Ny2 . Then Nx1 × Ny1 is a neighborhood of r1
which is disjoint to Nx2 × Ny2 , a neighborhood of r2 .
3.
Take any a1 ∈ A and a2 ∈ A\ {a1 }. Since X is Hausdorff, ∃ N1 , N2 –two
disjoint neighborhoods of a1 , a2 respectively in X. By the definition of the
subspace topology, N1 ∩ A and N2 ∩ A are two disjoint neighborhoods of a1 and
a2 in A respectively. Thus A is Hausdorff.

1.5 §18 Continuous Functions

1.5.1 Definition of a Continous Function
Let X and Y be two topological spaces. A function f : X → Y is called
continuous (relative to Open (X) and Open (Y )) if ∀ V ∈ Open (Y ) , f −1 (V ) ∈
Open (X).

1.5.2 Characterization of Continuity by Basis of a Topology

If Open (Y ) is generated by a basis B, then f : X → Y is continuous iff ∀B ∈
B, f −1 (B) ∈ Open (X).
Proof: The ⇒ direction is trivial.
Next, assume that ∀B ∈ B, f −1 (B) S ∈ Open (X). Now take any V ∈
Open (Y ). By Lemma_TODO V = α∈A Bα for some set A. But then
−1 −1 −1
S
S
f (V ) = f α∈A Bα = α∈A f (Bα ) ∈ Open (X). So that f is con-
tinuous.

1.5.3 Characterization of Continuity by Subbasis of a Topology

If Open (Y ) is generated by a subbasis S, then f : X → Y is continuous iff
∀S ∈ S, f −1 (S) ∈ Open (X).
Proof: The ⇒ direction is trivial.
Next, assume that ∀S ∈ S, f −1S (S) ∈T Open (X). Now take any V ∈
nα
Open (Y ). By Lemma_TODO
V = α∈A i=1 Siα for some set A. But then
−1 −1 nα α nα −1
= α∈A i=1 f (Siα ) ∈ Open (X). So that f
S T S T
f (V ) = f α∈A i=1 Si
is continuous.

1.5.4 Equivalence to Continuity of real functions of a real variable

with the ε − δ defintion
TODO

1.5.5 {x} ∈ Open (X)

(HWS02E04)
If {x} ∈ Open (X) then f ∈ Y X is continuous at x.

16
1.5.6 Example
(HWS02E05)
(
x x∈Q
f (x) := is continuous only at 0.
0 x∈
/Q

1.5.7 Example—Discontinuous Identities in the Lower Limit Topol-

ogy
The function f : R → Rl given by f (x) := x∀x ∈ R is not continuous: by
1.5.2, take a basis element from the basis that generates Rl , {[a, b)}}, [a, b),
and examine its inverse image: f −1 ([a, b)) = [a, b) ∈/ Open (R).
On the other hand, the inverse map, gS: Rl → R given by g (x) := x∀x ∈ R
is continuous for the reason that (a, b) = a0 >a [a0 , b) ∈ Open (Rl ).

1.5.8 Definition of Continuity at a Single Point

Let X, Y be topological spaces and let f : X → Y be a map. Let x0 ∈ X. f is
(defined to be) continuous at x0 iff ∀ neighborhood V of f (x0 ), ∃ a neighborhood
U of x0 such that f (U ) ⊂ V .

1.5.9 Theorem 18.1—Characterization of Continuity

Let X, Y be topological spaces and let f : X → Y be a map. Then th efollowing
are equivalent:
1. f is continuous (according to 1.5.1).
2. ∀A ∈ 2X , f (closureX (A)) ⊂ closureY (f (A)).

3. ∀B ∈ Closed (Y ) , f −1 (B) ∈ Closed (X).

4. f is continuous at x, ∀x ∈ X (according to 1.5.8).
Proof:
1. ⇒ 2.
(compare with Rudin’s PMA Exercise 4.2)
Take some x ∈ closureX (A).
Take a neighborhood V of f (x). By 1.5.1 f −1 (V ) ∈ Open (X) and observe
that x ∈ f −1 (V ). Thus f −1 (V ) is a neighborhood of x. But x ∈ closureX (A),
so by 1.4.16, ∃x0 ∈ Xsuch that x0 ∈ f −1 (V ) ∩ A. Thus f (x0 ) ∈ V ∩ f (A) ⇒
V ∩ f (A) 6= ∅. So we conclude that every neighborhood of f (x) intersects
f (A). So that again by 1.4.16 f (x) ∈ closureY (f (A)).
2. =⇒ 3.
Take some B ∈ Closed (Y ).
Claim: closureX f −1 (B) ⊂ f −1 (B)

Proof: Take some

x ∈ closureX f −1 (B) . By the previous part, f (x) ∈

closureY f f −1 (B) .

17
 But it is a fact of set theory that f f −1 (B) ⊂ B (in fact if f is surjective

there will be an equality).

In addition, by 1.4.12, it is clear that any

closed set that contains a set

is
going to contain its closure, so if f f −1 (B) ⊂ B ⇒ closureY f f −1 (B) ⊂
closureY (B) = B.
So we conclude that f (x) ∈ B ⇒ x ∈ f −1 (B).
Thus f −1 (B) ∈ Closed (X).
3. ⇒ 1.
Take some V ∈ Open (Y ). Thus Y \V ∈ Closed (Y ). Thus f −1 (Y \V ) ∈
Closed (X). But f −1 (Y \V ) = f −1 (Y ) \f −1 (V ) = X\f −1 (V ). Thus f −1 (V ) ∈
Open (X). Thus 1.5.1.
1. ⇒ 4.
Every neighborhood of f (x), V has f −1 (V ) as a neighborhood of x, such
−1


that f f (V ) ⊂ V .
4. ⇒ 1.
Take V ∈ Open (Y ). If f −1 (V ) = ∅ we are done as ∅ ∈ Open (X). Other-
wise, ∃x ∈ f −1 (V ). Then f (x) ∈ V and so V is a neighborhood of f (x). By
4., ∃Ux , a neighbhorhood of x suchS that f (Ux ) ⊂ V . Observe that this leads to
Ux ⊂ f −1 (V ). Thus f −1 (V ) = x∈f −1 (V ) Ux ∈ Open (X).

1.5.10 The Initial Topology

(HWS03E03)
Let Y, I be sets and let (Xi , Ti ) be a topological space, and fi : Y → Xi be
a map ∀i ∈ I.
S topology on Y with respect to fi as the topology generated
Define the initial
by the subbasis i∈I fi−1 (Ti ).

• The initial topology is the coarsest topology such that fi : (Y, T ) →

(Xi , Ti ) is continuous ∀i ∈ I.
• If (Z, V) is a topological space and f : Z → Y is a map, then f is
continuous ⇐⇒ fi ◦ f is continuous ∀i ∈ I.
• If A ⊂ X, then the initial topology on A with respect to the inclusion map
is the subspace topology on A.
• If (X
Qi , Ti )i∈I is a family of topological spaces and
Q fi are the projections
fi : j∈I Xj → Xi , then the initial topology on j∈I Xj with respect to
the projections fi is the product topology.

1.5.11 The Final Topology

(HWS03E04)
Let Y, I be sets and let (Xi , Ti ) be a topological space, and fi : Xi → Y be
a map ∀i ∈ I.
Define the final topology on Y with respect to fi as T := O ⊂ Y : ∀i ∈ I, fi−1 (O) ∈ Ti .


18
 • The final topology is the finest topology on Y such that fi is continuous
∀i ∈ I.
• If (Z, V) is a topological space and f : Y → Z is a map, then f is
continuous ⇐⇒ f ◦ fi is continuous ∀i ∈ I.

• If X is a set Ti are a bunch of topologies on

Tit, and idi : (X, Ti ) → X, the
final topology on X with respect to idi is i∈I Ti .

1.5.12 The Co-Product Topology (Disjoint Union Topology)

(HWS03E04) (c) and (d)   
x
The disjoint union of two topological spaces X1 and X2 is defined as : x ∈ X1 ∪
1
  
x ` `
: x ∈ X2 . If f1 : X1 ,→ X1 X2 and f2 : X2 ,→ X1 X2 are the two in-
2 ` `
clusion maps, then the final topology with respect to f1 and f2 is {U1 U2 ⊂ X1 X2 : U1 ∈ Open (X1 ) ∧ U2 ∈
Similarly, for any collection 
of`topological
` spaces, the final topology with
respect to the inclusion maps is U
i∈I i ⊂ i∈I Xi : Ui ∈ Open (Xi ) ∀i ∈ I .
This is defined as the coproduct topology.
`
• Example: Take {0, 1} with the discrete topology. n∈N {0, 1} with the
coproduct topology is discrete.

1.5.13 Counter Example

(HWS03E03) (d)
N
Take {0, 1} with the discrete topology. Observe that {0, 1} with the prod-
uct topology is not discrete!

1.5.14 Homeomorphism
Let X, Y be topological spaces and let f : X → Y be a bijection. If both f and
f −1 : Y → X are continuous, then f is called a homeomorphism.
Equivalently, a bijection f : X → Y is a homeomorphism if the following is
true: U ∈ Open (X) ⇔ f (U ) ∈ Open (Y ).

1.5.15 Example
(HWS02E05)
(a, b) is homeomorphic to (0, 1)
[a, b] is homeomorphic to [0, 1] if a, b are finite.

1.5.16 Example
(HWS03E02)
N
The cantor set is homeomorphic to {0, 1} .

19
1.5.17 Example
()
The map idX : (X, T ) → (X, T 0 ) is continuous iff T is finer than T 0 , and
idX is a homeomorphism iff T = T 0 .

1.5.18 Topological Imbedding (=Embedding)

Let X, Y be topological spaces and let f : X → Y be an injection. If the
bijection g : X → f (X) defined by g (x) := f (x) ∀x ∈ X happens to be a
homeomorphism, f is called a topological imbedding of X in Y .

1.5.19 Counterexample—Bijective continuous non-homeomorphism

One example is g defined in 1.5.7.
Another example is the following (Compare with Example 4.21 in Rudin’s
PMA):   
x
Let S 1 := ∈ R2 | x2 + y 2 = 1 be a topological space endowed with
y
the subspace topology from R2 .
Let [0, 1) be a topological space endowed
 with
 the order topology.
1 cos (2πt)
Define f : [0, 1) → S by f (t) := ∀t ∈ [0, 1).
sin (2πt)
Observe that f is bijective, and continuous.
However, f −1 is not continuous:
Take the set [0,
14 ) ∈ Open ([0,
1)).
Claim: f [0, 14 ) ∈ / Open S 1 Proof: Every basis element containing the
1
is not contained inside of f [0, 14 ) .


point f (0) =
0
In addition, if we define a new function g : [0, 1) → R2 by g (t) := f (t) ∀t ∈
[0, 1) then g is a counterexample for a continuous injective map that is not a
topological imbedding.

1.5.20 Theorem 18.2—Rules for Constructing Continuous Functions

Let X, Y, and Z be topological spaces.
ConstantFunction If y0 ∈ Y and f : X → Y is defined such that f (X) = {y0 }
then f is continuous.
Proof: Take any V ∈ Open (Y ).
Case 1: y0 ∈ V , then f −1 (V ) = X ∈ Open (X).
/ V , then f −1 (V ) = ∅ ∈ Open (X).
Case 2: y0 ∈
Inclusion If A is a subspace of X, then the inclusion function j : A → X
defined by j (a) := a ∀ a ∈ A is continuous.
Proof: Take any V ∈ Open (X). Then j −1 (V ) = A ∩ V . By definition of the
subspace topology, this set is ∈ Open (A).

20
Composition If f : X → Y and g : Y → Z are two continuous functions then
the map g ◦ f : X → Z is continous. (HWS02E04)
−1
Proof: Observe that f −1 g −1 (V ) = (g ◦ f ) (V ).

DomainRestriction If f : X → Y is continous and if A is a subspace of X

then the restricted function f |A : A → Y is continous.
Proof: (HWS03E01) f |A = f ◦j, where j is the inclusion map. Apply the above.

DomainRestrictionOrRangeExpansion Let f : X → Y be continuous. If

f (X) ⊂ Z ⊂ Y then the function g : X → Z obtained by restricting the
range of f is continuous. If Z ⊃ Y then the function h : X → Z obtained
by expanding the range of f , is continous.
Proof: Take some B ∈ Open (Z). Thus ∃U ∈ Open (Y ) such that B = Z ∩ U .
Since Z ⊃ f (X), observe that g −1 (B) = f −1 (U ) ∈ Open (X).
Next, observe that h = j ◦ f where j : Y → Z is the inclusion map.
LocalFormulationOfContinuity
S f : X → Y is continous if ∃ {Uα ∈ Open (X)}
such that X = α Uα and f |Uα is continuous ∀α.
Proof: Take some V ∈ Open (Y ). S
By assumption, ∃ {Uα ∈ Open (X)} such that X = α Uα and f |Uα is con-
tinuous ∀α.
S 
−1 
Observe that f −1 (V ) = α f −1 (V ) ∩ Uα = α f |Uα
S
(V )

−1
But since f |Uα are continuous, f |Uα (V ) ∈ Open (Uα ). Thus ∃Vα ∈

−1
−1
Open (X) such that f |Uα (V ) = Uα ∩ Vα . Thus f |Uα (V ) ∈ Open (X).
S 
−1 
So that α f |Uα (V ) ∈ Open (X).

1.5.21 Theorem 18.3—The Pasting Lemma

(HWS03E01 (b) the gluing lemma)
Let X = A ∪ B where A, B ∈ Closed (X). Let f : A → Y and g : B → Y be
( If f (x) = g (x) ∀x ∈ A ∩ B then the function h : X → Y defined by
continous.
f (x) x ∈ A
h (x) := is continuous.
g (x) x ∈ B
Proof:
Take any V ∈ Open (Y ). Observe that h−1 (V ) = f −1 (V ) ∪ g −1 (V ) ∈
Open (X).
Note: This theorem would’ve worked with open sets as well, in which case
it would’ve been a special case of the “local formulation of continuity” in 1.5.20.
Note: We needed f (x) = g (x) ∀x ∈ A ∩ B for h to be well-defined.

21
1.5.22 Counterexample—Discontinxuous Because the Domains are
not both closed or both open
The map ( l : R → R, both endowed with the standard topology, defined by
x − 2 x ∈ (−∞, 0)
l (x) := is not continuous. (1, 3) ∈ Open (R) but l−1 ((1, 3)) =
x + 2 x ∈ [0, ∞)
[0, 1) ∈
/ Open (R).

1.5.23 Theorem 18.4—Products of Continuous Functions

Let fx : A → X, fy: A →Y be maps and let there be a map f : A → X × Y
f (a)
defined by f (a) := x ∀a ∈ A.
fy (a)
Then f is continuous ⇔ both fx and fy are continuous.
The maps fx , fy are called coordinate functions of f .
Proof:
⇒
Let
  πx : X × Y →  Xand πy : X × Y → Y be the projection maps:
x x x
πx = x and πy =y ∀ ∈X ×Y.
y y y
Claim: πx : X × Y → X is continuous
Proof: Take U ∈ Open (X). Then πx−1 (U ) = U × Y . Since Y ∈ Open (Y ),
by the definition of the product topology, U × Y ∈ Open (X × Y ).
Simiarly, πy is continuous.
Observe that fx = πx ◦ f and fy = πy ◦ f .
⇐
Following 1.5.2, take a basis element of X × Y : Ux × Uy such that Ux ∈
Open (X) and Uy ∈ Open (Y ). Observe that f −1 (Ux × Uy ) = fx−1 (Ux ) ∩
fy−1 (Uy ) ∈ Open (A).
Note: @ useful criterion for when f : A × B → X is continuous.

1.6 §16 The Subspace Topology

1.6.1 Definition
Let X be a topological space and let Y ∈ 2X . The subspace topology on
Y is defined as follows: Open (Y ) := {Y ∩ U | U ∈ Open (X)}. With such a
topology, Y is now called a subspace of X.

1.6.2 Characterization with Closed Sets

Let X be a topological space and let Y ∈ 2X have the subspace topology.
Claim: If K ∈ Closed (Y ) then K = F ∩ Y where F ∈ Closed (X).
Proof: Let K ∈ Closed (Y ).
⇒ Y \K ∈ Open (Y )
⇒ Y \K = Y ∩ U where U ∈ Open (X).
⇒ K = Y \ (Y ∩ U ) = Y ∩ (Y \U ).

22
 

1.6.3 Lemma 16.1—Subspace Topology by Basis

If B is a basis that generates Open (X) then BY := {B ∩ Y | B ∈ B} is a basis
that generates the subspace topology on Y .
Proof:
Following 1.2.7: Take any U ∈ Open (X). Then U ∩Y is an arbitrary element
in Open (Y ). Take any y ∈ U ∩ Y . Since U ∈ Open (X), there exists By ∈ B
such that y ∈ By ⊂ U . But that means y ∈ By ∩ Y ⊂ U ∩ Y .

1.6.4 Lemma 16.2—Open Subspaces are Cool

Let Y be a subspace of X. If U ∈ Open (Y ) and Y ∈ Open (X) then U ∈
Open (X).
Proof:
U ∈ Open (Y ), so ∃Ux ∈ Open (X) such that U = Ux ∩ Y . So U is the finite
intersection of two open sets. ’nuff said.

1.6.5 Lemma 16.3—The Subspace topology on a Product is the

product topology of two subspaces
If A is a subspace of X and B is a subspace of Y , then the product topology
on A × B where A and B are both endowed with the subpsace topology is the
same as the subspace topology on A × B as a subspace in X × Y .
Proof:
Claim: The basis that generates the subspace topology on A × B is the same
as the basis that generates the product topology on A × B ⊂ X × Y .
Proof: The general basis element for X × Y is U × V where U ∈ Open (X)
and V ∈ Open (Y ). Thus (U × V ) ∩ (A × B) is the general basis element of the
subspace topology of A × B ⊂ X × Y .
Observe that (U × V ) ∩ (A × B) = (U ∩ A) × (V ∩ B).
Observe that U ∩ A is the general open set in the subspace topology of
A ⊂ X and V ∩ B is the general open set in the subpsace topology of B ⊂ Y ,
(U ∩ A)×(V ∩ B) is the general basis element in the product topology on A×B.
Thus since the two basis are the same they generate the same topology.

1.6.6 The Order Topology and the Subspace Topology Do Not Al-
ways Agree!
1.6.6.1 Example of Agreement Take R with the standard topology and
Y := [0, 1] with the subspace topology. Thus a general basis element in the


 (a, b) a, b ∈ Y

[ 0, b ) b ∈ Y, a ∈ /Y
subpsace topology is of the form: (a, b) ∩ Y := .


 ( a, 1 ] a ∈ Y, b ∈/Y
Y or ∅ a ∈ / Y, b ∈
/Y


23
 But this basis generates the order topology, so we have agreement.

1.6.6.2 Example of Disagreement Take R with the standard topology

and Y := [ 0, 1 ) ∪ {2}. In the subpsace topology on Y , {2} is open, because
{2} = Y ∩(1.999, 2.0001). However in the order topology on Y , {2} is not open:
any basis element of the order topology which would contain 2 is of the form
(a, 2] ∩ Y where a ∈ Y , which necessarily cannot be contained in {2}, because
it necessarily contains points below it.

1.6.6.3 Example of Disagreement in the Plane Let I = [0, 1]. The

dictionary order on I × I gives rise to a topology different than the subspace
topology on I × I from R × R with the order topology of the dictionary order.
For example, the set 12 × ( 12 , 1] isopen in I × I in
the subspace
topology
(it is, for example, the intersection of 12 × 12 , 1.111 ∈ Open R2 and 2
 I ,
1/2
which is open in the subspace topology), but not in the order topology, as
1
is not the largest element in I × I with the dictionary order topology, and so
the set cannot be open.
The set I × I in the dictionary order topology is called the ordered square
and is denoted by I02 .

1.6.7 Defintion of Convex Sets

Let X be an ordered set. Y ∈ 2X is called convex in X if for each a, b ∈ Y such
that a < b, {x ∈ X | a < x < b} ⊂ Y .
Note that intervals and rays in X are convex in X.

1.6.8 Lemma 16.4—Convex Subspaces have the same order topology

Let X be an ordered set in the order topology. Let Y ∈ 2X such that Y convex
in X. Then the order topology on Y is the same as the topology Y inherits as
a subspace of X.
Proof:
Claim: The order Topology on Y contains the subspace topology S on Y
S Proof: Take any open element in X, it will look like this: α∈A (aα , ∞) ∪
β∈B (−∞, a β ) where aα ∈ X∀α ∈ A and aβ ∈ X∀β ∈ B.
S Thus any openSelement in thesubspace topology on Y will look like this:
S S
α∈A (aα , ∞) ∪ β∈B (−∞, aβ ) ∩Y = α∈A ((aα , ∞) ∩ Y )∪ β∈B ((−∞, aβ ) ∩ Y ).
But (aα , ∞) ∩ Y and (−∞, aβ ) ∩ Y are both open in the order topology on
Y , because they are either open rays in Y (in case aα or aβ are in Y ) or they
are the empty set of they are the whole of Y , in either one of these cases, they
would be open sets.
Claim: The subspace topology on Y contains the order topology
S on Y
S Proof: Take a general open set in the order topology on Y : α∈A {x ∈ Y | x > aα }∪
β∈B {x ∈ Y | x > aβ } where aα ∈ Y ∀α ∈ A and aβ ∈ Y ∀β ∈ B.

24
 S
S can write this as an open set in the subspace topology on Y :
We α∈A {x ∈ X | x > aα }∩
Y ∪ β∈B {x ∈ X | x > aβ } ∩ Y .

1.7 §19 The Product Topology

Q
Let α∈A Xα be the Cartesian product (where A may be finite, countable, or
uncountable) of the topological spaces Xα .

1.7.1 Definition of the Box Topology

Q
The box topology is generated by the basis Bbox := α∈A Uα | Uα ∈ Open (Xα ) .

1.7.2 Definition of the Product Topology

In analogy with 1.8.5, the collection
[ 
S := πα−1 (Uα ) | Uα ∈ Open (Xα )
α∈A
Q
is a subbasis for the product topology on α∈A Xα .
From this it follows that the basis for the product topology is such that there
are only finitely many open sets in the product which are not the entire space,
and the rest of the product has to be the entire space.
Thus a typical basis element has the form:
Y
πα−1 −1 −1




1
(U α1 ) ∩ π α2
(Uα2 ) ∩ · · · ∩ π αn
(Uα n ) = Uα
α∈A

where Uα = Xα whenever α ∈
/ {αi | i ∈ {1, . . . , n}}.

1.7.3 Definition of tuple notation

Let J be an index set (finite, countable, or uncountable), and let X be any set.
Let x : J → X a map, thus (xα )α∈J is the set of points in X indexed by J. All
J-tuples of elements of X are denoted by X J .

1.7.4 Theorem 19.1—Comparison of the box and product topologies

Observe: for finite sets the box and product topology are identical. Secondly,
the box topology is finer than the product topology in general.

1.7.5 Theorem 19.2—Basis generating product spaces

Q
Let α∈A Xα be the Cartesian product (where A may be finite, countable, or
uncountable) of the topological spaces Xα , each of which is generated by the
basis Bα . Q
The basis  α∈A Bα | Bα ∈ Bα generates the box topology.
Q
The basis α∈A Bα | Bα ∈ Bα for finitely manyα ∈ A, otherwise Bα = Xα
generates the product topology.

25
 Proof:
TODO

1.7.6 Theorem 19.3—Subspace product topology

Q
Let α∈A Xα be the Cartesian product (where A may be finite, countable, or
uncountable)
Q of the topological spacesQ Xα , each of which has a subspace Aα .
Thus α∈A Aα is a subspace of α∈A Xα if both products are given the box
topology, or if both products are given the product topology.
Proof:
TODO

1.7.7 Theorem 19.4—Hausdorff Products

Q
α∈AXα is Hausdorff in both the box and product topologies.
Proof:
TODO

1.7.8 Theorem 19.5—Product of Closures

Let Aα ⊂ Xα for all α ∈ A. Then in both the product and box topologies,
!
Y Y
closureXα (Aα ) = closure α∈A Xα
Q Aα
α∈A α∈A

Proof:
TODO

1.7.9 Theorem 19.6—Continuous functions in the product topology

Q
Let f : A → α∈J Xα be given by Q the equation f (a) := (fα (a))a∈J where
fα : A → Xα for each α ∈ J. Let α∈J Xα have the product topology. Then
the function f is continuous ⇔ each function fα is continuous.
Proof:
⇒
Take some β ∈ J.
πβ is continuous because if Uβ ∈ Open (Xβ ), then πβ−1 (Uβ ) is a subbasis
element in the product topology and thus is open.
fb = πβ ◦ f .
Since both these functions are continuous, their composition is also contin-
uous.
⇐
Take any subbasis element from the subbasis of the product topology:
πβ−1 (Uβ ) where β ∈ J and Uβ ∈ Open (Xβ ).
 
Observe that because fb = πβ ◦ f , f −1 πβ−1 (Uβ ) = fβ−1 (Uβ ) ∈ Open (A)
as fβ is continuous.
So by 1.5.3 f is continuous.

26
1.7.10 Example—Discontinuous in Product Topology
Q
RN = n∈N R  
t
Define f : R → RN by the equation f (t) :=  t .
...
Thus fn (t) := t.
Each fn : R → R is continuous.
By the above theorem, RN with the product topology is continuous then.
Claim: f is discontinuous if RN is given the box topology.

Proof: Take the basis element B := (−1, 1) × − 12 , 12 × − 13 , 13 × . . .

in the box topology basis. Assume f −1 (B) ∈ Open (R). Thus ∃δ > 0 such
that (−δ, δ) ⊂ f −1 (B). Thus f ((−δ, δ)) ⊂ B. Apply πn to both sides of the
inclusion to get:
fn ((−δ, δ)) = (−δ, δ) ⊂ − n1 , n1 for any n ∈ N

Which is clearly a contradiction.

Thus f −1 (B) ∈/ Open (R), so that f is discontinuous.

1.8 §15 The Product Topology on X × Y

1.8.1 Definition of the Product Topology on X × Y
Let X and Y be topological spaces. The product topology on X ×Y is the topol-
ogy generated by the basis Bproduct := {U × V | U ∈ Open (X) ∧ V ∈ Open (Y )}.
Claim: Bproduct is a basis for a topology.
Proof: Following
  1.2.1:
x
Take any ∈ X × Y . Since X × Y ∈ Bproduct (as X and Y are open in
y  
x
their respective topologies), X × Y would be the basis element containing .
y
This satisfies the first condition for a basis.
Next, take any two basis elements from Bproduct : U1 × V1 , and U2 × V2 .
Observe that (U1 × V1 ) ∩ (U2 × V2 ) = (U1 ∩ U2 ) × (V1 × V2 ) is a basis element
as the intersection of two open sets is again open. Thus this element fulfils the
requirement for the second condition.

1.8.2 Theorem 15.1—Going to the basis of the product topology

from the constituent basis
If B is the basis for the topology of X and C is the basis for the topology of Y
then the collection D := {B × C | B ∈ B ∧ C ∈ C} is a basis for the topology
of X × Y .
Proof:
D ⊂ Open (X × Y ) because any basis S element in the constituent spaces is
open. Thus take any open set in X × Y , α∈A  Uα × Vα (that is, Uα ∈ Open (X)
x S
and Vα ∈ Open (Y )). Next take any point ∈ α∈A Uα × Vα (that is x ∈ Uβ
y

27
and y ∈ Vβ for some β ∈ A). Since B is the basis for Open (X) ∃B ∈ B such
 ∈ B ⊂ Uβ . Similarly, ∃C ∈ C such that y ∈ C ⊂ Vβ . Thus ∃B × C such
that x
x S
that ∈ B × C ⊂ Uβ × Vβ ⊂ α∈A Uα × Vα . But B × C ∈ D, so according
y
to 1.2.7 we are done.


1.8.3 Example 1—The basis for Open R2


Thanks to 1.8.2, we can ascertain that Open R2 is generated by sets of open
rectangles, rather than products of arbitrary open sets.

1.8.4 Projections
   
x x
Let π1 : X × Y → X be defined by π1 := x ∀ ∈X ×Y.
y  y  
x x
Let π2 : X × Y → X be defined by π2 := y ∀ ∈X ×Y.
y y
The maps π1 and π2 are called projections of X × Y onto its first and second
factors, respectively.
These maps are surjective.

1.8.5 Theorem 15.2—Subbasis for the Product Topology

The collection
2
[  −1
S := πi (U ) | U ∈ Open (Xi )
i=1

is a subbasis for the product topology on X1 × X2 .

Proof:
Let T denote the product topology on X × Y . Let T 0 denote the topology
generated by S.
T0⊂T
Note that every element of S belongs to T , and so every arbitrary unions of
finite intersections of elements of S (which represent arbitrary members of T 0 ).
T0⊃T
On the other hand, every basis element U × V for the topology T is a finite
intersection of elements of S: U ×V = π1−1 (U )∩π2−1 (V ). Therefore, U ×V ∈ T 0 .
But an arbitrary element of T is an arbitrary union of basis elements, whichT 0
is closed under, being a topology. Thus T ⊂ T 0 .

1.9 §22 The Quotient Topology

1.9.1 Definition of a Quotient Map
Let X and Y be topological spaces. Let p : X → Y be a surjective map. A
quotient map is a map p such that U ∈ Open (Y ) ⇔ p−1 (U ) ∈ Open (X).
Note: This condition is stronger than continuity.

28
 Equivalently we could define that U ∈ Closed (Y ) ⇔ p−1 (U ) ∈ Closed (X),
because f −1 (Y \B) = X\f −1 (B).

1.9.2 Definition of a Saturated Subset

A subset C ∈ 2X is saturated (with respect to the surjective map p : X → Y )
iff C ⊃ p−1 ({y}) ∀y ∈ Y such that C ∩ p−1 ({y}) 6= ∅.

1.9.3 Another Characterization of Saturated Set

Claim: C is saturated ⇔ C = p−1 (B) for some B ∈ 2Y .
Proof:
⇐
Let C = p−1 (B) for some B ∈ 2Y .
Take some y ∈ Y such that C ∩ p−1 ({y}) 6= ∅ ⇔ p−1 (B) ∩ p−1 ({y}) 6= ∅.
Thus y ∈ B.
Take some x ∈ p−1 ({y}). Thus p (x) = y. Thus p (x) ∈ B. Thus x ∈
−1
p (B) = C. So C is saturated.
⇒
Now assume that C is saturated.
Define B := y ∈ Y | p−1 ({y}) ∩ C 6= ∅ .
Claim: C = p−1 (B).
⊆
Take any x ∈ C.
Then p (x) ∈ B, because p−1 ({p (x)}) ∩ C ⊃ {x} 6= ∅.
Thus x ∈ p−1 (B).
⊇
Take any x ∈ p−1 (B).
Thus p (x) ∈ B. Thus p−1 ({p (x)}) ∩ C 6= ∅.
But {x} ⊂ p−1 ({p (x)}) ∩ C. So x ∈ C.

1.9.4 Another Characterization of Saturated Set

Claim: C is saturated ⇔ C = p−1 (p (C)).
This follows

from 1.9.3 together with the fact that for any set, p−1 (B) =
−1
p p p−1 (B) .

1.9.5 Quotient Maps of Saturated Sets

To say that p is a quotient map is equivalent to saying that p is surjective,
continuous and p maps saturated open sets of X to open sets of Y (or saturated
closed sets of X to closed sets of Y ).
Proof:
⇒
Assume that p is a quotient map. Thus p is continuous by definition.

29
 Take any saturated set C ∈ Open (X). We need to show that p (C) ∈
Open (Y ).
But C is saturated, so that C = p−1 (B) for some B ∈ Y . But since
C ∈ Open (X), p−1 (B)
∈ Open (X). But p is a quotient map, so B ∈ Open (Y ).
But p (C) = p p−1 (B) . Observe that p p−1 (B) = B since p is surjective.

Thus p (C) ∈ Open (Y ).

⇐
Assume that p is continuous and p maps saturated open sets of X to open
sets of Y .
Take any U ∈ Open (Y ). Since p is continuous, p−1 (U ) ∈ Open (X).
Next, let U ∈ 2Y such that p−1 (U ) ∈ Open (X).
But p−1 (U ) is exactly an open saturated set.
Thus p p−1 (U ) = U (because p is surjective) and so by our assumption

U ∈ Open (Y ).

1.9.6 Open Maps and Closed Maps

1.9.6.1 Closed maps and Open maps A map is open if ∀U ∈ Open (X) , f (U ) ∈
Open (Y ).
A map is closed if ∀U ∈ Closed (X) , f (U ) ∈ Closed (Y ).
• (HWS04)

– The product of two closed maps is not necessarily closed: for example,
R
let x0 ∈ R, f ∈ {x0 } then f × idR : R × R → R × {x0 } is not closed.

1.9.6.2 Connection to Quotient Maps If p : X → Y is a surjective

continuous map that is either open or closed then p is a quotient map.
• However, there are quotient maps that are neither open nor closed:

– (17:25) (+zeno) The identification map R -> R/Q (where all points
of Q are identified together) is a quotient map that sends the open
set (0,1) to a nonempty proper subset of R/Q, and also sends the
closed set {0} to a nonempty proper subset of R/Q. But R/Q has no
open nonempty proper subsets, as you can check.
  
x 2
– (HWS04) (Munkres 22.3) Define A := ∈R : x≥0∨y =0 .
y
Then define q : A → R by q := π1 |A . Observe that q is a quotient
map that is neither open nor closed.

1.9.7 Example
Let X := [0,(1] ∪ [2, 3] ⊂ R. Define Y := [0, 2] ⊂ R. Define p : X → Y as:
x x ∈ [0, 1]
p (x) :=
x − 1 x ∈ [2, 3]
Observe that p is surjective, continuous (pasting lemma).

30
 • Claim: p is closed.

– Proof 1:
p is continuous, so that it takes compact sets to compact sets.
But any closed subset of X is going to be compact, so that p (K) will
also be compact and so closed in Y .
– Proof 2:
Take any U ∈ Closed (X).
Thus by 1.6.2 U = X ∩ K where K ∈ Closed (R).
⇒ U = ([0, 1] ∪ [2, 3]) ∩ K = ([0, 1] ∩ K) ∪ ([2, 3] ∩ K).
Since ([0, 1] ∩ K) , ([2, 3] ∩ K) are two disjoint sets,
p (U ) = p ([0, 1] ∩ K) ∪ p ([2, 3] ∩ K)
By definition, p ([0, 1] ∩ K) = [0, 1]∩K = Y ∩([0, 1] ∩ K) ∈ Closed (Y ).
and
p ([2, 3] ∩ K) = [1, 2] ∩ K − 1 = Y ∩ ([1, 2] ∩ K − 1) ∈ Closed (Y ).
So that the union of two closed sets in Y is again closed and thus
p (U ) ∈ Closed (Y ).


Thus p is a quotient map. However, p is not open, because [0, 1] ∈ Open (X)
gets mapped to [0, 1] ∈
/ Open (Y ).

1.9.7.1 Counterexample Let A := [0, 1) ∪ [2, 3] ⊂ R. Define Y := [0, 2] ⊂

R. Define p (: A → Y as:
x x ∈ [0, 1)
q (x) :=
x − 1 x ∈ [2, 3]
Observe that q is surjective, continuous (pasting lemma), but not a quotient
map: [2, 3] ∈ Open (A) is saturated (q −1 ([1, 2]) = [2, 3]) and gets mapped to
[1, 2] ∈
/ Open (Y ).

1.9.8 Example
Let π1 : R2 → R. Then π1 (projection onto the first coordinate) is continuous
and surjective. Furthermore, π1 is an open map. Thus it is a quotient map.
• But π1 is not a closed map.
  
x 2
Define C := ∈ R | xy = 1
y


– Claim: C ∈ Closed R2
 
a
Take any ∈ R2 \C.
b
 
a
We could always pick an open rectangle around which doesn’t

b
intersect C. Thus R2 \C ∈ Open R2 .

31
 However, π1 (C) = R\ {0} ∈ Open (R).
 
0
1.9.8.1 Counterexample Let A := C ∪ , then the map q : A → R ob-
0
 by restricting π1 is continuous and surjective, but it is
tained not
a quotient map:
0 0
∈ Open (A) is saturated in A with respect to q ( = π1−1 ({0})),
0 0
but its image is not open in R.
TODO: Example of non-quotient map.

1.9.9 Definition of Quotient Topology

If X is a topological space and A is a set and if p : X → A is a surjective map,
then there exists exactly one topology T on A relative to which p is a quotient
map. It is called the quotient topology induced by p.
• Claim: T = U ∈ 2A | p−1 (U ) ∈ Open (X)


Proof:
Define R := U ∈ 2A | p−1 (U ) ∈ Open (X) .


– Step 1: R is a topology.
∗ Since p is surjective, p1− (∅) = ∅ and p−1 (A) = X and so
∅, A ∈ R.
−1
S
S


∗ Let
S Uα −1 ∈ R∀α ∈ J. Then α∈J Uα ∈ R because p α∈J Uα =
α∈J p (Uα ) ∈ Open (X).
T 
∗ Let Ui ∈ R∀i ∈ {1, . . . , n} , n ∈ N. Then i∈{1, ..., n} i ∈ R
U
T  T
because p−1 i∈{1, ..., n} Ui = i∈{1, ..., n} p
−1
(Ui ) ∈ Open (X).
– Step 2: If A is endowed with the topology defined by R then p is a
quotient map.
∗ p is continuous:
Take any U ∈ R. Thus p−1 (U ) ∈ Open (X).
∗ Take some U ∈ Open (X) such that U is saturated in X with
respect to p. That is, U = p−1 (B) for some B ∈ A. Then B ∈ R
because p−1 (B) = U ∈ Open (X).
– Step 3: If p is a quotient map then Open (A) = R.
∗ ⊆
Take any U ∈ Open (A). Since p is continuous, p−1 (U ) ∈
Open (X). Thus U ∈ R.
∗ ⊇
Take any U ∈ R. Thus p−1 (U ) ∈ Open (X). But since p is a
quotient map that means that U ∈ Open (A).


32
1.9.10 Example

a
 x>0
Let p : R → {a, b, c} be a surjective map defined by p (x) := b x<0 .

c x=0

The quotient topology on {a, b, c} is thus {∅, {a, b, c} , {a} , {b} , {a, b}}.

1.9.11 Definition of Quotient Space

Let X be a topological space, and let X ∗ be a partition of X into disjoint subsets
whose union is X. Let p : X → X ∗ be a surjective map that carries each point
of X to an element of X ∗ containing it. In the quotient topology induced by p,
the space X ∗ is called a quotient space of X. (Thus a quotient space is always
taken with respect to a certain partition of the a given topological space).
X ∗ is also called an identification space, or a decomposition space.
∗
Thus U ∈ 2X is a collection of “chunks” of X, and p−1 (U ) is the union of
all the elements belonging to any one of those “chunks”, all together. So the
open sets in X ∗ would be sets of classes whose union is an open set in X.

1.9.12 Example
  
x
∈ R | x + y ≤ 1 and let X ∗ := S 1 ∪
2 2 2

Let X be the closed unit ball
   y
x 2
∈ 2R | x2 + y 2 < 1 .
y
X ∗ w S2
by the following map:
f : S2 → X ∗     

 x 0
 1
S =
  0
y

    



z 2
 
x


 q ! 
f y  :=  x2 +y 2 +z 2
 2 arcsin
 2




z y
 
cos arctan

 



 qπ ! x 
otherwise
 


   2 arcsin x2 +y2 +z2 
2
 
 

sin arctan xy



π

1.9.13 Example–R/Q
(HWS04E01)
The quotient topology on R/Q is the trivial topology {∅, R/Q}.

1.9.14 Theorem 22.1–Restricted Maps of Quotient Maps

Let p : X → Y be a quotient map, and let A ∈ 2X be a saturated set of X. Let
q : A → p (A) to defined as the restriction of p to A.

33
1.9.14.1 a A ∈ Open (X) ∨ A ∈ Closed (X) =⇒ q is a quotient map.

1.9.14.2 b If p is either an open map or a closed map then q is a quotient

map. HOW??

1.9.15 Some Claims about Quotient Maps and Spaces

• The composite map of two quotient maps is again a quotient map
• The cartesian product of two quotient maps need not be a quotient map.
(see example 7)
• If p and q are both open maps then p × q is an open map, thus a quotient
map.
• The quotient space of a Hausdorff space need not be Hausdorff. (HWS04E01
– Zariski topology on C2 )

– S/R is Hausdorff ⇐⇒ ∀ [x]R , [y]R ∈ S/R : [x]R 6= [y]R , ∃Ax , Ay ∈

Open (S) : Ax ∩ Ay = ∅, Ax and Ay are saturated, and [x]R ⊂ Ax
and [y]R ⊂ Ay .
−1
– If ∼R : S×S → {0, 1}, S/R is Hausdorff =⇒ (∼R ) ({1}) ∈ Closed (S × S)
−1
– If (∼R ) ({1}) ∈ Closed (S × S) ∧ q : S → S/R is an open map =⇒
S/R is Hausdorff.

• If each elemen of a quotient space X ∗ is a closed subset of X then X ∗ will

satisfy the T1 axiom.

1.9.16 Theorem 22.2–Continuous maps out of Quotient Spaces

Let p : X → Y be a quotient map.
Let Z be a space and let g : X → Z be a map that is constant on each set
p−1 ({y}) for all y ∈ Y .
• Then:

1. g induces a map, f : Y → Z such that f ◦ p = g.

2. f is continuous ⇔ g is continuous.
3. f is a quotient map ⇔ g is a quotient map.

• Proof:

1. Since ∀y ∈ Y , g p−1 ({y}) is a single-point

set in Z, define f (y) to
be that point inside of g p−1 ({y}) .

2. TODO
3. TODO

34
1.9.17 Universal Property of the Quotient Topology
(HWS04)
Let S be a topological space, R be an equivalence relation on S and q : S →
S/R be the quotient map.
∀ topological space T and ∀f ∈ T S : f (x) = f (y) ⇐⇒ x ∼R y ∃!ϕ ∈
C (S/R, T ) : f = ϕ ◦ q.
The map is ϕ ([x]R ) := f (x).

1.9.18 Connectedness
(HWS06E01) Any quotient of a connected topological space is connected.

1.9.19 Homeomorphisms
Every homeomorphism is also a quotient map, and the quotient topology is
unique.

1.10 §20, §21 The Metric Topology

1.10.1 Definition of a Metric
A metric on a set X is a function

d:X ×X →R
having the following properties:
1. d (x, y) ≥ 0∀x, y ∈ X
2. d (x, y) = 0 ⇔ x = y∀x, y ∈ X
3. d (x, y) = d (y, x) ∀x, y ∈ X
4. d (x, y) + d (y, z) ≥ d (x, z) ∀x, y, z ∈ X

1.10.2 Definition of the -ball centered at x

For any  > 0 and any x ∈ X, define:

Bd (x, ) := {y | d (x, y) < }

1.10.3 Definition of the Metric Topology

If d is a metric on a set X, then the topology induced by the basis {Bd (x, ) | x ∈ X ∧  > 0}
is called the metric topology induced by d.
• Proof that this is a basis for a topology:

– The first point of 1.2.1 is fulfilled because the -ball centered at any
x ∈ X, for any  > 0 will be a basis element containing x.

35
 – For the second point, let Bd (x1 , 1 ) , Bd (x2 , 2 ) be two basis ele-
ments and take some x3 ∈ Bd (x1 , 1 ) ∩ Bd (x2 , 2 ).
Define 3 := min ({1 − d (x1 , x3 ) , 2 − d (x2 , x3 )}).
First, observe that 3 > 0 because d (xi , x3 ) < i ∀i ∈ {1, 2}, which
is true as x3 ∈ Bd (x1 , 1 ) ∩ Bd (x2 , 2 ).
∗ Claim: x3 ∈ Bd (x3 , 3 ) ⊂ B (x1 , 1 ) ∩ Bd (x2 , 2 )
∗ Proof: First by definition x3 ∈ Bd (x3 , 3 ).
Pick some i ∈ {1, 2}.
Next, take any x ∈ Bd (x3 , 3 ). Thus d (x, x3 ) < 3 < i −
d (xi , x3 ).
So that d (x, x3 ) + d (xi , x3 ) < i .
Invoke the triangle inequality to get: d (x, xi ) < i ⇒ x ∈
Bd (xi , i ).

• (HWS05E02) the topology induced by a metric is the coarsest such that

the distance is a continuous function.

1.10.4 Another characterization of Metric Topology


The metric topology is T := U ∈ 2X | ∀x ∈ U ∃ > 0 : Bd (x, ) ⊂ U .

1.10.5 Examples
(
1 x 6= y
1.10.5.1 The Discrete Topology The metric d (x, y) := in-
0 x=y
duces the discrete topology.Bd (x, 1) = {x}.

1.10.5.2 The Order Topology The metric d (x, y) := |x − y| on R induces

the order topology.

1.10.6 Metrizable Spaces

If X is a topological space, X is said to be metrizable if ∃ a metric d on the set
X of which the metric topology is equal to the original topology on X.
A metric space is a metrizable space X together with a specific metric d that
induces the topology of X.

1.10.7 The Standard Bounded Metric

Let X be a metric space with the metric d. Define d¯ : X × X → R by the
equation
d¯(x, y) := min ({d (x, y) , 1})
Then d¯ is a metric that induces the same topology as d.
Under this metric, every subset is bounded.

36
1.10.8 The Square Metric (The Product Metric with d∞ )
The metric ρ (x, y) := max ({|xi − yi | | i ∈ {1, . . . , n}}) on Rn is the square
metric.
Basis elements are squares in R2 for example.
It induces the same topology as the standard topology on Rn .

1.10.9 Lemma 20.2–Comparing topologies based on comparing met-

rics
Let d and d0 be two metrics on the set X, and let T and T 0 be the two topologies
induced by d and d0 respectively. Then T 0 is finer than T if and only if ∀x ∈
X, ε > 0 ∃δ > 0 such that Bd0 (x, δ) ⊂ Bd (x, ε).
Proof: TODO (by Theorem 13.3 of comparing bases for a topology)

1.10.10 Theorem 20.3–Metric topologies and Product Topologies

The topologies on Rn induced by the Euclidean metric d and the square metric
ρ are the same as the product topology on Rn .

1.10.11 The Uniform Metric

Let J be some index set, and let x := (xα )α∈J and y := (yα )α∈J be two points
of RJ .
Define the metric ρ (x, y) := sup ({min ({|xα − yα | , 1}) | α ∈ J}) on RJ ,
called the uniform metric.
The topology it induces is called the uniform topology.

• (HWS03E05)
  by the basis Bu :=
The uniform topology on RN is generated
Br ({xn }) : r > 0, {xn } ∈ RN where Br ({xn }) := {yn } ∈ RN : sup ({|xn − yn | : n ∈ N}) < r .
Observe that Tbox ⊃ Tunif orm ⊃ Tproduct .

1.10.12 Theorem 20.4–Relation between the uniform, box and prod-

uct topologies
1. The uniform topology on RJ is finer than the product topology and coarser
than the box topology.
Proof: TODO (in handwriting)
2. These three topologies are all different if J is infinite.
Proof: TODO

1.10.13 RR is not metrizable

(16:50) (+zeno) I think one way is: K uncountable, if R^K is metrizable then
so is {0,1}^K is also, where {0,1} is discrete, but then show this space does not
have a countable local base at each point (the way a metrizable space should).

37
 (16:52) (+zeno) any subspace of a metric space is metrized by the induced
("same") metric

1.10.14 Theorem 20.5–RN is metrizable

two points of RN .
Let x := (xn )n∈N and y := (yn )n∈N ben o
min({|xn −yn |, 1})
Define the metric D (x, y) := sup n | n ∈ N .

• Claim: D is a metric that induces the product topology on RN .

• Proof: TODO (in handwriting)

1.10.15 Metric Topology vs. Subspace Topology

If A is a subspace of a topological space X and d is a metric for X, then the
restriction of d to A × A is a metric for the topology of A.
Proof: TODO

1.10.16 Metric Topology vs. Order Topology

Some order topologies are metrizable (N and R for instance) and others are not.

1.10.17 Metric Topology vs. The Hausdorff Axiom

The Hausdorff axiom is satisfied by every metric topology.

1.10.18 Metric on Finite Product of Metric Spaces

Let (Si , di ) be a metric space ∀i ∈ {1, . . . , n}. Let p ≥ 1.
Pn p 1 Qn
Define Dp (x, y) := { i=1Q [di (xi , yi )] } p for all x, y ∈ i=1 Si .
n
Qn Then Dp is a metric on i=1 Si which induces the product topology on
i=1 Si .

1.10.19 Metric Topology on a Countable Product Topology

The countable product topology is metrizable.

1.10.20 Theorem 21.1–ε−δ continuity equivalent to continuous func-

tions
Let f : X → Y be a map from two metrizable spaces X and Y with metrics dX
and dY .
Claim: f is continuous ⇐⇒ [ ∀x ∈ X, ∀ε > 0∃δ > 0 such that dX (x, y) <
δ =⇒ dY (f (x) , f (y)) < ε ]
Proof: TODO

38
1.10.21 Countable Basis at a point
A space X is said to have a countable basis at the point x ∈ X if ∃ a countable
collection {Un }n∈N of neighborhoods of x such that any neighborhood U of x
contains at least one of the sets Un . A space X that has a countable basis at
each of its points is said to satisfy the first countability axiom.
• (HWS05E01) Every metrizable space has a countable basis at every point.

1.10.22 Theorme 21.2–The Sequence Lemma

Let X be a topological space, x ∈ X be some point, and let A ∈ 2X . If there is
a sequence of points of A converging to x, then x ∈ closureX (A). The converse
(x ∈ closureX (A) implies that ∃ a sequences of points of A convering to x)
holds if X has a countable basis at x.
Proof: TODO

1.10.23 Theorem 21.3–Continuity and Sequences

Let f : X → Y .
• Claim: f is continuous =⇒ ∀ {xn } → x in X, {f (xn )} → f (x)
• Proof: (HWS02E06) (b)
• Claim: For every convergence sequence {xn } → x in X, {f (xn )} → f (x)
and X has a countable basis at x =⇒ f is continuous.
• Proof: TODO
• Counterexample: (HWS02E06) (c) Observe that P (R) is finer than Tcocountable (R).
The identity map on idX : (R, Tcocountable ) → (R, P (R)) is not continuous
by 1.5.17. Any sequence which converges in (R, Tcocountable ) converges in
(R, P (R)) and so {ϕ (xn )} converges, yet ϕ is not continuous!

1.10.24 Lemma 21.4–Constructing Continous Functions

Addition, subtraction, multiplication operations are continuous functions from
R × R → R.
The quotient (division) operation is a continuous function from R×R\ {0} →
R.

1.10.25 Theorem 21.5–Sum/Multiplication of Continuous Functions

is Continuous
If Xis a topological space and f : X → R and g : X → R are two continuous
functions, then f +g, f −g and f ·g are continuous functions. If g (x) 6= 0∀x ∈ X
then f /g is continuous.
Proof: TODO

39
1.10.26 Uniform Convergence Definition
Let fn : X → Y be a sequence of functions from the set X to the metric space Y .
Led d be the metric on Y . We say that the sequence {fn } converges uniformly
to the function f : X → Y if:
∀ε > 0∃mε ∈ N such that d (fn (x) , f (x)) < ε ∀n > mε , x ∈ X

1.10.27 Theorem 21.6–Uniform Limit Theorem

Let fn : X → Y be a sequence of continuous functions from the set X to the
metric space Y . If {fn } converges uniformly to f then f is continuous.
Proof: TODO (in handwriting)

1.10.28 Uniform Metric and Uniform Convergence

Let RX be the space of all functions f : X → R, with the uniform metric ρ.
Claim: The sequence of functions fn : X → R converges uniformly to f
⇐⇒ {f
n } → f when these two are considered as elements of the metric space
RX , ρ .
Proof: TODO

1.10.29 Example–RN with the box topology is not metrizable

Proof: (HWS05) If it were then it would have had a countable basis and then by
21.2 for any set, and any point in the set’s closure, there would be a sequence
converging to that point in the set, but if we take the set of all points with
positive coordinates, zero is in the closure of the set, but no sequence converges
to it.

1.10.30 Example–An uncountable product of R with itself is not

metrizable

Proof: (HWS05) Similar to above, defined A = xα ∈ RJ : xα = 1 for all but finitely many α ∈ J .
Again, zero is in the closure, but no sequence converges to zero.

1.10.31 Ultrametrics
(HWS05)
A metric space (S, d) is said to be ultrametric ⇐⇒ ∀x, y, z ∈ S, d (x, z) ≤
max ({d (x, y) , d (y, z)}).
(
1 x 6= y
• Example: d (x, y) := is ultrametric.
0 x=y
• In an ultrametric space, every triangle is an isosceles (∀x, y, z ∈ S, d (x, y) =
d (y, z) ∨ d (x, y) = d (x, z) ∨ d (y, z) = d (x, z)).
• In an ultrametric space, ∀x ∈ S, ε > 0, Bd (x, ε) = Bd (y, ε) ∀y ∈
Bd (x, ε).

40
2 Chapter 3 Connectedness and Compactness
2.1 §23 Connected Spaces
2.1.1 Separation Definition
Let X be a topological space. A separation of X is two subsets, U, V ∈
Open (X) \ {∅} such that:

1. U ∩ V = ∅
2. U ∪ V = X

2.1.2 Connected Space Definition

A space X is connected if @ a separation of X.

2.1.3 Theorem–Homeomorphism Preserves Connectedness

Proof:
Let X be a connected topological space, let Y be any topological space, and
let f : X → Y be a homeomorphism. TODO (in handwriting)
(HWS06E02)

2.1.4 Theorem–Connected Spaces Characterized by Clopen Subsets


• Claim: A space X is connected ⇐⇒ U ∈ 2X | U ∈ Open (X) ∧ U ∈ Closed (X) =
{X, ∅}
• Proof: TODO (in handwriting)

2.1.5 Lemma 23.1–Separation (another definition)

If Y is a subspace of X, a separation of Y is two sets A, B ∈ 2Y \ {∅} such that:

1. closureX (A) ∩ B = A ∩ closureX (B) = ∅

2. A ∪ B = Y

• Claim: @ a separation of Y ⇐⇒ Y is connected.

• Proof: TODO (in handwriting–unsolved!)

2.1.6 Lemma 23.2

If {C, D} is a separation of X, and if Y is a connected subspace of X, then
Y ⊂ C Y Y ⊂ D.
Proof: TODO (in handwriting)

41
2.1.7 [I] Lemma 23.3–Union of Not-Disjoint Connected Subspaces is
Connected
T
Let {Uα }α∈J be a collection of connected subspaces of X such that α∈J Uα 6=

∅.
S
Claim: α∈J Uα is a connected subspace.
Proof: TODO (in handwriting)

2.1.8 [I] Theorem 23.4

Let A be a connected subspace of X. If A ⊂ B ⊂ closureX (A) then B is a
connected subspace of X.
(A∪ some (or all) of its limit points is connected)
(Thus closure of connected space is connected!)
Proof: TODO (in handwriting)

2.1.9 [I] Theorem 23.5–The image of a connected space under a con-

tinuous map is connected
Let X be a connected space and let Y be any space. Let f : X → Y be a
continuous map.
Claim: f (X) is a connected subspace of Y .
Proof: TODO (in handwriting)

2.1.10 Theorem 23.6–A finite cartesian product of connected spaces

is connected
Let n ∈ N and let {Xi }i∈{1,..., n} be a finite collection of connected spaces.
Q
Claim: i∈{1,..., n} Xi is a connected space.

2.1.11 Example 6–RN is not connected in the box topology

Separate RN to two open sets: all bounded sequences in RN and all unbounded
sequences in RN .

2.1.12 Example 7–RN is connected in the product topology

What is Rω ??? TODO unsolved
(19:16) (+Polytope) Munkres says that R^omega is the set of all sequences
of real numbers, and R^oo is the subspace consisting of all sequences that are
eventually zero.
(19:22) (+Psycho_pr) Hrmmm (19:24) (+Psycho_pr) R^oo = U_{n\in N}
R^n (19:24) (+Psycho_pr) R^omega = \prod_{n\in N} R

42
2.1.13 Totally Disconnected Definition
A space is called totally disconnected if its only connected subspaces are one-
point sets.
Claim: If X has the discrete topology, then X is totally disconnected.
Claim: ∃ totally disconnected spaces with a topology different from the
discrete topology: Q (HWS06EC).

2.1.14 Example
(HWS06)
R and R2 are not homeomorphic because if they were, then R\ {x0 } and
2
R \ {f (x0 )} would be homeomorphic (where f is the homeomorphism). How-
ever, the former is not connected whereas the second is.

2.1.15 Product of Connected Spaces

(HWS06E03) The (countable, uncountable) product of connected spaces in the
product topology is connected.

2.1.16 R\ {0} is not connected

Thus it is not path-connected. R\ {0} = (−∞, 0) ∪ (0, ∞) are its connected
components.

2.1.17 Q is not connected

S
Thus it is not path-connected. Q = a∈Q {a} are its connected components.

2.2 §24 Connected Subspaces of the Real Line (and path

connectedness)
2.2.1 Linear Continuum Definition
A simply ordered set (1.3) L having more than one element, such that:
1. L has the least upper bound property.
2. ∀x, y ∈ L such that x < y∃z ∈ L such thatx < z < y.

is called a linear continuum.

2.2.2 Theorem 24.1–Linear Continuums in the Order Topology are

Connected
Let L be a linear continuum with the order topology.
1. Claim: L is connected.

2. Claim: (a, b) , [a, b] , [a, b), (a, b] are connected ∀a, b ∈ L such that a < b.

43
 3. Claim: (a, ∞) and (−∞, a) are connected ∀a ∈ L.
Proof: TODO (in handwriting)
Thus R, as well as rays and intervals in R are connected.

2.2.3 [I] Theorem 24.3–Intermediate Value Theorem

Let X be a connected space and Y be an ordered set with the order topology.
Let f : X → Y be a continuous map. If a, b ∈ X such that a 6= b, and if
r ∈ (f (a) , f (b)) then ∃c ∈ X such that f (c) = r.
Proof: TODO (in handwriting)

2.2.4 Example 1–The Ordered Square is a Linear Continuum

(HWS06E03) The order square is not locally path-connected. It is however
connected.

2.2.5 Example 2–The product of a well-ordered with [0, 1) is a linear

continuum
2.2.6 Definition of a Path
Let X be a topological space and let there be two points x, y ∈ X. A path in X
from x to y is a continuous map f : [a, b] → X where a, b ∈ R and a < b, such
that f (a) = x and f (b) = y.

2.2.7 Definition of Path Connectedness

A space X is said to be path connected if ∀x, y ∈ X∃ a path in X from x to y.

2.2.8 [I] Path Connectedness means Connectedness

Proof: TODO (in handwriting)

2.2.9 The continuous image of a path connected space is path con-

nected
2.2.10 Example 3–The unit ball in Rn is connected
2.2.11 Example 4–Punctured Euclidean Space
Punctured Euclidean space Rn \ {0} is path connected ∀n ∈ N\ {1}.

2.2.12 Example 5–The Unit Sphere S n−1 ⊂ Rn

S n−1 := {x | ||x|| = 1}
x
If n > 1, it is path connected: g : Rn \ {0} → S n−1 defined by g (x) := ||x||
is continuous and surjective.

44
2.2.13 A connected space need not be path connected
2.2.13.1 Example 6 The ordered square is connected (because it is a linear
continuum).
Claim: It is not path connected.
Proof: TODO (in handwriting)
  
x
2
2.2.13.2 [I] Example 7–The Topologist’s Sine Curve S := ∈ R | x ∈ (0, 1]
sin x1
Because S is the image of a connected set (0, 1] under a continuous map, S
is connected. Therefore its closure closureR2 (S) is also connected by Theorem
23.4 is also connected.
The topologist’s Sine
Curve
 is: 
0 2
closureR2 (S) = S ∪ ∈ R | x ∈ [−1, 1]
x
Claim: closureR2 (S) is not path connected. (HWS06E03)
Proof:
  TODO (in handwriting).
 Its two path-connected components are S
0
and ∈ R2 | x ∈ [−1, 1]
x

2.2.14 Example–Orthogonal Group

(HWS06E04) On (R) is not connected, because there is a continuous function
from it to {0, 1} which is not constant, namely, the function δ1 ◦ det. SO (n) is
bnc
connected because SO (n) ' (SO (2)) 2 and SO (2) ' S 1 . Its two connected
components are SOn (R) and On̄ (R) ' SOn (R).

2.2.15 Example–SU (2)

(HWS06E05) SU (2) is isomorphic to S 3 which is connected.

2.2.16 A riddle
(HWS06)
There cannot be an f ∈ C (R, R) such that f (R\Q) ⊂ f (Q) and f (Q) ⊂
f (R\Q). If there were, then |f (R\Q)| ≤ ℵ0 =⇒ |f (R)| ≤ ℵ0 . But R is
connected, so f (R) is connected. So f (R) is connected countable subset of R,
so it is a point, so f (x) = x0 , which is a contradiction.

2.3 §25 Components and Local Connectedness

2.3.1 Components
Let X be a topological space. Define an equivalence relation on X be setting
x ∼ y ⇐⇒ ∃A ∈ 2X such that A is connected and x, y ∈ A. The equivalence
classes of X are called components or connected components of X.
Proof that ∼ is an equivalence Relation: TODO (in handwriting)

45
2.3.2 Theorem 25.1–Description of Components
The components of X are connected disjoint subspaces of X whose union is X,
such that each nonempty connected subspace of X intersects only one of them.
Proof: TODO
• Each component is a closed subset of X. (Since closures are connected
and each connected subspace intersects only one component.
• If there are only finitely many components then each component is also
open.

2.3.3 Path Components

Let X be a topological space. Define an equivalence relation on X be setting
x ∼ y ⇐⇒ ∃ path in X from x to y. The equivalence classes of X are called
path components of X.
Proof that ∼ is an equivalence Relation: TODO (in handwriting)

2.3.4 Theorem 25.2

The path components of X are path-connected disjoint subspaces of X whose
union is X such that each nonempty path connected subspace of X intersects
only one of these components.
Proof: TOOD (in handwriting)
• Path components need not be closed or open.

2.3.5 Example 1–Components of Q

The components of Q are all of its single points. None of the components are
open.

2.3.6 Example 2–The Topologist’s Sine Curve

The Topologist’s sine curve has one component (since it is connected) and two
path components. One path component is S and another is 0 × [−1, 1].
• Claim: S ∈ Open (closureR2 (S)) but S ∈
/ Closed (closureR2 (S))
• Claim: 0×[−1, 1] ∈
/ Open (closureR2 (S)) yet 0×[−1, 1] ∈ Closed (closureR2 (S)).
  
0
• Claim: closureR2 (S) \ ∈ R2 | x ∈ [−1, 1] ∩ Q has only one com-
x
ponent but uncountably many path components.

2.3.7 Local Connectedness at x

Let X be a space and let x ∈ X.
A space X is locally connected at x if ∀U ∈ Open (X) : x ∈ U , ∃V ∈
Open (X) : x ∈ V ⊂ U such that V is connected.

46
2.3.8 Locally Connected Space
If X is locally connected at each of its points, it is locally connected.

2.3.9 Local Path Connectedness at x

Let X be a space and let x ∈ X.
A space X is locally path connected at x if ∀U ∈ Open (X) : x ∈ U ,
∃V ∈ Open (X) : x ∈ V ⊂ U such that V is path connected.

2.3.10 Locally Path Connected Space

If X is locally connected at each of its points, it is locally path connected.

2.3.11 Example 3
• Each interval and each ray in R is locally connected (and as stated previ-
ously also connected).
• [−1, 0) ∪ (0, 1] is not connected but it is locally connected.
• The topologists sine curve is connected but not locally connected.
• Q is neither connected nor locally connected.

2.3.12 Theorem 25.3–A Criterion for Local Connectedness

A space X is locally connected ⇐⇒ ∀U ∈ Open (X), each component of U is
open in X.
Proof: TODO (in handwriting)

2.3.13 Theorem 25.4–A criterion for Local path-connectedness

A space X is locally path connected ⇐⇒ ∀U ∈ Open (X), each path component
of U is open in X.
Proof: TODO

2.3.14 Theorem 25.5–A relation between path components and comp-

nents
Let X be a topological space.
1. Each path component of X lies in a component of X.
Proof: Each path component is path connected and is thus connected and
so lies in exactly one component.
2. If X is locally path connected, then the components and the path compo-
nents of X are the same.
Proof: TODO (in handwriting)

47
2.4 §26 Compact Spaces
“Every metric space (for instance, Y) is Hausdorff. Then, you can use the fact
that every continuous and bijective map f : X → Y is a homeomorphism (that
is, f1 is continuous), if X is compact and Y Hausdorff.
To prove that claim, you can do as follows: in order to see that a map
g : X → Y is continuous, you can prove that, for every closed subset C ⊂ X,
g −1 (C) ⊂ Y is closed. In our case, g=f1, so it suffices to show that, for every

−1
closed subset C ⊂ X, f −1 (C) = f (C ⊂ Y ) is closed. Right?
But, if C is a closed subset of a compact space X, then it’s itself compact and
the image of a compact set by a continuous map, f, is compact. Hence f(C) is a
compact subset of a Hausdorff space, Y. Every compact subset of a Hausdorff
space is closed. Thus, f(C) is closed. qed.”

2.4.1 Cover
A collection A ⊂ 2X covers X, or “is a covering of X”, if
S
A∈A A = X.

2.4.2 Open Covering

A is an open covering if it is a covering such that A ∈ Open (X) ∀A ∈ A.

2.4.3 Compact Space

A space X is said to be compact if every open covering A of X contains a finite
subcollection that also covers X.

2.4.4 Example 1–R is not compact

A = {(n, n + 2) | n ∈ Z} is an open cover of R which has no finite subcollection
that covers R.

2.4.5 Example 2
X = {0} ∪ n1 | n ∈ N is compact.


Each cover of X will contain 0, and so will contain the infinite number of
points that are near it. All the rest of the points are finite.

2.4.6 Example 3–Finite Spaces are Compact

2.4.7 Example 4
(0, 1] is not compact, since the open cover A := ( n1 , 1] | n ∈ N contains no


finite subcollection covering (0, 1].

(0, 1) is not compact (same reason as above) with A := ( n1 , 1 − n1 ) | n ∈ N .


2.4.8 A cover for a subspace

If Y is a subspace of X, a collection A ⊂ 2X is said to cover Y if
S
A∈A A⊃Y.

48
2.4.9 Lemma 26.1–Compact Subspaces
Let Y be a subspace of X.

• Claim: Y is compact ⇐⇒ ∀ covering of Y by sets in Open (X), ∃ a finite

subcollection which covers Y .
• Proof: TODO (in handwriting)

2.4.10 [I] Lemma 26.2–Every closed subspace of a compact space is

compact
Proof: TODO (Proven in Rudin PMA chapter 2.)

2.4.11 Lemma 26.3–Every compact subspace of a Hausdorff space is

closed
Proof: TODO (in handwriting)

2.4.12 Lemma 26.4

If Y is a compact subspace of the Hausdorff space X, and x0 ∈ X\Y , then
∃U, V ∈ Open (X) such that x0 ∈ U , Y ⊂ V and U ∩ V = ∅.

2.4.13 Example 6–Theorem 26.3 breaks down without Hausdorff

Condition
Take X = R with the co-finite topology. This space is not Hausdorff.
Closed (X) = U ∈ 2X | |U | < ∞ ∪ {X}
Claim: Every subset of R is compact in X.
Proof: TODO (in handwriting)

2.4.14 Theorem 26.5–The image of a compact space under a contin-

uous map is compact
Proof: TODO (Proven in Rudin PMA chapter 4)

2.4.15 [I] Theorem 26.6–Verifying that a map is a homeomorphism

Let f : X → Y be a bijective continuous function. If X is compact and Y is
Hausdorff then f is a homeomorphism.
Proof: TODO (in handwriting)

2.4.16 Theorem 26.7–The product of finitely many compact spaces

is compact
Proof: TODO (in handwriting)

49
2.4.17 Lemma 26.8–The Tube Lemma
• Assumptions:

– Let X be a space and Y be a compact space.

– Let x0 ∈ X.
– Let N ∈ Open (X × Y ) such that x0 × Y ⊂ N .

• Claim: ∃W ∈ Open (X) such that:

– x0 ∈ W
– W × Y ⊂ N.

• Proof: Proven as part of Theorem 26.7.

2.4.18 Example 7–The Tube Lemma does not hold if Y is not com-
pact
n o
Let Y be the y-axis of R2 . Define N := x × y : |x| < y21+1 .


Then 0 × R ∈ N ∈ Open R2 , but it contains no tube about 0 × R.

2.4.19 Tychonoff Theorem–The Infinite Product of Compact Spaces

is Compact
Proof: TODO
(HWS09E06)

2.4.20 Finite Intersection Property

A collection C of subsets of X is said to have the finite intersection property if
for every B ∈ 2C such that |B| < ∞,
\
C 6= ∅
C∈B

2.4.21 Theorem 26.9–Criterion for Compactness Formulated in terms

of Closed Sets
Let X be a topological space.
X is compact
T ⇐⇒ ∀C ∈ 2Closed(X) such that C has “the finite intersection
property”, C∈C C 6= ∅.
Proof: TODO (Rudin PMA chapter 2, theorem 2.36–anyway, in handwrit-
ing)
Note: If we have a nested sequence, Ci ⊃ Ci+1 ∀iT∈ N and Ci 6= ∅∀i ∈ N,
then {Ci } has the finite intersection property. Thus i∈N Ci 6= ∅.

50
2.5 §27 Compact Subspaces of the Real Line
2.5.1 Theorme 27.1–Every closed interval is compact
Let X be a simply ordered set having the least upper bound property.
Claim: X in the order topology has the property that each closed interval in X
is compact.
Proof: TODO (in handwriting)
Corollary 27.2: Every closed interval in R is compact.

2.5.2 [I] Theorem 27.3–Characterization of compact subspaces of Rn

(Heine-Borel Theorem)
Let d be the Euclidean metric on Rn and ρ be the square metric on Rn .
Claim: A subspace A of Rn is compact ⇐⇒ A ∈ Closed (Rn )∧(diamd (A) < ∞ ∨ diamρ (A) < ∞)
Proof: (Rudin Chapter 2 Theorem 2.34?) Consider only one metric (why?). For
=⇒, build a covering by boxes of length in N. For ⇐=, A is contained inside
some compact box of some finite size, thus, being closed, it is compact also.

2.5.3 Example 1–The unit sphere and the closed unit ball are com-
pact in Rn
Because they are closed and bounded.

2.5.4 Example
• A = x × x1 |x ∈ (0, 1] is closed but not bounded, thus it is not com-


pact.
• S = x × sin x1 |x ∈ (0, 1] is bounded but not closed, thus it is not


compact.

2.5.5 Theorem 27.4–Extreme Value Theorem

Let f : X → Y be continuous, where Y is an ordered set in the order topology.
Claim: If X is compact then ∃ points xmin , xmax ∈ X such that f (xmin ) ≤
f (x) ≤ f (xmax ) for any x ∈ X.
Proof: TODO (in handwriting)

2.5.6 Distance from a point to a set

Let (X, d) be a metric space and let A ∈ 2X \ {∅}. ∀x ∈ X define the distance
from x to A by the equation:
d (x, A) := inf ({d (x, a) |a ∈ A})
Observe that for a fixed A, d : X → R is a continuous function, because
d (x, A) − d (y, A) ≤ d (x, y).

51
2.5.7 Lemma 27.5–The Lebesgue Number
Let A be an open covering of the metric space (X, d).
Claim: If X is compact, ∃δA > 0 such that ∀S ∈ 2X such that diam (S) < δA ,
∃AS ∈ A such that S ⊂ AS .
The number δA is called the Lebesgue number of A.
Proof: TODO (in handwriting)

2.5.8 Uniform Continuity

Let (X, dX ) and (Y, dY ) be two metric spaces. A function f : X → Y is said to
be uniformly continuous if ∀ε > 0∃δ > 0 : ∀ (x0 , x1 ) ∈ X 2 , dX (x0 , x1 ) < δ =⇒
dY (f (x0 ) , f (x1 )) < ε.

2.5.9 Theorem 27.6–Uniform Continuity Theorem

Let (X, dX ) be a compact metric space and (Y, dY ) a metric space, and let
f : X → Y be a continuous function.
Claim: f is uniformly continuous.
Proof: TODO (Rudin’s PMA chapter 4–also in handwriting)

2.5.10 Isolated Point

If X is a space, x ∈ X is said to be an isolated point of X if {x} ∈ Open (X).

2.5.11 Theorem 27.7–A set without isolated points is uncountable

Let X be a nonempty compact Hausdorff space. If X has no isolated points,
then X is uncountable.
Proof: TODO (in handwriting)
Corollary 27.8: Every closed interval in R (thus compact with no isolated points)
is uncountable.
But every superset of au uncountable set is uncountable, thus R is uncountable.

2.5.12 Examples
(HWS07)
• Q ∩ [0, 1] is not compact. If it were, then Q ∩ [0, 1] would have been in
Closed (R) by 26.3, however, it is not closed, since it does not contain its
limit points: (R\Q) ∩ [0, 1].
• O (n) is closed, and bounded, and thus, it is compact.

• A finite space is always compact.

• The Cantor set is an inifinite intersection of closed intervals, hence closed.
It is also bounded, hence, compact.

52
2.6 §28 Limit Point Compactness
2.6.1 Limit Point Compactness
A space X is said to be limit point compact if every infinite subset of X has a
limit point.

2.6.2 Theorme 28.1–Compactness implies limit point compactness

Proof: TODO (in handwriting)

2.6.3 Example 1
Y := {a, b}, Open (Y ) := {Y, ∅}
X := N × Y
• Claim: X is limit point compact.

– Proof: every nonempty subset of X has a limit point.

• Claim:X is not compact.

– Proof: the covering of X by open sets Un = {n} × Y has no finite

subcollection covering X.

2.6.4 Example 2
Let SΩ be the minimal uncountable well-order set, in the order topology.
TODO!!!
• Claim: SΩ is not compact.

– Proof: It has no largest element.

2.6.5 Example
n
N N
• (HWS09E02) In [0, 1] with the uniform topology, The set a ∈ [0, 1] : ∃i0 ∈ N : πi0 (a) = 1 ∧ πi (a) = 0∀
is infinite but has no limit point.
• (HWS09E03) [0, 1] is not limit point compact in the lower limit topology.
Take the set 1 − n1 : n ∈ N , it has no limit points. The neighborhood
[1, 2) of 1 does not intersect the set at all.

2.6.6 Limit Point Compactness and Continuous Functions

• (HWS09E04) If f ∈ C (X, Y ) and X is limit point compact, f (X) is not
necessarily limit point compact


– Take X := N, 2N × ({a, b} , {∅, {a, b}}) and let the continuous
function be the projection

onto the first coordinate. X is limit point
compact, but N, 2N is not.

53
2.6.7 Limit Point Compactness and Closedness
(HWS09E04) If X is limit point compact and A ∈ Closed (X) then A is limit
point compact.

2.6.8 Sequential Compactness

Let X be a topological space. If every sequence of points of X has a convergent
subsequence then X is sequentially compact.

2.6.9 Counterexample–Non-Sequentially Compact Space

(HWS09E05)
 N
[0, 1] [0, 1]
[0, 1] is not sequentially compact. If x ∈ [0, 1] then define fn ∈ [0, 1]
as fn (x) := the nth binary digit of x. If there were a subsequence of fn con-
verging, then we could take the number defined as alternating digits of that
subsequence, and the function on that number wouldn’t converge.

2.6.10 Theorem 28.2

Let X be a metrizable space. Then the following are equivalent:
1. X is compact.
2. X is limit point compact.

3. X is sequentially compact.
Proof: TODO (in handwriting)

2.7 §29 Local Compactness (note: not in material for test–

utility for measure theory)
2.7.1 Local Compactness at a point
A space X is called locally compact at x ∈ X iff ∃ some compact Cx ∈ 2X such
that ∃UCx ∈ Open (X) such that x ∈ UCx ⊂ Cx .

2.7.2 Local Compactness of a space

A space X is locally compact iff it is locally compact ∀x ∈ X.
Observe: A compact space is automatically locally compact, as the space itself
would contain any neighborhood of any point.

54
2.7.3 Example 1–R is locally compact, Q is not
Take any x ∈ R. Then ∃a, b ∈ R such that x ∈ (a, b) ⊂ [a, b]. [a, b] is compact.
Q is not locally compact: Take any rational r ∈ Q. If Q were compact, we’d be
able to find a neighborhood Ur whose closure would be compact, hence sequen-
tially compact, hence, contain a sequence which has a convergent subsequence.
But we can take then a sequence of rationals in Ur convering to an irrational in
closureR (Ur ).

2.7.4 Example 2–Rn is locally compact, RN is not.

The same argument as above applies for Rn . For RN , any basis element will not
be contained in a compact subspace.

2.7.5 Example 3–Simply Ordered set with Sup Property

Every simply ordered set X with the least upper bound property is locally
compact: Given a basis element for X, it is contained in a closed interval in X,
which is compact.

2.7.6 Theorem 29.1–One Point Compactification

(HWS07) “Alexandroff Compactification”
Let X be a space.
Claim: X is locally compact Hausdroff ⇐⇒ ∃ a space Y such that:
1. X ⊂ Y with the subspace topology.

2. |Y \X| = 1
3. Y is a compact Hausdorff space.
If Y and Y 0 are two spaces satisfying these conditions, then there is a homeo-
morphism of Y to Y 0 that equals the identity map on X.
Proof: TODO (in handwriting)
Observe: If X is itself compact, then Y is simply X together with a single
point. If X is not compact, Y \X contains a limit point of X in Y , so that
closureY (X) = Y .
Define Y = X∪{∞} and Open (Y ) := Open (X)∪{{∞} ∪ X\K : K ⊂ X is compact}.

2.7.7 Compactification
If Y is a compact Hausdorff space, and X is a proper subspace of Y such that
closureY (X) = Y , then Y is called the compactification of X. If |Y \X| = 1,
then Y is called the one point compactification of X.

55
2.7.8 Example 4–R and R2
(HWS08)
The one-point compactification of R is homeomorphic to S 1 .
The one-point compactification of R2 is homeomorphic to S 2 .
If R2 is considered as C, then the one-point compactification, C ∪ {∞} is called
the Riemann Sphere or the extended complex plane.

Observe that S 2 is not homeomorphic

to S 1 × S 1 (π1 S 1 × S 1 ' Z2 which
is not homomorphic to π1 S 2 = {[ex0 ]}.

2.7.9 Theorem 29.2–Another Characterization of Local Compact-

ness
(HWS07)
Let X be a Hausdorff space.
X is locally compact ⇐⇒ ∀x ∈ X∀ neighborhood of U of x, ∃ a neighborhood
V of x such that closureX (V ) is compact and closureX (V ) ⊂ U .
Proof: TODO (in handwriting)

2.7.10 Corollary 29.3

Let X be locally compact Hausdorff and let A ∈ 2X . If A ∈ Closed (X) or
A ∈ Open (X) then A is locally compact.
Proof: TODO (in handwriting)

2.7.11 Corollary 29.4

A space X is homeomorphic to an open subspace of a compact Hausdorff space
⇐⇒ X is locally compact Hausdorff.
Proof: Theorem 29.1 and Corollary 29.3.

2.7.12 Proper Maps

(HWS07)
A map of topological spaces f : S → T is said to be proper ⇐⇒ ∀ topological
space Z, f × idZ : S × Z → T × Z is closed.
• S is a compact topological space ⇐⇒ S → {∗} is proper.

• The composition and product of proper maps is proper.

• Let S be Hausdorff and T be locally compact Hausdorff. f is proper ⇐⇒
f is closed ∧ f −1 ({t}) is compact ∀t ∈ T ⇐⇒ ∀K ⊂ T which is compact,
f −1 (K) is compact.

56
3 Chapter 4 Countability and Separation Axioms
(not part of the curriculum for the test–for
competeness reasons)
3.1 §31 The Separation Axioms
3.1.1 Regular and Normal Spaces
Assume X is a space such that {x} ∈ Closed (X) ∀x ∈ X.
 
x


• If ∀ ∈ X × Closed (X) ∩ 2X\{x} ∃U, V ∈ Open (X) such that:
B
– x∈U
– B⊂V
– U ∩V =∅

Then X is said to be regular.

 
A

• If ∀ ∈ Closed (X) × Closed (X) ∩ 2X\A ∃U, V ∈ Open (X) such
B
that:

– A⊂U
– B⊂V
– U ∩V =∅

Then X is said to be normal.

Observe: A regular space is Hausdorff, and a normal space is regular.

3.2 §35 Imbeddings of Manifolds

3.2.1 m-Manifold
Let m ∈ N. An m-manifold is a Hausdorff space X with a countable ba-
sis (1.10.21) such that ∀x ∈ X, ∃Nx ∈ Open (X) : x ∈ Nx such that ∃V ∈
Open (Rm ) such that Nx is homeomorphic with V .
1. A curve is a 1-manifold.
2. A surface is a 2-manifold.
It can be proven that a manifold can be imbedded in a finite-dimensional Eu-
clidean space.

3.2.2 Support of a Function

Let φ ∈ RX . Define the support of φ, supp (φ) := closureX φ−1 (R\ {0}) .

Observe: If x ∈
/ supp (φ) then ∃U ∈ Open (X) : x ∈ U such that φ (U ) = {0}.

57
3.2.3 Partition of Unity
n n
Let {Ui }i=1 ⊂ Open (X) be an open covering of the space X. {φi }i=1 ⊂
n
C (X, [0, 1]) is said to be a partition of unity dominated by {Ui }i=1 if:
1. supp (φi ) ⊂ Ui ∀i ∈ {1, . . . , n}
Pn
2. i=1 φ (x) = 1∀x ∈ X

3.2.4 Theorem 36.1–Existence of Finite Partitions of Unity

n
Let {Ui }i=1 ⊂ Open (X) be an open covering of the normal space X. Then ∃ a
n
partition of unity dominated by {Ui }i=1 .
Proof: TODO

4 Chapter 7 Complete Metric Spaces and Func-

tion Spaces
4.1 §43 Complete Metric Spaces
4.1.1 Complete Metric Spaces
The metric space (X, d) is said to be complete if every Cauchy sequence in X
converges.

4.1.2 Lemma 43.1

A metric space X is complete if every Cauchy sequence in X has a convergent
subsequence.
Proof: TODO (in handwriting)

4.1.3 Theorem 43.2–Rk is complete

Euclidean space Rk is complete in either of its usual metrics, the Euclidean
metric d or the square metric ρ.
Proof: TODO (in handwriting)

4.1.4 Lemma 43.3

Q
Let X be the product space X = Xα .
Let xn be a sequence of points of X.
• Claim: xn → x ⇐⇒ πα (xn ) → πα (x) ∀α
• Proof: TODO (in handwriting)

58
4.1.5 Theorem 43.4–Metric on the product space RN such that it is
complete
Proof: TODO (in handwriting)
d (a, b) := min ({|a − b| , 1}) is the standard bounded
n metric on R.
o
d(xi , yi )
N
Define the metric D on R by D (x, y) := sup i |i∈N
D induces the product

topology on RN .
Claim: RN , D is complete.

4.1.6 Example 1–Non-complete Metric Space

y) := |x − y| is not complete. Take any sequence of rationals
Q with d (x, √
convering to 2, which is Cauchy in Q but does not converge in Q.

4.1.7 Example 2–Non-complete Metric Space

(−1, 1) ∈ Open (R) with the metric d (x, y) := |x − y| is not complete. The
sequence xn := 1 − n1 is Cauchy but does not converge.
However, (−1, 1) is homeomorphic to R, and R is complete, thus, com-
pleteness is not preserved under homeomorphisms. Thus completeness is not a
“topological quality”.

4.1.8 The Uniform Metric

Let (Y, d) be a metric space. Define d (a, b) := min ({d (a, b) , 1}) to be the
standard bounded metric on Y derived from d.
Let x := (xα )α∈J and y := (yα )α∈J be two points in the cartesian product Y J .
Define:


ρ (x, y) := sup d (xα , yα ) | α ∈ J

ρ is a metric, called the uniform metric on Y J corresponding to the metric

d on Y .

4.1.9 Theorem 43.5–Complete Cartesian Product Spaces

If Y is complete in the metric d, then Y J is complete in the uniform metric ρ
corresponding to d.
Proof: TODO (in handwriting)

4.1.10 Theorem 43.6–All Continuous and Bounded Functions

Let X be a topological space Y be a metric space.

1. Define C (X, Y ) ⊂ Y X to be the space of all continuous functions f : X →

Y.

59


• Claim: C (X, Y ) ∈ Closed Y X , ρ
• Claim: If Y is complete, (C (X, Y ) , ρ) is complete.

2. Define B (X, Y ) ⊂ Y X to be the space of all bounded functions f : X → Y .

• Claim: B (X, Y ) ∈ Closed Y X , ρ
• Claim: If Y is complete, (B (X, Y ) , ρ) is complete.

4.1.11 The Sup Metric

Let (Y, d) be a metric space, and define a metric–the sup metric–on B (X, Y )
by: ρ (g, f ) := sup ({d (f (x) , g (x)) | x ∈ X})
• Observe: ρ (f, g) = min ({ρ (f, g) , 1})

• Observe: If X is compact, C (X, Y ) ⊂ B (X, Y ) and so the sup metric is

defined
on C (X, Y ). Thus if Y is complete, C (X, Y ) is also complete in
Y X, ρ .

4.1.12 Completion of a Metric Space

Let X be a metric space. If h ∈ Y X is an isometric imbedding of X into a
complete metric space Y , then the subspace h (X) of Y is a complete metric
space, called the completion of X. The completion is uniquely determined up
to an isometry.
• (HW10E05) The completion of Q is R with respect to the standard metric.
The completion of Z is Z with respect to the standard metric.

4.1.13 Theorem 43.7–Isometric Imbedding in a Complete Metric

Space
Let (X, d) be a metric space.
• Claim: ∃ an isometric imbedding of X into a complete metric space.
• Proof: TODO (in Rudin PMA (chapter 3 and 8–twice!)

4.1.14 l2 (R) space

P∞
(HWS10E01) The set l2 (R) of all sequences qP in R such that i=1 x2i converges
∞ 2
is complete in hte l2 -metric: d (x, y) := i=1 (xi − yi ) .

• If an ∈ l2 (R) is Cauchy then, then each coordinate separately is Cauchy,

N

and so converges to something. The sequence an then converges to the

point composed of all those coordinates.

60
4.2 §44 Theorem 44.1–The Peano Space-Filling Curve
Let I = [0, 1].


• Claim: ∃f ∈ C I, I 2 such that f (I) = I 2 .’
• Proof: TODO (in handwriting–also in Rudin PMA chapter 8 in exercises)
(HWS10E02)

• There is also a continuous surjective map g : I → I n ∀n ∈ N, but of course

it is not injective, because then it could be made into a homeomorphism,
but I without a point is disconnected and I n isn’t.
• Thus we can create such a map from R → Rn , as a limit of maps into
compact subspaces.

• RN with the product topology has no continuous surjective map from R

to it.

– RN is not the countable union of compact subspaces.

4.3 §45 Compactness in Metric Spaces

4.3.1 Total Boundedness
A metric space (X, d) is totally bounded if ∀ε > 0, ∃ a finite covering by ε-balls.
Observe: Total boundedness implies boundedness.
Observe: Boundedness does not imply total boundedness! In the bounded met-
ric, R is bounded but not totally bounded.

4.3.2 Example 2
Under the metric |a − b|, R is complete, but not totally bounded, whereas
(−1, 1) is totally bounded but not complete. [−1, 1] is both complete and
totally bounded.

4.3.3 Theorem 45.1–Complete Metric Spaces

A metric space (X, d) is compact ⇐⇒ X is complete and totally bounded.
Proof: TODO (in handwriting)

4.3.4 Products of Compact Metric Spaces

(HWS10E03) n o
di (xi , yi )
If (Xi , di ) are metric spaces ∀i ∈ N then D (x, y) := sup i : i ∈ N
Q
is a metric on i∈N Xi . X is totally bounded if each Xi is totally bounded. Thus
by 45.1, since X is complete (TODO), then X is complete and totally bounded,
and so, compact.

61
4.3.5 Equicontinuity
Let (Y, dY ) be a metric space. Let X be a topological space. Let F ⊂ C (X, Y ).
Let x0 ∈ X.
F is equicontinuous at x0 iff ∀ε > 0∃U ∈ Open (X) : x0 ∈ U such that ∀x ∈
U, ∀f ∈ F:

dY (f (x) , f (x0 )) < ε

If F is equicontinuous at x0 ∀x0 ∈ X, then F is equicontinuous.

4.3.6 Lemma 45.2

Let (Y, dY ) be a metric space. Let X be a topological space. Let F ⊂ C (X, Y ).
• Claim: If F is totally bounded is totally bounded under the uniform metric
corresponding to dY , then F is equicontinuous under dY .
• Proof: TODO (in handwriting)

4.3.7 Lemma 45.3

• Assumptions:

1. Let (Y, dY ) be a compact metric space.

2. Let X be a compact topological space.
3. Let F ⊂ C (X, Y ) be equicontinuous under dY .

• Claim: F is totally bounded under the uniform and sup metrics corre-
sponding to dY .
• Proof: TODO (in handwriting)

4.3.8 Pointwise Boundedness

Let (Y, dY ) be a metric space. Let F ⊂ C (X, Y ). F is called pointwise bounded
under dY if ∀x ∈ X, {f (x) | f ∈ F} is bounded under dY .

4.3.9 Theorem 45.4–Ascoli’s Theorem, the Classical Version

Let X be a compact topological space. Let d be the Euclidean metric or the
square metric on Rn . Let (C (X, Rn ) , ρ) be the corresponding metric space with
the uniform metric. Let F ⊂ C (X, Y ).

• Claim: Closure(C(X, Rn ), ρ) (F) is compact ⇐⇒ F is equicontinuous and

pointwise bounded under d.
• Proof: TODO (in handwriting)

62
4.3.10 Corollary 45.5
Let X be a compact topological space. Let d be the square metric or the
Euclidean metric on Rn . Let (C (X, Rn ) , ρ) be the corresponding metric space
with the uniform metric. Let F ⊂ C (X, Y ).
• Claim: F is compact ⇐⇒ F is closed, bounded under the sup metric ρ
and equicontinuous under d.

• Proof: TODO (in handwriting)

4.3.11 Equicontinuity Criterions

(HWS10E04)
Let (Y, d) be a metric space. Let F ∈ 2C(X, Y ) .
• |F| < ∞ =⇒ F is equicontinuous.
N
• If fn ∈ C (X, Y ) converges uniformly, then {fn } is equicontinous (com-
pare with Rudin about compact condition).
• If F is a collection of differentiable functions f ∈ RR such that each x ∈ R
lies in a neighborhood U on which the derivatives of the functions in F
are uniformly bounded, then F is equicontinous.

4.4 §46 Pointwise and Compact Convergence

(HWS07E01)

4.4.1 The Compact-Open Topology

Let X and Y be topological spaces. If C is a compact subspace of X and U ∈
Open (Y ), define S (C, U ) := {f ∈ C (X, Y ) : f (C) ⊂ U }. The sets S (C, U )
form a subbasis for a topology on C (X, Y ) that is called the compact-open
topology.
• C ({∗} , Y ) ' Y

• If X is compact and Y is metrizable, then the compact-open topology is

metrizable.
• If Y is Hausdorff, then the compact-open topology is Hausdorff.

63
Part II
Algebraic Topology
5 Chapter 9 The Fundamental Group
The goal: to be able to show that two given spaces are not homeomorphic.
Thus we need to generate more “topological” properties. One such property is
the free group.
TODO: The “long line”.

• “Two spaces that are homeomorphic have fundamental groups that are
isomorphic.”
• Simple-connectedness means the fundamental group is trivial.
Throughout, I := [0, 1].

5.1 §51 Homotopy of Paths

5.1.1 Groupoid
A group without closure (the composition is not defined on any two elements).

5.1.2 Homotopy of Maps

Let f, f 0 ∈ C (X, Y ).
f is homotopic to f 0 if ∃ F ∈ C (X × I, Y ) such that F (x, 0) = f (x)∧F (x, 1) =
f 0 (x) ∀x ∈ X.
F is called a homotopy between f and f 0 .
f being homotopic to f 0 is denoted by f ' f 0 .

5.1.2.1 Nulhomotopy If f ' f 0 and f 0 is a constant map, then f is nul-

homotopic.

5.1.3 Path Homotopy

Let x0 , x1 ∈ X.
Let f and f 0 be two paths from x0 to x1 .
f and f 0 are path homotopic f ∃F ∈ C I 2 , X such that:

• ∀s ∈ I

– F (s, 0) = f (s)
– F (s, 1) = f 0 (s)

(s can be thought of as “where we are” at a given path)

64
 • ∀t ∈ I

– F (0, t) = x0
– F (1, t) = x1

(t can be thought of as a label for the set of possible paths)

F is a path homotopy between f and f 0 .
f being path homotopic to f 0 is denoted by f 'p f 0 .
Thus a path homotopy between two paths is also a homotopy, but a stronger
one, in the sense that the end points are kept fixed for all the variations of paths
in between (so that they are also paths).

5.1.4 Lemma 51.1–Homotopy and Path Homotopy are Equivalence

Relations
The path homotopy class of f will be denoted by [f ].
Proof: TODO (in handwriting)

5.1.5 Example 1–Straight-Line Homotopy

Let f, g ∈ C X, R2 .
Then f ' g: ∃F (x, t) := (1 − t) f (x) + tg (x), which is a homotopy between f
and g.
If f and g are paths from x0 to x1 , then F is a path homotopy from f to g.

5.1.5.1 A convex set in Rn A convex set in Rn contains the straight line

between any two points inside of it.
If A is a convex subspace of Rn then any two paths f, g in A from x0 to x1
are path homotopic in A, for the straight-line homotopy F between them has
its image set in A.

5.1.5.2 Star Convex (HWS11E01)

A ⊂ Rn is star convex if ∃ a point a0 ∈ A such that all line segments joining
a0 to the other points of A lie in A.
• A star convex set which is not convex is star shaped subset of R2 .
• A star convex set is simply connected.

5.1.6 Example 2–The Punctured Plane

   
2 cos (πs) cos (πs)
Let X := R \ {0}. Then f (s) := and g (s) := are path
sin (πs) 2 sin (πs)
homotopic.
 The straight-line
 homotopy between them exists. But the path
cos (−πs)
h (s) := is not homotopic to f , for example (this is not a proof)
sin (−πs)
the straight line homotopy will necessarily pass through the origin.

65
5.1.7 The Product of Paths
• Let x0 , x1 , x2 ∈ X.

• Let f ∈ C (I, X) such that f (0) = x0 and f (1) = x1 .

• Let g ∈ C (I, X) such that g (0) = x1 and g (1) = x2 .

Define a path from x0 to x2 , denoted by f ∗ g, called the product of the paths

f and g as follows:
(
s ∈ 0, 12
 
f (2s)
(f ∗ g) (s) :=
g (2s − 1) s ∈ 12 , 1
 

f ∗ g ∈ C (I, X) (by the pasting lemma)

5.1.8 The Product of Path Homotopy Classes

Define a composition law between two path homotopy classes [f ] and [g] as
[f ] ∗ [g] := [f ∗ g].
Claim: This product is well-defined.
Proof: TODO (in handwriting).

5.1.9 Theorem 51.2–The Path-Homotopy Classes Form a Groupoid

5.1.9.1 Definition of Identity Map Define ex ∈ C (I, X) by ex (s) :=
x ∀s ∈ I.

5.1.9.2 Definition of the Reverse Map If f is a path from x0 to x1 then

define path from x1 to x0 , denoted by f , called the reverse of f , by f (s) :=
f (1 − s) ∀s ∈ I.

5.1.9.3 Definition of the Positive Linear Map If a, b, c, d ∈ R such that

d−c d−c
a < b ∧ c < d then ∃!p ∈ C ([a, b] , [c, d]) of the form p (x) = b−a · x − a b−a +c
such that p (a) = c ∧ p (b) = d.
p is called the positive linear map from [a, b] to [c, d]. The inverse of a positive
linear map and the composite of two positive linear maps are positive linear
maps.

5.1.9.4 Associativity of ∗ If [f ] ∗ ([g] ∗ [h]) is defined, then ([f ] ∗ [g]) ∗ [h]

is also defined and [f ] ∗ ([g] ∗ [h]) = ([f ] ∗ [g]) ∗ [h].

5.1.9.5 Right and Left Inverses for ∗ If f is a path from x0 to x1 then

[ex0 ] ∗ [f ] = [f ] and [f ] ∗ [ex1 ] = [f ].

66
  
5.1.9.6 Inverse for ∗ If f is a path from x0 to x1 , then [f ] ∗ f = [ex0 ]
 
and f ∗ [f ] = [ex1 ].
Proof: TODO (in handwriting).

5.1.10 Theorem 51.3

n
Let f be a path in X and let {ai }i=1 be n numbers such that 0 = a0 < a1 <
· · · < an = 1. Let fi ∈ C (I, X) be the path that equals the positive linear map
of I onto [ai−1 , ai ] followed by f . Then

[f ] = [f1 ] ∗ · · · ∗ [fn ]
Proof: TODO.

5.2 §52 The Fundamental Group

5.2.1 Loops
Let X be a space, and let x0 ∈ X. A path in X from x0 to x0 is called a loop
based at x0 .

5.2.2 The Fundamental Group

The set of path homotopy classes of loops based at the point x0 ∈ X, together
with the operation ∗, is called the fundamental group of X relative to the base
point x0 . It is denoted by π1 (X, x0 ). Alternatively it is called the first homotopy
group of X.
Given two loops f and g with base at x0 , the product f ∗ g is always defined
and is a loop based at x0 . Thus ∗ is closed on the set of path homotopy classes
and so by 51.2 we have indeed a group.

5.2.3 Example with Topological Groups

(HWS11E04) If G is a topological group then π1 (G, x0 ) is Abelian. This follows
from observing that [f ] ∗ [g] = [m (f (t) , g (t))] where m is the group compo-
sition and also by observing that m ((a ∗ b) (t) , (c ∗ d) (t)) = m (a (t) , c (t)) ∗
m (b (t) , d (t)) and using tricks with the identity element.

5.2.4 Example 1–Rn

Let x0 ∈ Rn be any point. Then π1 (Rn , x0 ) is trivial. This is because the
straight-line homotopy is a path homotopy between any loop based at x0 and
the constant path at x0 .
Generally, if X is any convex subset of Rn , then π1 (X, x0 ) = {[ex0 ]}. This, in
particular, includes the unit ball.

67
5.2.5 Fundamental Groups at Different Base Points
Let α be a path in X from x0 to x1 .
π (X, x0 )
Define a map α̂ ∈ π1 (X, x1 ) 1 by the equation α̂ ([f ]) := [α] ∗ [f ] ∗ [α] for
any [f ] ∈ π1 (X, x0 ).

5.2.6 Theorem 52.1

The map α̂ is a group isomorphism.
Proof: TODO (in handwriting).

5.2.7 Corollary 52.2

If X is path connected and x0 , x1 ∈ X , then π1 (X, x0 ) is isomorphic to
π1 (X, x1 ).
It is easy to verify then that if C is a path component of X containing x0 then
π1 (C, x0 ) = π1 (X, x0 ), since all loops and homotopies in X that are based
at x0 must lie in a subspace of C. Thus π1 (X, x0 ) depends on only the path
component of X containing x0 and gives no information about the rest of X.

5.2.8 Lemma–Exercise 3–Path connected and Abelianness

(HWS11E02)
The isomorphism of π1 (X, x0 ) with π1 (X, x1 ) is independent of path if and
only if the fundamental group is Abelian. That is:
Let x0 , x1 ∈ X, where X is path connected.
π1 (X, x0 ) is Abelian ⇐⇒ ∀α, β paths from x0 to x1 , α̂ = β̂.

5.2.9 Simply Connectedness

A space X is said to be simply connected if it is a path-connected space and
if π1 (X, x0 ) is trivial for some x0 ∈ X (and hence for every x ∈ X). This is
denoted by π1 (X, x0 ) = 0.

5.2.10 Lemma 52.3

If X is simply connected, any two paths from the same initial to final points are
path homotopic.
Proof: TODO (in handwriting).

5.2.11 The Homomorphism Induced by a Continuous Map

π (X, x )
Let h ∈ C (X, Y ) such that h (x0 ) = y0 . Define (hx0 )∗ ∈ π1 (Y, y0 ) 1 0
by
the equation (hx0 )∗ ([f ]) := [h ◦ f ].
The map h∗ is called the homomorphism induced by h, relative to the base point
x0 (once x0 is chosen y0 = h (x0 )).
h∗ is indeed a group homomorphism, because (h ◦ f ) ∗ (h ◦ g) = h ◦ (f ∗ g).

68
5.2.12 Theorem 52.4–Functorial Properties of the Induced Homo-
morphism
Let h ∈ C (X, Y ) such that h (x0 ) = y0 and let k ∈ C (Y, Z) such that k (y0 ) =
z0 .
Then:
• (k ◦ h)∗ = k∗ ◦ h∗ .
• (idX )∗ is the identity homomorphism.
Proof: TODO (in handwriting).

5.2.13 Corollary 52.5

If h : (X, x0 ) → (Y, y0 ) is a homeomorphism of X with Y , then h∗ is a group
isomorphism of of π1 (X, x0 ) with π1 (Y, y0 ).
Proof: TODO (in handwriting).

5.2.14 Example
(HWS11E03)
Let x0 , x1 ∈ X, y0 , y1 ∈ Y . Let h ∈ C (X, Y ) such that h (x0 ) = y0 ∧h (x1 ) =
y1 . Let α be a path from x0 to x1 and define β := h ◦ α (a path from y0 to y1 ).
Then β̂ ◦ h∗ = h∗ ◦ α̂.

5.3 §53 Covering Spaces

5.3.1 Evenly Covered Open Sets
Let p ∈ C (E, B) be surjective. U ∈ Open (B) is evenly covered by p if:
• ∃ {Vα } ⊂ Open (E) such that Vα ∩ Vβ = ∅∀α 6= β.
• p−1 (U ) = α Vα .
S

• ∀α, p|Vα is a homeomorphism of Vα onto U .

{Vα } is called a partition of p−1 (U ) into slices.
Observe: If U is evenly covered by p and W ∈ Open (B) ∩ 2U , then W is also
evenly covered by p.

5.3.2 Covering Map and Covering Space

Let p ∈ C (E, B) be surjective. If ∀b ∈ B∃U ∈ Open (B) : b ∈ U such that U
is evenly covered by p, then p is called a covering map, and E is said to be a
covering space of B.
Observe: If p : E → B is a covering map, then
• ∀b ∈ B, p−1 ({b}) ⊂ E with the subspace topology has the discrete topol-
ogy.
• p is an open map.

69
5.3.3 Example 1
Let X be any space. idX is a covering map.
More generally, let E = X × {1, . . . , n} (n disjoint copies of X). The map
p : E → X given by p (x, i) = x∀i ∈ {1, . . . , n} is a covering map.

5.3.4 Theorem 53.1

 
1 R

cos (2πx)
The map p ∈ S given by the equation p (x) := is a covering
sin (2πx)
map.
Proof: TODO (in handwriting).

5.3.5 Local Homeomorphism

If p : E → B is a covering map, then p is a local homeomorphism of E with B:
Each point e ∈ E has a neighborhood that is mapped homeomorphically by p
onto an open subset of B. The condition that p be a local homeomorphism does
not suffice, however, to ensure that p is a covering map.

1
5.3.5.1 Example 2–Counter
  Example The map p : R+ → S given by
cos (2πx)
the equation p (x) := is surjective, and it is a local homeomorphism.
sin (2πx)
 
1
But it is not a covering map: has no neighborhood U that is evenly covered
0 
1
by p. Each neighborhood of has a preimage that contains an interval of
0
the form (0, ε) (along with neighborhoods of n∀n ∈ N) but the interval is not
homeomorphically mapped onto U , only imbedded in U by p (not surjective).

5.3.6 Example 3
The map p : S 1 → S 1 given by p (z) := z 2 is a covering map.

5.3.7 Theorem 53.2

Let p : E → B be a covering map. If B0 is a subspace of B, and if E0 = p−1 (B0 )
then the map p0 : E0 → B0 obtained by restricting p is a covering map.
Proof: TODO (in handwriting).

5.3.8 Theorem 53.3

If p : E → B and p0 : E 0 → B 0 are covering maps, then p × p0 : E × E 0 → B × B 0
is a covering map.
Proof: TODO

70
5.3.9 Example 4–The Torus
The space T = S 1 × S 1 is called the torus. The product map: p × p : R × R →
S 1 → S 1 is a covering map of the torus by the plane R2 , where p denotes the
covering map of Theorme 53.1.

5.3.10 Example 5–The Figure-Eight Space

     
1 1
B0 := S 1 × ∪ × S 1 is the figure-eight space.
0 0
The space E0 = p−1 (B0 ) = (R × Z) ∪ (Z × R) is the infinite grid. It is only one
covering space of the figure eight.

5.3.11 Example 6
The map p × idR+ : R × R+ → S 1 × R+ where p is the map of theorem 53.1 is
a covering map.
We can thus form a covering of R2 \ {0} by compsing p × idR+ with the standard
homeomorphism of R × R+ with R2 \ {0}.

5.4 §54 The Fundamental Group of the Circle

5.4.1 A Lifting
Let p ∈ B E . If f ∈ C (X, B), a lifting of f is some f˜ ∈ E X such that p ◦ f˜ = f .

5.4.2 Example 1
Let p : R → S 1 be as defined in Theorem 53.1.    
1 cos (πs)
Let f : [0, 1] → S 1 be a path beginning at from given by f (s) = .
0 sin (πs)
f lifts to the path f˜ (s) = 2s beginning at 0 and ending at 12 .

5.4.3 Lemma 54.1

Let p ∈ B E be a covering map, such that p (e0 ) = b0 . Any path f : [0, 1] → B
beginning at b0 has a unique lifting to a path f˜ in E beginning at e0 .
Proof: TODO (in handwriting).

5.4.4 Lemma 54.2

Let p ∈ B E be a covering map, let p (e0 ) = b0 . Let the map F ∈ C I 2 , B be
such that F (0, 0) = b0 .
• Claim:


– ∃! a lifting of F to a continuous map F̃ ∈ C I 2 , E such that
F̃ (0, 0) = e0 .

71
 – If F is a path homotopy, then F̃ is a path homotopy.

• Proof: TODO (in handwriting).

5.4.5 Theorem 54.3

• Assumptions:

– Let p ∈ C (E, B) be a covering map.

– Let p (e0 ) = b0 .
– Let f, g ∈ C (I, B) be two paths from b0 to b1 .
– Let f˜, g̃ be their respective liftings to paths in E beginning at e0 .

• Claim: If f 'p g then f˜ and g̃ end at the same point of E and f˜ 'p g̃.
• Proof: TODO (in handwriting).

5.4.6 Lifting Correspondence

• Assumptions:

– Let p ∈ B E be a covering map.

– Let b0 ∈ B. Choose e0 so that p (e0 ) = b0 .
– Let [f ] ∈ π1 (B, b0 ).
– Let f˜ be the lifting of f to a path in E that begins at e0 .
– Define φ ([f ]) := f˜ (1).

• Claim: φ is a well-defined set map: φ : π1 (B, b0 ) → p−1 ({b0 }), called the
lifting correspondence derived from the covering map p.

5.4.7 Theorem 54.4

• Assumptions:

– Let p ∈ B E be a covering map.

– Let b0 ∈ B. Choose e0 so that p (e0 ) = b0 .

• Claims:

– E is path connected =⇒ φ is surjective.

– E is simply connected =⇒ φ is bijective.

• Proof: TODO (in handwriting).

72
 5.4.8 Theorem 54.5

The fundamental group of S 1 is isomorphic to (Z, +).

Proof: TODO (in handwriting).
5.4.9 Theorem 54.6
• Assumptions:

– Let p ∈ B E be a covering map.

– Let b0 ∈ B. Choose e0 so that p (e0 ) = b0 .

• Claim:

– The homomorphism induced by p is injective.

– Define H := p∗ (π1 (E, e0 )). The lifting correspondence φ induces an
injective map:
Φ : π1 (B, b0 ) /H → p−1 ({b0 }) of the collection of right cosets of H
into p1 ({b0 }), which is bijective if E is path connected.
– If f is a loop in B based at b0 , then [f ] ∈ H ⇐⇒ f lifts to a loop
in E based at e0 .

5.5 §55 Retractions and Fixed Points

5.5.1 A Retraction
If A ∈ 2X , a retraction of X onto A is r ∈ C (X, A) (surjective) such that
r|A = idA .

5.5.2 A Retract of a Space

If A ∈ 2X and ∃ a retraction of X onto A, then A is a retract of X.

5.5.3 Lemma 55.1

If A is a retract of X, then the homomorphism of fundamental groups induced
by inclusion j : A → X is injective.
Proof: TODO (in handwriting).

5.5.4 Theorem 55.2–No-Retraction Theorem

@ retraction of B 2 onto S 1 , where B 2 is the closed unit ball in R2 .
Proof: TODO (in handwriting).

73
5.5.5 Theorem 55.3


Let h ∈ C S 1 , X .

• Claim: The following conditions are equivalent:

1. h is nulhomotopic.


2. h extends to some k ∈ C B 2 , X .
3. h∗ is the trivial homomorphism of fundamental groups.

• Proof: TODO (in handwriting).

5.5.6 Corollary 55.4

• The inclusion map j : S 1 → R2 \ {0} is not nulhomotopic.
• idS 1 is not nulhomotopic.

Proof: TODO (in handwriting).

5.5.7 Theorem 55.5

Given a nonvanishing vector field on B 2 , v : B 2 → R2 \ {0}, ∃ z ∈ S 1 where the
vector field points directly inward and ∃z 0 ∈ S 1 where the vector field points
directly outward.
Proof: TODO (in handwriting).

5.5.8 Theorem 55.6–Brouwer fixed-point theorem for the disc

If f ∈ C B 2 , B 2 then ∃x ∈ B 2 such that f (x) = x.
Proof: TODO (in handwriting).

5.5.9 Theorem 55.7–A Classical Theorem by Frobenius

Let A ∈ M at3×3 (R+ ). Then A has a positive real eigenvalue (characteristic
value).
Proof: TODO (in handwriting).

5.5.10 Theorem 55.8


Let T = (x, y) ∈ R2 : x ≥ 0 ∧ y ≥ 0 ∧ x + y ≤ 1 .
• Claim: ∃ε > 0 such that ∀ open covering A of T by sets of diameter less
than ε, some point of T belongs to at least three elements of A.
• Proof: TODO.

74
5.6 §58 Deformation Retracts and Homotopy Type
5.6.1 Lemma 58.1
Let h, k ∈ C (X, Y ) be such that h (x0 ) = y0 ∧ k (x0 ) = y0 . If h and k are
homotopic, and if the image of the base point x0 of X remains fixed at y0
during the homotopy, then the homomorphisms h∗ and k∗ are equal.

5.6.2 Theorem 58.2

The inclusion map j : S n → Rn+1 \ {0} induces an isomorphism of fundamental
groups.
Proof: TODO (in handwriting).

5.6.3 Deformation Retract

Let A be a subspace of X. A is a deformation retract of X if idX is homotopic
to a map that carries all of X into A, such that each point of A remains fixed
during the homotopy.
That is, ∃H ∈ C (X × I, X) such that H (x, 0) = x and H (x, 1) ∈ A ∀x ∈ X,
and H (a, t) = a ∀a ∈ A.
The homotopy H is called a deformation retraction of X onto A, and H is a
homotopy between idX and j ◦ r where j is the inclusion map.

5.6.4 Theorem 58.3

Let A be a deformation retract of X; Let x0 ∈ A. The the inclusion map
j : (A, x0 ) → (X, x0 ) induces an isomorphism of fundamental groups.
Proof: As in 58.2.

5.6.5 Example 1


R3 \ (0, 0, z) ∈ R3 : z ∈ R has the punctured xy-plane, R2 \ {0} × 0, as a
deformation retract. The map H (x, y, z, t) := (x, y, (1 − t) z) is a deformation
retraction.

5.6.6 Example 2–The Doubly Punctured Plane

R2 \ {p} \ {q}, the doubly punctured plane, has the “figure eight” space as a
deformation retract.

5.6.7 Example 3–Theta Space

The doubly punctured plane has the theta space as a deformation retract: θ :=
S 1 ∪ (0 × [−1, 1]).

75
5.6.8 Homotopy Equivalences and Homotopy Inverses
Let f ∈ C (X, Y ) and g ∈ C (Y, X). Suppose g ◦ f ' idX and f ◦ g ' idY . Then
the maps f and g are called homotopy equivalences, and each is said to be a
homotopy inverse of the other.

5.6.9 Homotopy Equivalent Spaces

Given two spaces X and Y , if there are homotopy equivalences between X and
Y then X and Y are called homotopy equivalent, or of the same homotopy type.
Homotopy equivalence is an equivalence relation.

• Observe: If A is a deformation retract of X, then A has the same homotopy

type as X.
• Proof: TODO (in handwriting).

5.6.10 Lemma 58.4

Let h, k ∈ C (X, Y ), and assume that h (x0 ) = y0 and k (x0 ) = y1 . If h ' k,
∃ a path α in Y from y0 to y1 such that k∗ = α̂ ◦ h∗ . In particular, if H ∈
C (X × I, Y ) is the homotopy between h and k, then α is given by α (t) =
H (x0 , t).

5.6.11 Corollary 58.5

Let h, k ∈ C (X, Y ) be homotopic. Let h (x0 ) = y0 and k (x0 ) = y1 . If h∗ is
injective, or surjective, or trivial, so is k∗ .
Proof: TODO (in handwriting).

5.6.12 Corollary 58.6

Let h ∈ Y X . If h is nulhomotopic, then h∗ is the trivial homomorphism.
Proof: TODO (in handwriting).

5.6.13 Theorem 58.7

Let f ∈ C (X, Y ) and let f (x0 ) = y0 . If f is a homotopy equivalence, then
f∗ : π1 (X, x0 ) → π1 (Y, y0 ) is an isomorphism.
Proof: TODO (in handwriting).

5.7 §59 The Fundamental Group of S n

5.7.1 Theorem 59.1–Special Case of the Seifert-van Kampen Theo-
rem
Suppose X = U ∪ V where U, V ∈ Open (X). Suppose that U ∩ V is path
connected, and that x0 ∈ U ∩ V . Let i and j be the inclusion mappings of U
and V , respectively, into X.

76
 • Claim: The images of the induces homomorphisms i∗ : π1 (U, x0 ) →
π1 (X, x0 ) and j∗ : π1 (V, x0 ) → π1 (X, x0 ) generate π1 (X, x0 ).
• Proof: TODO (in handwriting).

5.7.2 Corollary 59.2

Suppose X = U ∪ V where U, V ∈ Open (X) and suppose that U ∩ V 6= ∅ is
path connected. If U and V are simply connected, then X is simply connected.

5.7.3 Theorem 59.3–S n is simply connected for n ≥ 2

Proof: TODO (in handwriting).

5.8 §60 Fundamental Groups of Some Surfaces

5.8.1 Theorem 60.1–Fundamental Groups Product Spaces
  
x ∼
• Claim: π1 X × Y, 0 = π1 (X, x0 ) × π1 (Y, y0 )
y0
• Proof: TODO (in handwriting).

5.8.2 Corollary 60.2–Fundamental Group of the Torus is isomorphic

to that of Z2
The torus is defined as T = S 1 × S 1 . Since π1 S 1 , z0 ∼


= Z, the result follows.

5.8.3 Antipodal Points

(HWS11E05)
The map σ : S 1 → S 1 sending any point to its antipodal point is the identity
homomorphism.

5.8.4 The Projective Plane

The projective plane P 2 is the quotient space obtained from S 2 by identifying
each point x ∈ S 2 with its antipodal point −x.

5.8.5 Theorem 60.3

The projective plane P 2 is a compact surface, and the quotient map p : S 2 → P
is a covering map.
Proof: TODO (in handwriting).

5.8.6 Corollary 60.4

π1 P 2 , y = 2
Proof: TODO (in handwriting).
Observe: Any group of order 2 is isomorphic to Z/2.

77
5.8.7 The Projective n-Space
Let n ∈ N. Define P n , the projective n-space, as the space obtained from S n
by identifying each point x with its antipode −x.
Note: Theorem 60.3 applies on P n without change. (HWS11E06)
Note: S n is simply connected ∀n ≥ 2, |π1 (P n , y)| = 2.

5.8.8 Lemma 60.5–The Figure Eight Fundamental Group

The fundamental group of the figure eight is not Abelian.
Proof: TODO (in handwriting).
(HWS12E01)
The fundamental group of the figure eight is in fact F2 (the free group of
two symbols). For a bouquet of n circles, the fundamental group is Fn (using
van Kampen theorem).

5.8.9 The Hawaiian Earring

  
S x 2 2 1 2

1
The Hawaiian Earring is defined as H := n∈N ∈ R : x + y − n = n2
y
endowed with the subspace topology. Observe that the Hawaiian earring is not
homeomorphic to B∞ (the Hawaiian earring is compact yet B∞ is not: the
complement of x0 is open, together with a small neighborhood of x0 is an open
cover with no finite subcover). Alternatively, π1 (H, x0 ) is uncountable, whereas
F∞ is countable.

5.8.10 Theorem 60.6–The Double Torus Fundamental Group

The double torus, T #T , is the surface obtained by taking two copies of the
torus, deleting a small open disc from each of them, and pasting the remaining
pieces together along their edges.
• Claim: The fundamental group of the double torus is not Abelian.
• Proof: TODO (in handwriting).

– Claim: The figure eight, X, is a retract of T #T .

– Proof: TODO (in handwriting).
– Observe: This implies that the inclusion map j : X → T #T induces
an injective homomorphism j∗ , so that π1 (T #T, x0 ) is not Abelian.

5.8.11 Corollary 60.7

The 2-sphere, torus, projective plane, and double torus are topologically dis-
tinct.

78
6 Chapter 11–The Seifert-van Kampen Theorem
6.1 §67 Direct Sums of Abelian Groups
Let G be an Abelian group and let {Gα }α∈J be a sequence of subgroups in it.

6.1.1 Generators
{Gα }α∈J generates G if ∀g ∈ G:
1. ∃ n ∈ N
n
2. ∃ a sequence {αi }i=1 such that αi ∈ J∀i ∈ {1, . . . , n}.

3. ∃ gαi ∈ Gαi ∀i ∈ {1, . . . , n}

• Such that g = gα1 gα2 · · · gαn .

In this case we say G is the sum of {Gα }.

Note: Since G is Abelian, then there is no significance for the order of {αi }.
Thus if αi = αj for some i 6= j ∈ {1, . . . , n} we could rearrange the sequence
so that this is never the case.

6.1.2 Direct Sums

Let {Gα } generate G. If ∀g ∈ G:
1. ∃! n ∈ N
n
2. ∃! a sequence {αi }i=1 such that αi ∈ J∀i ∈ {1, . . . , n}.
3. ∃! gαi ∈ Gαi ∀i ∈ {1, . . . , n}

• Such that g = gα1 gα2 · · · gαn .

L
then G is the direct sum of {Gα }. This is denoted by G = α∈J Gα .

6.1.3 Example 1–RN

Observe that RN is an Abelian group under the operation of coordinate-wise
addition. 
Define Gn := {xi }i∈N : xi = 0∀i 6= n . This is a subgroup of RN which hap-
pens to be isomorphic to R.
Observe: {Gn }n∈N generates the subgroup R∞ of RN . In fact, R∞ = n∈N Gn .
L

6.2 §68 Free Products of Groups

Let G be a (not necessarily Abelian) group. Let {Gα }α∈J be a family of sub-
groups of G.

79
6.2.1 Generators
{Gα }α∈J generates G if ∀g ∈ G:

1. ∃ n ∈ N
n
2. ∃ a sequence {αi }i=1 such that αi ∈ J∀i ∈ {1, . . . , n}.
3. ∃ gαi ∈ Gαi ∀i ∈ {1, . . . , n}

• Such that g = gα1 gα2 · · · gαn .

n
6.2.1.1 Words The sequence {gαι }i=1 is called a word of length n in the
groups {Gα }. It is said to represent g ∈ G.
Note:
1. If for some i ∈ {1, . . . , n}, gαi = e, then we could omit this term from the
word, thereby obtaining a word of length n − 1.

2. If for some i ∈ {1, . . . , n}, αi = αi+1 , then since Gαi is a group, gαi gαi+1 ∈
Gαi and so we can redefine the word: gα0 i := gαi gαi+1 and omit gαi+1 .
n
If we apply the two rules repeatedly we obtain a reduced word: {gαi }i=1 where
(gαi 6= e ∧ αi 6= αi+1 ) ∀i ∈ {1, . . . , n}.
We define the representation of e to be ∅ by convention. Thus if {Gα }α∈J
generates G, then every element of G can be represented by a reduced word in
the elements of the groups {Gα }α∈J .
n m
( Note: If {xi }i=1 represents x and {yi }i=1 represents y then define zi =
xi i≤n n+m
∀i {1, . . . , n + m}. In this case {zi }i=1 represents xy. However,
yi−n i > n
n+m n
do note that {zi }i=1 is not necessarily a reduced word, even if both {xi }i=1
m n+m
and {yi }i=1 are reduced words. The condition for {zi }i=1 to be a reduced word
is that @α in the list of two words such that xn , y1 ∈ Gα .

6.2.2 Free Product

Let G be a group, and let {Gα }α∈J be a family of subgroups of G that generates
G. Suppose that Gα ∩ Gβ = {e} ∀α 6= β. G is the free product of the groups
Gα if ∀x ∈ G, ∃! reduced word in the groups Gα that represents x. In that case
we write:
∗
Y
G= Gα
α∈J

Or in the finite case G = G1 ∗ · · · ∗ Gn .

80
6.3 §69 Free Groups
Let G be a group, let {aα } be a family of elements of G, ∀α ∈ J. We say {aα }
generate G if every element of G can be written as a product of powers of the
elements aα . If the family {aα } is finite, we say G is finitely generated.

6.3.1 Free Group

Let G be a group, let {aα } be a family of elements of G, ∀α ∈ J. Suppose each
aα generates an infinite cyclic subgroup Gα of G. If G is the free product of the
groups {Gα }, then G is said to be a free group, and the family {aα } is called a
system of free generators for G.

6.4 §70 The Seifert-van Kampen Theorem

6.4.1 Theorem 70.1–The Seifert-van Kampen Theorem
• Assumptions:

1. U, V ∈ Open (X)
2. X = U ∪ V
3. U, V, U ∩ V are path connected.
4. x0 ∈ U ∩ V
5. H is any group.
6. φ1 ∈ H π1 (U, x0 ) is a homomorphism.
7. φ2 ∈ H π1 (V, x0 ) is a homomorphism.
8. i1 is the inclusion map of U ∩ V in U , (i1 )∗ is the induced homomor-
phism.
9. i2 is the inclusion map of U ∩ V in V , (i2 )∗ is the induced homomor-
phism.
10. j1 is the inclusion map of U in X, (j1 )∗ is the induced homomorphism.
11. j2 is the inclusion map of V in X, (j2 )∗ is the induced homomorphism.

• Claim: If φ1 ◦i1 = φ2 ◦i2 =⇒ ∃! a unique homomorphism Φ : π1 (X, x0 ) →

H such that Φ ◦ j1 = φ1 ∧ Φ ◦ j2 = φ2 .
• Proof: TODO.

6.5 §71 The Fundamental Group of a Wedge of Circles

6.5.1 A Wedge of Circles
Sn
Let X be a Hausdorff space such that X = i=1 Si where each Si is homeo-
morphic to S 1 (the unit circle). Assume that there is a point p ∈ X such that
Si ∩ Sj = {p} ∀i 6= j. Then X is called a wedge of the circles S1 , . . . , Sn .

81
7 Chapter 13–Classification of Surfaces
Convention: throughout this chapter, p : E → B is a covering map means
that E and B are locally path connected and path connected, unless otherwise
stated.

7.1 §79 Equivalence of Covering Spaces

7.1.1 Equivalent Covering Spaces
Let p ∈ C (E, B) and p0 ∈ C (E 0 , B) be two covering maps. They are said to
be equivalent if ∃ a homeomorphism h : E → E 0 such that p = p0 ◦ h. The
homeomorphism h is called an equivalence of covering maps or an equivalence
of covering spaces.

7.1.2 Lemma 79.1–The General Lifting Lemma

Let p : E → B be a covering map; Let p (e0 ) = b0 . Let f : Y → B be a
continuous map, with f (y0 ) = b0 . Suppose Y is path connected and locally
path connected. The map f can be lifted to a map f˜ : Y → E such that
f˜ (y0 ) = e0 if and only if f∗ (π1 (Y, y0 )) ⊂ p∗ (π1 (E, e0 )).
If such a lifting exists, it is unique.

7.1.3 Theorem 79.2

Let p : E → B and p0 : E 0 → B be covering maps; Let p (e0 ) = p0 (e00 ) = b0 .
∃ an equivalence h : E → E 0 such that h (e0 ) = e00 if and only if the groups
0
H0 = p∗ (π1 (E, e0 )) and H00 = p∗ (π1 (E 0 , e00 )) are equal. If h exists, it is unique.

7.2 §80 The Universal Covering Space

Suppose p : E → B is a covering map, with p (e0 ) = b0 . If E is simply connected,
then E is called a universal covering space of B. By theorem 79.4, any two
universal covering spaces of B are equivalent.

7.2.1 Lemma 80.1

Let B be path connected and locally path connected. Let p : E → B be a cover-
ing map in the former sense (so that E is not required to be path connected). If
E0 is a path component of E, then the map p0 : E0 → B obtained by restricting
p is a covering map.

7.2.2 Lemma 80.2

Let p ∈ C (X, Z), q ∈ C (X, Y ) and r ∈ C (Y, Z), such that p = r ◦ q.
1. If p and r are covering maps, so is q.
2. If p and q are covering maps, so is r.

82
7.2.3 Theorem 80.3
Let p : E → B be a covering map, with E being simply connected. Given any
covering map r : Y → B, ∃ a covering map q : E → Y such that r ◦ q = p.

7.2.4 Lemma 80.4

Let p : E → B be a covering map. Let p (e0 ) = b0 . If E is simply connected,
then b0 has a neighborhood U such that inclusion i : U → B induces the trivial
homomorphism i∗ : π1 (U, b0 ) → π1 (B, b0 ).

7.2.5 Example 1–The Hawaiian Earring

  
x
2
2
Cn := ∈ R2 : x − n1 + y 2 = n1
S y
X :=  n∈N
 Cn
0
b0 :=
0
• Claim: If U is any neighborhood of b0 in X, then the homomorphism of
fundamental groups induced by inclusion i : U → X is not trivial.

7.3 §81 Covering Transformations (Deck Transformations)

A covering transformation is the set of all equivalences of a covering map p :
E → B with itself. Composites and inverses of covering transformations are
covering transformations, so this set forms a group, called the group of covering
transformations, denoted by C (E, p, B).
In this section we assume that p : E → B is a covering map, with p (e0 ) = b0
and H0 := p∗ (π1 (E, e0 )).

7.3.1 Defintion

 ≤ G, then the
H normalizer of H in G is the subset of G defined by N (H) :=
g ∈ G : gHg −1 = H .
• Claim: N ≤ G.

• HEN
• N is the largest such subgroup of G.

7.3.2 Definition
Given p : E → B with p (e0 ) = b0 , let F := p−1 ({e0 }). Let Φ : π1 (B, b0 ) /H0 →
F be the lifting correspondence of theorem 54.6. It is a bijection. Define also
a correspondence Ψ : C (E, p, B) → F by setting Ψ (h) := h (e0 ) ∀ covering
transformation h : E → E. Since h is uniquely determined once its value at e0
is known, the correspondence Ψ is injective.

83
7.3.3 Lemma 81.1
The image of the map Ψ equals the image under Φ of the subgroup N (H0 ) /H0
of π1 (B, b0 ) /H0 .

7.3.4 Theorem 81.2

The bijection Φ−1 ◦ Ψ : C (E, p, B) → N (H0 ) /H0 is an isomorphism of groups.

7.3.5 Corollary 81.3

H0 E π1 (B, b0 ) ⇐⇒ ∀ pair of points e1 and e2 of p−1 ({b0 }), ∃ a covering
transformation h : E → E such that h (e1 ) = e2 . In this case ∃ an isomorphism
Φ−1 ◦ Ψ : C (E, p, B) → π1 (B, b0 ) /H0 .

7.3.6 Corollary 81.4

Let p : E → B be a covering map. If E is simply connected, then C (E, p B) ∼ =
π1 (B, b0 ). If H0 E π1 (B, b0 ) then p : E → B is called a regular covering map.

7.3.7 Definition–The Orbit

Let X be a space, and let G be a subgroup of the group of homemorphisms of
X with itself. The orbit space X/G is defined to be the quotient space obtained
from X by means of th equivalence relation x ∼ g (x) ∀x ∈ X, ∀g ∈ G. The
equivalence class of x is called the orbit of x.

7.3.8 Definition–Properly Discontinuous

(HWS11E06)
If G is a group of homemorphisms of X, the action of G on X is said
to be properly discontinuous if ∀x ∈ X ∃U ∈ Open (X) : x ∈ U such that
g (U ) ∩ U = ∅ whenever g 6= eG . It follows that g0 (U ) ∩ g1 (U ) = ∅ whenever
g0 6= g1 , for otherwise U and g0−1 g1 (U ) would not be disjoint.

7.3.9 Theorem 81.5

Let X be path connected and locally path connected. Let G be a group of
homeomoprhims of X. The quotient map π : X → X/G is a covering map if
and only if the action of G is properly discontinuous. In this case, the covering
map π is regular and G is its group of covering transformations.

7.3.10 Theorem 81.6

If p : X → B is a regular covering map and G is its group of covering trans-
formations, then ∃ a homeomorphism k : X/G → B such that p = k ◦ π where
π : X → X/G is the projection.

84