Unit 07: Metric Topology

Unit 07: Metric Topology

Contents
Expected Learning Outcomes ....................................................................................................................... 1
Introduction ..................................................................................................................................................... 1
Summary ........................................................................................................................................................ 16
Keywords ....................................................................................................................................................... 16
Self-assessment.............................................................................................................................................. 16
Review Questions ......................................................................................................................................... 19
Further Readings .......................................................................................................................................... 20
Weblinks ........................................................................................................................................................ 20

Expected Learning Outcomes

After studying this unit, you will be able to

• define metric topology and explain metric topology with examples

• define metric topology on ℝ𝑛 and compare the Euclidean, square and product topology on
ℝ𝑛
• define uniform topology on ℝ𝐽 and compare product, uniform, and box topology on ℝ𝐽
• find the definition of continuity in terms of open balls equivalent to the 𝜖 − 𝛿 definition
• state and prove sequence lemma
• define first countability axiom and understand new ways to construct continuous functions
• give examples of some non-metrizable spaces

Introduction
The ability to measure and compare distances between elements of a set is often crucial, and it
provides more structure than a general topological space possesses. One of the most important and
frequently used ways of imposing a topology on a set is to define the topology in terms of a metric
on the set.
In this chapter, we will define metric topology and explain metric topology with examples. Further,
the metric topology will be defined on ℝ𝑛 and compare the Euclidean, square and product topology
on ℝ𝑛 . Uniform topology on ℝ 𝐽 will be defined. Further, product, uniform, and box topology will be
compared on ℝ𝐽 . Definition of continuity will be explained in terms of open balls equivalent to the
𝜖 − 𝛿 definition. The sequence lemma will be proved. The first countability axiom will be defined,
and new ways to construct continuous functions will be explained. Further, examples of some non-
metrizable spaces will be given.

7.1 Metric Topology

Definition 7.1.1: A metric on a set 𝑋 is a function 𝑑: 𝑋 × 𝑋 → ℝ having the following properties:
(1) 𝑑(𝑥, 𝑦) ≥ 0 for all 𝑥, 𝑦 ∈ 𝑋; equality holds if and only if 𝑥 = 𝑦.
(2) 𝑑(𝑥, 𝑦) = 𝑑(𝑦, 𝑥)∀ 𝑥, 𝑦 ∈ 𝑋
(3) (Triangle inequality) 𝑑(𝑥, 𝑦) + 𝑑(𝑦, 𝑧) ≥ 𝑑(𝑥, 𝑧), ∀ 𝑥, 𝑦, 𝑧 ∈ 𝑋.
Given a metric 𝑑 on 𝑋, the number 𝑑(𝑥, 𝑦) is often called the distance between 𝑥 and 𝑦 in the metric
𝑑.

LOVELY PROFESSIONAL UNIVERSITY 1

Topology

Definition 7.1.2: Given 𝜖 > 0, consider the set 𝐵𝑑 (𝑥, 𝜖) = {𝑦| 𝑑(𝑥, 𝑦) < 𝜖} of all points 𝑦 whose
distance from 𝑥 is less than 𝜖. It is called the 𝜖 −ball centred at 𝑥.

Note:
Sometimes we omit the metric 𝑑 from the notation and write this ball simply as 𝐵(𝑥, 𝜖)
when no confusion will arise.

Definition 7.1.3: If 𝑑 is a metric on the set 𝑋, then the collection of all 𝜖 −balls 𝐵𝑑 (𝑥, 𝜖), for 𝑥 ∈ 𝑋 and
𝜖 > 0, is a basis for a topology on 𝑋, called the metric topology induced by 𝑑.
The topological space, thus created by 𝑋 and 𝑑 is called Euclidean metric space.
Theorem 7.1.4: The set 𝔅 = {𝐵𝑑 (𝑥, 𝜖) |𝑥 ∈ 𝑋, 𝜖 > 0} is a basis for some topology on 𝑋.
Proof: ∀ 𝑥 ∈ 𝑋,
𝐵𝑑 (𝑥, 𝜖) = {𝑦|𝑑(𝑥, 𝑦) < 𝜖}
Since 𝑑(𝑥, 𝑥) = 0 < 𝜖, 𝑥 ∈ 𝐵𝑑 (𝑥, 𝜖)
Therefore, ∃ 𝐵𝑑 (𝑥, 𝜖) such that 𝑥 ∈ 𝐵𝑑 (𝑥, 𝜖), 𝐵𝑑 (𝑥, 𝜖) ∈ 𝔅
Consider 𝐵1 , 𝐵2 ∈ 𝔅
We will show that if 𝑥 ∈ 𝐵1 ∩ 𝐵2 , then there exists 𝐵3 ∈ 𝔅 such that 𝑥 ∈ 𝐵3 ⊆ 𝐵1 ∩ 𝐵2 . To prove this,
first, we need to establish a claim
Claim: For 𝑦 ∈ 𝐵𝑑 (𝑥, 𝜖), ∃ 𝐵𝑑 (𝑦, 𝛿) ∈ 𝔅 such that 𝐵𝑑 (𝑦, 𝛿) ⊂ 𝐵𝑑 (𝑥, 𝜖)
Let 𝑧 ∈ 𝐵𝑑 (𝑦, 𝛿)
⇒ 𝑑(𝑧, 𝑦) < 𝛿
Now 𝑦 ∈ 𝐵𝑑 (𝑥, 𝜖), 𝑑(𝑥, 𝑦) < 𝜖 so that 𝜖 − 𝑑(𝑥, 𝑦) > 0.
Choose 𝛿 = 𝜖 − 𝑑(𝑥, 𝑦)
⇒ 𝑑(𝑧, 𝑦) < 𝜖 − 𝑑(𝑥, 𝑦)
⇒ 𝑑(𝑧, 𝑦) + 𝑑(𝑥, 𝑦) < 𝜖
⇒ 𝑑(𝑥, 𝑧) ≤ 𝑑(𝑧, 𝑦) + 𝑑(𝑥, 𝑦) < 𝜖
⇒ 𝑧 ∈ 𝐵𝑑 (𝑥, 𝜖)
So that 𝐵𝑑 (𝑦, 𝛿) ⊆ 𝐵𝑑 (𝑥, 𝜖)
Now 𝑥 ∈ 𝐵1 ∩ 𝐵2
There exists 𝐵𝑑 (𝑥, 𝛿1 ), 𝐵𝑑 (𝑥, 𝛿2 ) ∈ 𝔅 such that 𝑥 ∈ 𝐵𝑑 (𝑥, 𝛿1 ) ⊆ 𝐵1 and 𝑥 ∈ 𝐵𝑑 (𝑥, 𝛿2 ) ⊆ 𝐵2
Choose 𝛿 = min{𝛿1 , 𝛿2 }
𝑥 ∈ 𝐵(𝑥, 𝛿) ⊆ 𝐵1 ∩ 𝐵2 ; 𝐵(𝑥, 𝛿) ∈ 𝔅
Therefore, 𝔅 is a basis.
Now we rephrase the definition of metric topology in terms of its basis elements
Definition 7.1.5: A set 𝑈 is open in the metric topology induced by 𝑑 if and only if for each 𝑦 ∈
𝑈, there is a 𝛿 > 0 such that 𝐵𝑑 (𝑦, 𝛿) ⊆ 𝑈.
Since 𝐵𝑑 (𝑦, 𝛿) ∈ 𝔅
A set 𝑈 is open in a topological space if for each 𝑦 ∈ 𝑈, ∃ 𝐵𝑑 (𝑦, 𝛿) ∈ 𝔅 such that 𝑦 ∈ 𝐵𝑑 (𝑦, 𝛿).
Conversely, if 𝑈 is an open set in this topology then by definition, there exists some 𝐵𝑑 (𝑥, 𝜖) ⊂ 𝑈 and
if 𝑦 ∈ 𝐵𝑑 (𝑥, 𝜖), we have proved that there exists some 𝐵𝑑 (𝑦, 𝛿) ⊂ 𝐵𝑑 (𝑥, 𝜖) ⊂ 𝑈
This implies, 𝐵𝑑 (𝑦, 𝛿) ⊂ 𝑈

Example 7.1.6: Given a set 𝑋, define

1 𝑖𝑓 𝑥 ≠ 𝑦
𝑑(𝑥, 𝑦) = {
0 𝑖𝑓 𝑥 = 𝑦

2 LOVELY PROFESSIONAL UNIVERSITY

 Unit 07: Metric Topology

Proof: (i) 𝑑(𝑥, 𝑦) = 1 or 0 ≥ 0

Also, 𝑑(𝑥, 𝑦) = 0 if and only if 𝑥 = 𝑦.
(ii) Consider 𝑥, 𝑦 ∈ 𝑋
If 𝑥 = 𝑦, 𝑑(𝑥, 𝑦) = 0. 𝑦 = 𝑥 hence 𝑑(𝑦, 𝑥) = 𝑑(𝑥, 𝑦)
If 𝑥 ≠ 𝑦, 𝑦 ≠ 𝑥
Hence, 𝑑(𝑦, 𝑥) = 𝑑(𝑥, 𝑦) = 1
Therefore, 𝑑(𝑥, 𝑦) = 𝑑(𝑦, 𝑥)
(iii) Consider 𝑥, 𝑦, 𝑧 ∈ 𝑋
Case I: 𝑥 = 𝑦 = 𝑧
𝑑(𝑥, 𝑦) + 𝑑(𝑦, 𝑧) = 0 + 0 = 𝑑(𝑥, 𝑧)
Case II: 𝑥 = 𝑦, 𝑦 ≠ 𝑧 ⇒ 𝑥 ≠ 𝑧
𝑑(𝑥, 𝑦) + 𝑑(𝑦, 𝑧) = 0 + 1 = 1 = 𝑑(𝑥, 𝑧)
Case III: 𝑥 ≠ 𝑦, 𝑦 ≠ 𝑧, 𝑥 = 𝑧
𝑑(𝑥, 𝑦) + 𝑑(𝑦, 𝑧) = 1 + 1 > 0 = 𝑑(𝑥, 𝑧)
Case IV: 𝑥 ≠ 𝑦, 𝑦 ≠ 𝑧, 𝑧 ≠ 𝑥
𝑑(𝑥, 𝑦) + 𝑑(𝑦, 𝑧) = 1 + 1 = 2 > 1 = 𝑑(𝑥, 𝑧)
Therefore, 𝑑 induces a metric on 𝑋. Hence, (𝑋, 𝑑) is a metric space.
For 𝑥 ∈ 𝑋, 𝜖 > 0
𝐵𝑑 (𝑥, 𝜖) = {𝑦|𝑑(𝑥, 𝑦) < 𝜖}
𝐵𝑑 (𝑥, 𝜖) = {𝑦|𝑑(𝑥, 𝑦) < 𝜖}
= {𝑦| 0 or 1 < 𝜖}
But 𝜖 > 0 is very small.
So, 𝐵𝑑 (𝑥, 𝜖) = {𝑦|𝑑(𝑥, 𝑦) = 0} = {𝑥}
The topology it induces is discrete topology. Its basis set is {𝐵(𝑥, 1)|𝑥 ∈ 𝑋}

Example 7.1.7: The standard metric on the real numbers ℝ is defined by the equation
𝑑(𝑥, 𝑦) = |𝑥 − 𝑦|. It is easy to check that 𝑑 is a metric.

Proof: 𝑑(𝑥, 𝑦) = |𝑥 − 𝑦| ≥ 0
|𝑥 − 𝑦| = 0 ⇔ 𝑥 − 𝑦 = 0 ⇔ 𝑥 = 𝑦
That is, 𝑑(𝑥, 𝑦) = 0 ⇔ 𝑥 = 𝑦
(ii) For all 𝑥, 𝑦 ∈ 𝑋,
𝑑(𝑥, 𝑦) = |𝑥 − 𝑦| = |(−1)(𝑦 − 𝑥)| = |𝑦 − 𝑥| = 𝑑(𝑦, 𝑥)
𝐵𝑑 (𝑥, 𝜖) = {𝑦|𝑑(𝑥, 𝑦) < 𝜖}
= {𝑦| 0 or 1 < 𝜖}
But 𝜖 > 0 is very small.
So, 𝐵𝑑 (𝑥, 𝜖) = {𝑦|𝑑(𝑥, 𝑦) = 0} = {𝑥}
The topology it induces is discrete topology. Its basis set is {𝐵(𝑥, 1)|𝑥 ∈ 𝑋}
(iii) For 𝑥, 𝑦, 𝑧 ∈ 𝑋
𝑑(𝑥, 𝑦) + 𝑑(𝑦, 𝑧) = |𝑥 − 𝑦| + |𝑦 − 𝑧| ≥ |𝑥 − 𝑧| = 𝑑(𝑥, 𝑧)
Hence, 𝑑 induces a metric on 𝑋 = ℝ.
Result: Order topology on the set of real numbers is the same as the metric topology on ℝ induced
by metric 𝑑, where 𝑑(𝑥, 𝑦) = |𝑥 − 𝑦| ∀ 𝑥, 𝑦 ∈ ℝ.

LOVELY PROFESSIONAL UNIVERSITY 3

Topology

Any basis element of metric topology is

𝐵𝑑 (𝑥, 𝜖) = {𝑦| 𝑑(𝑥, 𝑦) < 𝜖}
= {𝑦||𝑥 − 𝑦| < 𝜖}
= {𝑦| − 𝜖 < 𝑦 − 𝑥 < 𝜖}
= {𝑦|𝑥 − 𝜖 < 𝑦 < 𝑥 + 𝜖} = (𝑥 − 𝜖, 𝑥 + 𝜖)
𝐵𝑑 (𝑥, 𝜖) being an open interval is a basis element of order topology on ℝ.
The arbitrary basis element of the order topology is (𝑎, 𝑏), 𝑎 < 𝑏, 𝑎, 𝑏 ∈ ℝ.
𝑎+𝑏 𝑏−𝑎
Let 𝑥 = , 𝜖=
2 2

𝐵𝑑 (𝑥, 𝜖) = {𝑦||𝑥 − 𝑦| < 𝜖}

= {𝑦|𝑥 − 𝜖 < 𝑦 < 𝑥 + 𝜖}
𝑎+𝑏 𝑏−𝑎 𝑎+𝑏 𝑏−𝑎
= {𝑦| − <𝑦< + }
2 2 2 2
= {𝑦|𝑎 < 𝑦 < 𝑏} = (𝑎, 𝑏)
Therefore, (𝑎, 𝑏) = 𝐵𝑑 (𝑥, 𝜖), that is a basis element of metric topology.
Therefore, the metric topology under 𝑑 is the same as the order topology on ℝ.
Definition 7.1.8: Let 𝑋 and 𝑌 be topological spaces. The product topology on 𝑋 × 𝑌 is the topology
having as basis 𝔅 the collection of all sets of the form 𝑈 × 𝑉, where 𝑈 is an open subset of 𝑋 and 𝑉 is
an open subset of 𝑌.
𝔅 = {𝑈 × 𝑉|𝑈 is open set in 𝑋, 𝑉 is open set in 𝑌}
Definition 7.1.9: Let 𝑋 be a metric space with metric 𝑑. A subset 𝐴 of 𝑋 is said to be bounded if there
is some number such that 𝑑(𝑎1 , 𝑎2 ) ≤ 𝑀 ∀ 𝑎1 , 𝑎2 ∈ 𝐴.
If 𝐴 is bounded and nonempty, the diameter of 𝐴 is defined to be the number
𝑑𝑖𝑎𝑚 𝐴 = sup{𝑑(𝑎1 , 𝑎2 )|𝑎1 , 𝑎2 ∈ 𝐴}

Note:
• The boundedness of a set is not a topological property, for it depends on the
metric 𝑑 that is used for 𝑋.
• For instance, if 𝑋 is a metric space with metric 𝑑, then there exists a metric 𝑑̅ that
gives the topology of 𝑋, relative to which every subset of 𝑋 is bounded. It is
defined as follows.

Definition 7.1.10: Let 𝑋 be a metric space with metric 𝑑. Define 𝑑̅ : 𝑋 × 𝑋 → ℝ by the equation
𝑑̅ (𝑥, 𝑦) = min{𝑑(𝑥, 𝑦), 1}
Then 𝑑̅ is a metric that induces the same topology as 𝑑. The metric 𝑑̅ is called the standard bounded
metric corresponding to 𝑑.
Note that 𝑑̅ (𝑥, 𝑦) ≤ 𝑑(𝑥, 𝑦) and 𝑑̅ (𝑥, 𝑦) ≤ 1
Theorem 7.1.11: We prove that 𝑑̅ defined earlier is a metric on 𝑋.
Proof: 𝑑̅ (𝑥, 𝑦) = min{𝑑(𝑥, 𝑦), 1}
𝑑(𝑥, 𝑦) is a metric on 𝑋.
This implies, 𝑑(𝑥, 𝑦) ≥ 0 ∀ 𝑥, 𝑦 ∈ 𝑋 and 𝑑(𝑥, 𝑦) = 0 if and only if 𝑥 = 𝑦.
Since 𝑑(𝑥, 𝑦) ≥ 0 and 1 > 0, min{𝑑(𝑥, 𝑦), 1} ≥ 0
This implies 𝑑̅ (𝑥, 𝑦) ≥ 0
𝑑̅ (𝑥, 𝑦) = 0
⇔ min{𝑑(𝑥, 𝑦), 1} = 0
⇔ 𝑑(𝑥, 𝑦) = 0

4 LOVELY PROFESSIONAL UNIVERSITY

 Unit 07: Metric Topology

⇔𝑥=𝑦
(ii) Let 𝑥, 𝑦 ∈ 𝑋
⇔ 𝑑(𝑥, 𝑦) = 𝑑(𝑦, 𝑥)
⇔ min{𝑑(𝑥, 𝑦), 1} = min{𝑑(𝑦, 𝑥), 1}
⇔ 𝑑̅ (𝑥, 𝑦) = 𝑑̅ (𝑦, 𝑥)
(iii) Let 𝑥, 𝑦, 𝑧 ∈ 𝑋
Case 1: 𝑑(𝑥, 𝑦) ≥ 1 or 𝑑(𝑦, 𝑧) ≥ 1
⇒ 𝑑̅ (𝑥, 𝑦) = 1 or 𝑑̅ (𝑦, 𝑧) = 1
⇒ 𝑑̅ (𝑥, 𝑦) + 𝑑̅ (𝑦, 𝑧) ≥ 1
⇒ 𝑑̅ (𝑥, 𝑦) + 𝑑̅ (𝑦, 𝑧) ≥ 1 ≥ 𝑑̅ (𝑥, 𝑧)
Case 2: 𝑑(𝑥, 𝑦) < 1 and 𝑑(𝑦, 𝑧) < 1
⇒ 𝑑̅ (𝑥, 𝑦) = 𝑑(𝑥, 𝑦) and 𝑑̅ (𝑦, 𝑧) = 𝑑(𝑦, 𝑧)
⇒ 𝑑̅ (𝑥, 𝑦) + 𝑑̅ (𝑦, 𝑧) = 𝑑(𝑥, 𝑦) + 𝑑(𝑦, 𝑧) ≥ 𝑑(𝑥, 𝑧) ≥ 𝑑̅ (𝑥, 𝑧)
Therefore, 𝑑̅ is a metric on 𝑋.

Note:
Now we see that the metric 𝑑 and 𝑑̅ induce the same topology.

Basis element for 𝑑̅ is 𝐵𝑑̅ (𝑥, 𝜖), 0 < 𝜖 < 1

𝐵𝑑̅ (𝑥, 𝜖) = {𝑦| 𝑑̅ (𝑥, 𝑦) < 𝜖}
Now 𝑑(𝑥, 𝑦) ≥ 𝑑̅ (𝑥, 𝑦)
𝑑̅ (𝑥, 𝑦) < 𝜖 < 1
⇒ 𝑑(𝑥, 𝑦) < 𝜖
⇒ 𝑦 ∈ 𝐵𝑑 (𝑥, 𝜖)
Again, if 𝑦 ∈ 𝐵𝑑 (𝑥, 𝜖)
⇒ 𝑑(𝑥, 𝑦) < 𝜖
Since 𝑑̅ (𝑥, 𝑦) ≤ 𝑑(𝑥, 𝑦) < 𝜖
⇒ 𝑑̅ (𝑥, 𝑦) < 𝜖
⇒ 𝑦 ∈ 𝐵𝑑̅ (𝑥, 𝜖)
𝐵𝑑 (𝑥, 𝜖) = 𝐵𝑑̅ (𝑥, 𝜖) ∀ 𝑥 ∈ 𝑋, 0 < 𝜖 < 1
⇒ 𝜖 −balls under 𝑑 and 𝑑̅ are the same.
⇒ 𝑑 and 𝑑̅ induce the same topology on 𝑋.

Task:

1. Prove or disprove “corresponding to every metric 𝑑 on 𝑋, there exists a bounded

metric 𝑑̅ on 𝑋. "
2. Comment on the uniqueness of this bounded metric 𝑑̅ .

Definition 7.1.12: Given 𝑥 = (𝑥1 , 𝑥2 , … , 𝑥𝑛 ) ∈ ℝ𝑛 , we define the norm of 𝑥 by the equation

‖𝑥‖ = (𝑥12 + 𝑥22 + ⋯ + 𝑥𝑛2 )1/2

Example 7.1.13: Consider 𝑥 = (1, − 2, 3) ∈ ℝ3

LOVELY PROFESSIONAL UNIVERSITY 5

Topology

1
‖𝑥‖ = (12 + (−2)2 + 32 )2
1
= (1 + 4 + 9)2 = √14

Definition 7.1.14: Given 𝑥 = (𝑥1 , 𝑥2 , … , 𝑥𝑛 ), 𝑦 = (𝑦1 , 𝑦2 , … , 𝑦𝑛 ) ∈ ℝ𝑛 , we define the Euclidean

metric 𝑑 on ℝ𝑛 by the equation
𝑑(𝑥, 𝑦) = ‖𝑥 − 𝑦‖

Example 7.1.15: 𝑥 = (1, − 2, 3), 𝑦 = (1, 2, 1) ∈ ℝ3

𝑥 − 𝑦 = (1, − 2, 3) − (1, 2, 1) = (0, − 4, 2)
1
Hence, ‖𝑥 − 𝑦‖ = (02 + (−4)2 + 22 )2 = √20

Definition 7.1.16: Given 𝑥 = (𝑥1 , 𝑥2 , … , 𝑥𝑛 ), 𝑦 = (𝑦1 , 𝑦2 , … , 𝑦𝑛 ) ∈ ℝ𝑛 , we define the square metric

𝜌 on ℝ𝑛 by the equation
𝜌(𝑥, 𝑦) = max {|𝑥1 − 𝑦1 |, |𝑥2 − 𝑦2 |, … |𝑥𝑛 − 𝑦𝑛 |}

Example 7.1.17: 𝑥 = (1, − 2, 3), 𝑦 = (1, 2, 1) ∈ ℝ3

𝜌(𝑥, 𝑦) = max{|1 − 1|, |−2 − 2|, |3 − 1|}
= max {0, 4, 2} = 4

Theorem 7.1.18: The function 𝜌 defined as 𝜌(𝑥, 𝑦) = max {|𝑥1 − 𝑦1 |, |𝑥2 − 𝑦2 |, … |𝑥𝑛 − 𝑦𝑛 |} for 𝑥 =
(𝑥1 , 𝑥2 , … , 𝑥𝑛 ), 𝑦 = (𝑦1 , 𝑦2 , … , 𝑦𝑛 ) ∈ ℝ𝑛 is a metric on ℝ𝑛 .
Proof: Since |𝑥𝑖 − 𝑦𝑖 | ≥ 0 ∀ 𝑖,
max {|𝑥𝑖 − 𝑦𝑖 |} ≥ 0
1≤𝑖≤𝑛

This implies, 𝜌(𝑥, 𝑦) ≥ 0

Again, 𝜌(𝑥, 𝑦) = 0
⇔ max {|𝑥𝑖 − 𝑦𝑖 |} = 0
1≤𝑖≤𝑛

⇔ |𝑥𝑖 − 𝑦𝑖 | = 0 ∀ 1 ≤ 𝑖 ≤ 𝑛
⇔ 𝑥𝑖 = 𝑦𝑖 ∀ 1 ≤ 𝑖 ≤ 𝑛
⇔𝑥=𝑦
(ii) 𝜌(𝑥, 𝑦) = max {|𝑥𝑖 − 𝑦𝑖 |} = max {|𝑦𝑖 − 𝑥𝑖 |}
1≤𝑖≤𝑛 1≤𝑖≤𝑛

= 𝜌(𝑦, 𝑥)
(iii) Consider 𝑥, 𝑦, 𝑧 ∈ ℝ𝑛 , 𝑥 = (𝑥𝑖 )1≤𝑖≤𝑛 , 𝑦 = (𝑦𝑖 )1≤𝑖≤𝑛 , 𝑧 = (𝑧𝑖 )1≤𝑖≤𝑛
|𝑥𝑖 − 𝑦𝑖 | + |𝑦𝑖 − 𝑧𝑖 | ≥ |𝑥𝑖 − 𝑧𝑖 |
That is, |𝑥𝑖 − 𝑧𝑖 | ≤ |𝑥𝑖 − 𝑦𝑖 | + |𝑦𝑖 − 𝑧𝑖 |
≤ max {|𝑥𝑖 − 𝑦𝑖 |} + max {|𝑦𝑖 − 𝑧𝑖 |}
1≤𝑖≤𝑛 1≤𝑖≤𝑛

= 𝜌(𝑥, 𝑦) + 𝜌(𝑦, 𝑧)
Taking maximum on both sides we get
𝜌(𝑥, 𝑧) ≤ 𝜌(𝑥, 𝑦) + 𝜌(𝑦, 𝑧)
Therefore, 𝜌 induces a metric on ℝ𝑛 .

Note:
On the real line ℝ, these two metrics coincide with the standard metric for ℝ.

6 LOVELY PROFESSIONAL UNIVERSITY

 Unit 07: Metric Topology

In the plane ℝ2 , the basis elements under 𝑑 can be pictured as circular regions, while the basis
elements under 𝜌 can be pictured as square regions.
We now show that each of these metrics induces the usual topology on ℝ. We need the following
lemma:
Lemma 7.1.19: Let 𝑑 and 𝑑′ be two metrics on the set 𝑋; let 𝜏 and 𝜏 ′ be the topologies they induce,
respectively. Then 𝜏′ is finer than 𝜏 if and only if for each 𝑥 ∈ 𝑋 and each 𝜖 > 0, there exists a 𝛿 > 0
such that 𝐵𝑑 ′ (𝑥, 𝛿) ⊆ 𝐵𝑑 (𝑥, 𝜖).
Proof: Suppose 𝜏′ is finer than 𝜏.
This implies, for all basis elements 𝐵𝑑 (𝑥, 𝜖) for 𝜏, there exists a basis element 𝐵′ ∈ 𝜏′ such that 𝑥 ∈
𝐵′ ⊆ 𝐵𝑑 (𝑥, 𝜖)
This implies that there exists some 𝛿 > 0 such that
𝑥 ∈ 𝐵𝑑 ′ (𝑥, 𝛿) ⊆ 𝐵′ ⊆ 𝐵𝑑 (𝑥, 𝜖)
⇒ 𝛿 > 0 such that𝐵𝑑 ′ (𝑥, 𝛿) ⊆ 𝐵𝑑 (𝑥, 𝜖)
Conversely, let the 𝜖 − 𝛿 definition holds.
Let 𝐵 be a basis element for 𝜏 such that 𝑥 ∈ 𝐵.
We need to find a basis element 𝐵′ of 𝜏′ such that 𝑥 ∈ 𝐵′ ⊆ 𝐵
Since 𝑥 ∈ 𝐵, there exists some 𝜖 > 0 such that 𝑥 ∈ 𝐵𝑑 (𝑥, 𝜖) ⊆ 𝐵.
By hypothesis, there exists 𝛿 > 0 such that
𝑥 ∈ 𝐵𝑑 ′ (𝑥, 𝛿) ⊆ 𝐵𝑑 (𝑥, 𝜖)
We can take 𝐵′ = 𝐵𝑑 ′ (𝑥, 𝛿), that is, a basis element for 𝜏 ′ .
This implies, 𝜏′ is finer than 𝜏.
Theorem 7.1.20: The topologies on ℝ𝑛 induced by the Euclidean metric 𝑑 and the square metric 𝜌 are
the same as the product topology on ℝ𝑛 .
Let 𝑥 = (𝑥1 , 𝑥2 , … , 𝑥𝑛 ) and 𝑦 = (𝑦1 , 𝑦2 , … , 𝑦𝑛 ) ∈ ℝ𝑛

Claim: 𝜌(𝑥, 𝑦) ≤ 𝑑(𝑥, 𝑦) ≤ √𝑛 𝜌(𝑥, 𝑦)

𝜌(𝑥, 𝑦) = 𝑚𝑎𝑥 {|𝑥𝑖 − 𝑦𝑖 |}
1≤𝑖≤𝑛

Let 𝑚𝑎𝑥 {|𝑥𝑖 − 𝑦𝑖 |} = |𝑥𝑗 − 𝑦𝑗 |

1≤𝑖≤𝑛

Note that
2
|𝑥𝑗 − 𝑦𝑗 | ≤ |𝑥1 − 𝑦1 |2 + |𝑥2 − 𝑦2 |2 + ⋯ + |𝑥𝑛 − 𝑦𝑛 |2
1
⇒ |𝑥𝑗 − 𝑦𝑗 | ≤ (|𝑥1 − 𝑦1 |2 + |𝑥2 − 𝑦2 |2 + ⋯ + |𝑥𝑛 − 𝑦𝑛 |2 )2
⇒ 𝜌 (𝑥, 𝑦) ≤ 𝑑(𝑥, 𝑦) … (1)

𝑛 𝑛
2
𝑑(𝑥, 𝑦) = √∑|𝑥𝑖 − 𝑦𝑖 |2 ≤ √∑|𝑥𝑗 − 𝑦𝑗 |
𝑖=1 𝑖=1

𝑛
2
𝑑(𝑥, 𝑦) ≤ √∑|𝑥𝑗 − 𝑦𝑗 | = √𝑛 |𝑥𝑗 − 𝑦𝑗 |
𝑖=1

= √𝑛 𝜌(𝑥, 𝑦) … (2)
Since 𝜌(𝑥, 𝑦) ≤ 𝑑(𝑥, 𝑦)
∀ 𝑦 ∈ 𝐵𝑑 (𝑥, 𝜖), 𝑑(𝑥, 𝑦) < 𝜖
So, 𝜌(𝑥, 𝑦) ≤ 𝑑(𝑥, 𝑦) < 𝜖
This implies, 𝑦 ∈ 𝐵𝜌 (𝑥, 𝜖)

LOVELY PROFESSIONAL UNIVERSITY 7

Topology

𝐵𝑑 (𝑥, 𝜖) ⊆ 𝐵𝜌 (𝑥, 𝜖)

Again, 𝑑(𝑥, 𝑦) ≤ √𝑛 𝜌(𝑥, 𝑦)

𝜖
Consider 𝐵𝜌 (𝑥, ),
√𝑛
𝜖
Let 𝑦 ∈ 𝐵𝜌 (𝑥, )
√𝑛

⇒ 𝜌(𝑥, 𝑦) < 𝜖/√𝑛

⇒ √𝑛 𝜌(𝑥, 𝑦) < 𝜖
⇒ 𝑑(𝑥, 𝑦) < 𝜖
So, by lemma, both the topologies induced by 𝑑 and 𝜌 are the same.
Now we show that product topology is the same as 𝜌.
First, let 𝐵 = (𝑎1 , 𝑏1 ) × (𝑎2 , 𝑏2 ) × … × (𝑎𝑛 , 𝑏𝑛 ) be a basis element for product topology.
Let 𝑥 = (𝑥1 , 𝑥2 , … , 𝑥𝑛 ) ∈ 𝐵
𝑥𝑖 ∈ (𝑎𝑖 , 𝑏𝑖 ), ∃ (𝑥𝑖 − 𝜖𝑖 , 𝑥𝑖 + 𝜖𝑖 ) such that
(𝑥𝑖 − 𝜖𝑖 , 𝑥𝑖 + 𝜖𝑖 ) ⊆ (𝑎𝑖 , 𝑏𝑖 )
Let 𝜖 = min{𝜖𝑖 } , 1 ≤ 𝑖 ≤ 𝑛
⇒ 𝜌 − topology is finer than the product topology.
Conversely, let 𝐵𝜌 (𝑥, 𝜖) be a basis element for 𝜌 −topology on ℝ𝑛 .
Given 𝑦 ∈ 𝐵𝜌 (𝑥, 𝜖) then 𝐵𝜌 (𝑥, 𝜖) is trivially
(𝑥1 − 𝜖1 , 𝑥1 + 𝜖1 ) × (𝑥2 − 𝜖2 , 𝑥2 + 𝜖2 ) × … × (𝑥𝑛 − 𝜖𝑛 , 𝑥𝑛 + 𝜖𝑛 )
which is again a basis element for product topology on ℝ𝑛 .
Therefore, product topology is finer than 𝜌 −topology.
Therefore, Euclidean topology, square topology and product topology are the same in ℝ𝑛 .
We consider the infinite cartesian product. Then it is natural to try to generalize the metrics 𝑑 and 𝜌
to this space.
For instance, one can attempt to define a metric 𝑑 on ℝ𝜔 by the equation
∞

𝑑(𝑥, 𝑦) = ∑|𝑥𝑖 − 𝑦𝑖 |
𝑖=1

But this equation does not always make sense, for the series in question need not converge.
Similarly, one can attempt to generalize the square metric 𝜌 to ℝ𝜔 by defining
𝜌(𝑥, 𝑦) = sup{|𝑥𝑛 − 𝑦𝑛 |}
Again, this formula does not always make sense.
However, if we replace the usual metric 𝑑(𝑥, 𝑦) = |𝑥 − 𝑦| on ℝ by its bounded counterpart 𝑑̅ (𝑥, 𝑦) =
min{𝑑(𝑥, 𝑦), 1} then this definition does make sense; it gives a metric on ℝ𝜔 called the uniform metric.
Definition 7.1.21: The uniform metric can be defined more generally on the cartesian product for
arbitrary 𝐽, as follows:
Given an index set J and given points 𝑥 = (𝑥𝛼 )𝛼∈𝐽 and 𝑦 = (𝑦𝛼 )𝛼∈𝐽 of ℝ𝐽 . Let us define a metric 𝜌̅ on
ℝ𝐽 by the equation 𝜌̅ (𝑥, 𝑦) = sup{𝑑̅ (𝑥𝛼 , 𝑦𝛼 )} where 𝑑̅ is the standard bounded metric on ℝ. It is called
𝛼∈𝐽
the uniform metric on ℝ𝐽 and the topology it induces is called the uniform topology.
Theorem 7.1.22: The uniform topology on ℝ𝐽 is finer than the product topology and coarser than the
box topology.
Proof: Suppose we are given a point 𝑥 = (𝑥𝛼 )𝛼∈𝐽
Let ∏ 𝑈𝛼 be the basis element of product topology on ℝ𝐽 such that 𝑥 ∈ ∏ 𝑈𝛼 .

8 LOVELY PROFESSIONAL UNIVERSITY

 Unit 07: Metric Topology

Let 𝛼1 , 𝛼2 , … , 𝛼𝑛 are the indices for which 𝑈𝛼 ≠ ℝ.

Since 𝑈𝛼𝑖 is open in ℝ ∀ 1 ≤ 𝑖 ≤ 𝑛.

There exists 𝜖𝑖 > 0 such that 𝐵𝑑̅ (𝑥𝛼𝑖 , 𝜖𝑖 ) ⊆ 𝑈𝛼𝑖

Let 𝜖 = min{𝜖1 , 𝜖2 , … , 𝜖𝑛 }
Claim: 𝐵𝜌̅ (𝑥, 𝜖) ⊆ ∏ 𝑈𝛼
Let 𝑧 ∈ 𝐵𝜌̅ (𝑥, 𝜖); 𝑧 = (𝑧𝛼 )𝛼∈𝐽
⇒ 𝜌̅ (𝑥, 𝑧) < 𝜖
⇒ 𝑑̅ (𝑥𝛼 , 𝑧𝛼 ) < 𝜖 ∀ 𝛼

(𝑧𝛼 )𝛼∈𝐽 ∈ ∏ 𝑈𝛼

⇒ 𝑧 ∈ ∏ 𝑈𝛼

Therefore, 𝐵𝜌̅ (𝑥, 𝜖) ⊆ ∏ 𝑈𝛼

Therefore, uniform topology is finer than product topology.
Again, let 𝐵 = 𝐵𝜌̅ (𝑥, 𝜖)
𝜖 𝜖
Consider box neighbourhood 𝑈 = ∏ (𝑥𝛼 − , 𝑥𝛼 + ) of 𝑥.
2 2
𝜖
As for 𝑦 ∈ 𝑈, 𝑑̅(𝑥𝛼 , 𝑦𝛼 ) < ∀𝛼
2
𝜖 𝜖
𝑈 = ∏ (𝑥𝛼 − , 𝑥𝛼 + )
2 2
Take 𝑦 ∈ 𝑈
𝑦 = (𝑦𝛼 )𝛼∈𝐽
𝜖 𝜖
𝑥𝛼 − < 𝑦𝛼 < 𝑥𝛼 +
2 2
𝜖
⇒ |𝑥𝛼 − 𝑦𝛼 | <
2
𝜖
⇒ 𝜌̅ (𝑥, 𝑦) <
2
Therefore, box topology is finer than uniform topology.
Theorem 7.1.23: Let 𝑑̅ (𝑎, 𝑏) = min{|𝑎 − 𝑏|, 1} be the standard bounded metric on ℝ. If 𝑥 and 𝑦 are
𝑑̅ (𝑥𝑖 , 𝑦𝑖 )
two points of ℝ𝜔 , define 𝐷(𝑥, 𝑦) = sup { }
𝑖

Then 𝐷 is a metric that induces the product topology on ℝ𝜔 .

Proof: 𝑑̅ is a metric.
Therefore, 𝑑̅ (𝑥𝑖 , 𝑦𝑖 ) ≥ 0 and 𝑑̅ (𝑥𝑖 , 𝑦𝑖 ) = 0 ⇔ 𝑥𝑖 = 𝑦𝑖 ∀ 𝑖
𝑑̅ (𝑥𝑖 , 𝑦𝑖 ) 𝑑̅ (𝑥𝑖 , 𝑦𝑖 )
⇒ ≥ 0 and = 0 ⇔ 𝑥𝑖 = 𝑦𝑖 ∀ 𝑖
𝑖 𝑖

⇒ 𝐷(𝑥, 𝑦) ≥ 0 and 𝐷(𝑥, 𝑦) = 0 ⇔ 𝑥 = 𝑦

(ii) 𝑑̅ is a metric on ℝ
Therefore, 𝑑̅ (𝑥𝑖 , 𝑦𝑖 ) = 𝑑̅ (𝑦𝑖 , 𝑥𝑖 ) ∀ 𝑖
𝑑̅ (𝑥𝑖 , 𝑦𝑖 ) 𝑑̅ (𝑦𝑖 , 𝑥𝑖 )
⇒ = ∀𝑖
𝑖 𝑖
⇒ 𝐷(𝑥, 𝑦) = 𝐷(𝑦, 𝑥)
(iii) By triangular inequality for 𝑑̅
𝑑̅(𝑥𝑖 , 𝑧𝑖 ) ≤ 𝑑̅ (𝑥𝑖 , 𝑦𝑖 ) + 𝑑̅ (𝑦𝑖 , 𝑧𝑖 )
Since 𝑖 > 0

LOVELY PROFESSIONAL UNIVERSITY 9

Topology

𝑑̅(𝑥𝑖 , 𝑧𝑖 ) 𝑑̅ (𝑥𝑖 , 𝑦𝑖 ) 𝑑̅ (𝑦𝑖 , 𝑧𝑖 )

≤ +
𝑖 𝑖 𝑖
̅
𝑑 (𝑥𝑖 , 𝑧𝑖 ) ̅
𝑑 (𝑥𝑖 , 𝑦𝑖 ) 𝑑̅ (𝑦𝑖 , 𝑧𝑖 )
≤ sup + sup
𝑖 𝑖 𝑖
̅
𝑑(𝑥𝑖 , 𝑧𝑖 )
≤ 𝐷(𝑥, 𝑦) + 𝐷(𝑦, 𝑧)
𝑖
Taking supremum,
𝐷(𝑥, 𝑧) ≤ 𝐷(𝑥, 𝑦) + 𝐷(𝑦, 𝑧) ∀ 𝑥, 𝑦, 𝑧 ∈ ℝ𝜔
Now we show that 𝐷 and the product topology are the same.
First, let 𝑈 be an open set in the metric topology and 𝑥 ∈ 𝑈.
We wish to find an open set 𝑉 in the product topology such that 𝑥 ∈ 𝑉 ⊆ 𝑈
1
Choose 𝜖 > 0, 𝐵𝐷 (𝑥, 𝜖) ⊆ 𝑈. Then choose 𝑁 large enough so that <𝜖
𝑁

Consider V = (𝑥1 − 𝜖, 𝑥1 + 𝜖) × (𝑥2 − 𝜖, 𝑥2 + 𝜖) × … × (𝑥𝑁 − 𝜖, 𝑥𝑁 + 𝜖) × ℝ × ℝ × … × ℝ

Claim: 𝑉 ⊂ 𝐵𝐷 (𝑥, 𝜖)
For 𝑦 ∈ ℝ𝜔
𝑑̅ (𝑥𝑖 , 𝑦𝑖 ) 1
≤ ∀𝑖 ≥𝑁
𝑖 𝑁
𝑑̅ (𝑥1 ,𝑦1 ) 𝑑̅ (𝑥𝑁 , 𝑦𝑁 ) 1
Therefore, 𝐷(𝑥, 𝑦) ≤ max { , …, , }
1 𝑁 𝑁

If 𝑦 ∈ 𝑉 then clearly 𝐷(𝑥, 𝑦) < 𝜖

⇒ 𝑉 ⊂ 𝐵𝐷 (𝑥, 𝜖)
Conversely, consider a basis element 𝑈 = ∏𝑖∈ℤ+ 𝑈𝑖 for the product topology where 𝑈𝑖 is an open set
in ℝ for 𝑖 = 𝛼1 , 𝛼2 , … , 𝛼𝑛 and 𝑈𝑖 = ℝ ∀ 𝑖 ≠ 𝛼1 , 𝛼2 , … , 𝛼𝑛 .
Given 𝑥 ∈ 𝑈
We try to find an open set 𝑉 of the metric topology such that 𝑥 ∈ 𝑉 ⊂ 𝑈
Choose an interval (𝑥𝑖 − 𝜖𝑖 , 𝑥𝑖 + 𝜖𝑖 ) in ℝ centred at 𝑥𝑖 such that (𝑥𝑖 − 𝜖𝑖 , 𝑥𝑖 + 𝜖𝑖 ) ⊆ 𝑈𝑖 for 𝑖 =
𝛼1 , 𝛼2 , … , 𝛼𝑛
Choose 𝜖𝑖 ≤ 1 ∀ 𝑖
𝜖
Define 𝜖 = min { 𝑖 | 𝑖 = 𝛼1 , 𝛼2 , … , 𝛼𝑛 }
𝑖

Claim: 𝑥 ∈ 𝐵𝐷 (𝑥, 𝜖) ⊂ 𝑈
Let 𝑦 ∈ 𝐵𝐷 (𝑥, 𝜖)
⇒ 𝐷(𝑥, 𝑦) < 𝜖
𝑑̅ (𝑥𝑖 , 𝑦𝑖 )
⇒ <𝜖
𝑖
𝜖𝑖
If 𝑖 = 𝛼1 , 𝛼2 , … , 𝛼𝑛 then 𝜖 ≤
𝑖

⇒ 𝑑̅ (𝑥𝑖 , 𝑦𝑖 ) < 𝜖𝑖
⇒ |𝑥𝑖 − 𝑦𝑖 | < 𝜖𝑖
Therefore, 𝑦 ∈ ∏ 𝑈𝑖
Hence, both the topologies are the same.
Theorem 7.1.24: The Hausdorff axiom is satisfied by every metric topology.
Proof: Let (𝑋, 𝑑) be a metric space.
Let 𝑥, 𝑦 ∈ 𝑋 such that 𝑥 ≠ 𝑦
⇒ 𝑑(𝑥, 𝑦) > 0

10 LOVELY PROFESSIONAL UNIVERSITY

 Unit 07: Metric Topology

Let 𝑑(𝑥, 𝑦) = 2𝜖; 𝜖 > 0

Consider open balls 𝐵𝑑 (𝑥, 𝜖) and 𝐵𝑑 (𝑦, 𝜖), neighbourhoods of 𝑥 and 𝑦 respectively.
Consider 𝑧 ∈ 𝐵𝑑 (𝑥, 𝜖) ∩ 𝐵𝑑 (𝑦, 𝜖)
⇒ 𝑑(𝑥, 𝑧) < 𝜖, 𝑑(𝑦, 𝑧) < 𝜖
⇒ 𝑑(𝑥, 𝑧) + 𝑑(𝑦, 𝑧) < 2𝜖
𝑑(𝑥, 𝑦) ≤ 𝑑(𝑥, 𝑧) + 𝑑(𝑦, 𝑧) < 2𝜖
⇒ 𝑑(𝑥, 𝑦) < 2𝜖
So, our supposition was wrong. Therefore, there exist open sets containing 𝑥 and 𝑦 respectively given
by
𝐵𝑑 (𝑥, 𝜖) ∩ 𝐵𝑑 (𝑦, 𝜖) = 𝜙
Therefore, (𝑋, 𝑑) is a Hausdorff space.

Notes:
• If 𝐴 is a subspace of the 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.
• Some of the order topologies are metrizable (for instance, ℤ+ and ℝ, and others
are not.
• The familiar "𝜖 − 𝛿" definition of continuity carries over to general metric
spaces, and so does the "convergent sequence definition" of continuity.

Theorem 7.1.25: Let 𝑓: 𝑋 → 𝑌; let 𝑋 and 𝑌 be metrizable with metrics 𝑑𝑋 and 𝑑𝑌 respectively. Then
continuity of 𝑓 is equivalent to the requirement that given 𝑥 ∈ 𝑋 and given 𝜖 > 0, there exists 𝛿 > 0
such that
𝑑𝑋 (𝑥, 𝑦) < 𝛿 ⇒ 𝑑𝑌 (𝑓(𝑥), 𝑓(𝑦)) < 𝜖
Proof: Suppose 𝑓 is continuous.
Given 𝑥 and 𝜖, consider the set 𝑉 = 𝑓 −1 (𝐵(𝑓(𝑥), 𝜖))
𝐵(𝑓(𝑥), 𝜖) is an open ball in 𝑌, hence it is an open set. Since 𝑓 is a continuous map, 𝑉 is an open set
in 𝑋.
Also, 𝑓(𝑥) ∈ 𝐵(𝑓(𝑥), 𝜖) implies, 𝑥 ∈ 𝑓 −1 (𝐵(𝑓(𝑥), 𝜖)) = 𝑉.
Hence, 𝑥 ∈ 𝑉.
Therefore, 𝑉 is an open set containing 𝑥.
By definition of an open set, there exists basis element 𝐵(𝑥, 𝛿) such that 𝑥 ∈ 𝐵(𝑥, 𝛿) ⊆ 𝑉.
We need to prove 𝑑𝑋 (𝑥, 𝑦) < 𝛿 ⇒ 𝑑𝑌 (𝑓(𝑥), 𝑓(𝑦)) < 𝜖.
Let 𝑑𝑋 (𝑥, 𝑦) < 𝛿
⇒ 𝑦 ∈ 𝐵(𝑥, 𝛿) ⊆ 𝑉
⇒ 𝑦 ∈ 𝑉 = 𝑓 −1 (𝐵(𝑓(𝑥), 𝜖))
⇒ 𝑓(𝑦) ∈ 𝐵(𝑓(𝑥), 𝜖)
⇒ 𝑑𝑌 (𝑓(𝑥), 𝑓(𝑦)) < 𝜖

Conversely, let 𝜖 − 𝛿 definition is satisfied. That is, 𝑑𝑋 (𝑥, 𝑦) < 𝛿 ⇒ 𝑑𝑌 (𝑓(𝑥), 𝑓(𝑦)) < 𝜖.
Let 𝑉 be an open set in 𝑌.
We need to show that 𝑓 −1 (𝑉) is an open set in 𝑋.
Let 𝑥 ∈ 𝑓 −1 (𝑉)
⇒ 𝑓(𝑥) ∈ 𝑉
Since 𝑉 is an open set, there exists an open ball 𝐵(𝑓(𝑥), 𝜖) ⊂ 𝑉.

LOVELY PROFESSIONAL UNIVERSITY 11

Topology

Claim: ∃ 𝛿 −ball 𝐵(𝑥, 𝛿) centred at 𝑥 such that 𝑓(𝐵(𝑥, 𝛿)) ⊂ 𝐵(𝑓(𝑥), 𝜖)

Let 𝑦 ∈ 𝑓(𝐵(𝑥, 𝛿))

⇒ 𝑦 = 𝑓(𝑧); 𝑧 ∈ 𝐵(𝑥, 𝛿)
Since 𝑧 ∈ 𝐵(𝑥, 𝛿), 𝑑𝑋 (𝑥, 𝑧) < 𝛿
By hypothesis, 𝑑𝑌 (𝑓(𝑥), 𝑓(𝑧)) < 𝜖
Or 𝑑𝑌 (𝑓(𝑥), 𝑦) < 𝜖 so that 𝑦 ∈ 𝐵(𝑓(𝑥), 𝜖)
𝐵(𝑥, 𝛿) is a neighbourhood of 𝑥.
Also, 𝑓(𝐵(𝑥, 𝛿)) ⊂ 𝐵(𝑓(𝑥), 𝜖) ⊂ 𝑉
⇒ 𝐵(𝑥, 𝛿) ⊂ 𝑓 −1 (𝑉)
∀ 𝑥 ∈ 𝑓 −1 (𝑉), ∃ 𝐵(𝑥, 𝛿) such that 𝑥 ∈ 𝐵(𝑥, 𝛿) ⊂ 𝑓 −1 (𝑉)
Hence, 𝑓 is a continuous function.
Lemma 7.1.26: The sequence lemma
Let 𝑋 be a topological space; let 𝐴 ⊆ 𝑋. If there is a sequence of points of 𝐴 converging to 𝑥, then 𝑥 ∈
𝐴̅ , the converse holds if 𝑋 is metrizable.
Proof: Suppose that {𝑥𝑛 } → 𝑥
𝑥𝑛 ∈ 𝐴
Let 𝑈 be a neighborhood of 𝑥.
Since {𝑥𝑛 } → 𝑥
Therefore, 𝑈 contains all but finitely many 𝑥𝑛 ; 𝑥𝑛 ∈ 𝐴
⇒ 𝑈∩𝐴 ≠ 𝜙
⇒ 𝑥 ∈ 𝐴̅.
Conversely, let 𝑋 is metrizable and 𝑥 ∈ 𝐴̅.
Let 𝑑 be the metric for a topology on 𝑋.
1 1
For each positive integer 𝑛, take the neighbourhood 𝐵𝑑 (𝑥, ) of radius of 𝑥.
𝑛 𝑛

Since 𝑥 ∈ 𝐴̅.
1
⇒ 𝐵𝑑 (𝑥, ) ∩ 𝐴 ≠ 𝜙
𝑛
1
Let 𝑥𝑛 ∈ 𝐵𝑑 (𝑥, ) ∩ 𝐴
𝑛

Claim: {𝑥𝑛 } → 𝑥
Let 𝑈 be an open set containing 𝑥. By definition, 𝑈 contains an 𝜖 −ball 𝐵𝑑 (𝑥, 𝜖) centred at 𝑥.
1
Choose 𝑁 so that <𝜖
𝑁

⇒ 𝑈 contains 𝑥𝑖 ∀ 𝑖 ≥ 𝑁
⇒ {𝑥𝑛 } → 𝑥
Theorem 7.1.27: Let 𝑓: 𝑋 → 𝑌. If the function 𝑓 is continuous, then for every convergent sequence
𝑥𝑛 → 𝑥 in 𝑋, the sequence 𝑓(𝑥𝑛 ) converges to 𝑓(𝑥). The converse holds if 𝑋 is metrizable.
Proof: Let 𝑓 be a continuous map.
Let 𝑥𝑛 → 𝑥
We need to prove that 𝑓(𝑥𝑛 ) → 𝑓(𝑥)
Let 𝑉 be a neighbourhood of 𝑓(𝑥) that is, 𝑓 −1 (𝑉) is a neighbourhood of 𝑥.
Since {𝑥𝑛 } → 𝑥, there exists 𝑁 such that 𝑥𝑛 ∈ 𝑓 −1 (𝑉) ∀ 𝑛 ≥ 𝑁
Then 𝑓(𝑥𝑛 ) ∈ 𝑉 ∀ 𝑛 ≥ 𝑁

12 LOVELY PROFESSIONAL UNIVERSITY

 Unit 07: Metric Topology

⇒ {𝑓(𝑥𝑛 )} → 𝑓(𝑥)
Conversely, assume that the convergent sequence definition is satisfied.
Let 𝐴 ⊆ 𝑋

Claim: 𝑓(𝐴̅) ⊆ ̅̅̅̅̅̅

𝑓(𝐴)
If 𝑥 ∈ 𝐴̅
By the previous lemma, there exists a sequence {𝑥𝑛 } in 𝐴 converging to 𝑥.
By hypothesis, {𝑓(𝑥𝑛 )} → 𝑓(𝑥)
Since 𝑓(𝑥𝑛 ) ∈ 𝑓(𝐴)
̅̅̅̅̅̅
Again, using the lemma, 𝑓(𝑥) ∈ 𝑓(𝐴)

This implies, 𝑓(𝐴̅) ⊆ ̅̅̅̅̅̅

𝑓(𝐴) and hence 𝑓 is a continuous map.
Definition 7.1.28: A space 𝑋 is said to have a countable basis at a point 𝑥 if there is a countable
collection {𝑈𝑛 }𝑛∈ℤ+ of neighbourhoods of 𝑥 such that any neighbourhood 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.

Notes:
• In the sequence lemma, if we take 𝑋 to be satisfying the first countability axiom,
1
then replacing 𝐵𝑑 (𝑥, ) by 𝐵𝑛 = 𝑈1 ∩ 𝑈2 ∩ … ∩ 𝑈𝑛 , we get the same result.
𝑛

• The proof of the theorem remains intact.

• This is because a metrizable space always satisfies the first countability axiom,
but the converse is not true.

Lemma 7.1.29: The addition, subtraction and multiplication operations are continuous functions from
ℝ × ℝ into ℝ. The quotient operation is a continuous function from ℝ × (ℝ − {0}) into ℝ.
Proof: Consider the metric 𝑑 on ℝ given by
𝑑(𝑥, 𝑦) = |𝑥 − 𝑦|
and 𝜌 on ℝ2 given by
𝜌((𝑥, 𝑦), (𝑥0 , 𝑦0 )) = max{|𝑥 − 𝑥0 |, |𝑦 − 𝑦0 |}
This implies,
𝜌((𝑥, 𝑦), (𝑥0 , 𝑦0 )) ≥ |𝑥 − 𝑥0 |
and
𝜌((𝑥, 𝑦), (𝑥0 , 𝑦0 )) ≥ |𝑦 − 𝑦0 |
To prove that a function 𝑓 is continuous from ℝ × ℝ to ℝ, we need to prove that for 𝜖 > 0, ∃ 𝛿 > 0
such that 𝑑(𝑓(𝑥, 𝑦), 𝑓(𝑥0 , 𝑦0 )) < 𝜖 whenever 𝜌((𝑥, 𝑦), (𝑥0 , 𝑦0 )) < 𝛿 for 𝑥, 𝑦, 𝑥0 , 𝑦0 ∈ ℝ.
1. Let 𝑓: ℝ2 → ℝ be defined as 𝑓(𝑥, 𝑦) = 𝑥 + 𝑦
𝑑(𝑓(𝑥, 𝑦), 𝑓(𝑥0 , 𝑦0 )) = 𝑑(𝑥 + 𝑦, 𝑥0 , +𝑦0 )
= |𝑥 + 𝑦 − (𝑥0 + 𝑦0 )|
= |𝑥 − 𝑥0 + 𝑦 − 𝑦0 |
≤ |𝑥 − 𝑥0 | + |𝑦 − 𝑦0 | … (1)
Let 𝜌((𝑥, 𝑦), (𝑥0 , 𝑦0 )) < 𝛿
⇒ |𝑥 − 𝑥0 |, |𝑦 − 𝑦0 | < 𝛿
Using (1)
𝑑(𝑓(𝑥, 𝑦), 𝑓(𝑥0 , 𝑦0 )) ≤ |𝑥 − 𝑥0 | + |𝑦 − 𝑦0 |

LOVELY PROFESSIONAL UNIVERSITY 13

Topology

< 𝛿 + 𝛿 = 2𝛿
𝜖
Choose 𝜖 = 2𝛿 or 𝛿 =
2

We get that
𝑑(𝑓(𝑥, 𝑦), 𝑓(𝑥0 , 𝑦0 )) < 𝜖 ∀ 𝜌((𝑥, 𝑦), (𝑥0 , 𝑦0 )) < 𝛿
This implies 𝑓 is continuous.
Similar proof works for subtraction.
Now consider 𝑓(𝑥, 𝑦) = 𝑥𝑦
𝑑(𝑓(𝑥, 𝑦), 𝑓(𝑥0 , 𝑦0 )) = 𝑑(𝑥𝑦, 𝑥0 𝑦0 )
= |𝑥𝑦 − 𝑥0 𝑦0 |
= |𝑥𝑦 − 𝑥0 𝑦 + 𝑥0 𝑦 − 𝑥𝑦0 + 𝑥𝑦0 + 𝑥0 𝑦0 − 𝑥0 𝑦0 − 𝑥0 𝑦0 |
≤ |𝑥0 ||𝑦 − 𝑦0 | + |𝑦0 ||𝑥 − 𝑥0 | + |𝑥 − 𝑥0 ||𝑦 − 𝑦0 | … (2)
Now 𝜌((𝑥, 𝑦), (𝑥0 , 𝑦0 )) < 𝛿
⇒ |𝑥 − 𝑥0 |, |𝑦 − 𝑦0 | < 𝛿
From (2) and choosing 𝛿 < 1, we get,
𝑑(𝑓(𝑥, 𝑦), 𝑓(𝑥0 , 𝑦0 )) < |𝑥0 |𝛿 + |𝑦0 |𝛿 + 𝛿
= (|𝑥0 | + |𝑦0 | + 1)𝛿
𝜖
Taking 𝛿 = |𝑥0 |+|𝑦0 |+1

We have 𝑑(𝑓(𝑥, 𝑦), 𝑓(𝑥0 , 𝑦0 )) < 𝜖 whenever 𝜌((𝑥, 𝑦), (𝑥0 , 𝑦0 )) < 𝛿.
This implies 𝑓 is a continuous map. A similar proof can be done for the quotient map.
Theorem 7.1.30: If 𝑋 is a topological space, and if 𝑓, 𝑔: 𝑋 → ℝ are continuous functions, then 𝑓 +
𝑓
𝑔, 𝑓 − 𝑔 and 𝑓𝑔 are continuous. If 𝑔(𝑥) ≠ 0 for all 𝑥, then is continuous.
𝑔

Proof: Consider the function ℎ: 𝑋 → ℝ × ℝ by ℎ(𝑥) = 𝑓(𝑥) × 𝑔(𝑥)

By maps into product theorem, we have ℎ(𝑥) is continuous.
Consider 𝑓 + 𝑔: 𝑋 → ℝ then 𝑓 + 𝑔 is a composite map of ℎ and +: ℝ × ℝ → ℝ.
⇒ 𝑓 + 𝑔 is continuous.
Similarly, 𝑓 − 𝑔 is a composite map of ℎ and −: ℝ × ℝ → ℝ. This implies 𝑓 − 𝑔 is continuous.
𝑓
Moreover, we can also observe in the same way that 𝑓𝑔 and (𝑔(𝑥) ≠ 0 ∀ 𝑥) are both continuous.
𝑔

Definition 7.1.31: Let 𝑓𝑛 : 𝑋 → 𝑌 be a sequence of functions from the set 𝑋 to the metric space 𝑌. Let 𝑑
be the metric for 𝑌. We say that the sequence {𝑓𝑛 } converges uniformly to the function 𝑓: 𝑋 → 𝑌 if
given 𝜖 > 0, there exists an integer 𝑁 such that 𝑑(𝑓𝑛 (𝑥), 𝑓(𝑥)) < 𝜖 for all 𝑛 > 𝑁 and all 𝑥 ∈ 𝑋.
Uniformity of convergence depends not only on the topology of 𝑌 but also on the metric.
Theorem 7.1.32: Uniform limit theorem: Let 𝑓𝑛 : 𝑋 → 𝑌 be a sequence of continuous functions from the
topological space 𝑋 to the metric space 𝑌. If {𝑓𝑛 } converges uniformly to 𝑓, then 𝑓 is continuous.
Proof: Let 𝑉 be an open set in 𝑌.
Let 𝑥0 ∈ 𝑓 −1 (𝑉)
We wish to find a neighbourhood 𝑈 of 𝑥0 such that 𝑈 ⊆ 𝑓 −1 (𝑉) or 𝑓(𝑈) ⊆ 𝑉.
Let 𝑦0 = 𝑓(𝑥0 )
Since 𝑉 is an open set in 𝑌, we can choose 𝜖 > 0 such that 𝐵(𝑦0 , 𝜖) ⊆ 𝑉.
𝜖
Then using uniform convergence, choose 𝑁 so that ∀ 𝑛 ≥ 𝑁, 𝑥 ∈ 𝑋, 𝑑(𝑓𝑛 (𝑥), 𝑓(𝑥)) < .
3
𝜖
Using continuity of 𝑓𝑁 , choose a neighbourhood 𝑈 of 𝑥0 such that 𝑓𝑁 carries 𝑈 into the ball in 𝑌
3
centred as 𝑓𝑁 (𝑥0 ).

14 LOVELY PROFESSIONAL UNIVERSITY

 Unit 07: Metric Topology

Claim: 𝑓(𝑈) ⊆ 𝐵(𝑦0 , 𝜖) ⊆ 𝑉

Note that if 𝑥 ∈ 𝑈, then by choice of 𝑁,
𝜖
𝑑(𝑓(𝑥), 𝑓𝑁 (𝑥)) <
3
By choice of 𝑈,
𝜖
𝑑(𝑓𝑁 (𝑥), 𝑓𝑁 (𝑥0 )) <
3
By choice of 𝑁,
𝜖
𝑑(𝑓𝑁 (𝑥0 ), 𝑓(𝑥0 )) <
3
Consider
𝑑(𝑓(𝑥), 𝑓(𝑥0 )) ≤ 𝑑(𝑓(𝑥), 𝑓𝑁 (𝑥)) + 𝑑(𝑓𝑁 (𝑥), 𝑓𝑁 (𝑥0 )) + 𝑑(𝑓𝑁 (𝑥0 ), 𝑓(𝑥0 ))
𝜖 𝜖 𝜖
< + + =𝜖
3 3 3
Therefore, 𝑓(𝑥) ∈ 𝐵(𝑓(𝑥0 ), 𝜖) = 𝐵(𝑦0 , 𝜖)
This implies, 𝑓(𝑈) ⊆ 𝐵(𝑦0 , 𝜖) ⊆ 𝑉
So, 𝑓 is a continuous map.

Example 7.1.33: ℝ𝜔 in the box topology is not metrizable.

Proof: We shall show that the sequence lemma does not hold for ℝ𝜔 .
Let 𝐴 ⊆ ℝ𝜔 such that 𝐴 consists of those points all of whose coordinates are positive.
𝐴 = {(𝑥1 , 𝑥2 , … )|𝑥𝑖 > 0 ∀ 𝑖 ∈ ℤ+ }
Let 𝑂 be the origin of ℝ𝜔 . Then 𝑂 = (0, 0, 0, … )
Claim 1: In box topology 0 ∈ 𝐴̅.
If 𝐵 = (𝑎1 , 𝑏1 ) × (𝑎2 , 𝑏2 ) × … is any element containing, 𝑂 then 𝐵 ∩ 𝐴 ≠ 𝜙.
𝑏 𝑏2
Because ( 1 , ,…) ∈ 𝐵 ∩ 𝐴
2 2

This implies 𝑂 ∈ 𝐴̅.

Claim 2: There does not exist any sequence of points of 𝐴 converging to 𝑂.
Let {𝑎𝑛 } be a sequence of points of 𝐴 where
𝑎𝑛 = (𝑥1𝑛 , 𝑥2𝑛 , … , 𝑥𝑖𝑛 , … ), 𝑥𝑖𝑛 > 0 ∀ 𝑖
So, we can construct a basis element 𝐵′ for box topology on ℝ as
𝐵′ = (−𝑥11 , 𝑥11 ) × (−𝑥22 , 𝑥22 ) × …
Since 𝑥𝑛𝑛 ∉ (−𝑥𝑛𝑛 , 𝑥𝑛𝑛 )
⇒ 𝑛𝑡ℎ coordinate is not a member of 𝑛𝑡ℎ set.
⇒ {𝑎𝑛 } does not tend to 0 in box topology.
Therefore, ℝ𝜔 is not a metrizable space.

Example 7.1.34: An uncountable product of ℝ with itself is not metrizable.

Proof: Let 𝐽 be an uncountable index set.

We show that ℝ𝐽 does not satisfy the sequence lemma.
Let 𝐴 ⊆ ℝ𝐽 such that
𝐴 = {(𝑥𝛼 )𝛼∈𝐽 |𝑥𝛼 = 1 for all but finitely many 𝛼}

LOVELY PROFESSIONAL UNIVERSITY 15

Topology

Let 𝑂 be the origin in ℝ𝐽 .

Claim: 𝑂 ∈ 𝐴̅
Let ∏ 𝑈𝛼 be a basis element containing 𝑂. Then 𝑈𝛼 ≠ ℝ for only finitely many values of 𝛼.
Let 𝛼 = 𝛼1 , 𝛼2 , … , 𝛼𝑛 ; 𝑈𝛼 ≠ ℝ.
Let (𝑥𝛼 ) be the point of 𝐴 defined by letting 𝑥𝛼 = 0 for 𝛼 = 𝛼1 , 𝛼2 , … , 𝛼𝑛 and 𝑥𝛼 = 1 ∀ 𝛼 ≠
𝛼1 , 𝛼2 , … , 𝛼𝑛 .
Then (𝑥𝛼 ) ∈ 𝐴 ∩ ∏ 𝑈𝛼

𝐴 ∩ ∏ 𝑈𝛼 ≠ 𝜙

This implies, 𝑂 ∈ 𝐴̅.

Claim 2: There does not exist any sequence of points of 𝐴 converging to 𝑂.
Let {𝑎𝑛 } be a sequence of points of 𝐴.
Given 𝑛, let 𝐽𝑛 denote the subset of 𝐽 consisting of those indices 𝛼 for which the 𝛼𝑡ℎ coordinate of
𝑎𝑛 ≠ 1.
The union of all 𝐽𝑛 is a countable union of finite sets is countable.
Because 𝐽 is not countable, there exists an index 𝛽 ∈ 𝐽 that does not lie in any of the sets 𝐽𝑛 .
This implies that for each of the points 𝑎𝑛 , its 𝛽𝑡ℎ coordinate is equal to 1.
Now let 𝑈𝛽 = (−1, 1) in ℝ, let 𝑈 = 𝜋𝛽−1 (𝑈𝛽 ) in ℝ𝐽 is an open set.

The set 𝑈 is a neighbourhood of 0 that contains none of 𝑎𝑛 . Therefore, {𝑎𝑛 } does not converge to 0.
This implies, ℝ𝐽 is not metrizable.

Summary
• Metric topology is defined and explained with the help of examples
• Metric topology is defined on ℝ𝑛 , and the Euclidean, square and product topology on ℝ𝑛
are compared
• Uniform topology is defined on ℝ𝐽 , and the product, uniform, and box topology on ℝ𝐽 are
compared
• The definition of continuity in terms of open balls is defined equivalent to the 𝜖 − 𝛿
definition
• The sequence lemma is proved
• The first countability axiom is defined, and new ways to construct continuous functions are
established
• Examples of some non-metrizable spaces are given

Keywords
• Metric topology
• Euclidean space
• Square metric
• Sequence Lemma
• First countability axiom
• Non-metrizable spaces

Self-assessment
Choose the most suitable answer from the options given with each question.

16 LOVELY PROFESSIONAL UNIVERSITY

 Unit 07: Metric Topology

𝟏 𝒙≠𝒚
Question 1: Let 𝑿 be a non-empty set. Then the map 𝒅: 𝑿 × 𝑿 → ℝ defined by 𝒅(𝒙, 𝒚) = { .
𝟎 𝒙=𝒚
Then 𝒅
A: does not induce a metric on 𝑋
B: induces a metric on 𝑋 and (𝑋, 𝑑) is called a discrete metric.
C: induces a metric on 𝑋 and (𝑋, 𝑑) is called an indiscrete metric.
D: induces a metric on 𝑋 and (𝑋, 𝑑) is called the usual metric.

Question 2: Which of the following function does not define a metric on the set of real numbers
A: 𝑑(𝑥, 𝑦) = |𝑥 − 𝑦|
B: 𝑑(𝑥, 𝑦) = 0
C: 𝑑(𝑥, 𝑦) = 𝑥 + 𝑦
1 𝑥≠𝑦
D: 𝑑(𝑥, 𝑦) = {
0 𝑥=𝑦

Question 3: Let 𝑿 be the set of real numbers. Then for 𝒙 ∈ 𝑿, 𝝐 > 𝟎, open ball 𝑩(𝒙, 𝝐) is defined as
A: (𝑥 − 𝜖, 𝑥 + 𝜖)
B: (𝑥 − 𝜖, 𝑥 + 𝜖]
C: [𝑥 − 𝜖, 𝑥 + 𝜖)
D: [𝑥 − 𝜖, 𝑥 + 𝜖]

Question 4: Let 𝑿 be the set of real numbers. Then the diameter of the set [𝟎, 𝟏] ⊂ 𝑿 is given by
A: 1
B: 0.5
C: 0
D: 2

Question 5: For a metric space (𝑿, 𝒅);

A: we can always find some metric (𝑋, 𝑑′ ) which is bounded.
B: There may not exist any metric (𝑋, 𝑑′ ) which is bounded.
C: There never exist any metric (𝑋, 𝑑′ ) which is bounded.
D: (𝑋, 𝑑) is always bounded.

Question 6: Consider the set of real numbers. Let 𝒅 and 𝝆 define Euclidean and square metrics on
ℝ𝟒 respectively. Then
A: The topology induced by 𝜌 is strictly finer than 𝑑
B: The topology induced by 𝑑 is strictly finer than 𝜌
C: The topology induced by 𝜌 is the same as that by 𝑑
D: The topologies generated by 𝑑 and 𝜌 are not comparable.

Question 7: Consider the statements

I: The uniform topology on ℝ𝑱 is finer than the product topology.

LOVELY PROFESSIONAL UNIVERSITY 17

Topology

II: The uniform topology on ℝ𝑱 is finer than the box topology.

A: Only statement I is correct.
B: Only statement II is correct.
C: Both statements I and II are correct.
D: None of the statements I and II is correct.

Question 8: The Hausdorff axiom is satisfied by

A: Every metric space
B: Only finite metric spaces
C: Only infinite metric spaces
D: No metric space

Question 9: Consider the two definitions given in calculus

D1: "𝝐 − 𝜹" definition of continuity
D2: “convergent sequence definition”
Then choose the correct option
A: Only D1 carries over to general metric spaces
B: Only D2 carries over to general metric spaces
C: Both D1 and D2 carry over to general metric spaces
D: Neither D1 nor D2 carry over to general metric spaces

Question 10: The "𝝐 − 𝜹" definition of continuity of a function 𝒇 from a metric space 𝑿 to another
space 𝒀 is equivalent to saying that given 𝒙, 𝒚 ∈ 𝑿 and given 𝝐 > 𝟎, there exists 𝜹 > 𝟎 such that
A: 𝑑𝑋 (𝑥, 𝑦) < 𝛿 ⇒ 𝑑𝑌 (𝑓(𝑥), 𝑓(𝑦)) < 𝜖

B: 𝑑𝑋 (𝑥, 𝑦) < 𝛿 ⇐ 𝑑𝑌 (𝑓(𝑥), 𝑓(𝑦)) < 𝜖

C: 𝑑𝑋 (𝑥, 𝑦) > 𝛿 ⇒ 𝑑𝑌 (𝑓(𝑥), 𝑓(𝑦)) < 𝜖

D: 𝑑𝑋 (𝑥, 𝑦) < 𝛿 ⇒ 𝑑𝑌 (𝑓(𝑥), 𝑓(𝑦)) > 𝜖

Question 11: Let 𝑿 be a metric space and 𝑨 ⊆ 𝑿, 𝑨 ≠ 𝝓. Then for a convergent sequence {𝒙𝒏 } ∈ 𝑨, if
{𝒙𝒏 } converges to 𝒙, then for any open set 𝑼 containing 𝒙,
A: 𝑈 ∩ 𝐴 = 𝜙
B: 𝑈 ∩ 𝐴 ≠ 𝜙
C: 𝑈 ∩ 𝐴 = {𝑥}
D: 𝑈 ∩ 𝐴 = 𝐴

Question 12: Let 𝒇: 𝑿 → 𝒀 be a continuous function. Then the statement “for every convergent
sequence 𝒙𝒏 → 𝒙 in 𝑿, the sequence 𝒇(𝒙𝒏 ) converges to 𝒇(𝒙) in 𝒀” is true for
A: any topological spaces 𝑋 and 𝑌
B: any topological space 𝑋 but for metrizable space 𝑌
C: any topological space 𝑌 but for metrizable space 𝑋

18 LOVELY PROFESSIONAL UNIVERSITY

 Unit 07: Metric Topology

D: both 𝑋 and 𝑌 should be non-metrizable

𝒇
Question 13: Let 𝒇 and 𝒈 are two continuous functions on ℝ. Then the function is
𝒈

A: always continuous
B: never continuous
C: continuous only if 𝑔(𝑥) = 0 for some 𝑥 ∈ ℝ
D: continuous only if 𝑔(𝑥) ≠ 0 ∀ 𝑥 ∈ ℝ

Question 14: Let 𝒇𝒏 : 𝑿 → 𝒀 be a sequence of functions from the set 𝑿 to the metric space 𝒀. Let 𝒅
be the metric for 𝒀. Uniformity of convergence of the sequence {𝒇𝒏 } depends
A: only on the topology of 𝑌 but not on the metric.
B: only on the metric of 𝑌 but not on the topology.
C: both on the topology and the metric of 𝑌
D: Neither on the topology of 𝑌 nor the metric.

Question 15: The sum of two continuous functions 𝒇 and 𝒈 on ℝ is

A: always continuous
B: never continuous
C: continuous only if 𝑓 + 𝑔 = 0
D: continuous only if 𝑓 + 𝑔 is invertible

Answers:

1) B 2) C 3) A 4) A 5) A

6) C 7) A 8) A 9) C 10) A

11) B 12) A 13) D 14) C 15) A

Review Questions
1. Let 𝑋 be a metric space with metric 𝑑. Show that 𝑑: 𝑋 × 𝑋 → ℝ is continuous.
2. Let 𝑋 be a metric space with metric 𝑑. Let 𝑋 ′ denote a space having the same underlying set
as 𝑋. Show that if 𝑑: 𝑋 ′ × 𝑋 ′ → ℝ is continuous, then the topology of 𝑋 ′ is finer than the
topology of 𝑋.
3. Show that the Euclidean metric 𝑑 on ℝ𝑛 is a metric, as follows: If 𝑥, 𝑦 ∈ ℝ𝑛 and 𝑐 ∈ ℝ, define
𝑥 + 𝑦 = (𝑥1 + 𝑦1 , 𝑥2 + 𝑦2 , … , 𝑥𝑛 + 𝑦𝑛 ), 𝑐𝑥 = (𝑐𝑥1 , 𝑐𝑥2 , … , 𝑐𝑥𝑛 ), 𝑥 ⋅ 𝑦 = 𝑥1 𝑦1 + 𝑥2 𝑦2 + ⋯ + 𝑥𝑛 𝑦𝑛 .
Show that 𝑥 ⋅ (𝑦 + 𝑧) = (𝑥 ⋅ 𝑦) + (𝑥 ⋅ 𝑧).
4. Show that the Euclidean metric 𝑑 on ℝ𝑛 is a metric, as follows: If 𝑥, 𝑦 ∈ ℝ𝑛 and 𝑐 ∈ ℝ, define
𝑥 + 𝑦 = (𝑥1 + 𝑦1 , 𝑥2 + 𝑦2 , … , 𝑥𝑛 + 𝑦𝑛 ), 𝑐𝑥 = (𝑐𝑥1 , 𝑐𝑥2 , … , 𝑐𝑥𝑛 ), 𝑥 ⋅ 𝑦 = 𝑥1 𝑦1 + 𝑥2 𝑦2 + ⋯ + 𝑥𝑛 𝑦𝑛 .
Show that |𝑥 ⋅ 𝑦| ≤ ‖𝑥‖‖𝑦‖.
5. Show that the Euclidean metric 𝑑 on ℝ𝑛 is a metric, as follows: If 𝑥, 𝑦 ∈ ℝ𝑛 and 𝑐 ∈ ℝ, define
𝑥 + 𝑦 = (𝑥1 + 𝑦1 , 𝑥2 + 𝑦2 , … , 𝑥𝑛 + 𝑦𝑛 ), 𝑐𝑥 = (𝑐𝑥1 , 𝑐𝑥2 , … , 𝑐𝑥𝑛 ), 𝑥 ⋅ 𝑦 = 𝑥1 𝑦1 + 𝑥2 𝑦2 + ⋯ + 𝑥𝑛 𝑦𝑛 .
Show that ‖𝑥 + 𝑦‖ ≤ ‖𝑥‖ + ‖𝑦‖ and verify that 𝑑 is a metric.

LOVELY PROFESSIONAL UNIVERSITY 19

Topology

Further Readings
1) TOPOLOGY by JAMES R. MUNKRES, PEARSON
2) INTRODUCTION TO TOPOLOGY PURE AND APPLIED by C. ADOMS AND R.
FRANZOSA, PEARSON

Weblinks
1) https://onlinecourses.nptel.ac.in/noc22_ma25/preview
2) https://www.google.com/url?q=https://onlinecourses.nptel.ac.in/noc22_ma61/
preview&sa=D&source=editors&ust=1658319329167901&usg=AOvVaw0ZXNJUG
ZeoRZEfXGTBleKa

20 LOVELY PROFESSIONAL UNIVERSITY