Topology

FIN
Contents

1 Topology Spaces 2
1.1 Topological Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Basis of a Given Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Topologizing of Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Elementary Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.5 Topologizing with Preassigned Elementary Operations . . . . . . . . . . . . . . . . . . . . . . 5
1.6 Gδ , Fσ , and Borel Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.7 Relativization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.8 Continuous Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.9 Piecewise Definition of Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.10 Continuous Maps into R . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

1
Chapter 1

Topology Spaces

1.1 Topological Spaces

Definition 1.1.1. Let X be a set. A topology in X is a family T of subsets of X such that:
(1) it is closed under arbitrary unions and is closed under finite intersections.
(2) ∅, X ∈ T .
A pair (X, T ) consisting of a set X and a topology T in X is called a topological space. The elements of
topological spaces are called points. The members of T are called the open sets of the topological space
Instead of “topological space (X, T )”, we also say “T is the topology of the space X”, or “X carries
topology T ”. When it is not necessary to specify T explicitly, we simply sat “X is a space”.
Definition 1.1.2. Let (X, T ) be a space. By a neighborhood (written nbd) of an x ∈ X is an open set
containing x. “U is a nbd of x” is written “nbd U (x)”; the points of U (x) are U -close to x.
A set X may have many topologies; with each, it is a distinct topological space. By regarding each
topology as a subset of P(X); the topologies in X are partially ordered by inclusion.
Definition 1.1.3. Suppose (X, T0 ) and (X, T1 ) are two topological spaces. We say that T1 is finer (or T0 is
coarser) if T0 ⊆ T1 .
T
Let {Tα : α ∈ A} be any family of topologies in X. Then it is, a fortiori, that α∈A Tα = {U : ∀α ∈
A, U ∈ Tα } is also a topology in X.

1.2 Basis of a Given Topology

Definition 1.2.1. Let (X, T ) be a topological space. A family B ⊆ T is called a basis for T if each open
set is the union of members of B.
Theorem 1.2.2 (characterization of bases). Let B ⊆ T . The following two properties of B are equivalent:
(1) B is a basis for T .
(2) For each G ∈ T and each x ∈ G, there is a U ∈ B with x ∈ U ⊆ G.
Proof. (1) ⇒ (2). Let x ∈ G. Since G is the union of members of B, there exists U ∈ B such
S that x ∈ U ⊆ G.
(2) ⇒ (1). Suppose G ∈ T . For each x ∈ G, find Ux ∈ B with x ∈ Ux ⊆ G; then G = {Ux : x ∈ G}.
Corollary 1.2.2.1. Let B ⊆ T be a basis for T . Then G ⊆ X is open if and only if for any x ∈ G, there
exists U ∈ B with x ∈ U ⊆ G.
Proof. If G
S is open, the condition follows from 1.2.2. Conversely, if the condition holds, then (as in 1.2.2) we
find G = {Uα : α ∈ A}, where each Uα ∈ B ⊆ T , Therefore, G is open from the definition of topological
spaces.

2
1.3 Topologizing of Sets
Since arbitrary intersections of topologies is still a topology and P(X) is a topology, for any collection
C ⊆ P(X), there exists the smallest topology T (C) containing C, that is, the intersection of all topologies
containing C, called the topology generated by C. C is called a subbasis for T (C). Conversely, for a
topology T , a collection C ⊆ T is called a subbasis for T whenever T = T (C).

Theorem 1.3.1. Let C = {Aα : α ∈ A} ⊆ P(X) be a collection. Then T (C) can be described as below: It
consists of ∅, X, all finite intersections of the Aα , and all arbitrary union of these finite intersections.
S
Proof. It is clear that T (C) contains those sets. On the other hand, because α distributes over ∩, the sets
listed actually do form a topology containing C, and which therefore contains T (C).
The construction of a topology from a subbasis loses some control over the open sets; they build up from
the finite intersections of the Aα rather than from the Aα themselves. In the second general method for
topologizing a set, which we will now describe, the open sets are constructed only by union from the given
family; that is, by specifying a family to be used as a basis for constructing the topology. Since intersections
are involved in topologies but no intersections are involved in forming open sets from a basis, it is to be
expected that not every family can serve as the basis for some topology.

Theorem 1.3.2. Let B ⊆ P(X) be a collection. If for each U1 , U2 ∈ B and x ∈ U1 ∩ U2 , there exists U3 ∈ B
such that x ∈ U3 ⊆ U1 ∩ U2 , then T (B) consists of ∅, X and all union of members of B.
S S
Proof. It is clear that the sets listed isScontainedSin T (B). On
S the other hand, suppose α1 Uα1 and α2 Uα2
are unions of members of B. Then α1 Uα1 ∩ α2 Uα2 = (α1 ,α2 ) Uα1 ∩ Uα2 and each Uα1 ∩ Uα2 and be
written as the union of members of B. Therefore, the mentioned sets forms a topology of B (it is trivial that
the collection consisting these sets is closed under arbitrary union) and thus contains T (C).
Remark. If a collection B satisfying the collection of 1.3.2, then we say that B is a basis for topology.
Similarly, for a topology T , a basis for topology B is a basis for T if T = T (B).
To specification of a topology by giving a basis generally accomplished by specifying for each x ∈ X a
family of nbds {Uα (x) : α ∈ A(x)} and verifying that the family B = {Uα (x) : α ∈ A(x), x ∈ X} is a basis
for topology.
Definition 1.3.3. Two bases B, B ′ in X are equivalent if T (B) = T (B ′ ).

Theorem 1.3.4. Two bases B, B ′ in X are equivalent if and only if both the following conditions hold:
(1) For each U ∈ B and each x ∈ U , there is a U ′ ∈ B ′ with x ∈ U ′ ⊆ U .
(2) For each U ′ ∈ B ′ and each x ∈ U ′ , there is a U ∈ B ′ with x ∈ U ⊆ U ′ .
Proof. We claim a stronger statement: T (B ′ ) ⊆ T (B) if and only if each U ′ ∈ B ′ and each x ∈ U ′ , there
exists U ∈ B such that x ∈ U ⊆ U ′ . Suppose T (B ′ ) ⊆ T (B). Then B ′ ⊆ T (B) and every U ′ ∈ B ′ can be
written as a union of members of B by 1.3.2. On the other hand, if every U ′ ∈ B ′ can be written as a union
of members of B, then B ′ ⊆ T (B) and therefore T (B ′ ) ⊆ T (B).

1.4 Elementary Concepts

Throughout this section, we consider a fixed space X and give some definitions.
Definition 1.4.1. F ⊆ X is closed if X\F is open.

By De Morgan’s law, we have the following:

(1) The intersection of any family of closed sets is a closed set.
(2) The union of finitely many closed sets is a closed set.

3
By (1), for any S ⊆ X, there is the smallest closed set S containing S, called the closure of S. Similarly,
since arbitrary union of open sets is open, there is the largest open set int S contained in S, which is the
interior of S. From the definitions, it is immediate that:
(1) F ⊆ X is closed if and only if F = F .

(2) G ⊆ X is open if and only if int G = G.

(3) X\ int (X\S) = S for all S ⊆ X.
(4) If A, B ⊆ X with A ⊆ B, then A ⊆ B and int A ⊆ int B.
(5) If A, B ⊆ X, then A ∪ B = A ∪ B and int(A ∩ B) = int A ∩ int B.

Proposition 1.4.2. Suppose S ⊆ X. Then x ∈ S if and only if for any nbd U (x), U ∩ S ̸= ∅.
Proof. x ∈
/ S ⇐⇒ x ∈ int (X\S) ⇐⇒ ∃nbd U (x) s.t. U ⊆ X\S.
Definition 1.4.3. For S ⊆ X, the boundary ∂A is S ∩ X\S.
From the definition and 1.4.2, it is clear that:

(1) ∂S = ∂(X\S).
(2) x ∈ ∂S if and only if for any nbd U (x), U ∩ S ̸= ∅ and U ∩ (X\S) = ∅.
(3) ∂S = S\ int S.

(4) X = int S ∪ ∂S ∪ int(X\S) is a pairwise disjoint union.

(5) S = S ∪ ∂S.
Proposition 1.4.4. Suppose S ⊆ X. Then ∂(∂S) ⊆ ∂S and the equality holds if and only if int(∂S) = ∅.
Proof. ∂(∂S) = ∂S ∩ X\∂S = ∂S ∩ X\∂S ⊆ ∂S since ∂S is closed. Suppose int(∂S) = ∅, then X =
X\ int(∂S) = X\∂S. Thus, ∂(∂S) = ∂S ∩ X = ∂S. On the other hand, if int(∂S) ̸= ∅, then there exists
x ∈ ∂S and a nbd U (x) such that x ∈ U ⊆ ∂S. Then X\∂S ⊆ X\U and therefore X\∂S ⊆ X\U as X\U is
closed. Thus, x ∈
/ X\∂S, x ∈
/ ∂(∂S), and ∂(∂S) ⊊ ∂(S).
Corollary 1.4.4.1. If S ⊆ X is open or closed, then ∂(∂S) = ∂S.

Proof. By 1.4.4, it suffices to show that int(∂S) = ∅. If S is closed, then

X\ int(∂S) = X\∂S = int S ∪ int(X\S)

= int S ∪ X\S
= int S ∪ X\S ∪ ∂(X\S)
⊇ int S ∪ int(X\S) ∪ ∂S = X.

Thus, int(∂S) = ∅. If S is open, then X\S is closed and ∅ = int(∂(X\S)) = int(∂S).

Definition 1.4.5. D ⊆ X is dense if D = X.

Proposition 1.4.6. G ⊆ X is open if and only if G ∩ S ⊆ G ∩ S for all S ⊆ X.

Proof. Suppose G∩S = G ∩ S for all S ⊆ X. Take S = X\G. Then G∩X\G = ∅ and therefore G∩∂G = ∅.
Since X = int G ∪ ∂G ∪ int(X\G), G = (G ∩ int G) ∪ (G ∩ ∂G) ∪ (G ∩ int(X\G)) = G ∩ int G. Thus, G = int G
| {z } | {z }
=∅ =∅
and therefore G is open. Conversely, suppose G is open. Let S ⊆ X be given and x ∈ G ∩ S. Then for any
nbd U (x), U ∩ G ∩ S ̸= ∅ =⇒ x ∈ G ∩ S.

Corollary 1.4.6.1. G ⊆ X is open if and only if G ∩ S = G ∩ S for all S ⊆ X.

Corollary 1.4.6.2. If D ⊆ X is dense, then G = G ∩ D for all open G ⊆ X.

4
Proposition 1.4.7. The following four statements are equivalent:
(1) D is dense in X.
(2) If F is any closed set in X and D ⊆ F , then F = X.
(3) Each nonempty open set in X contains an element of D.
(4) The complement of D has empty interior.
Proof. (1) ⇒ (2). Since D = X, X = D ⊆ F = F .
(2) ⇒ (3). Suppose U ⊆ X is nonempty and open. If U ∩ D = ∅, then D ⊆ X\U . However, X\U is
closed and X\U ̸= X, a contradiction.
(3) ⇒ (4). Suppose U ⊆ X\D is open. Then U = ∅ and therefore int(X\D) = ∅.
(4) ⇒ (1). X = X\ int(X\D) = D.

1.5 Topologizing with Preassigned Elementary Operations

Theorem 1.5.1. Let X be a set and cl : P(X) → P(X) be a map with the properties:
(1) cl(∅) = ∅.
(2) S ⊆ cl(S) for each S ⊆ X.
(3) cl is idempotent.
(4) cl(A ∪ B) = cl(A) ∪ cl(B) for each A, B ⊆ X.
Then T = {X\cl(S) : S ∈ P(X)} is a topology, and S = cl(S) for each S ⊆ X.
Proof. Since cl(∅) = ∅ and cl(X) = X (by (2)), ∅, X ∈ T .
Suppose {Sα : α ∈ A} ⊆ T . Then
[ \
X\ Sα = X\Sα ⊆ X\Sα
α∈A α∈A

for any α ∈ A. Then

! !
[ \ (3) \ (2) [ [
cl X\ Sα ⊆ cl(X\Sα ) = X\Sα =⇒ X\ Sα = cl X\ Sα .
α∈A α∈A α∈A α∈A α∈A
S S

Therefore, α∈A Sα = X\cl X\ α∈A Sα ∈ T .
Suppose A, B ∈ T . Then X\(A ∩ B) = (X\A) ∪ (X\B). There exist C, D ⊆ X such that cl(C) = X\A
(4)
and cl(D) = X\B. Therefore, X\(A ∩ B) = cl(C) ∪ cl(D) = cl(C ∪ D) and A ∩ B = X\cl(C ∪ D) ∈ T .
Finally, in T , cl(S) is closed for all S ⊆ X. Therefore, by (2), S ⊆ cl(S). To show that cl(S) ⊆ S, note
(3)
that since X\S is open, we have S = cl(A) for some A ⊆ X. Then cl(S) ⊆ cl(S) = cl(A) = S. (For A ⊆ B,
(4)
we have cl(B) = cl(B ∪ A) = cl(B) ∪ cl(A) =⇒ cl(A) ⊆ cl(B).)
Theorem 1.5.2. Let X be a set and (·)◦ : P(X) → P(X) be a map such that
(1) X ◦ = X.
(2) S ◦ ⊆ S for all S ⊆ X.
(3) (·)◦ is idempotent.
(4) (A ∩ B)◦ = A◦ ∩ B ◦ for all A, B ⊆ X.
Then T = {S ◦ : S ∈ P(X)} is a topology and S ◦ = int S for each S ⊆ X.
Proof. Set η : P(X) → P(X) by η(S) = X\(X\S)◦ . Then

5
 (a) η(∅) = ∅.
(b) S ⊆ η(S) ⇐⇒ (X\S)◦ ⊆ X\S and the latter statement holds by (2).
(c) η(η(S)) = X\(X\η(S))◦ = X\((X\S)◦ )◦ = X\(X\S)◦ = η(S).
(d) η(A ∪ B) = X\(X\(A ∪ B))◦ = X\((X\A) ∩ (X\B))◦ = X\((X\A)◦ ∩ (X\B)◦ ) = η(A) ∪ η(B).
Therefore, by 1.5.1, T = {η(S) : S ∈ P(X)} is a topology with η(S) = S. Then X\(X\S)◦ = S ⇐⇒
(X\S)◦ = X\S = int(X\S) for all S ⊆ X.
Theorem 1.5.3. Let X be a set, and β : P(X) → P(X) a map such that:
(1) β(∅) = ∅.
(2) β(S) = β(X\S) for all S ⊆ X.
(3) β(β(S)) ⊆ β(S) for all S ⊆ X.
(4) A ∩ B ∩ β(A ∩ B) = A ∩ B ∩ (β(A) ∪ β(B)) for all A, B ⊆ X.
Then T = {X\(S ∪ β(S)) : S ∈ P(X)} is a topology, and β(S) = ∂S for all S ⊆ X.
Proof. Set η : P(X) → P(X) by η(S) = S ∪ β(S). Then
(a) η(∅) = ∅.
(b) S ⊆ η(S).
(2)
(c) η(η(S)) = η(S) ∪ β(η(S)). Since β(η(S)) = β(X\η(S)) = β((X\S) ∩ (X\β(S))),
(4)
β((X\S) ∩ (X\β(S))) ∩ (X\S) ∩ (X\β(S)) = (X\S) ∩ (X\β(S)) ∩ (β(X\S) ∪ β(X\β(S)))
(2)
= (X\S) ∩ (X\β(S)) ∩ (β(X\S) ∪ β(β(S)))
(3)
⊆ (X\S) ∩ (X\β(S)) ∩ (β(X\S) ∪ β(S))
(2)
= (X\S) ∩ (X\β(S)) ∩ (β(S) ∪ β(S))
= (X\S) ∩ (X\β(S)) ∩ β(S) = ∅.

Therefore, β(η(S)) ⊆ X\((X\S) ∩ (X\β(S))) = S ∪ β(S) = η(S). Then η(η(S)) = η(S).

(d) Since η(A ∪ B) = A ∪ B ∪ β(A ∪ B), then

X\η(A ∪ B) = X\A ∩ X\B ∩ X\(β(A ∪ B)).

Since
(2)
β(A ∪ B) = β(X\A ∩ X\B),
and
(4)
β(X\A ∩ X\B) ∩ (X\A) ∩ (X\B) = (X\A) ∩ (X\B) ∩ (β(X\A) ∪ β(X\B)),

it follows that
(2)
X\(β(A ∪ B)) = A ∪ B ∪ (X\β(A) ∩ X\β(B)).
Therefore,
X\η(A ∪ B) = X\A ∩ X\B ∩ X\β(A) ∩ X\β(B)
and η(A ∪ B) = A ∪ B ∪ β(A) ∪ β(B) = η(A) ∪ η(B).
Therefore, by 1.5.1, T = {X\η(S) : S ∈ P(X)} = {X\(S ∪ β(S)) : S ∈ P(X)} is a topology with
(2)
η(S) = S for all S ⊆ X. Hence, ∂S = S ∩ X\S = η(S) ∩ η(X\S) = (S ∪ β(S)) ∩ ((X\S) ∪ β(X\S)) =
(S ∩ β(S)) ∪ ((X\S) ∩ β(S)) ∪ β(S) = (S ∪ (X\S) ∪ X) ∩ β(S) = X ∩ β(S) = β(S).

6
1.6 Gδ , Fσ , and Borel Sets
Definition 1.6.1. A set F is called an Fσ -set if it is the union of at most countably many closed sets. A
set G is called a Gδ -set if it is the intersection of at most countably many open sets.

From definitions, it is, a fortiori, that

S∞
(1) If F is an Fσ -set, then there is an increasing sequence F1 ⊆ F2 ⊆ · · · of closed sets with F = 1 Fj .
T∞
(2) If G is a Gδ -set, then there is a decreasing sequence G1 ⊇ G2 ⊇ · · · of open sets with G = 1 Gj .

(3) The countable union and finite intersection of Fσ -sets is an Fσ .

(4) The countable intersection and finite union of Fσ -sets is an Fσ .
(5) The complement of an Fσ is a Gδ , and conversely.
Definition 1.6.2. A nonempty collection Σ ⊆ P(X) is called a σ-algebra if it closed under complement
and countable union.
It is clear arbitrary intersection of σ-algebras is a σ-algebra. Therefore, for any collection C ⊆ P(X), there
exists the smallest σ-algebra σ(C) containing C as P(X) is a σ-algebra. σ(T ) is the collection of Borel sets
in X. By definition, all Fσ -sets and Gδ -sets are Borel sets.

1.7 Relativization
Definition 1.7.1. Let (X, T ) be a topological space, and Y ⊆ X. The induced topology TY on Y is
{Y ∩ U : U ∈ T }. (Y, TY ) is called a subspace of (X, T ).
To verify TY is actually a topology on Y is trivial.
Let Y be a subspace of X and A ⊆ Y . Since (Y, TY ) is a space, we can form the closure of A, using TY ,
to obtain AY ; but also A ⊆ X, so we can form A, using T . We now determine the relation between A and
AY , as well as that for the other operations in section 1.4.
Theorem 1.7.2. Let (X, T ) be a space and (Y, TY ) a subspace. Then:
(1) If {Uα : α ∈ A} is a basis (subbasis) for T , {Y ∩ Uα : α ∈ A} is a basis (subbasis) for TY .

(2) Let A ⊆ Y . Then A is TY -closed if and only if A = Y ∩ F , where F is T -closed.

(3) AY = Y ∩ A; Y ∩ int A ⊆ intY A; ∂Y A ⊆ Y ∩ ∂A.
Proof. (1) Suppose {Uα : α ∈ A} is a basis for T . Let S ∈ TY . Then there exists G ∈ T such that S = G ∩ Y .
Since G is a union of members of Uα ’s, S is a union of members of Uα ∩ Y ’s, showing that {Uα ∩ Y : α ∈ A}
is a basis. If {Uα : α S∈ A} is a subbasis for T , by 1.3.1, any open set is a union of finite intersections of
members of Uα , since α and ∩ is distributive, T ({Uα ∩ Y : α ∈ A}) = TY .
(2) A is TY -closed ⇐⇒ Y \A ∈ TY ⇐⇒ Y \A = Y ∩ G for some G ∈ T ⇐⇒ A = Y ∩ (X\G) for some
G∈T.
(3) Suppose x ∈ AY . Let U (x) be a T -nbd. Then (U ∩Y )(x) is a TY -nbd, therefore U ∩Y ∩A ̸= ∅, therefore
x ∈ A ∩ Y (it is clear that x ∈ Y ) by 1.4.2. Conversely, suppose x ∈ Y ∩ A. Let U (x) be a TY -nbd, then there
exists G ∈ T such that U = G ∩ Y . Therefore, G ∩ A ̸= ∅. Since A ⊆ Y , G ∩ A = G ∩ Y ∩ A = U ∩ A ̸= ∅
and thus x ∈ AY by 1.4.2.
Suppose x ∈ Y ∩ int A. Then there exists a T -nbd U (x) such that x ∈ U ⊆ A. Then x ∈ U ∩ Y ⊆ A
and therefore x ∈ intY A. Finally, since ∂Y (A) = AY \ intY A ⊆ (Y ∩ A)\(Y ∩ int A) = Y ∩ A ∩ (X\ int A) =
Y ∩ ∂A.

Corollary 1.7.2.1. Let Y be a subspace of X. If A ⊆ Y is closed (open) in Y , and Y is closed (open) in

X, then A is closed (open) in X.

7
1.8 Continuous Maps
Definition 1.8.1. Let (X, TX ) and (Y, TY ) be spaces. A map f : X → Y is called continuous if f −1 (TY ) :=
{f −1 (U ) : U ∈ TY } ⊆ TX .
Proposition 1.8.2.
(1) If f : X → Y and g : Y → Z are continuous, so also is g ◦ f : X → Z.
(2) If f : X → Y is continuous and A ⊆ X is taken with the subspace topology, then f |A : A → Y is
continuous.
(3) If f : X → Y is continuous and f (X) is taken with the subspace topology, then f : X → f (X) is
continuous.
Proof. (1) (g ◦ f )−1 = f −1 ◦ g −1 . (2) (f |A )−1 (U ) = f −1 (U ) ∩ A. (3) f −1 (f (X) ∩ U ) = f −1 (U ).

Theorem 1.8.3. Let X, Y be topological spaces, and f : X → Y a map. The following statements are
equivalent:
(1) f is continuous.
(2) The inverse image of each closed set in Y is closed in X.

(3) The inverse image of each member of a subbasis (basis) for Y is open in X.
(4) For each x ∈ X and each nbd W (f (x)) in Y , there exists a nbd V (x) in X such that f (V ) ⊆ W .
(5) f (A) ⊆ f (A) for every A ⊆ X.
(6) f −1 (B) ⊆ f −1 (B) for every B ⊆ Y .

Proof. (1) ⇔ (2), since f −1 (Y \U ) = X\f −1 (U ) for any U ⊆ X.

(1) ⇔ (3). Let {Uα : α ∈ A} be a subbasis for Y . If f is continuous, each f −1 (Uα ) is open. Conversely,
if each f −1 (Uα ) is open, then because any open U ⊆ Y can be written
[
U = {Uα1 ∩ · · · ∩ Uαn : {α1 , · · · , αn } ⊆ A},

we have that [
f −1 (U ) = {f −1 (Uα1 ) ∩ · · · ∩ f −1 (Uαn )}
is a union of open sets and so is open.
(1) ⇒ (4). Trivial. (4) ⇒ (5). Suppose x ∈ A. Let W (f (x)) be a nbd. Then there exists nbd V (x) such
that f (V ) ⊆ W . Since x ∈ A, V ∩ A ̸= ∅ and therefore W ∩ f (A) ̸= ∅. Since W is arbitrary, f (x) ∈ f (A).
Since x ∈ A is arbitrary, f (A) ⊆ f (A).
(5) ⇒ (6). Let B ⊆ Y . Then f (f −1 (B)) ⊆ f (f −1 (B)) ⊆ B. Therefore, f −1 (B) ⊆ f −1 (B).
(6) ⇒ (1). Suppose F is closed in TY . Then f −1 (F ) ⊆ f −1 (F ) = f −1 (F ). Thus, f −1 (F ) = f −1 (F ) and
f −1 (F ) is closed in TX .

Definition 1.8.4. An f : X → Y is continuous at x0 ∈ X if 1.8.3(4) is satisfied at x0 .

1.9 Piecewise Definition of Maps

Definition 1.9.1. A family {Aα : α ∈ A} of sets in a space is called nbd-finite if each point of X has a
nbd V such that V ∩ Aα = ∅ for all but finitely many α.
Proposition 1.9.2. Let {Aα : α ∈ A} be a nbd-finite family in X. Then:
(1) {Aα : α ∈ A} is also nbd-finite.
S
(2) For each B ⊆ A, {Aα : α ∈ B} is closed.

8
Proof. (1) Given x, there exists nbd U (x) such that Aα ∩ U = ∅ for all but finitely many α; since

Aα ∩ U = ∅ =⇒ Aα ⊆ X\U =⇒ Aα ⊆ X\U =⇒ Aα ∩ U = ∅.
S S T
(2) Suppose x ∈/ {Aα : α ∈ B}. Then x ∈ X\ α∈B Aα = α∈B X\Aα . By (1), a nbd U (x) meeting at
Tn S
most finitely many Aβ say, Aβ1 , · · · , Aβn ; then U ∩ 1 X\Aβj is a nbd of x not meeting {Aα : α ∈ B}.
S
Theorem 1.9.3. Let {Aα : α ∈ A} be a family of set that cover the space X; that is, X = α∈A Aα .
Assume that either:
(1) All the Aα are open, or
(2) All the Aα are closed, and form a nbd-finite family.

Then B ⊆ X is open (resp. closed) if and only if each B ∩ Aα is open (resp. closed) in the subspace Aα .
Proof. The sufficiency is clear from the definition and 1.7.2. Suppose Aα ’s are open and B ∩ Aα is open
for each α ∈ A. S Then for each S α ∈ A, there exists open Uα such that B ∩ Aα = Uα ∩ Aα . Therefore,
B = B ∩ X = α∈A B ∩ Aα = α∈A Uα ∩ Aα is open.
Suppose all Aα are closed and form a nbd-finite family.
S If B ∩ Aα is closed for each α, there exists closed
Fα such that B ∩ Aα = Fα ∩ Aα . Then B = B ∩ X = α∈A Fα ∩ Aα . Since {Fα ∩ Aα : α ∈ A} is nbd-finite,
it is closed by 1.9.2.
Theorem 1.9.4. Let {Aα : α ∈ A} be a covering of a space X such that either:
(1) The set Aα are all open, or

(2) The set Aα are all closed, and form a nbd-finite family.
For each α ∈ A, let fα : Aα → Y be continuous and assume that fα |Aα ∩Aβ = fβ |Aα ∩Aβ for each (α, β) ∈
A × A. Then there exists a unique continuous map f : X → Y , which is an extension of each fα .
S
Proof. Define f : X → Y by f |Aα = fα . It is well-defined as X = α Aα and fα ’s agree on their intersections.
We shall prove that f is open. If Aα are all open, let U ⊆ Y be an open set. Then f −1 (U ) = α fα−1 (U )
S
and f −1 (U ) ∩ Aα = fα−1 (U ) is open in Aα for each α. Then by 1.9.3, f −1 (U ) is open.
S Suppose Aα are all closed, and form a nbd-finite family. Let F ⊆ Y be a closed set. Then f −1 (F ) =
−1
f
α α (F ) and f −1 (F ) ∩ Aα = fα−1 (F ) is closed in Aα for each α. Then by 1.9.3 again, f −1 (F ) is closed.

1.10 Continuous Maps into R

Proposition 1.10.1. f : X → R is continuous of and only if for each real b, the set {f > b} := {x ∈ E :
f (x) > b} and {f < b} := {x ∈ E : f (x) < b} are open.
Proof. Since (−∞, b) and (b, ∞) are both open, if f is continuous, {f > b} and {f < b} are open. On the
other hand, since {(b, ∞) : b ∈ R} ∪ {(−∞, b) : b ∈ R} form a subbasis of R, by 1.8.3, we are done.

Corollary 1.10.1.1. Let f, g : X → R be continuous. Then:

(1) af + bg are continuous for any a, b ∈ R;
α
(2) |f | is continuous for each α ≥ 0;
(3) f g is continuous;

(4) If f (x) ̸= 0 on X, then 1/f is continuous.

9
Proof. (1) We first show that af is continuous: the case a = 0 is trivial; if a > 0, then {af > b} = {f > b/a}
and {af < b} = {f < b/a} are bothSopen; if a < 0, then {af > b} = {f < b/a} andS{af < b} = {f > b/a} are
both open. Second, {f + g > b} = λ∈R {f > b − λ} ∩ {g > λ} and {f + g < b} = λ∈R {f < b − λ} ∩ {g < λ}
are open.
α α α
(2) If b < 0, then {|f | < b} = ∅ and {|f | > b} = X are open. If b ≥ 0, then {|f | > b} = {f >
1/α 1/α α 1/α 1/α
b } ∪ {f < −b } and {|f | < b} = {f < b } ∩ {f > −b } are open.
2 2
(3) Since f g = 14 (|f + g| − |f − g| ), f g is continuous.
(4) Since f (x) ̸= 0, {1/f > b} = ({f > 0} ∩ {bf < 1}) ∪ ({f < 0} ∩ {bf > 1}) and {1/f < b} = ({f >
0} ∩ {bf > 1}) ∪ ({f < 0} ∩ {bf > 1}) are open.
Definition 1.10.2. An f : X → R is upper semicontinuous if for each real b, {f < b} is open; it is lower
semicontinuous if for each real b, {f > b} is open.

For maps into the extended real line R, the criterion for continuity is evidently the same as 1.10.1, so we
can speak of upper and of lower semicontinuous maps. Furthermore, R has the feature that for any A ⊆ R,
sup A and inf A always exists.
Corollary 1.10.2.1. Let {fα : α ∈ A} be any family of continuous maps fα : X → R. Then:

(1) M = supα fα is lower semicontinuous;

(2) m = inf α fα is upper semicontinuous.
S S
Proof. {M > b} = α {fα > b} and {m < b} = α {fα < b} are open.
Theorem P1.10.3. Let f1 , f2 , · · · : X → R be a sequence of continuous
P∞ maps such that |fi (x)| ≤ Mi for each
∞
i, where i=1 Mi is a convergent series of reals. Then f (x) = i=1 fi (x) exists and is a continuous map
f : X → R.

10