General Topology Lecture Notes

William S. Neal February 24, 2014

Contents
1 Lecture 01 1.1 History of Topology . . . . . . . . 1.2 Set Theoretic Preliminaries . . . . 1.3 Relations and Functions . . . . . . 1.4 Cardinality and Operations on Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 3 3 4 6

2 Lecture 02 7 2.1 Logic and Techniques of Proof . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.2 Topology on a Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.3 Open and Closed Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 3 Lecture 03 3.1 Set Theoretic Preliminaries 3.2 Interior and Closure . . . . 3.3 Exterior and Boundary . . . 3.4 Examples . . . . . . . . . . 12 12 13 14 16

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

4 Lecture 04 18 4.1 Interior, Closure, Exterior, and Boundary Points . . . . . . . . . . . . . . . . . . . . . . . 18 4.2 Cluster Points and Isolated Points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 5 Lecture 05 23 5.1 The Isolated Points of a Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 5.2 Introduction to Connectedness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 5.3 Introduction to Separation Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 6 Lecture 06 6.1 Dense Subsets . . . . . . . . . . . . . . . . . . . 6.2 Nowhere Dense Subsets . . . . . . . . . . . . . 6.3 Separable Spaces . . . . . . . . . . . . . . . . . 6.4 Equivalent Statements About Open and Closed 6.5 More on Cluster Points and Isolated Points . . 28 28 30 31 32 32

. . . . . . . . . Sets . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

7 Lecture 07 33 7.1 Basis for a Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 7.2 The Digital Line Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 8 Lecture 08 38 8.1 More Properties of a Basis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 8.2 Calculating the Interior and Closure of Unions and Intersections . . . . . . . . . . . . . . 39 8.3 Properties of the Digital Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 9 Lecture 09 9.1 More Properties of the Digital Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2 Relations on a Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3 Introduction to the Order Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 43 45 47

10 Lecture 10 48 10.1 More Properties of Posets and Simply Ordered Sets . . . . . . . . . . . . . . . . . . . . . 48 10.2 Properties of the Order Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 11 Lecture 11 53 11.1 Compositions, Inverses, and Restrictions of Functions . . . . . . . . . . . . . . . . . . . . 53 11.2 Finite Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 12 Lecture 12 12.1 The Conite Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.2 Countable Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.3 Introduction to the Cocountable Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 57 60 62

13 Lecture 13 62 13.1 Properties of the Conite and Cocountable Topologies . . . . . . . . . . . . . . . . . . . . 62 13.2 More Countability Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 13.3 The Order Topology on Q . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 14 Lecture 14 68 14.1 More on Local Bases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 14.2 The Order Topology on Q and R . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 14.3 Introduction to the Subspace Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 15 Lecture 15 73 15.1 Properties of the Subspace Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 15.2 Convex Subsets and Subbases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 16 Lecture 16 78 16.1 Inherited Properties of Subspaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 16.2 Functions Acting on Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 16.3 Introduction to Continuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

1
1.1

Lecture 01
History of Topology

Remark 1.1. The original source for the following historical excerpts is the 6th edition of An Introduction to the History of Mathematics by Howard Eves. Topology started as a branch of geometry, but is now more properly considered, along with geometry, algebra, and analysis, as a fundamental division of mathematics (Eves, 617). From the [geometric] point of view, topology may be regarded as the study of those properties of geometric gures that remain invariant under single-valued continuous mappings possessing single-valued continuous inverses (Eves, 617). The notion of a geometric gure as made up of a nite set of joined fundamental pieces gradually gave way to the concept [that] an arbitrary set can constitute a topological space. This latter, and very general, viewpoint of topology has become known as set topology (Eves, 619). The earlier [geometric] viewpoint has become known as algebraic topology (Eves, 619). Remark 1.2. The functions mentioned in the second historical excerpt are bijections where both the function and its inverse are continuous. These functions have a special name. They are called homeomorphisms. Remark 1.3. Set topology is more commonly called general topology. Remark 1.4. General topology has evolved into the study of the properties of abstract spaces. Of particular interest are those properties which are preserved under homeomorphisms. Such properties are called topological properties. Fundamentally, general topology is concerned with abstract spaces formed from arbitrary sets. Accordingly, the foundation underlyining topology, and indeed all of modern mathematics, is set theory.

1.2

Set Theoretic Preliminaries

Remark 1.5. We accept as a fundamental intuitive fact the primative notion of a set : A set is a collection of objects called the elements or members of the set. Let A be a set. If the element x is in A, we write x A. If x is not in A, we write x / A. Denition 1.6. Let A and B be sets. Then A is a subset of B , denoted A B , if every element of A is also an element of B . That is, if x A x B , then A B . Remark 1.7. We accept as an axiom from set theory the existence of a set = {} which has no elements. is called the empty set or null set. If A is any arbitrary set, then A. The empty set satises the dention of subset vacuously. There are no elements in that must be contained in the other set. 3

Denition 1.8. Let A and B be sets. Then A = B if and only if A B and B A. In notation, A = B A B and B A. Remark AB AB A B 1.9. (Explanation of Notation) means x A, x B means A B or A = B means A B and A = B B , then A is called a proper subset of B .

Denition 1.10. If A

Lemma 1.11. Every set is a subset of itself. Proof. Let A be a set and let x A. Since x A x A, A A. Lemma 1.12. (Transitivity of Subsets) If A B and B C , then A C . Proof. Suppose that A B and B C . Let x A. Then as A B , x B . Now as B C , x C . Hence, A C . So we can represent this fact in one statement: A B C . Denition 1.13. Let A and B be sets. The Cartesian product of A and B is the set A B = {(a, b) : a A and b B }. Example 1.14. Let A = {a, b} and B = {1, 2, 3}. Then A B = {(a, 1), (a, 2), (a, 3), (b, 1), (b, 2), (b, 3)}. In particular, (2, b) / A B but (2, b) B A.

1.3

Relations and Functions

Denition 1.15. Let A and B be sets. A relation between A and B is any subset R A B . In particular, R contains ordered pairs. Remark 1.16. If (a, b) R, this is also denoted aRb. Denition 1.17. Let A and B be sets. A function between A and B is a nonempty relation A B with the following property: If (a, b) and (a, c) , then b = c. Remark 1.18. A function is also called a map or mapping. Denition 1.19. Let be a function between sets A and B . The domain of the function is the set dom = {a A : b B such that (a, b) }. The range of the function is the set rng = {b B : a A such that (a, b) }. Denition 1.20. Let be a function between sets A and B . If dom = A, then is a function from A into B , denoted : A B . In this case, the set A is called the domain of the function and the set B is called the codomain of the function. 4

Remark 1.21. Let : A B . If (a, b) , this is denoted as (a) = b. Denition 1.22. The function : A B is called an injection, or an injective function, if every element in the domain A is mapped into a unique element in the codomain B . If is injective, then [(a) = (b)] [a = b]. Example 1.23. The function : Z Z dened by (a) = 2a is injective. Proof. Suppose (a) = (b). Then 2a = 2b. So a = b. is an injection. Remark 1.24. An injection is sometimes called a 1 1 function (read one to one). Denition 1.25. The function : A B is called a surjection, or a surjective function, if each element in the codomain B is mapped onto by at least one element from the domain A. If is a surjection, then [b B ] [a A such that (a) = b]. Example 1.26. The function : Z Z dened by (a) = 2a is not surjective. Proof. Notice 1 Z and a Z such that 1 = (a). Example 1.27. The function : Z 2Z dened by (a) = 2a is surjective. Proof. Let b 2Z. Then b = 2a for some a Z. That is, b 2Z a Z such that b = (a). is a surjection. Remark 1.28. A surjection is sometimes called an onto function. Remark 1.29. A necessary condition for : A B to be a surjection is rng = B . Denition 1.30. A bijection is a function : A B that is both injective and surjective. Remark 1.31. A bijection is sometimes called a 1 1 correspondence. Example 1.32. The function : Z 2Z dened by (a) = 2a is a bijection. Proof. We have already demonstrated that is a surjection. Suppose (a) = (b). Then 2a = 2b. So a = b. Hence, is an injection. is a bijection as it is both injective and surjective. Remark 1.33. A necessary condition for a function : A B to be a bijection is that the number of elements in the domain and codomain must be equal. That is, |A| = |B |.

1.4

Cardinality and Operations on Sets

Denition 1.34. The cardinality of a set is the number of distinct elements in the set. The cardinality of the set A is denoted by |A|. Example 1.35. The set A = {1, 1, 1, 2} has two distinct elements. So |A| = 2. Denition 1.36. A set A is nite if a bijection : A {1, 2, . . . , n} for some positive integer n. In this case, |A| = n. Denition 1.37. A set which is not nite is innite. Denition 1.38. A set A is countably innite if a bijection : A Z+ . In this case, |A| = 0 . Denition 1.39. A set is countable if it is either nite or countably innite. Denition 1.40. An innite set which is not countably innite is uncountable. Denition 1.41. Let A and B be sets. The union of the sets A and B , denoted A B , is the set that contains those elements that are either in A or in B , or in both. Remark 1.42. Notice that A A B and B A B . Further, if A B , then A B = B . So A = A. Denition 1.43. Let A and B be sets. The intersection of the sets A and B , denoted A B , is the set containing those elements that are simultaneously in both A and B . Remark 1.44. Notice A B A and A B B . Further, if A B , then A B = A. So A = . Denition 1.45. Sets are called disjoint if their intersection is the empty set. Denition 1.46. Given a countable set A, the power set of A, denoted P (A), is the set of all subsets of the set A. Remark 1.47. If A is a nite set with cardinality |A| = n, then the cardinality of the power set of A is |P (A)| = 2n . If |A| = 0 , then |P (A)| = 20 = c, where c is a transnite number which stands for continuum. Example 1.48. Let A = {a, b, c}. P (A) = {, {a}, {b}, {c}, {a, b}, {a, c}, {b, c}, {a, b, c}}. Notice |A| = 3 and |P (A)| = 23 = 8. Denition 1.49. A collection of sets is called a family of sets. Remark 1.50. Notice that P (A) is a family of sets containing the subsets of A. Denition 1.51. Let I be a non-empty set, called an indexing set, such that for each index i I there corresponds a set Ai . Then the family A = {Ai : i I } is an indexed family of sets. Denition 1.52. Let {Ai : i I } be an indexed family of sets. The arbitrary union of sets in the indexed family of sets over the index i is the set iI Ai = {x : x Ai for at least one i I }. Denition 1.53. Let {Ai : i I } be an indexed family of sets. The arbitrary intersection of sets in the indexed family of sets over the index i is the set iI Ai = {x : x Ai for every i I }. 6

 Remark 1.54. If |I | = 0 , then iI Ai = i=1 Ai and is called a countable union. Similarly, i=1 Ai is called a countable intersection. n If |I | = n, then iI Ai = n i=1 Ai is called a nite i=1 Ai and is called a nite union. Similarly, intersection.

Remark 1.55. (Arbitrary Unions and Intersections) If x iI Ai , then x Ai for at least one index i I . In notation, we have x iI Ai i I such that x Ai . If x / iI Ai , then x / Ai for any index i I . In notation, we have x / iI Ai i I such that x Ai . If x iI Ai , then x Ai for every index i I . In notation, we have x iI Ai x Ai i I . / Ai . In notation, we have If x / iI Ai , then at lest one i I such that x x / iI Ai i I such that x / Ai . Denition 1.56. Let A X . The complement of A in X is the set X \ A = {x X : x / A}. Remark 1.57. Notice that if A X and x / A, then x X \ A. Theorem 1.58. (De Morgans Laws) Let {Ai : i I } be an indexed family of sets where each Ai X . Then (1) X \ iI Ai = iI (X \ Ai ) (2) X \ iI Ai = iI (X \ Ai ) / Ai . Thus, x X \ Ai for at / iI Ai . So i I such that x Proof. (1) Let x X \ iI Ai . Then x least one index i I . Hence, x iI (X \ Ai ). Conversely, let x iI (X \ Ai ). Then x X \ Ai for at least one index i I . So i I such that x / Ai . Thus x / iI Ai . Hence, x X \ iI Ai . (2) Let x X \ iI Ai . Then x / iI Ai . So i I such that x Ai . Thus, x X \ Ai i I . Hence, x iI (X \ Ai ). Conversely, let x iI (X \ Ai ). Then x X \ Ai i I . So i I such that x Ai . Thus, x / iI Ai . Hence, x X \ iI AI .

2
2.1

Lecture 02
Logic and Techniques of Proof

Denition 2.1. An implication is a compound statement of the form p q . When written out, it has the form If p, then q. In the implication, the statement p is called the hypothesis and the statement q is called the conclusion. Remark 2.2. To prove an implication directly, we assume that the hypothesis is true and show that this logically leads to the conclusion. Remark 2.3. Given an implication, we can form other statements, namely, the converse, the contrapositive, and the inverse. Denition 2.4. Given an implication p q , the converse of the implication is the compound statement q p. 7

Remark 2.5. Notice that in the converse we have a reversal of roles. What was the hypothesis in the inplication is now the conclusion in the converse and what was the conclusion in the implication is now the hypothesis in the converse. Denition 2.6. Given an implication p q , the contrapositive of the implication is the compound statement q p. Denition 2.7. Given an implication p q , the inverse of the implication is the compound statement p q . Remark 2.8. We can group the four statements in pairs that are logically equivalent: The implication is logially equivalent to its contrapositive: p q q p The converse is logically equivalent to the inverse: q p p q Notice that the inverse is the contrapositive of the converse. Remark 2.9. (Indirect Proof of an Implication by its Contrapositive) We can prove an implication indirectly by proving its contrapositive. To do this, we assume the negation of the conclusion and show that this leads logically to the negation of the hypothesis. Remark 2.10. (Proof by Contradiction) Proof by contradiction is based on the logical equivalence p q (p q ) Contradiction. So to prove an implication by contradiction, we assume that both the hypothesis and the negation of the conclusion are true and show that this leads to a contradiction. Remark 2.11. (Proof by Cases) Proof by cases is based on the logical equivalence (p q ) r p r q r. We use this technique of proof when we must consider two or more cases in the hypothesis. Example 2.12. Suppose we wish to prove that A1 A2 B . Then for any arbitrary element x A1 A2 , the element can be in the set A1 or the set A2 . So we must prove that both cases lead to the conclusion that the element is also in the set B . That is, we prove both x A1 x B and x A2 x B . When both implications are true, the implication (x A1 x A2 ) B is also true. Hence, A1 A2 B . Denition 2.13. A biconditional is a statement of the form p q . When written out, it has the form p if and only if q. The phrase if and only if is abbreviated i. Remark 2.14. The biconditional p q is logically equivalent to the implication p q and its converse q p: [p q ] [p q q p] So to prove a biconditional directly, we must prove that both the implication and its converse are true. Remark 2.15. (Indirect Proof of the Biconditional by its Contrapositive) Just as we can prove an implication indirectly by proving its contrapositive, we can prove a biconditional indirectly by proving its contrapositive. The contrapositive of a biconditional p q is the statement p q . It is formed by taking the contrapositive of both the implication and its converse.

2.2

Topology on a Set

Denition 2.16. A topology on a set X is a collection of subsets of X such that (1) , X . (2) If {Ui : i I } is an indexed family of subsets of X with Ui , then iI Ui . (3) If U1 , U2 , . . . , Un , then n i=1 Ui . Denition 2.17. A set X together with a topology specied on X is called a topological space and is often denoted (X, ). Remark 2.18. In mathematics, a space is a set which admits some type of structure. In a topological space, the structure of interest is the topology itself. Remark 2.19. Given a topological space X , the empty set and the entire set are in the space. Arbitrary unions of elements in the space remain in the space and nite intersections of elements in the space remain in the space. Denition 2.20. (Indiscrete Topology) Let X be any arbitrary set and let = {, X }. Then is the indiscrete topology on X . Proof. (1) Clearly, , X . (2) X = X . (3) X = . is a topology on X . Remark 2.21. The indiscrete topology is also called the trivial topology. Denition 2.22. (Discrete Topology) Let X be a countable set and let = P (X ). Then is the discrete topology on X . Proof. (1) Since X and X X , , X . (2) An arbitrary union of subsets of X is itself a subset of X . (3) A nite intersection of subsets of X is itself a subset of X . is a topology on X . Denition 2.23. (Particular Point Topology) Let X be any set and let p X be a specied point. Then = {} {U X : p U } is called the particular point topology. Proof. (1) since {} . Similarly, X since X {U X : p U } . (2) Let {Ui : i I } be an indexed family of sets such that Ui i I . Notice Ui = or Ui {U X : p U }. Suppose p Ui for at least one index i. Then p iI Ui . Hence, iI Ui {U X : p U } . Otherwise, iI Ui = . In either case, iI Ui . (3) Let U1 , . . . , Un . Again, either Ui = or Ui {U X : p U }. Suppose p Ui for every index i. n Then p n i=1 Ui . Hence, i=1 Ui {U X : p U } . 9

Otherwise, n i=1 Ui = . In either case, n i=1 Ui . is a topology on X . Example 2.24. (Excluded Point Topology) Let X be any set and let p X be a specied point. Then = {X } {U X : p / U } is called the excluded point topology. Proof. (1) X since X {X } . Similarly, since {U X : p / U} . (2) Let {Ui : i I } be an indexed family of sets such that Ui i I . Notice Ui = X or Ui {U X : p / U }. Suppose p / Ui for any index i. / U} . Then p / iI Ui . Hence, iI Ui {U X : p Otherwise, iI Ui = X . In either case, iI Ui . (3) Let U1 , . . . , Un . Again, either Ui = X or Ui {U X : p / U }. Suppose p / Ui for at least one index i. n / U} . Then p / n i=1 Ui {U X : p i=1 Ui . Hence, n Otherwise, i=1 Ui = X . In either case, n i=1 Ui . is a topology on X . Denition 2.25. (Comparable Topologies) Let and be two topologies dened on the same underlying set X . If , then is said to be ner than . Equivalently, is said to be coarser than . If , then is said to be strictly ner than . If either or , then the two topologies are said to be comparable. Otherwise, the two topologies are incomparable. Remark 2.26. Evidently, the indiscrete topology is the coarsest topology possible while the discrete topology is the nest topology possible. Remark 2.27. Recall from the previous lecture that if A B , then A B = A. So if , then = . Equivalently, if = , then . Similarly, if = , then . So if the intersection is not one of the topologies, then the two topologies are incomparable. Example 2.28. Let p X be the specied point, let P be the particular point topology on X , and let E be the excluded point topology on X . Then P E = [{} {U X : p U }] [{X } {U X : p / U }] = {, X } So P and E are incomparable. Example 2.29. Let X = {a, b} and let a be the specied point. The indiscrete topology on X is I = {, {a, b}}. The discrete topology on X is D = {, {a}, {b}, {a, b}}. 10

The particular point topology on X is P = {, {a}, {a, b}}. The excluded point topology on X is E = {, {b}, {a, b}}. Notice I P D and I E D . Remark 2.30. Notice that if in example 2.29 X = {a}, then I = P = E = D . In particular, since I = D , any topology dened on a singleton set is indistinguishable from the indiscrete topology. Denition 2.31. (Sierpinski Space) A topological space consisting of a two point set and either the particular point topology or the excluded point topology is called a Sierpinski space. Remark 2.32. A Sierpinski space is the smallest topological space that is neither indiscrete nor discrete.

2.3

Open and Closed Sets

Denition 2.33. Let (X, ) be a topological space. A subset U X is open in X if U . Denition 2.34. Let (X, ) be a topological space. A subset A X is closed in X if (X \ A) . That is, a subset A X is closed if its complement X \ A is open. Remark 2.35. Within a topological space, the complement of an open set is closed and the complement of a closed set is open. Example 2.36. The subset (0, 1) R is called an open interval in R because there exists a topology on R, called the standard topology on R, such that (0, 1) = {x R : 0 < x < 1} . Similarly, [0, 1] = {x R : 0 x 1} is called a closed interval in R since its complement R \ [0, 1] = (, 0) (1, ) . Remark 2.37. Let (X, ) be a topological space. Then (1) and X are open. (2) Arbitrary unions of open sets are open. (3) Finite intersections of open sets are open. Theorem 2.38. Let (X, ) be a topological space. Then (1) and X are closed. (2) Arbitrary intersections of closed sets are closed. (3) Finite unions of closed sets are closed. Proof. (1) Since X \ = X , is closed. Similarly, since X \ X = , X is closed. (2) Let {Ui : i I } be an indexed family of sets such that Ui i I . Then X \ Ui is closed i I . Since iI (X \ Ui ) = X \ iI Ui and iI Ui , iI (X \ Ui ) is closed. n (3) Let U1 , . . . , Un . Then X \ Ui is closed i {1, . . . , n}. Since n i=1 (X \ Ui ) = X \ i=1 Ui and n n i=1 Ui , i=1 (X \ Ui ) is closed. Remark 2.39. In the rst statement of Theorem 2.38, we are confronted right away with a fundamental fact about open and closed sets. Not only are the two not opposites, but they are also not mutually exclusive. In particular, and X are both open and closed in any topological space X . In general, a subset of the underlying set in a given topological space can be (1) both open and closed. (2) open but not closed. (3) closed but not open. (4) neither open nor closed. 11

Example 2.40. Let X = {a, b} and let a be the specied point. The particular point topology on X is P = {, {a}, {a, b}}. {a} is open in (X, p ) but is not closed since {a, b} \ {a} = {b} / P The excluded point topology on X is E = {, {b}, {a, b}}. {b} is open in (X, E ) but is not closed since {a, b} \ {b} = {a} / E . Remark 2.41. Notice that in a Sierpinski space the only sets that are both open and closed are and X . As we will see, this is a characteristic of connected spaces. A Sierpinski space is thus the smallest nontrivial connected space. Remark 2.42. In contrast, every subset in a discrete topological space (X, ) is both open and closed since the complement of any subset is again a subset of X . As we will see, a discrete topological space is completely disconnected.

3
3.1

Lecture 03
Set Theoretic Preliminaries

Theorem 3.1. Let A B and let C be a set. Then (1) A C B C . (2) A C B C . Proof. (1) Let x A C . Then x A or x C . Suppose x A. Then since A B , x B . Thus x B C . Now suppose x C . Then x B C . In either case, if A B , then A C B C . (2) Let x A C . Then x A and x C . Since x A and A B , x B . Thus x B C . If A B , then A C B C . Theorem 3.2. (Distributive Laws) Let A, B , and C be sets. Then (1) A (B C ) = (A B ) (A C ) (2) A (B C ) = (A B ) (A C ) Proof. (1) Let x A (B C ). Then x A or x B C . Suppose, x A. Then x A B . It is also true that x A C . Thus, x (A B ) (A C ). Now suppose x B C . Then x B and x C . Since x B , x A B . Similarly, since x C , x A C . Thus, x (A B ) (A C ). Hence, in either case, A (B C ) (A B ) (A C ). Conversely, let x (A B ) (A C ). Then x A B and x A C . Since x A B , x A or x B and since x A C , x A or x C . Suppose x A. Then x A (B C ). Now suppose x / A. Then x B and x C . That is, x B C . Thus x A (B C ). Hence, (A B ) (A C ) A (B C ). (2) Let x A (B C ). Then x A and x B C . Since x B C , x B or x C . Suppose x B . Then since x A, x A B . Thus, x (A B ) (A C ). Now suppose x C . Then since x A, x A C . Thus, x (A B ) (A C ). 12

Hence, A (B C ) (A B ) (A C ). Conversely, let x (A B ) (A C ). Then x A B or x A C . Suppose x A B , then x A and x B . Since x B , x B C . Thus, x A (B C ). Now suppose x A C . Then x A and x C . Since x C , x B C . Thus, x A (B C ). Hence, (A B ) (A C ) A (B C ). Theorem 3.3. Let A B X . Then (1) B \ A = B (X \ A). (2) X \ B X \ A. Proof. (1) Let x B \ A. Then x B and x / A. As x / A, x X \ A. So x B (X \ A). Hence, B \ A B (X \ A). Conversely, let x B (X \ A). Then x B and x X \ A. As x X \ A, x / A. So x B \ A. Hence, B (X \ A) B \ A. B \ A = B (X \ A). (2) Let x X \ B . Then x / B . Since A B , x / A. Thus, x X \ A. Hence, X \ B X \ A.

3.2

Interior and Closure

Denition 3.4. Let (X, ) be a topological space and let A X . The interior of A is the set Int(A) = {Ui : Ui A and Ui }. That is, Int(A) is the union of all open sets contained in A. Remark 3.5. Since Int(A) is an arbitrary union of open sets, it is open. It is in fact the largest (i.e. most inclusive) open set contained in A. So if U is an open set such that U A, then U Int(A) A. Denition 3.6. Let (X, ) be a topological space and let A X . The closure of A is the set Cl(A) = {Ci : A Ci and X \ Ci }. That is, Cl(A) is the intersection of all closed sets containing A. Remark 3.7. Since Cl(A) is an arbitrary intersection of closed sets, it is closed. It is in fact the smallest (i.e. most exclusive) closed set containing A. So if C is a closed set such that A C , then A Cl(A) C . Remark 3.8. The closure of A is sometimes denoted A. Remark 3.9. Let (X, ) be a topological space. For any subset A X , Int(A) A Cl(A). Theorem 3.10. Let (X, ) be a topological space and let A X . Then (1) A is open i A = Int(A) (2) A is closed i A = Cl(A).

13

Proof. (1) Suppose A is open. Then A is an open set contained in itself. So A Int(A). Since Int(A) A, we have Int(A) A Int(A). Hence, A = Int(A). Conversely, suppose A = Int(A). Then since the interior of a set is always open, A is open. (2) Suppose A is closed. Then A is a closed set containing itself. So Cl(A) A. Since A Cl(A), we have Cl(A) A Cl(A). Hence, A = Cl(A). Conversely, suppose A = Cl(A). Then since the closure of a set is always closed, A is closed. Denition 3.11. (Neighborhood of a Point) Let (X, ) be a topological space and let x X . Any subset A X with x Int(A) is called a neighborhood of x. If A is open, then A is called an open neighborhood of the point x X .

3.3

Exterior and Boundary

Denition 3.12. Let (X, ) be a topological space and let A X . The exterior of A is the set Ext(A) = X \ Cl(A). Remark 3.13. Notice Ext(A) is open as it is the complement of a closed set. Further, since Ext(A) = X \ Cl(A), we have X \ (X \ Cl(A)) = X \ Ext(A) Cl(A) = X \ Ext(A) Denition 3.14. Let (X, ) be a topological space and let A X . The boundary of A is the set A = X \ [Int(A) Ext(A)]. Remark 3.15. Notice Int(A)Ext(A) is open as it is a union of open sets. So A = X \[Int(A)Ext(A)] is closed as it is the complement of an open set. Remark 3.16. Let (X, ) be a topological space and let A X . We are accustomed to considering the subset A as a collection of points in the set X . We can now begin to consider the subset A as a geometric object in the topological space X that has a boundary, an interior, and an exterior. For simplicity, lets suppose that the space resembles, at least in the near vicinity of the subset A, a 2-dimensional plane that is very similar to the real plane. Suppose that the boundary of the subset A is the closed curve shown below.

14

Remark 3.17. (The Subset A as a Geometric Object) The interior of A is everything inside the boundary (not including the boundary itself). The exterior of A is everything outside the boundary (not including the boundary itself). The boundary can indeed be described as what remains when both the interior and exterior of A are removed. The closure of A can be described as what remains when the exterior is removed: it is the boundary and everything inside the boundary. Denition 3.18. Let X be a set. An algebra over X is a collection F of subsets of X closed under (1) Arbitrary Unions (2) Arbitrary Intersections (3) Complements of all subsets in F Remark 3.19. The collection F = {, X } is the smallest collection of subsets that constitutes an anlgebra over X . This is the trivial algebra over X . Remark 3.20. In contrast, a topology over X is a collection of subsets of X closed under (1) Arbitrary Unions (2) Finite Intersections (3) Complements of closed sets (i.e. double complements of subsets in ). Remark 3.21. As we already know, = {, X } is the smallest collection of subsets of X that constitutes a topology over X . It is called the trivial topology. Remark 3.22. Notice that not every topology is an algebra but every algebra is a topology. However, as a topology every subset A F is both open and closed. This generally undesireable condition results in a disconnected space. As we will see, the less restrictive condition of requiring that a topology be closed under nite intersections 15

and double complements gives an orientation to the space. In this way topology is much more like geometry than algebra. Indeed, it gives us a dierent way of thinking about the geometry of the space X . Remark 3.23. Let (X, ) be a topological space and let A X . Recall that A = X \ (Int(A) Ext(A)). A is a closed set since its complement is a union of open sets and arbitrary unions of open sets are open. Int(A) is the largest subset of A that is open (i.e. in the topology). Ext(A) is the subset of X \ A that is open. Cl(A) is the smallest superset of A that is closed. Hence, its complement is open. Theorem 3.24. Let (X, ) be a topological space and let A X . Then A = Cl(A) \ Int(A). Proof. A = X \ [Int(A) Ext(A)] = (X \ Int(A)) (X \ Ext(A)) = (X \ Int(A)) Cl(A) Since Int(A) Cl(A) X , Cl(A) (X \ Int(A)) = Cl(A) \ Int(A). Hence, A = Cl(A) \ Int(A). Remark 3.25. Theorem 3.24 not only gives us another way to compute the boundary of a subset A but also gives us another interpretation of the boundary: the boundary of a set A is that portion of the closure of A that completely encloses the subset of A that is open.

3.4

Examples

Example 3.26. Let I = {, R}. Consider (a, b) R where (a, b) = {x R : a < x < b}. (a, b) / . R \ (a, b) = (, a] [b, ] / . So (a, b) is neither open nor closed in (R, I ). The only open set contained in (a, b) is . So Int((a, b)) = . The only closed set that contains (a, b) is R. So Cl((a, b)) = R. Hence, (a, b) = Cl((a, b)) \ Int((a, b)) = R \ = R Ext((a, b)) = R \ Cl((a, b)) = R \ R = R \ (a, b) = R \ R =

16

Example 3.27. Let be the indiscrete topology on X and let A be a proper nonempty subset of X . Subset Interior Closure Boundary Exterior X A X X X X X Example 3.28. Let E = {R} {U R : 0 / U }. Consider (a, b) R. There are two cases. Case 1: 0 (a, b) (a, b) is not open since 0 (a, b) and (a, b) R. (a, b) is closed since 0 / R \ (a, b) R. So (a, b), where a < 0 < b, is closed but not open in (R, E ). Cl((a, b)) = (a, b) Int((a, b)) = (a, b) \ {0} = (a, 0) (0, b) Hence, (a, b) = (a, b) \ [(a, 0) (0, b)] = {0} Ext((a, b)) = R \ (a, b) = (, a] [b, ) R \ (a, b) = R \ {0} = (, 0) (0, ) Case 2: 0 / (a, b). We will consider the subcase a > 0. (a, b) is open since 0 / (a, b) and (a, b) R. (a, b) is not closed since 0 R \ (a, b) and R \ (a, b) R. So (a, b), where a > 0, is open but not closed in (R, E ). Int((a, b)) = (a, b) Cl((a, b)) = (a, b) {0} Hence, (a, b) = [(a, b) {0}] \ (a, b) = {0} Ext((a, b)) = R \ [(a, b) {0}] = (, 0) (0, a] [b, ] R \ (a, b) = R \ {0} = (, 0) (0, )

17

Example 3.29. Let be the excluded point topology on X with p X the specied point. Let U be open but not closed. Let C be closed but not open. There are no subsets that are neither open nor closed. The only subsets that are both open and closed are and X . Subset U C X Interior U C \ { p} X Closure U {p} C X Boundary {p} {p } Exterior X X \ [U { p }] X \C

4
4.1

Lecture 04
Interior, Closure, Exterior, and Boundary Points

Lemma 4.1. Let (X, ) be a topological space and let A X . Then (1) Cl(A) = A Int(A) (2) A Cl(A) (3) Int(A) A = (4) Int(A) = Cl(A) \ A Proof. (1) A = Cl(A) \ Int(A) So (A) Int(A) = [Cl(A) \ Int(A)] Int(A) = Cl(A). (2) A A Int(A) = Cl(A) (3) Int(A) A = Int(A) Cl(A) \ Int(A) = (4) A = Cl(A) \ Int(A) Cl(A) So Cl(A) \ A = Cl(A) \ [Cl(A) \ Int(A)] = Int(A).

18

Theorem 4.2. Let (X, ) be a topological space and let A X . Then (1) x Int(A) i an open neighborhood U of x such that U A. (2) x Cl(A) i open neighborhood U of x, U A = . Proof. (1) Let x Int(A). Then Int(A) is itself an open neighborhood of x where Int(A) A. Conversely, suppose an open neighborhood U of x such that U A. Then U is an open set contained in A. So U Int(A). As x U and U Int(A), x Int(A). (2) We will prove the contrapositive: x / Cl(A) i an open neighborhood U of x such that U A = . Suppose x / Cl(A). Then x X \ Cl(A). Notice that X \ Cl(A) is an open neighborhood of x. Now since A Cl(A), X \ Cl(A) X \ A. So [X \ Cl(A)] A (X \ A) A = . Since [X \ Cl(A)] A, we have [X \ Cl(A)] A = . Conversely, suppose an open neighborhood U of x such that U A = . Then X \ (U A) = X \ . Equivalently, X \ U X \ A = X . So A [(X \ U ) (X \ A)] = X A (A (X \ U )) = A A (X \ U ) = A Hence, A (X \ U ). Now since U is open, X \ U is closed. That is, X \ U is a closed set containing A. So Cl(A) X \ U . Equivalently, U X \ Cl(A). As x U and U X \ Cl(A), x X \ Cl(A). Thus x / Cl(A). Remark 4.3. Theorem 4.2 states the following: (1) A point is in Int(A) i it has an open neighborhood contained in A. (2) A point is in Cl(A) i every open neighborhood of it intersects A. That is, every open neighborhood of it contains a point of A. Remark 4.4. (Interior Points) x Int(A) an open neighborhood U of x such that U A x / Int(A) open nighborhood U of x, U A Remark 4.5. (Closure Points) x Cl(A) open neighborhood U of x, U A = x / Cl(A) an open neighborhood U of x such that U A = Remark 4.6. A closure point is more commonly called an adherent point. Theorem 4.7. Let (X, ) be a topological space and let A X . Then (1) Int(X \ A) = X \ Cl(A). (2) Cl(X \ A) = X \ Int(A). Proof. (1) Since A Cl(A), we have X \ Cl(A) X \ A. As Cl(A) is closed, X \ Cl(A) is open. So X \ Cl(A) is an open set contained in X \ A. Hence, X \ Cl(A) Int(X \ A). 19

Conversely, let x Int(X \ A). Notice that Int(X \ A) is open neighborhood of x. Since Int(X \ A) X \ A, Int(X \ A) A (X \ A) A = . Since Int(X \ A) A, Int(X \ A) A = . So x / Cl(A). Thus x X \ Cl(A). Hence, Int(X \ A) X \ Cl(A). Int(X \ A) = X \ Cl(A). (2) Since Int(A) A, we have X \ A X \ Int(A). As Int(A) is open, X \ Int(A) is closed. So X \ Int(A) is a closed set containing X \ A. Hence, Cl(X \ A) X \ Int(A). Conversely, let x X \ Int(A). We need to show x Cl(X \ A). Suppose to the contrary that x / Cl(X \ A). Then x X \ Cl(X \ A). By part (1), X \ Cl(X \ A) = Int(X \ (X \ A)) = Int(A). That is x Int(A). But this contradicts the fact that x X \ Int(A). So, by contradiction, x Cl(X \ A). Hence, X \ Int(A) Cl(X \ A). Cl(X \ A) = X \ Int(A). Corollary 4.8. Let (X, ) be a topological space and let A X . Then (1)Ext(A) = Int(X \ A). (2) A = Cl(A) Cl(X \ A). Proof. (1) Ext(A) = X \ Cl(A) = Int(X \ A) (2) A = Cl(A) \ Int(A) = Cl(A) X \ Int(A) = Cl(A) Cl(X \ A)

Remark 4.9. Notice that by part (1) of corollary 4.8, Ext(A) is the union of all open sets contained in X \ A and is therefore the largest (i.e. most inclusive) open set contained in X \ A. So if U is an open set such that U X \ A, then U Ext(A). Theorem 4.10. Let (X, ) be a topological space, let A X , and let x X . Then (1) x Ext(A) i an open neighborhood U of x such that U X \ A. (2) x A i open neighborhood U of x, U A = and U (X \ A) = . Proof. (1) Let x Ext(A). Then Ext(A) is an open neighborhood of x where Ext(A) X \ (A). Conversely, suppose an open neighborhood U of x such that U X \ A. Then as U is an open set contained in X \ A, U Ext(A). Since x U and U Ext(A), x Ext(A). (2) Let x A. Then x Cl(A) Cl(X \ A). That is, x Cl(A) and x Cl(X \ A). As x Cl(A), open neighborhood U of x, U A = . 20

As x Cl(X \ A), open neighborhood U of x, U (X \ A) = . Conversely, suppose open neighborhood U of x, U A = and U (X \ A) = . Then as it is true that open neighborhood U of x, U A = , x Cl(A). Similarly, since open neighborhood U of x, U (X \ A) = , x CL(X \ A). That is, x Cl(A) Cl(X \ A) = A. Remark 4.11. Theorem 4.10 states the following: (1) A point is in Ext(A) i it has an open neighborhood contained in X \ A. (2) A point is on A i every open neighborhood ot it intersects A and X \ A. That is, every open neighborhood of it contains a point of A and X \ A. Remark 4.12. (Exterior Points) x Ext(A) an open neighborhood U of x such that U X \ A x / Ext(A) open neighborhood U of x, U X \ A Remark 4.13. (Boundary Points) x A open neighborhood U of x, U A = and U (X \ A) = x / A an open neighborhood U of x such that U A = or U (X \ A) = Remark 4.14. Let (X, ) be a topological space and let A X . Then X = A (X \ A) = A (X \ [X \ (Int(A) Ext(A)]) = Int(A) A Ext(A) Exercise 4.15. We have already demonstrated that Int(A) A = . Show Int(A) Ext(A) = and A Ext(A) = .

4.2

Cluster Points and Isolated Points

Denition 4.16. Let (X, ) be a topological space and let A X . Then a point x X is a cluster point of A i every open neighborhood U of x intersects A in at least one point other than x itself. That is, x is a cluster point of A i open neighborhood U of x, U (A \ {x}) = . Remark 4.17. In light of Theorem 4.2, we have x is a cluster point of A i x Cl(A \ {x}). Remark 4.18. A cluster point is also called a limit point or accumulation point. Denition 4.19. Let (X, ) be a topological space, let A X , and let U be an open neighborhood of the point x X . If U is not a singleton, then U \ {x} is called a deleted neighborhood of the point x. Exercise 4.20. Prove that a point x is a cluster point of a set A i every deleted neighborhood of x intersects A. Denition 4.21. The set of all cluster points of a set A is called the derived set of A and is denoted by A . Example 4.22. Let X be a countable set and let D = P (X ). Then A X , A = .

21

Proof. In the discrete topology on X , every singleton set is open. So {x} is an open neighborhood of x where {x} (A \ {x}) = for any subset A X . Hence, x is not a cluster point of the set A. Since this is true x X , A has no cluster points. That is, A = . Theorem 4.23. Let (X, ) be a topological space and let A X . Then Cl(A) = A A . Proof. Let x Cl(A). Since A Cl(A), we must consider two cases: x A or x / A. Suppose x A. Then x A A . Now suppose x / A. Then since x Cl(A), open neighborhood U of x, U A = . Further, as x / A, U intersects A in a point dierent from x. That is, U (A \ {x}) = . So x is a cluster point of A. That is, x A . Thus, x A A . Hence, in either case, Cl(A) A A . Conversely, let x A A . Then x A or x A . Suppose x A. Then as A Cl(A), x Cl(A). Now suppose x A . Then x is a cluster point of A and so open neighborhood U of x, U (A \{x}) = . That is, y X such that y U (A \ {x}). Since A \ {x} A, we have U (A \ {x}) U A. As y U (A \ {x}) U A, we have U A = . So x Cl(A). Hence, in either case, A A Cl(A). Cl(A) = A A . Corollary 4.24. Let (X, ) be a topological space. A subset A X is closed i A A. Proof. Suppose A is closed. Then A = Cl(A) = A A . So A A. Conversely, suppose A A. Then A A A A = A. That is, Cl(A) A. Since Cl(A) = A A , we have A = Cl(A). Hence A is closed. Example 4.25. Let (X, ) be a topological space and let A X . Then Cl(A) = A A = A A. However, in general, A = A. We will give a counterexample in a Sierpinski space: Let X = {a, b} with = {, {a}, X }. Consider A = {b}. A is closed since X \ A = {a, b} \ {b} = {a} . So Cl(A) = A, Int(A) = , and A = A \ = A. Notice that the only open neighborhood of b is X . Since A X = {b}, an open neighborhood of b that intersects A in a point dierent from b. So b is not a cluster point of A and thus b / A. Now {a} is an open neighborhood of a such that {a} A = . So a / Cl(A) = A A and thus a / A. Hence, A = . Clearly, A = A. Denition 4.26. Let (X, ) be a topological space and let A X . A point x A is an isolated point of A i an open neighborhood U of x such that U A = {x}. Example 4.27. Let X be a countable set with D = P (X ) and let A X . Then every point in A is an isolated point of A.

22

Proof. In the discrete topology on X , every singleton set is open. So x A, {x} is an open neighborhood of x such that {x} A = {x}. That is, every point in A is an isolated point of A. Theorem 4.28. Let (X, ) be a topological space and let A X . A point x A is either a cluster point of A or an isolated point of A but not both. Proof. Let x A and suppose x is a cluster point of A. Then open neighborhood U of x, U (A \ {x}) = . So y X such that y U (A \ {x}). Since A \ {x} A, U (A \ {x}) U A. So y U A open neighborhood U of x. Thus an open neighborhood U of x such that U A = {x}. Hence x is not an isolated point of A. Now suppose x is not a cluster point of A. Then an open neighborhood U of x such that U (A \ {x}) = . So [U (A \ {x}] {x} = {x} U A = {x} Hence x is an isolated point of A. Example 4.29. Let I = {, R}. Consider [a, b] R. [a, b] is neither open nor closed in this space. Notice r R, the only open neighborhood of r is R. Since r R, R ([a, b] \ {r}) = , every point r R is a cluster point of [a, b]. That is, [a, b] = R. So [a, b] has no isolated points. Cl([a, b]) = R and [a, b] = R. Notice that in this case [a, b] = [a, b] . Exercise 4.30. Let E = {R} {U R : 0 / U }. Consider [a, b] where a > 0. Determine the cluster points and isolated points of [a, b].

5
5.1

Lecture 05
The Isolated Points of a Space

Denition 5.1. Let A and B be sets such that A B . The relative complement of A in B is the set B \ A = {x B : x / A}. That is, B \ A = B \ (B A). Remark 5.2. If A B , the relative complement of A in B is the absolute complement of A B in B . Theorem 5.3. Let (X, ) be a topological space and let A X . A point x X is a cluster point of A i deleted neighborhood U \ {x} of x, (U \ {x}) A = . Proof. Suppose x is a cluster point of A. Then open neighborhood U of x, U (A \ {x}) = . So y = x X such that y U (A \ {x}). 23

That is, y U and y A \ {x}. As A \ {x} A, y A. As y = x, y U \ {x}. So y (U \ {x}) A. Hence, (U \ {x}) A = . Conversely, suppose deleted neighborhood U \ {x} of x, (U \ {x}) A = . Then y = x X such that y (U \ {x}) A. That is, y U \ {x} and y A. As U \ {x} U , y U . As y = x, y A \ {x}. So y U (A \ {x}). That is, U (A \ {x}) = . Hence, x is a cluster point of A. Example 5.4. Let E = {R} {U R : 0 / U }. Consider [a, b] where a > 0. Notice that any subset that contains 0 is closed. So the only open neighborhood of 0 is R. Clearly, R ([a, b] \ {0}) = or, equivalently, (R \ {0}) [a, b] = . Hence 0 is a cluster point of [a, b]. Notice r [a, b], {r} is an open neighborhood of r such that {r} [a, b] = {r}. That is, every point in [a, b] is an isolated point of [a, b]. Cl([a, b]) = [a, b] {0}. Notice that in this case [a, b] = [a, b] = {0}. Example 5.5. Let |X | 2, let I = {, X }, and let x X . The only open neighborhood of the point x is the entire space X . As |X | 2, X (X \ {x}) = since X must contain at least one point distinct from x. So x is a cluster point of X . As the point x was chosen abitrarily, every point x X is a cluster point of X . Hence there are no isolated points of X . Example 5.6. Let |X | 2 and let P = {} {U X : p U }. Notice {p} is an open neighborhood of p such that {p} X = {p}. Hence p is an isolated point of X . Now let x = p X . Then every open neighborhood U of x contains the point p. That is, p U (X \ {x}). Thus U (X \ {x}) = . Hence every point x = p X is a cluster point of X . Example 5.7. Let |X | 2 and let E = {X } {U X : p / U }. Notice that the only open neighborhood of p is the entire space X . As |X | 2, X \ {p} = . So X (X \ {p}) = . Hence, p is a cluster point of X . Now let x = p X . Then {x} is an open neighborhood of x such that {x} X = {x}. Hence every point x = p X is an isolated point in X . 24

Example 5.8. Let X be a countable set and let D = P (X ). Then x X , {x} is an open neighborhood of the point x such that {x} X = {x}. Hence every point x X is an isolated point of X . Remark 5.9. As we will see, the number of isolated points in a given space is a topological invariant. That is, any space that is topologically equivalent (i.e. homeomorphic) to a given space must have the same number of isolated points.

5.2

Introduction to Connectedness

Denition 5.10. Let X be any set. A partition of the set X is any collection {Xi : i I } of nonempty subsets of X such that (1) X = iI Xi and (2) Xi Xj = i = j . Lemma 5.11. Let A, B, C and D be sets such that A B and C D. Then A C B D. Proof. Let x A C . Then x A and x C . As x A and A B , x B . As x C and C D, x D. That is, x B D. Hence A C B D. Example 5.12. Let (X, ) be a topological space and let A X . We have already demonstrated that A Int(A) = . Now Ext(A) A = Ext(A) [X \ (Int(A) Ext(A))] = and Ext(A) Int(A) = Int(X \ A) Int(A). As Int(X \ A) X \ A and Int(A) A, Int(X \ A) Int(A) (X \ A) A = . That is, Ext(A) Int(A) . Since Ext(A) Int(A), Ext(A) Int(A) = . Remark 5.13. Let (X, ) be a topological space and let A X . Then X = A Int(A) Ext(A). So if A = , Int(A) = , and Ext(A) = , then {A, Int(A), Ext(A)} is a partition of X . Denition 5.14. Let (X, ) be a topological space. A separation of X is a pair U, V of nonempty proper open subsets of X that form a partition of X . Denition 5.15. A topological space X is connected i a separation of X . Otherwise, the space is disconnected. Theorem 5.16. A topological space X is connected i the only subsets of X that are both open and closed are and X . Proof. We will prove the contrapositive: A topological space X is disconnected i a nonempty proper subset of X that is both open and closed. Suppose X is disconnected. Then nonempty, proper, open subsets U and V that partition X . That is, X = U V and U V = . 25

Notice X \ U = V is open. So U is both open and closed. Hence, a nonempty proper subset U X that is both open and closed. Conversely, suppose U is a nonempty proper subset of X that is both open and closed. Then as U is closed, X \ U is open. Further, X \ U = . Since X = U X \ U , U and X \ U is a separation of X . Hence X is disconnected. Remark 5.17. As we will see, connectedness is also a topological invariant of a given topological space. Example 5.18. We have already demonstrated that there are no nonempty proper subsets that are both open and closed in either the particualr point topology or the excluded point topology. So spaces with these topologies are connected spaces. Remark 5.19. Notice that a Sierpinski space is indeed the smallest nontrivial connected space as it is a two point set with either the particular point topology or the excluded point topology. In either topology, the space has one cluster point and one isolated point. Example 5.20. Let X be a countable set and let D = P (X ). Then every nonempty proper subset U X is both open and closed. Hence a discrete topological space is completely disconnected.

5.3

Introduction to Separation Conditions

Denition 5.21. (T0 Space) A topological space X for which an open neigborhood for at leat one point in each pair x1 , x2 of distinct points not containing the other is called a T0 , or Kolmogorov, space. That is, x1 = x2 X , open neighborhoods U1 of x1 and U2 of x2 such that either x2 / U1 or x1 / U2 . Denition 5.22. (T1 Space) A topological space X for which open neigborhoods for both points in each pair x1 , x2 of distinct points not containing the other is called a T1 , or Fr echet, space. That is, x1 = x2 X , open neighborhoods U1 of x1 and U2 of x2 such that x2 / U1 and x1 / U2 . Remark 5.23. Notice that not every T0 space is T1 . However, every T1 space is T0 . The contrapositive is also true. That is, if a space is not T0 , then it is not T1 . Example 5.24. Let P = {} {U X : p U }. Notice that every open set contains the point p. So an open neighborhood U of x = p X such that p / U. Hence the space X is not T1 . However, pair of distinct points x1 = x2 = p X , {x1 , p} is an open neighborhood of x1 not containing x2 and {x2 , p} is an open neighborhood of x2 not containing x1 . Further, x = p X , {p} is an open neighborhood of p not containing x. Hence the space is T0 . Theorem 5.25. A topological space X is T1 i every singleton set is closed. Proof. Let X be a T1 space. Then x = y X an open neigborhood Uy of y such that x / Uy . Let {Ui : i I } be the collection of all open neighborhoods of all points y = x (i.e. all points y X \{x}) such that x / Ui . 26

Then iI Ui is open as it is a union of open sets. Further, iI Ui = X \ {x}. That is, {x} is closed. This is true x X . Hence, every singleton set in a T1 space is closed. Conversely, suppose that all singleton sets are closed. Let x, y X such that x = y . Then X \ {x} is an open neighborhood of y not containing x. Similarly, X \ {y } is an open neighborhood of x not containing y . This is true x = y X . Hence the space X is T1 . Example 5.26. Let E = {X } {U X : p / U }. Then x = p X , {x} is open but not closed. Hence the space X is not T1 . Exercise 5.27. Show that a topological space X with the excluded point topology is T0 . Example 5.28. Let I = {, X }. Then x X , the only open neighborhood of x is the entire space X . So an open neigborhood that does not contain any given point. Hence the space X is not T0 and therefore is also not T1 . Corollary 5.29. Every nite point set in a T1 space is closed. Proof. Let X be a T1 space and let F = {x1 , . . . , xn } be a nite point set in the space X . Then F = n i=1 {xi } where {xi } is closed i {1, . . . , n}. Hence F is closed. Denition 5.30. (T2 Space) A topological space X for which disjoint open neigborhoods for both points in each pair x1 , x2 of distinct points is called a T2 , or Hausdor, space. That is, x1 = x2 X , open neighborhoods U1 of x1 and U2 of x2 such that U1 U2 = . Remark 5.31. Some denitions of a Hausdor space allow for any neighborhoods of the points. That is, the neighborhoods need not necessarily be open. Remark 5.32. A T1 space is not necessarily Hausdor since it is not required that the neighborhoods be disjoint. However, every Hausdor space is T1 . The contrapositive is also true. That is, if a space is not T1 , then it is not Hausdor. Example 5.33. Let X be a countable space and let D = P (X ). Then every singleton set is both open and closed. So x = y X , {x} is an open neighborhood of x and {y } is an open neighborhood of y such that {x} {y } = . Hence the space X is T2 (Hausdor). Remark 5.34. In a Hausdor space, every distinct pair of points can be separated by disjoint neighborhoods (not necessarily open). Today, we see this as a separation condition that may or may not be satised by a given space. One of the earliest denitions of a topological space included what we now consider the Hausdor condition. This denition was stated by none other than Felix Hausdor. 27

Remark 5.35. In a T1 or Fr echet space, every distinct pair of points can be separated by open neighborhoods that are not necessarily disjoint. Remark 5.36. In a T0 or Kolmogorov space, at least one of the points in any distinct pair of points can be separated from the other by an open neighborhood. Remark 5.37. In the early days of topology, these separation conditions were taken as axioms. The T is for the German word Trennungsaxiom, which means separation axiom. The trend in modern mathematical literature is to move away from the use of the T terms in favor of more descriptive terms that do not refer to an axiom since we no longer take these separation conditions as axioms. This is especially true for the Hausdor condition forward. Remark 5.38. As we will see, the separation condition statised by a given topological space is also a topological invariant. Remark 5.39. The topological invariants which we have explored in this lecture are summarized below: Indiscrete Topology 0 Connected None PPT 1 Connected T0 EPT |X | 1 Connected T0 Discrete Topology |X | Disconnected T2

# of Isolated Points Connectedness Separation Condition Satised

Exercise 5.40. Determine the subsets A X for which Cl(A) = X in the (1) indiscrete topology (2) particular point topology (PPT) (3) excluded point topology (EPT) (4) discrete topology For each topology assume |X | 2. For the discrete topology assume |X | 0 .

6
6.1

Lecture 06
Dense Subsets

Example 6.1. Let E = {X } {U X : p / U }. Then the space X is T0 . Proof. Notice that the only open neighborhood of the excluded point p is the entire space X . So an open neighborhood U of p not containing a point x distinct from p. Hence the space X is not T1 . However, x = p X , {x} is an open neighborhood of x such that p / {x}. Further, x = y = p X , {x} is an open neighborhood of x and {y } is an open neighborhood of y such that x / {y } and y / {x}. Hence the space X is T0 , or Kolmogorov. Denition 6.2. Let (X, ) be a topological space. A proper subset D everywhere dense, in X i Cl(D) = X . Example 6.3. Let I = {, X } and let A be a nonempty proper subset A Subset A X Interior X 28 Closure X X X is said to be dense, or X.

So every nonempty proper subset A

X is dense in X . X.

Example 6.4. Let |X | 0 , let D = P (X ), and let A be a nonempty proper subset A Subset A X So a proper subset U topological space. Interior A X Closure A X

X such that Cl(U ) = X . Hence there are no dense subsets in a discrete

Example 6.5. Let P = {} {U X : p U } and let U and C be nonempty proper subsets of X such that U, X \ C P . Subset U C X So every nonempty proper open subset U Interior U X Closure X C X

X is dense in X .

Example 6.6. Let E = {X } {U X : p / U } and let U and C be nonempty proper subsets of X such that U, X \ C E . Subset U C X Interior U C \ {p} X Closure U { p} C X

So the only proper subset that is dense in X is the open set X \ {p}. Theorem 6.7. Let (X, ) be a topological space. A proper subset D nonempty open subset U X intersects D. Proof. Suppose D is dense in X . Then Cl(D) = X . So x X , x Cl(D). Hence open neighborhood U of x, U D = . That is, every nonempty open seubset U X intersects D. Conversely, suppose that every nonempty open subset U X intersects D. Then x X , open neighborhood U of x, U D = . That is, every point x X is in Cl(D). Hence Cl(D) = X . Thus D is dense in X . Remark 6.8. Notice that theorem 6.7 states that if D open subset U X contains at least one point x D. X is dense in the space X , then every nonempty X is dense in X i every

29

6.2

Nowhere Dense Subsets

Denition 6.9. (Nowhere Dense Set) Let (X, ) be a topological space. A nonempty subset A X is nowhere dense in X i Int(Cl(A)) = . That is, a subset A X is nowhere dense in X if there are no nonempty open sets contained in Cl(A). Exercise 6.10. Determine the validity of the following conjecture: if A X is nowhere dense in X , then nonempty proper open subset U X , U A = . Remark 6.11. The complement of a dense subset is not always a nowhere dense subset. For example, a nonempty proper subset A and X \ A are both dense in X with the indiscrete topology. However, the complement of a nowhere dense subset is a dense subset. Theorem 6.12. Let (X, ) be a topological space. If A X is nowhere dense in X , then X \ A is dense in X . The converse is true if X \ A is open. Proof. Suppose A is nowhere dense in X . Then Int(Cl(A)) = . So a nonempty open set contained in Cl(A). As Int(A) A Cl(A), Int(A) = . Since Int(A) = X \ Cl(X \ A) = , Cl(X \ A) = X . Hence, X \ A is dense in X . Now suppose that X \ A is both open and dense in X . Then since X \ A is open, A is closed. Since X \ A is dense in X , Cl(X \ A) = X . So A = Cl(A) Cl(X \ A) = A X = A. Now since A = Cl(A) \ Int(A) = A \ Int(A) = A, Int(A) = . So Int(Cl(A)) = Int(A) = . Hence, A is nowhere dense in X . Theorem 6.13. Subsets of nowhere dense sets are nowhere dense. Proof. Let A be nowhere dense and let B A. As B A Cl(A), Cl(B ) Cl(A). So Int(Cl(B )) Cl(B ) Cl(A). Hence, Int(Cl(B )) Int(Cl(A)) = as A is nowhere dense. Thus Int(Cl(B )) = . That is, B is nowhere dense. Example 6.14. Let I = {, X } and let A Subset A X So a nonempty proper subset A subsets in an indiscrete space. X such that A = . Closure X X Int(Closure) X X

X such that Int(Cl(A)) = . Hence there are no nowhere dense X such that A = . Int(Closure) A X

Example 6.15. Let |X | 0 , let D = P (X ), and let A Subset A X So a nonempty proper subset A subsets in a discrete space. Closure A X

X such that Int(Cl(A)) = . Hence there are no nowhere dense

Example 6.16. Let P = {} {U X : p U } and let U and C be nonempty proper subsets of X such that U, X \ C P . 30

Subset U C X So every nonempty proper closed subset C

Closure X C X

Int(Closure) X X

X is nowhere dense in X .

Example 6.17. Let E = {X } {U X : p / U } and let U and C be nonempty proper subsets of X such that U, X \ C E . Subset U C X Closure U {p} C X Int(Closure) U C \ { p} X

So the only proper subset that is nowhere dense in X is the closed set {p}.

6.3

Separable Spaces

Denition 6.18. A topological space X is separable i it contains a countable dense subset. Remark 6.19. The term separable is known to have originated from Maurice Fr echet but it is not clear why he chose this term to denote a space with a countable dense subset. In the collected works of Felix Hausdor, Hausdor comments that the term was well established but that he was unsure of the reason for the choice of the term. Remark 6.20. The condition of separability is related to the countability of a space. In fact, separability is considered a countability condition that may or may not be satised by a given space. As we will see, the countability condition satised by a given space is another topological invariant. Example 6.21. In the indiscrete topological space X , every singleton set {x} is dense in X . So a nite dense subset in X . Hence an indiscret space is separable. Example 6.22. In the discrete topological space X , there are no dense subsets. Hence a discrete topological space is not separable. Example 6.23. Let P = {} {U X : p U }. Then the singleton set {p} is dense in X . Hence a topological space with the particular point topology is separable. Example 6.24. Let E = {X } {U X : p / U }. The only dense subset in this space is X \ {p}. So a topological space with the excluded point topology is separable only if the underlying set is countable. Remark 6.25. The topological invariants which we have explored so far are summarized below: Indiscrete Topology 0 Connected None Separable PPT 1 Connected T0 Separable 31 EPT |X | 1 Connected T0 Separable only if |X | 0 Discrete Topology |X | Disconnected T2 Not Separable

# of Isolated Points Connectedness Separation Condition Countability Condition

6.4

Equivalent Statements About Open and Closed Sets

Remark 6.26. (Open Sets) A subset A X is open in a topological space (X, ) i (1) A (2) A = Int(A) = Cl(A) \ A (3) A A = (4) x A an open neighborhood U of x such that U A (5) X \ A is closed Remark 6.27. (Closed Sets) A subset A X is closed in a topological space (X, ) i (1) X \ A (2) A = Cl(A) = A A = A A (3) A A (4) A A (5) x A open neighborhood U of x, U A = (6) X \ A is open Theorem 6.28. Let (X, ) be a topologial space and let A X . Then A = i A is both open and closed. Proof. Suppose A = . Then Cl(A) \ Int(A) = . As Int(A) Cl(A), Int(A) = Cl(A) \ = Cl(A). Since Int(A) A Cl(A) and Int(A) = Cl(A), A = Int(A) and A = Cl(A). That is, A is both open and closed. Conversely, suppose that A is both open and closed. The A = Int(A) and A = Cl(A). So A = Cl(A) \ Int(A) = A \ A = . Remark 6.29. (Properties of A) Let (X, ) be a topological space and let A X . The following are properties of the boundary of A. (1) A = X \ (Int(A) Ext(A)) (2) A = Cl(A) \ Int(A) (3) A = Cl(A) Cl(X \ A) (4) A Cl(A) (5) A Int(A) = (6) A Int(A) = Cl(A) (7) A is closed. (8) A = i A is both open and closed.

6.5

More on Cluster Points and Isolated Points

Theorem 6.30. Let X be a T1 (Fr echet) space and let A X . A point x X is a cluster point of A i every open neighborhood of x contains innitely many points of A. Proof. Let x X be a cluster point of A. We need to show that every open neighborhood of x contains innitely many points of A. 32

Suppose to the contrary that an open neighborhood U of x that contains only a nite number of points of A. Then U A \ {x} = {x1 , . . . , xm } is closed as every nite point set is closed in a T1 space. So X \ {x1 , . . . , xm } is open and U X \ {x1 , . . . , xm } is open as it is a nite intersection of open sets. Further U X \{x1 , . . . , xm } is an open neighborhood of x such that (U X \{x1 , . . . , xm }) A \{x} = . But this contradicts the fact that x is a cluster point of A. Hence, by contradiction, every open neighborhood of x contains innitely many points of A. Conversely, suppose that every open neighborhood U of x contains innitely many points of A. Then U A \ {x} also contains innitely many points of A. That is, open neighborhood U of x, U A \ {x} = . Hence x is a cluster point of A. Theorem 6.31. Let (X, ) be a topological space where |X | 2. Then x X is an isolated point of X i {x} is open. Proof. Suppose x X is an isolated point of X . Then an open neighborhood U of x such that U X = {x}. So U X \ {x} = . As |X | 2, X \ {x} = . Hence U = {x}. Conversely, suppose {x} is open. Then {x} is an open neighborhood of x such that {x} X = {x}. Hence, x is an isolated point of X . Exercise 6.32. Let B = {U X : p U } and let U1 , U2 B . Show that if x U1 and x U2 , then U3 B such that x U3 U1 U2 . Consider two cases: x = p and x = p.

7
7.1

Lecture 07
Basis for a Topology

Remark 7.1. Recall that a proper subset A of a topological space X is dense in X i every nonempty open subset U X contains at least one point x A. Example 7.2. We give a counterexample to the conjecture: If A X is nowhere dense in X , then nonempty proper open subset U X , U A = . Counterexample: Let |X | > 2, let P = {} {U X : p U }, and let x = p X . Then {x} is nowhere dense in X . However, {x, p} is a nonempty proper open subset of X such that {x, p} {x} = {x} = . Remark 7.3. The previous conjecture asks If A is nowhere dense in a space X , is A a set such that every open set distinct from the entire space X has no points in common with it? From the counterexample, we see that the answer is No. Remark 7.4. The contrapositive of theorem 6.7 is also a true statement: A proper subset D not dense in X i at least one open subset U X such that U D = . Let A be nowhere dense in X . Then Int(Cl(A)) = . 33 X is

So Cl(A) = X . Hence A is not dense in X . However, a nowhere dense set is more than a set which is simply not dense in X . It is a set for which the only open set contained in its closure is the empty set. In addition, we know that the complement of a nowhere dense set is always a dense set, while the complement of a dense set is nowhere dense only if it is open. Remark 7.5. Let {Ai : i I } be an indexed family of subsets of X . Recall that the arbitrary union is iI Ai = {x : x Ai for at least one i I }. If the indexing set I = Z+ , we have a countable union + iZ+ Ai = i=1 Ai = {x : x Ai for at least one i Z } If the indexing set I = {1, 2, . . . , n}, we have a nite union n i=1 Ai = A1 A2 An i I Ai = There are two degenerate cases: (1) Suppose I = {1}. Then iI Ai = 1 i=1 Ai = A1 . (2) Suppose I = . Then i Ai = {x : x Ai for at least one i }. Since i , x / Ai for any index i. So i Ai = . The degenerate case (2) is called the empty union. Denition 7.6. (Basis) Let (X, ) be a topological space. A basis for is a subcollection B such that every member of is a (possibly empty) union of some members of B . That is, a basis B is a collection {Bi : i I } of open subsets of X such that open subset U X J I such that U = j J Bj . Remark 7.7. Clearly, can be its own basis. Remark 7.8. A basis is sometimes called a base. The elements of a basis are called basis elements, base elements, or basic open sets. Lemma 7.9. Let (X, ) be a topological space and let A X . (1) If A Int(A), then A is open. (2) If Cl(A) A, then A is closed. Proof. (1) Suppose A X such that A Int(A). Then since Int(A) A, A = Int(A). Hence A is open. (2) Suppose A X such that Cl(A) A. Then since A Cl(A), A = Cl(A). Hence A is closed. Theorem 7.10. Let (X, ) be a topological space and let B = {Bi : i I } be a basis for . Then U X is open i x U Bi B such that x Bi U . Proof. Suppose U X is open. Then U = Case 1: J = . Let x U = j J Bj . Then x Bj for at least one index j J . Further, Bj j J Bj = U j J . So x Bj U .
j J

Bj for some J I . We consider two cases.

34

That is, x U Bi B such that x Bi U . Case 2: J = . Then U = j Bj = and the result follows vacuously. Conversely, suppose U X satisfying the condition that x U Bi B such that x Bi U . Let x U . Then Bi B such that x Bi U . As Bi U X , X \ U X \ Bi where X \ Bi is closed. So Cl(X \ U ) X \ Bi . As x Bi , x / X \ Bi . So x / Cl(X \ U ) = X \ Int(U ). Hence x Int(U ). Thus U Int(U ). That is, U is open. Remark 7.11. In any topology, the entire set X is open. Hence, by Theorem 7.10, x X Bi B , where B is a basis for the topology, such that x Bi X . That is, x X , there is at least one basis element containing x. Corollary 7.12. Let (X, ) be a topological space and let B = {Bi : i I } be a basis for . Then X = iI Bi . Proof. Since x X at least one Bi B such that x Bi X , x iI Bi x X . Hence, X iI Bi . The reverse inclusion is obvious since Bi X i I . Thus X = iI Bi . Denition 7.13. A collection A = {Ai : i I } of subsets of a space X is said to cover X , or to be a covering of X , i X = iI Ai . It is called an open covering of X if its elements are open subsets of X. Remark 7.14. Notice that a basis for a given topological space X is an open covering of X . This is one importnat property of a basis. Corollary 7.15. Let (X, ) be a topological space, let B = {Bi : i I } be a basis for , and let x X . If B1 , B2 B such that x B1 B2 , then B3 B such that x B3 B1 B2 . Proof. Each set Bi B is open. So B1 B2 is open as it is a nite intersection of open sets. Hence, by Theorem 7.10, x B1 B2 at least one basis element, call it B3 , such that x B3 B1 B2 . Example 7.16. Let B = {U X : p U } and let U1 , U2 B . Suppose x U1 and x U2 . Then x U1 U2 . We consider two cases. Case 1: x = p Then {p} B such that p {p} U1 U2 since p U1 and p U2 . Case 2: x = p Then {x, p} B such that x {x, p} U1 U2 since x, p U1 and x, p U2 . So if x U1 U2 , where U1 , U2 B , then there exists a third element in B , call it U3 , such that x U3 U1 U2 .

35

Theorem 7.17. Let (X, ) be a topological space. Suppose that B is a collection of open subsets of X such that open U X and x U , there is an Bx B such that x Bx U . Then B is a basis for . Proof. We need to show that every open subset U X is a (possibly empty) union of some elements of B. If U = , then U is the empty union of elements in B . Now suppose U = . Then x U Bx B such that x Bx U . Since Bx xU Bx x U , x xU Bx . That is, U xU Bx . Now Bx xU Bx , Bx U . So xU Bx U . Hence U = xU Bx . B is a basis for by denition. Remark 7.18. As an alternate proof, notice that
x U

Bx =
xU

Bx

= Int(U ) As U is open, U = Int(U ) =

Bx .

Denition 7.19. Let X be a set and let B = {Bi : i I } be a collection of subsets of X such that (1) x X , x Bi for at least one index i I (2) If x B1 B2 , then B3 B such that x B3 B1 B2 . Then the topology generated by B is dened as follows: A subset U X is open i x U Bi B such that x Bi U . Proof. (1) Notice satises the condition for openness vacuously. Now x X Bi B such that x Bi X . Hence, X . (2) Let {U : J } be an indexed family of sets in and let V = J U . Suppose x V . Then x U for at least one J where U is open. So Bi B such that x Bi U . Hence x V Bi B such that x Bi V . Thus V = J U . (3) Suppose U1 , U2 and let x U1 U2 . Then x U1 and x U2 . As x U1 , B1 B such that x B1 U1 . As x U2 , B2 B such that x B2 U2 . That is, x B1 B2 U1 U2 . So B3 B such that x B3 B1 B2 U1 U2 . Hence U1 U2 . Let U1 , . . . , Un . We will show by induction that n =1 U . The case n = 1 is the statement U1 , which is true. Now suppose the case n 1 is true. 36

That is, suppose U1 Un1 . Then n =1 U = (U1 Un1 ) Un is the intersection of two open sets. Hence n =1 U . Exercise 7.20. Let B = {U X : p U }. (1) Show B satises the two conditions for a topology generated by a basis. (2) Show that the topology generated by B is the particular point topology on X . Exercise 7.21. Let (X, ) be a topological space and let B = {Bi : i I } be a collection of open subsets of X such that (1) X = iI Bi . (2) If x B1 B2 , then B3 B such that x B3 B1 B2 . Show that is ner than the topology generated by B .

7.2

The Digital Line Topology

Denition 7.22. (Digital Line Topology) Let X = Z and n Z dene Bn by Bn = {n} {n 1, n, n + 1} if n is odd if n is even

Then B = {Bn : n Z} generates a topology on Z called the digital line topology. Proof. (1) Clearly, n Z, Bn B containing n. (2) If Bm Bn = for distinct m, n Z, then at least one of the m, n Z must be even. Lets choose m to be even. Then Bm = {m 1, m, m + 1}. Suppose n is odd. Then Bn = {n}. So Bm Bn = implies n = m 1 or n = m + 1. In either case, if n Bm Bn , then Bn B such that n Bn Bm Bn . Now suppose n is also even. Then Bn = {n 1, n, n + 1}. Let k Bm Bn . Then k = n + 1 = m 1 or k = n 1 = m + 1. So Bk = {k }. In either case, if k Bm Bn , then Bk sucht that k Bk Bm Bn . Remark 7.23. Explicitly, the digital line topology is = {U Z : n U Bn B such that n Bn U }. The topological space (Z, ) is called the digital line. Remark 7.24. For the remainder of this lecture, Z is assumed to have the digital line topology. Example 7.25. Let Z \ 2Z = O. That is, O is the set of all odd integers. Then O is dense in Z. Proof. Notice that every Bn B contains at least one odd integer. So every nonempty open subset U Z, which is a union of some Bn B , contains at least one odd integer. That is, nonempty open subset U Z, U O = . Hence, O is dense in Z. Example 7.26. The digital line Z is separable.

37

Proof. We will show that the dense subset O Z is countable. Let : Z+ O be dened by (n) = n n1 if n is odd if n is even

Then m O, where m > 0, n 2Z+ such that m = n 1. Similarly, m O, where m < 0, n Z+ \ 2Z+ such that m = n. So m O n Z+ such that m = (n). Hence is surjective. Now suppose (m) = (n). Then m, n Z+ are either both even or both odd. Suppose m, n Z+ are both even. Then (m) = (n) implies m 1 = n 1. So m = n. Now suppose m, n Z+ are both odd. Then (m) = (n) implies m = n. So m = n. In either case, if (m) = (n), then m = n. That is, : Z+ O is a bijection. Thus O is countably innite. Hence it is countable. As Z contains a countable dense subset, it is separable. Example 7.27. Let 2Z = E. Then E is nowhere dense in Z. Proof. Notice O = i=1 {mi } where mi is an odd integer. That is, O is an arbitrary union of open sets and is therefore open. As O is both dense and open, Z \ O = E is nowhere dense in Z.

8
8.1

Lecture 08
More Properties of a Basis

Example 8.1. Let B = {U X : p U }. Then B is a basis for the particular point topology on X . Proof. We have already demonstrated that if x U1 U2 , where Ui B , then U3 B such that x U3 U1 U2 . Let x X . Then {x, p} B . So x X Ux = {x, p} B such that x Ux X . Hence B satises the two conditions for a topology geberated by a basis. So B is a basis for B = {V X : x V Ui B such that x Ui V }. Let P = {} {U X : p U }. Then nonempty U P , U B B . Further, B vacuously. So P B . Now let V B . Then V = j J Uj where Uj P j J . Hence B P . B = P .

38

Lemma 8.2. Let (X, ) be a topological space and let B = {Bi : i I } be a collection of open subsets of X such that (1) X = iI Bi (2) If x B1 B2 , then B3 B such that x B3 B1 B2 . Then the topology is ner than the topology generated by B . Proof. Notice B and that B satises the two conditions for a topology generated by a basis. So B is a basis for B = {U X : x U Bi B such that x Bi U }. Let U B . Then U = j J Bj where Bj j J . So U . Hence B . Remark 8.3. Notice that in order for the reverse inclusion, B , to be true, U and x U , it must be true that Bi B such that x Bi U . Theorem 8.4. Let and be two topologies on the same underlying set X with bases B and B , respectively. If , then x X and B B such that x B B B such that x B B . Proof. Let x X . Then as B is a basis for , B B such that x B X . Notice B B . That is, B is open in (X, ). As B is a basis for , x B B B such that x B B . Hence x X and B B such that x B B B such that x B B .

8.2

Calculating the Interior and Closure of Unions and Intersections

Lemma 8.5. Let (X, ) be a topological space and let A B X . Then (1) Int(A) Int(B ) (2) Cl(A) Cl(B ) Proof. (1) Since Int(A) A and A B , Int(A) B . That is, Int(A) is an open set contained in B . Since Int(B ) is the largest open set contained in B , Int(A) Int(B ). (2) Since A B and B Cl(B ), A Cl(B ). That is, Cl(B ) is a closed set containing A. Since Cl(A) is the smallest closed set containing A, Cl(A) Cl(B ). Theorem 8.6. Let (X, ) be a topological space and let A X and B X . Then (1) Int(A) Int(B ) Int(A B ) (2) Cl(A B ) Cl(A) Cl(B ) Proof. (1) Since Int(A) A A B , Int(A) is an open set contained in A B . As Int(A B ) is the largest open set contained in A B , we must have Int(A) Int(A B ). Now since Int(B ) B A B , Int(B ) is an open set contained in A B . As Int(A B ) is the largest open set containted in A B , we must have Int(B ) Int(A B ). Int(A) Int(B ) Int(A B ). (2) Since A B A Cl(A) and A B B Cl(B ), both Cl(A) and Cl(B ) are closed sets containing 39

A B . As Cl(A B ) is the smallest closed set containing A B , we must have Cl(A B ) Cl(A) and Cl(A B ) Cl(B ). Thus, Cl(A B ) Cl(A) Cl(B ). Remark 8.7. The inclusions can be strict. That is, its possible for Int(A) Int(B ) Cl(A B ) Cl(A) Cl(B ) . Int(A B ) and

Remark 8.8. Recall from the previous lecture that the set O = Z \ 2Z of all odd integers is both open and dense. As a result, the set E = 2Z of all even integers is closed and nowhere dense. Example 8.9. As O is open, Int(O) = O. As O is dense in Z, Cl(O) = Z. As E is closed, Cl(E) = E. As E is nowhere dense in Z, Int(Cl(E)) = Int(E) = . So Int(O) Int(E) = O = O while Int(O E) = Int(Z) = Z. Indeed, Int(O) Int(E) Int(O E). Now Cl(O) Cl(E) = Z E = E while Cl(O E) = Cl() = . Indeed, Cl(O E) Cl(O) Cl(E). Remark 8.10. Notice O is open but not closed while E is closed but not open. Lemma 8.11. Let (X, ) be a topological space and let A X and B X . Then (1) Int(A B ) = Int(A) Int(B ) (2) Cl(A B ) = Cl(A) Cl(B ) Proof. (1) As A B A, Int(A B ) Int(A). As A B B , Int(A B ) Int(B ). So Int(A B ) Int(A) Int(B ). Conversely, let x Int(A) Int(B ). We need to show that x Int(A B ). Suppose to the contrary that x / Int(A B ). Then open neighborhood U of x, U A B . But if U A B , then either U A or U B . Since Int(A) Int(B ) Int(A) A and Int(A) Int(B ) Int(B ) B , in either case, x / Int(A) Int(B ). This is a contradiction. So x Int(A B ) by contradiction. Thus, Int(A) Int(B ) Int(A B ). Int(A B ) = Int(A) Int(B ). (2) As A A B , Cl(A) Cl(A B ). As B A B , Cl(B ) Cl(A B ). So Cl(A) Cl(B ) Cl(A B ). Conversely, let x Cl(A B ). Then open neighborhood U of x, U (A B ) = . That is, (U A) (U B ) = . So either U A = or U B = . In the former case x Cl(A) and in the latter case x Cl(B ). In either case, x Cl(A) Cl(B ). Thus, Cl(A B ) Cl(A) Cl(B ). Cl(A B ) = Cl(A) Cl(B ). Theorem 8.12. Let (X, ) be a topological space and let A1 , . . . , An be a nite collection of subsets of X . Then n (1) Int( n i=1 Ai ) = i=1 Int(Ai ) n n (2) Cl( i=1 Ai ) = i=1 Cl(Ai ) Proof. We will use induction: n (1) Let P (n) : Int( n i=1 Ai ) = i=1 Int(Ai ). Then P (1) : Int(A1 ) = Int(A1 ), which is true. 1 n1 Suppose P (n 1) : Int( n i=1 Ai ) = i=1 Int(Ai ) is true. n n1 Then Int( i=1 Ai ) = Int([ i=1 Ai ] An ). 1 n1 By lemma 8.11, Int([ n i=1 Ai ] An ) = Int( i=1 Ai ) Int(An ). 1 n1 By the induction hypothesis, Int( n i=1 Ai ) = i=1 Int(Ai ). 40

n1 n So Int( n i=1 Ai ) = [ i=1 Int(Ai )] Int(An ) = i=1 Int(Ai ). Hence, P (n 1) implies P (n). n + By induction, Int( n i=1 Int(Ai ) is true n Z . i=1 Ai ) = n (2) Let P (n) : Cl( n i=1 Cl(Ai ). i=1 Ai ) = Then P (1) : Cl(A1 ) = Cl(A1 ), which is true. 1 n1 Suppose P (n 1) : Cl( n i=1 Ai ) = i=1 Cl(Ai ) is true. n1 Then Cl( n A ) = Cl ([ A ] A i i n ). i=1 i=1 n1 1 A ] A ) = Cl( i By lemma 8.11, Cl([ n i n =1 Ai ) Cl(An ). i=1 1 n1 By the induction hypothesis, Cl( i=1 Ai ) = n i=1 Cl(Ai ). n n1 n So Cl( i=1 Ai ) = [ i=1 Cl(Ai )] Cl(An ) = i=1 Cl(An ). Hence, P (n 1) implies P (n). n + By induction, Cl( n i=1 Cl(Ai ) is true n Z . i=1 Ai ) =

8.3

Properties of the Digital Line

Remark 8.13. Recall from theorem 6.31, that in a topological space X with |X | 2, a point x X is an isolated point i the singleton containing that point, {x}, is open. Example 8.14. Every odd integer is an isolated point of the digital line Z. Proof. Notice n O, {n} is open as it is a basic open set. As |Z| > 2, every n O is an isolated point of Z. Example 8.15. Every even integer is a cluster point of the digital line Z. Proof. Let n E. Then open neighborhood U of n, Bn = {n 1, n, n + 1} U . So {n 1, n + 1} U Z \ {n}. That is, open neighborhood U of n, U Z \ {n} = . Hence every n E is a cluster point of Z. Remark 8.16. Recall that a point in a topological space is either an isolated point or a cluster point, but not both. So the set of all isolated points of the digital line Z is the set O of all odd integers while the set of all cluster points of the digial line, the derived set Z , is the set E of all even integers. Hence the number of isolated points in Z is |O| = 0 . That is, there is a countably innite number of isolated points in the digital line Z. Remark 8.17. Recall that subsets of a nowhere dense set are also nowhere dense. So any subset of E is nowhere dense in Z. In particular, every singleton containing an even integer is nowhere dense in Z. Example 8.18. Every singleton set containing an even integer is closed but not open. Further, it is its own boundary. Proof. Let m E = 2Z and let Be = {Bn : n E}. Using interval notation, {n Z : n m 1} = (, m 1].

41

Notice that (, m 1] is open since (, m 1] = {m 3, m 2, m 1} {m 5, m 4, m 3} = Bm2 Bm4 =

Bn where Bn Be

Similarly, [m + 1, ) is open since [m + 1, ) = {m + 1, m + 2, m + 3} {m + 3, m + 4, m + 5} = Bm+2 Bm+4 =

Bn where Bn Be

So Z \ {m} = (, m 1] [m + 1, ) is open as it is a union of open sets. Hence {m} is closed. So Cl({m}) = {m}. As {m} is nowhere dense, Int(Cl({m})) = Int({m}) = . Since {m} = Int({m}), {m} is not open. {m} = Cl({m}) \ Int({m}) = {m} \ = {m}

Exercise 8.19. Show Be = {Bn : n 2Z} is an open cover but is not a basis for the digital line. Example 8.20. Every singleton set containing an odd integer is open but not closed. Proof. Let m O = Z \ 2Z and let Be = {Bn : n 2Z}. Clearly, {m} is open as it is a basis element. Notice (, m 2] = n<m2 Bn where Bn Be is open. Similarly, [m + 2, ) = n>m+2 Bn where Bn Be is open. So Z \ {m 1, m, m + 1} = (, m 2] [m + 2, ) is open as it is a union of open sets. Further, it is the largest open set not containing m since any open set containing the even integer m 1 must contain Bm1 = {m 2, m 1, m} and any open set containing the even integer m + 1 must contain Bm+1 = {m, m + 1, m + 2}. So {m 1, m, m + 1} is the smallest closed set containing m. That is, Cl({m}) = {m 1, m, m + 1}. Since {m} = Cl({m}), {m} is not closed. Example 8.21. Let m O. Then the boundary of {m} is the set containing the pair of consecutive even integers about m. Proof. {m} = Cl({m}) \ Int({m}) = {m 1, m, m + 1} \ {m} = {m 1, m + 1}

42

Remark 8.22. Since it is not true that every singleton set is closed, the digital line Z is not T1 (Fr echet). Hence it is also not T2 (Hausdor). Exercise 8.23. Show that the digital line Z is T0 (Kolmogorov). Example 8.24. {O, E} is a partition but is not a separation of the digital line Z. Further, E is the boundary of both sets. Proof. Clearly, O and E are nonempty subsets of Z such that O E = and O E = Z. Hence {O, E} is a partition of Z. However, it is not a separation since E is not open. O = Cl(O) \ Int(O) = Z\O = E E = Cl(E) \ Int(E) = E\ = E

Example 8.25. The digital line Z is connected. Proof. Suppose {U, V } is a separation of Z. Then U and V are nonempty open subsets of Z such that U V = and U V = Z. Since {O, E} is not a separation of Z, neither U nor V consists enitrely of odd integers. That is, m, n E such that m U and n V . As U is open, Bm = {m 1, m, m + 1} U . Notice m +2 / V since otherwise Bm+2 = {m +1, m +2, m +3} V , which would require that m +1 V . This is not possible since U V = and m + 1 U . So m + 2 U and hence Bm+2 = {m + 1, m + 2, m + 3} U . But then m + 4 / V. We can continue ad innitum so that V contains no even integers. But this contradicts the fact that n E such that n V . Hence, by contradition, there does not exist a separation of Z. The digital line Z is connected. Exercise 8.26. Let (a, b) = {n Z : a < n < b} and let B = {(a, b) : a, b Z}. Show that the topology generated by B is the discrete topology on Z.

9
9.1

Lecture 09
More Properties of the Digital Line

Example 9.1. Be = {Bn : n 2Z} is an open cover but is not a basis for the digital line.

43

Proof. Clearly, Be is a collection of open sets in the digital line Z. Let m, m + 2 2Z. Then Bm = {m 1, m, m + 1} and Bm+2 = {m + 1, m + 2, m + 3}. Notice Bm Bm+2 = {m + 1}. So m, m + 2 2Z, m, m + 1, m + 2 Bm Bm+2 where Bm , Bm+2 Be . Hence n2Z Bn = Z. That is, Be is an open covering of Z. However, Be is not a basis for the digital line since for the odd integer m + 1, m + 1 Bm Bm+2 but B Be such that m + 1 B Bm Bm+2 . Example 9.2. The digital line Z is a T0 (Kolmogorov) space. Proof. Consider two consecutive integers m and m + 1. Notice that one is even while the other is odd. Without loss of generality, we may assume m is odd. Then Bm = {m} is an open neighborhood of m such that m + 1 / Bm . However, the smallest open set containing m + 1 is Bm+1 = {m, m + 1, m + 2} where m Bm+1 . Now consider m, m + 2 Z. Notice that these integers are either both even or both odd. Suppose both are even. Then Bm = {m 1, m, m +1} is an open neighborhood of m such that m +2 / Bm and Bm+2 = {m + 1, m + 2, m + 3} is an open neighborhood of m + 2 such that m / Bm+2 . Now suppose both are odd. Then Bm = {m} is an open neighborhood of m such that m + 2 / Bm and Bm+2 = {m + 2} is an open neighborhood of m + 2 such that m / Bm+2 . More generally, if m, n Z such that |m n| 2, then Bm is an open neighborhood of m such that n / Bm and Bn is an open neighborhood of n such that m / Bn . So m = n Z, an open neighborhood of at least one of the integers not containing the other. Hence Z is T0 (Kolmogorov). Remark 9.3. The topological invariants which we have explored so far are summarized below for the digital line topology (DLT): # of Isolated Points Connectedness Separation Condition Countability Condition 0 Connected T0 Separable

Remark 9.4. As the digital line Z is connected, the only sets which are both open and closed are and Z. We now consider whether there exist subsets that are neither open nor closed. We will begin by looking at a necessary condition for an open set which contains an even integer. Example 9.5. Let A Z. If A is open, then for every even integer m A, the consecutive odd integers m 1, m + 1 A. Proof. Suppose A is open and let m A where m is an even integer. Then as A is is open, Bm = {m 1, m, m + 1} B such that m Bm A. But then m 1, m + 1 A. This is a contradiction if either m 1 / A or m + 1 / A. So A must contain both m 1 and m + 1. Example 9.6. Let m, m + 1 Z. Then {m, m + 1} is neither open nor closed.

44

Proof. Notice one of the integers m and m + 1 is even. So {m, m + 1} does not contain both consecutive odd integers centered about the even integer. Hence {m, m + 1} is not open. Without loss of generality, we may assume m is even. Then Cl({m, m + 1}) = Cl({m}) Cl({m + 1}) = {m} {m, m + 1, m + 2} = {m, m + 1, m + 2} Hence {m, m + 1} is not closed. So {m, m + 1} is neither open nor closed.

9.2

Relations on a Set

Remark 9.7. Recall that a relation between two sets X and Y is a subset R X Y . Denition 9.8. A relation on a set X is a subset R X X . Denition 9.9. A relation R on a set X is reexive i x X , (x, x) R. That is, xRx. Denition 9.10. A relation R on a set X is irreexive i x X , (x, x) / R. That is, x Rx. Denition 9.11. A relation R on a set X is symmetric i x, y X , if(x, y ) R, then (y, x) R. That is, if xRy , then yRx. Denition 9.12. A relation R on a set X is antisymmetric i x, y X , if (x, y ) R and (y, x) R, then x = y . That is, if xRy and yRx, then x = y . Remark 9.13. Notice that symmetric and antisymmetric are not opposite terms. Example 9.14. Let R = {(x, x) : x X }. Clearly, R is reexive. Notice that R is also both symmetric and antisymmetric. Denition 9.15. A relation R on a set X is asymmetric i x, y X , if (x, y ) R, then (y, x) / R. That is, if xRy , then y Rx. Remark 9.16. Notice that if a relation R is asymmetric, then it is also irreexive. Further, an asymmetric relation satises the conditions for antisymmetry vacuously. Denition 9.17. A relation R on a set X is transitive i x, y, z X , if (x, y ) R and (y, z ) R, then (x, z ) R. That is, if xRy and yRz , then xRz . Denition 9.18. Let R be a relation on the set X and let x, y X . If either (x, y ) R or (y, x) R, then the elements x and y are said to be comparable. If neither (x, y ) R nor (y, x) R, then the elements x and y are said to be incomparable. Denition 9.19. Let R be a relation on the set X . If every pair of distinct elements x = y X are comparable, then the relation R is called a total ordering on X . If at least one pair of distinct elements in X are comparable, then the relation R is called a partial ordering on X . 45

Remark 9.20. Notice that every total ordering is a partial ordering, but not every partial ordering is a total ordering. Further, a total ordering can be irreexive since we require only that all distinct elements be comparable. Denition 9.21. Let R be a partial ordering on X such that R is reexive, antisymmetric, and transitive. Then the set X together with this relation, (X, R), is called a partially ordered set or poset. If in addition R is a total ordering on X , then (X, R) is called a totally ordered set. Remark 9.22. A totally ordered set is also called a linearly ordered set or a chain. Example 9.23. (Z, ) is a totally ordered set. Proof. a, b Z, either a b or b a. So is a total ordering on Z. Now we show that (Z, ) is a poset. (1) a Z, it is true that a a since a = a. Hence is reexive. (2) Let a, b Z and suppose a b and b a. Then a b a. So a = b. Hence is antisymmetric. (3) Let a, b, c Z and suppose a b and b c. Then a b c. That is, a c. Hence is transitive. Denition 9.24. Let (X, R) be a poset. An element x X is called a minimal element of X if a distinct element y X such that yRx. An element x X is called the least element of X , denoted x = min X , if y X , xRy . Example 9.25. Let A = {2, 4, 5, 10, 12, 20, 25}. Then (1) (A, ) has one minimal element wich is also the least element, namely 2. (2) (A, |) has two minimal elements, namely 2 and 5, but no least element. Proof. For (A, ): 2 4 5 10 12 20 25. x = 2 A such that x 2. Hence 2 is a minimal element. Further, y A, 2 y . Hence 2 is the least element. For (A, |): 2|4|12, 2|10|20, 5|10|20, and 5|25. x = 2 A such that x|2 and y = 5 A such that y |5. Hence 2 and 5 are minimal elements. x A such that y A, x|y . Hence there is no least element. Lemma 9.26. Let (X, R) be a poset. Then the least element is unique if it exists. Proof. Suppose x and y are both least elements of X . Then as x is a least element, it is also a minimal element. So y = x X such that yRx. Hence if yRx, then y = x. As y is a least element, yRx. Thus y = x. Denition 9.27. Let (X, R) be a poset. An element x X is called a maximal element of X if a distinct element y X such that xRy . An element x X is called the greatest element of X , denoted x = max X , if y X , yRx. Remark 9.28. Notice that the greatest element, if it exists, is also a maximal element.

46

Exercise 9.29. Let A = {2, 4, 5, 10, 12, 20, 25}. For (A, ) and (A, |), determine the maximal elements and the greatest elements (if they exist). Exercise 9.30. Let (X, R) be a poset. Prove the greatest element is unique if it exists. Denition 9.31. Let R be a toal ordering on the set X such that R is irreexive and transitive. Then the relation R is called a simple ordering on X and (X, R) is called a simply ordered set. Remark 9.32. Some mathematicians, and the corresponding literature, refer to a simple ordering as a strict partial ordering. Example 9.33. (Z, <) is a simply ordered set. Proof. (1) a = b Z, either a < b or b < a. So < is a total ordering on Z. (2) a Z, a < a. Hence, < is irreexive. (3) a, b, c Z, if a < b and b < c, then a < c. Hence, < is transitive. Lemma 9.34. (Trichotomy Law) Let (X, <) be a simply ordered set. Then x, y X exactly one of the following is true. (1) x < y (2) x = y (3) x > y Proof. As (X, <) is a simply ordered set, < is a total ordering on X . So x = y X , either x < y or y < x. That is, either x < y or x > y . So for every pair of not necessarily distinct elements x, y X , either x < y or x = y or x > y , where these are pairwise mutually exclusive. Remark 9.35. Notice that if (X, <) is a simply ordered set, then is also a relation on X which we can dene in terms of the relation <: ab a < b or a = b

Further, the relation is a total ordering on X . Exercise 9.36. Show (X, <) is a simply ordered set i (X, ) is a totally ordered set.

9.3

Introduction to the Order Topology

Denition 9.37. Let (X, <) be a simply ordered set. Given elements a, b X such that a < b, there are four subsets of X called the intervals determined by a and b. They are the following: (a, b) = {x X : a < x < b} (a, b] = {x X : a < x b} [a, b) = {x X : a x < b} [a, b] = {x X : a x b} Denition 9.38. (The Order Topology) Let (X, <) is a simply ordered set such that |X | 2 and let B be the collection of all sets of the following types: (1) All intervals (a, b) X . 47

(2) All intervals [a0 , b) X where a0 = min X (if it exists). (3) All intervals (a, bo ] X where b0 = max X (if it exists). Then collection B is a basis for a topology on X called the order topology. Exercise 9.39. Show B satises the two conditions for a topology generated by a basis. Remark 9.40. Intervals of the form (a, b) are called open intervals because they are always in the order topology. The half-open intervals [a0 , b) and (a, b0 ] are only in the order topology if a0 = min X exists and b0 = max X exists. Example 9.41. Let B = {(m, n) : m, n Z}. Then the topology generated by B , which is the order topology on Z, is the discrete topology on Z. Proof. Notice (n 1, n + 1) = {x Z : n 1 < x < n + 1} = {n}. Hence every singleton set is open in the order topology on Z. So U Z, U = nU {n} is open. So the topology generated by B contains all subsets of Z. That is, B = {U Z} is the discrete topology on Z. Denition 9.42. Let (X, <) be a simply ordered set. Given an element a X , there are four subsets of X called the rays determined by a. They are the following: (a, ) = {x X : x > a} (, a) = {x X : x < a} [a, ) = {x X : x a} (, a] = {x X : x a} Example 9.43. The rays of the form (a, ) and (, a) are called open rays since they are always open in the order topology on X . Proof. If X has a largest element b0 , then (a, ) = (a, b0 ]. Otherwise, (a, ) = n>a (a, n). In either case, (a, ) is open. Similarly, if X has a smallest element a0 , then (, a) = [a0 , a). Otherwise, (, b) = n<a (n, a). In either case, (, a) is open. Open rays are always open in the order topology. Example 9.44. The intervals of the form [a, b] are called closed intervals since they are always closed in the order topology on X . Proof. Since X \ [a, b] = (, a) (b, ) is the union of open sets, it is open. Hence, [a, b] is closd. Intervals of the form [a, b] are always closed in an order topology.

10
10.1

Lecture 10
More Properties of Posets and Simply Ordered Sets

Example 10.1. Let A = {2, 4, 5, 10, 12, 20, 25}. Then (1) (A, ) has one maximal element which is also the greatest element, namely 25. (2) (A, |) has three maximal elements, namely 12, 20, and 25, but no greatest element.

48

Proof. For (A, ): 2 4 5 10 12 20 25. x = 25 A such that 25 x. Hence 25 is a maximal element. Further, y A, y 25. Hence 25 is the greatest element. For (A, |): 2|4|12, 2|10|20, 5|10|20, and 5|25. x = 12 A such that 12|x, y = 20 A such that 20|y , and z = 25 A such that 25|z . Hence 12, 20, and 25 are maximal elements. x A such that y A, y |x. Hence there is no greatest element. Lemma 10.2. Let (X, R) be a poset. Then the greatest element is unique if it exists. Proof. Suppose x and y are both greatest elements of X . Then as x is a greatest element, it is also a maximal element. So y = x X such that xRy . Hence if xRy , then y = x. As y is a greatest element, xRy . Thus y = x. Denition 10.3. Let Y be a subset of the poset (X, R). If x X such that y Y , yRx, then the element x is called an upper bound of Y and Y is said to be bounded above. If Y is bounded above, then the element u X is called the least upper bound, or supremum, of Y if u is an upper bound of Y such that uRz whenever z is an upper bound of Y . Remark 10.4. Notice that the supremum of Y , denoted sup Y , is the least element, w.r.t. the relation R, in the set of all upper bounds of Y . Further, if sup Y Y , then sup Y = max Y . If Y is not bounded above, we write sup Y = . Example 10.5. Let A = {2, 4, 5, 10, 12, 20, 25}. Then (1) As a subset of (Z+ , ), sup A = 25. (2) As a subset of (Z+ , |), sup A = 300. Proof. (1) Notice that any x {u Z+ : u 25} is an upper bound of A. The least element in this set is 25. So sup A = 25 = max A. (2) Clearly, p = 2 4 5 10 12 20 25 Z+ is an upper bound of A. The smallest integer divisible by all the elements in the set A is called the least common multiple of A. Notice 2 = 21 , 4 = 22 , 5 = 51 , 10 = 2 5, 12 = 22 3, 20 = 22 5, and 25 = 52 . So lcm(A) = 22 3 52 = 300. Hence sup A = 300. Denition 10.6. Let Y be a subset of the poset (X, R). If x X such that y Y , xRy , then the element x is called a lower bound of Y and we say Y is bounded below. If Y is bounded below, then the element l X is called the greatest lower bound, or inmum, of Y if l is a lower bound of Y such that zRl whenever z is a lower bound of Y . Remark 10.7. Notice that the inmum of Y , denoted inf Y , is the greatest element, w.r.t. the relation R, in the set of all lower bounds of Y . Further, if inf Y Y , then inf Y = min Y . If Y is not bounded below, we write inf Y = . Example 10.8. Let A = {2, 4, 5, 10, 12, 20, 25}. Then (1) As a subset of (Z+ , ), inf A = 2. (2) As a subset of (Z+ , |), inf A = 1. 49

Proof. (1) Notice L = {1, 2} is the set of all lower bounds of A. The greatest element in L is 2. So Inf A = 2 = min A. (2) Notice that the only positive integer that divides all elements in A is 1. So the set of all lower bounds of A is the singleton L = {1}. Hence inf A = 1. Exercise 10.9. Let Y be a subset of the poset (X, R). Prove (1) sup Y is unique when it exists (2) inf Y is unique when it exists Denition 10.10. Let Y be a subset of the poset (X, R). Then Y is bounded if it is both bounded below and bounded above. Example 10.11. (X, <) is a simply ordered set i (X, ) is a totally ordered set. Proof. Suppose (X, <) is a simply ordered set. Then < is a total ordering on X . So x, y X , either x < y or y < x. If x < y , then x y . Similarly, if y < x, then x y . So x, y X , either x y or y x. Hence is a total ordering on X . We have alreay established that is reexive, antisymmetric, and transitive. Thus (X, ) is a totally ordered set. Conversely, suppose (X, ) is a totally ordered set. Then is a total ordering on X . So x, y X , either x y or y x. That is, x < y or x = y or y < x. So x = y X , either x < y or y < x. Hence < is a total ordering on X . We have alreay established that < is irreexive and transitive. Thus (X, <) is a simply ordered set. Remark 10.12. Let (X, <) be a simply ordered set. If X = {a, b}, then either a < b or b < a. In either case, X has both a least element and a greatest element. If |X | > 2 and x X such that x = min X and x = max X , then a, b X such that a < x < b.

10.2

Properties of the Order Topology

Example 10.13. Let (X, <) be a simply orded set where |X | 2 and let B be the set of all intervals of the form (1) (a, b) X (2) [a0 , b) X where a0 = min X . (3) (a, b0 ] X where b0 = max X . Then B satises the two conditions for a topology generated by a basis. Proof. (1) Let x X . If x = a0 = min X , then b X such that a0 < b since |X | 2. So x [a0 , b]. If x = b0 = max X , then a X such that a < b0 since |X | X . So x (a, b0 ]. Otherwise, x = min X and x = max X and so |X | > 2. Hence a, b X such that a < x < b. So x (a, b). 50

Thus x X B B such that x B . (2) Let x B1 B2 where B1 , B2 B . Then x B1 and x B2 . Suppose x = a0 = min X . Then B1 = [a0 , b) and B2 = [a0 , c) for some b, c X . Exactly one of the following is true: b < c, b = c, or c < b. If b < c, then a0 < b < c. So x [a0 , b) [a0 , b) [a0 , c). If b = c, then a0 < b = c. So x [a0 , b) [a0 , b) [a0 , c). If c < b, then a0 < c < b. So x [a0 , c) [a0 , b) [a0 , c). Now suppose x = b0 = max X . Then B1 = (a, b0 ] and B2 = (c, b0 ] for some a, c X . Exactly one of the following is true: a < c, a = c, or c < a. If a < c, then a < c < b0 . So x (c, b0 ] (a, b0 ] (c, b0 ]. If a = c, then a = c < b0 . So x (c, b0 ] (a, b0 ] (c, b0 ]. If c < a, then c < a < b0 . So x (a, b0 ] (a, b0 ] (c, b0 ]. Otherwise, x = min X and x = max X . Without loss of generality, we may assume B1 = (a, b) and B2 = (c, d) for some a, b, c, d X . So a < x < b and c < x < d. As a < x and x < d, a < d. As c < x and x < b, c < b. Exactly one of the following is true: a < c, a = c, or c < a. Similarly, exactly one of the following is true: b < d, b = d, or d < b. There are nine cases. Case 1i: a < c and b < d. So a < c < x < b < d. Hence x (c, b) (a, b) (c, d). Case 1ii: a < c and b = d. So a < c < x < b = d. Hence x (c, b) (a, b) (c, d). Case 1iii: a < c and d < b. So a < c < x < d < b. Hence x (c, d) (a, b) (c, d). Case 2i: a = c and b < d. So a = c < x < b < d. Hence x (c, b) (a, b) (c, d). Case 2ii: a = c and b = d. So a = c < x < b = d. Hence x (c, b) (a, b) (c, d). Case 2iii: a = c and d < b. So a = c < x < d < b. Hence x (c, d) (a, b) (c, d). Case 3i: c < a and b < d. So c < a < x < b < d. Hence x (a, b) (a, b) (c, d). Case 3ii: c < a and b = d. So c < a < x < b = d. Hence x (a, b) (a, b) (c, d). Case 3iii: c < a abd d < b. 51

So c < a < x < d < b. Hence x (a, d) (a, b) (c, d). Thus, in any case, if x B1 B2 where B1 , B2 B , then B3 B such that x B3 B1 B2 . Remark 10.14. Intervals of the form [a0 , b) are open in the order topology on X only if a0 = min X . That is, only if X is bounded below and inf X X . Intervals of the form (a, b0 ] are open in the order topology on X only if b0 = max X . That is, only if X is bounded above and sup X X . Exercise 10.15. Let B = {[a, b) R} {(a, b] R}. Show that the topology generated by B is the discrete topology on R. Remark 10.16. Recall that open intervals (a, b) and the open rays (, a) and (a, ) are always open in the order topology on X while closed intervals [a, b] are always closed in the order topology on X . Example 10.17. The half-open rays (, a] and [a, ) are always closed in the order topology on X. Proof. Since X \ [a, ) = (, a) is open, [a, ) is closed. Similarly, (, a] is closed since X \ (, a] = (a, ) is open. Hence, the half-open rays are always closed in the order topology. Remark 10.18. If X has a least element a0 , then [a0 , ) = X . Similarly, if X has a greatest element b0 , then (, b0 ] = X . So if X has both a least element a0 and a greatest element b0 , then [a0 , b0 ] = X . Remark 10.19. The interval notation used in the previous remark may be misleading. The notation X = [a0 , ) does not necessarliy mean that X contains all elements greater than or equal to a0 . Rather, the notation [a0 , ) means the set of all elements x X such that x a0 . Example 10.20. Let X = [a0 , b) (b, ). Then as a subset of X , [a0 , ) = [a0 , b) (b, ). Now suppose X R. Then as a subset of R, [a0 , ) = X since b [a0 , ) and b / X. Remark 10.21. Unless otherwise stated, all intervals and rays are considered as subsets of the space X. Remark 10.22. If X has neither a least element nor a greatest, then the half-open intervals are neither open nor closed in the order topology on X . Example 10.23. Suppose X has neither a least element nor a greatest element. Subset (a, b) [a, b) (a, b] [a, b] (, a) (, a] (a, ) [a, ) Interior (a, b) (a, b) (a, b) (a, b) (, a) (, a) (a, ) (a, ) Closure [a, b] [a, b] [a, b] [a, b] (, a] (, a] [a, ) [a, ) Boundary {a, b} {a, b} {a, b} {a, b} {a} {a} {a} {a}

Example 10.24. Every singleton set is closed in the order topology on X . 52

Proof. Let a X . Then {a} = (, a] [a, ) is closed as it is the intersection of two closed sets. Hence all singleton sets are closed in the order topology on X . Remark 10.25. Since every singleton set is closed in the order topology on X , the space X is T1 (Fr echet). Further, every nite point set is closed in the order topology on X as it is a nite union of closed singleton sets. Example 10.26. The space X with the order topology is Hausdor. Proof. Notice that at least one pair of distinct elements exist in X since |X | 2. Let a = b X . Without loss of generality, we may assume a < b. There are two cases. Suppose c X such that a < c < b. Then a (, c) and b (c, ) where (, c) (c, ) = . If no such c exists, then a (, b) and b (a, ) where (, b) (a, ) = (a, b) = . Hence a = b X disjoint open neighborhoods of a and b. X is T2 (Hausdor). Remark 10.27. We have seen that the order topology on Z is indistinguishable from the discrete topology on Z. This is true in general for any subset of Z. Many properties of the order topology depend upon the underlying set X .

11
11.1

Lecture 11
Compositions, Inverses, and Restrictions of Functions

Denition 11.1. Let f : A B and g : B C . The function g f : A C dened by the equation (g f )(a) = g (f (a)) is called a composition of f and g . That is, g f is the set {(a, c) : b rng f, f (a) = b and g (b) = c} . Denition 11.2. The function iA : A A dened by iA (a) = a is called the identity function on A. Remark 11.3. Clearly, the identity function is a bijection. Denition 11.4. Let f : A B . Then g : B A is the inverse of f i g f = iA and f g = iB . If g is the inverse of f , we write g = f 1 . Theorem 11.5. Let f : A B . Then f has an inverse i f is bijective. Proof. Suppose f : A B has an inverse f 1 : B A. Let x, y A and suppose f (x) = f (y ). Then f 1 (f (x)) = f 1 (f (x)) f 1 f (x) = f 1 f (y ) iA (x) = iA (y ) x = y Hence f is injective. Now let b B . Then b = iB (b) = f f 1 (b) = f (f 1 (b)) 53

where f 1 (b) A. So b B a A such that b = f (a). Hence f is surjective. Thus f is bijective. Conversely, suppose f : A B is bijective. Dene g : B A by g (b) = a i f (a) = b. Then g f : A A where g f (a) = g (f (a)) = g (b) = a So g f = iA . Similarly, f g : B B where f g (b) = f (g (b)) = f (a) = b So f g = iB . Hence g = f 1 . Remark 11.6. Notice that if f : A B is a bijection, then it has an inverse f 1 : B A. Further, f 1 is also a bijection since it has an inverse, namely f . Theorem 11.7. Let f : A B and g : B C . (1) If f and g are both surjective, then g f is surjective. (2) If f and g are both injective, then g f is injective. (3) If f and g are both bijective, then g f is bijective. Proof. (1) Since g is surjective, rng g = C . That is, c C b B such that g (b) = c. Similarly, since f is surjective, rng f = B and so b B a A such that f (a) = b. Hence, c C a A such that (g f )(a) = g (f (a)) = c. Thus, g f is surjective. (2) Suppose (g f )(a1 ) = (g f )(a2 ). Then g (f (a1 )) = g (f (a2 )). Since g is injective, f (a1 ) = f (a2 ). Now since f is injective, a1 = a2 . Hence, g f is injective. (3) The result follows immediately from (1) and (2). Exercise 11.8. Let f : A B and g : B C be bijections. Show (g f )1 = f 1 g 1 . Denition 11.9. Let : A B and let A0 A and B0 B . Then (1) The set (A0 ) = {(a) B : a A0 } is called the image, or direct image, of A0 under . (2) The set 1 (B0 ) = {a A : (a) B0 } is called the inverse image, or preimage, of B0 under . Remark 11.10. Notice (A0 ) rng B (A) = rng 1 (B0 ) A = dom 1 (B ) = A = dom 54

Denition 11.11. Let : A B and let b B . The inverse image of a singleton set, 1 ({b}), is called a bre of the map . Remark 11.12. Notice that a function : A B is (1) injective i every bre is either or a singleton. (2) surjective i every bre is nonempty. (3) bijective i every bre is a singleton. Further, if b1 = b2 , then 1 ({b1 }) 1 ({b2 }) = since otherwise a A such that (a, b1 ) and (a, b2 ) and hence is not a function. Denition 11.13. Let : A B and let A0 A. The restriction of to A0 , denoted |A0 , is a function mapping A0 into B such that |A0 = {(a, (a)) : a A0 }. Remark 11.14. Notice that if : A B is an injection, then |A0 is also an injection.

11.2

Finite Sets

Remark 11.15. Recall that a set A is nite i there exists a bijection : A {1, , n}, where n Z+ . In this case, the cardinality of A is n. By convention, is nite and has cardinality 0. Lemma 11.16. Let n Z+ , let A be a set, and let a0 A. Then there exists a bijection f : A {1, . . . , n + 1} i there exists a bijection g : A \ {a0 } {1, . . . , n}. Proof. Suppose g : A \ {a0 } {1, . . . , n} is a bijection. Then we dene a function f : A {1, . . . , n + 1} by g (a) a A \ {a0 } f (a) = n + 1 if a = a0 . Clearly, f is the desired bijection mapping A into {1, . . . , n + 1}. Conversely, suppose f : A {1, . . . , n + 1} is a bijection. If f (a0 ) = n + 1, then f |A\{a0 } is the desired bijection mapping A \ {a0 } into {1, . . . , n}. Otherwise, let f (a0 ) = m and f (a1 ) = n + 1. Then we dene a new function h : A {1, . . . , n + 1} dened by n + 1 if a = a0 h(a) = m if a = a1 f (a) a A \ {a0 , a1 } Now Let g = h|A\{a0 } . Clearly, g is the desired bijection mapping A \ {a0 } into {1, . . . , n}. Theorem 11.17. Let m, n Z+ such that m < n and let B A such that B = . Suppose that there exists a bijection f : A {1, . . . , n} for n 2. Then there exists a bijection g : B {1, . . . , m} for 1 m < n. Proof. Consider the bijection f : A {1, . . . , n}. We choose an element a0 A and form the nonempty proper subset An1 = A \{a0 }. Then, by lemma 11.16, there exists a bijection g : An1 {1, . . . , n 1}. If |An1 | > 1, we can choose a dierent element a1 An1 and form the nonempty proper subset An2 = An1 \ {a1 } = A \ {a0 , a1 }. Then there exists a bijection h : An2 {1, . . . , n 2}. We can continue this process a nite number of times until we obtain the nonempty proper subset A1 which 55

contains a single element for which there exists a bijection mapping this element into {1}. Hence, for any nonempty proper subset B A, if there exists a bijection f : A {1, . . . , n} for n 2, then there exists a bijection g : B {1, . . . , m} for 1 m < n. Corollary 11.18. Let A be a nite set and let B A. Then B is nite. Proof. Let A have cardinality n 2. Suppose B = A. Then B is nite and has the same cardinality as A. Now suppose B = . Then, by convention, B is nite with cardinality 0. We now consider nonempty proper subsets B A. Since A is nite, there exists a bijection f : A {1, . . . , n} for n 2. Hence, by theorem 11.17, there exists a bijection g : B {1, . . . , m} for 1 m < n. Thus B is nite with cardinality m. If A has cardinality n 1, then A is either a singleton set or is empty. Any subset of A is therefore the same singleton set or is empty. In either case, B A is clearly nite. In any case, if A is nite, then any subset B A is also nite. Corollary 11.19. Arbitrary intersections of nite sets are nite. Proof. Let {Ai : i I } be an indexed family of nite sets. Then iI Ai is a subset of each set Ai . Hence, by corollary 11.18, the arbitrary intersection of nite sets is nite. Theorem 11.20. Let B be a nonempty set and let n Z+ . Then the following are equivalent: (1) B is nite. (2) There exists a surjection mapping {1, . . . , n} onto B . (3) There exists an injection mapping B into {1, . . . , n}. Proof. (1) (2) If B is nite, there exists a bijection f : B {1, . . . , n}. Hence, f 1 : {1, . . . , n} B is also a bijection. (2) (3) Let f : {1, . . . , n} B be surjective. Dene g : B {1, . . . , n} by the equation g (b) = min f 1 ({b}). Since f is surjective, the bre f 1 ({b}) is not empty for any b B . If b = b , then the bres f 1 ({b}) and f 1 ({b }) are disjoint. So min f 1 ({b}) = min f 1 ({b }). That is g (b) = g (b ). Hence, g : B {1, . . . , n} is injective. (3) (1) Let g : B {1, . . . , n} is injective. If g is also surjective, then g is a bijection and hence B is nite. Otherwise, there exists a superset A of B for which we can form a bijection f : A \ B {1, . . . , n} \ rng g . Now dene a function h : A {1, . . . , n} by h(a) = g (a) for a B f (a) for a A \ B

Then h is a bijection and hence A is nite. Since B A, B is nite. Corollary 11.21. Finite unions of nite sets are nite. Proof. First, we show that if A and B are nite, then so is A B . The result is trivial if either A or B is empty. Otherwise, there are bijections f : {1, . . . , m} A and g : {1, . . . , n} B for some m, n Z+ . Dene h : {1, . . . , m + n} A B by h(i) = f (i) g (i m) for 1 i m for m + 1 i m + n 56

Notice that if A B = , then h is a bijection and if A B = , then h is surjective but is not an injection. In either case, h is surjective. Hence, by theorem 11.20, A B is nite. Now we show by induction that if the sets A1 , . . . , An are nite, then so is their union. The result is trivial for n = 1. Now assume it is true for n 1. Then the nite union A1 An is the union of two nite sets A1 An1 and An . Hence, the union of the these two nite sets is nite. Thus, the case n 1 implies the case n. So, by induction, the nite union of nite sets is nite. Exercise 11.22. Let = {U X : U = or X \ U is nite }. Show that is a topology on X . Theorem 11.23. Let Y be a subset of the poset (X, R). Then (1) sup Y is unique when it exists. (2) inf Y is unique when it exists. Proof. (1) Suppose Y is bounded above and let U X be the set of all upper bounds of Y . If (U, R) has a least element, then it is unique. Hence, sup Y = min L is unique. If (U, R) has no least element, then sup Y does not exist. If Y is not bounded above, then sup Y = (i.e. the supremum does not exist). (2) Suppose Y is bounded below and let L X be the set of all lower bounds of Y . If (L, R) has a greatest element, then it is unique. Hence inf Y = max L is unique. If (L, R) has no greatest element, then inf Y does not exist. If Y is not bounded below, then inf Y = (i.e. the inmum does not exist). Remark 11.24. Notice that it is not true that every set that is bounded above has a least upper bound and it is not true that every set that is bounded below has a greatest lower bound. Example 11.25. Let B be the collection of all half-open intervals of the form (1) (a, b] R (2) [a, b) R. Then the topology generated by B is the discrete topology on R. Proof. Let a, b, c R such that a < b < c. Notice {b} = (a, b] [b, c) is open. Hence, every subset U R is open as it is a union of singleton sets. Thus the topology generated by B is the discrete topology on R.

12
12.1

Lecture 12
The Conite Topology

Example 12.1. Let f : A B and g : B C be bijections. Then f 1 : B A dened by f 1 (b) = a i f (a) = b and g 1 : C B dened by g 1 (c) = b i g (b) = c. Further, f 1 and g 1 are both bijections. So g f : A C and f 1 g 1 : C A are both bijections.

57

Notice that (f 1 g 1 ) (g f ) : A A where (f 1 g 1 ) (g f )(a) = (f 1 g 1 ) (g (f (a))) = (f 1 g 1 )(g (b)) = (f 1 g 1 )(c) = f 1 (g 1 (c)) = f 1 (b) = a So (f 1 g 1 ) (g f ) = iA . Further, (g f ) (f 1 g 1 ) : C C where (g f ) (f 1 g 1 )(c) = (g f )((f 1 (g 1 (c))) = (g f )((f 1 (b)) = (g f )(a) = g (f (a)) = g (b) = c So (g f ) (f 1 g 1 ) = iC . Hence (g f )1 = f 1 g 1 . Denition 12.2. A subset A X is conite i X \ A is nite. That is, A contains all but nitely many elements of X . Example 12.3. Let = {U X : U = or X \ U is nite}. Then is called the conite topology, or nite-complement topology, on X . Proof. (1) Clearly, . Since X \ X = is nite, X . (2) Let {Ui : i I } be an indexed family of nonempty sets in . The X \ Ui is nite i I . So iI X \ Ui is nite since arbitrary intersections of nite sets are nite. Since iI X \ Ui = X \ iI Ui , the arbitrary union iI Ui . Clearly, iI Ui = iI Ui . Hence arbitrary unions of sets in are also in . (3) Let U1 , . . . , Un be nonempty sets in . Then X \ Ui is nite i {1, . . . , n}. So n i=1 X \ Ui is nite since nite unions of nite sets are nite. n n Since n i=1 X \ Ui = X \ i=1 Ui , the nite intersection i=1 Ui . n Clearly, i=1 Ui = . Hence nite intersections of sets in are also in . is a topology on X . Remark 12.4. Notice that if X is nite, then every subset of X is also nite. So the conite topology on a nite set X is the same as the discrete topology on the set X . Example 12.5. Every conite topological space is T1 (Fr echet). 58

Proof. Let be the conite topology on a set X and let x X . Since X \ (X \ {x}) = {x} is nite, X \ {x} is open. So {x} is closed. Hence every singleton set is closed. Thus X is T1 (Fr echet). Exercise 12.6. Let X be a connite topological space where |X | 0 . Show that X is not T2 (Hausdor). Example 12.7. If the underlying set X is not nite, there are no isolated points in the conite topological space X . Proof. Let be the connite topology on X where |X | 0 , let x X , and let U be an open neighborhood of x. Then U contains all but nitely many elements of X as it is conite. That is, every open neighborhood U of x contains innitely many points of X . Since a conite topological space is T1 (Fr echet), by therorem 6.30 x is a cluster point of X . Hence there are no isolated points in the conite topological space where the underlying set is not nite. Example 12.8. Let X be a conite topological space. If A X is closed, then either A = X or A is nite. Proof. Suppose A X is closed. The X \ A is open. So either X \ A = or X \ (X \ A) is nite. That is, either A = X or A is nite. Example 12.9. Let X be a conite topological space. A Proof. Suppose A X is closed. Then by example 12.8, A is nite. Conversely, suppose A X is nite. Then since X \ (X \ A) is nite, X \ A is open. Hence A is closed. Example 12.10. Let X be a conite topological space where |X | 0 . Then the only nite set that is open is and the only innite set that is closed is X . Let F X be nite and let U X be open. Subset F U X Interior U X Closure F X X X is closed i A is nite.

Since the only sets that are both open and closed are and X , X is connected. Further, every nonempty proper open subset U X is dense in X since Cl(U ) = X . If |X | = 0 , a countably innite open subset D X . That is, D is a countable dense subset. Hence X is separable if |X | = 0 . 59

Exercise 12.11. Let X be a conite topological space where |X | 0 . Show that every nite set F X is nowhere dense in X .

12.2

Countable Sets

Remark 12.12. Recall that a set A is countably innite i a bijection : A Z+ . Evidently, Z+ is itself a countably innite set since iZ+ is a bijection. Lemma 12.13. Every innite subset of Z+ is countably innite. Proof. Let C be an innite subset of Z+ . We dene a map h : Z+ C by induction. Dene h(1) to be the smallest element in C . Clearly, h(1) is uniquely dened since (C, ) is bounded below by 1 Z+ . Then assumming h(2), . . . , h(n 1) are dened, dene h(n) to be the smallest element of C \ h({1, . . . , n 1}). The set C \ h({1, . . . , n 1}) is not empty since if it were, then h : {1, . . . , n 1} C would be a surjection and so C would be nite. Hence, by induction, h is dened for all n Z+ . Thus h is a map from Z+ into C . Supose m = n. Without loss of generality, we may assume m < n. Notice h(m) belongs to the set h({1, . . . , n 1}) whereas h(n), by denition, does not. Hence h(n) = h(m). Thus h is injective. Let c C . Notice that h(Z+ ) cannot be contained in the nite set {1, . . . , c} since h(Z+ ) is innite. So there exists an n Z+ such that h(n) > c. Let m be the smallest element of Z+ such that h(m) c. Then for all i < m, we must have h(i) < c. Thus, c does not belong to the set h({1, . . . , m 1}). Since h(m) is the smallest element in C \ h({1, . . . , m 1}), we must have h(m) c. Hence, h(m) c h(m) and so h(m) = c. Thus, h is surjective. Since h : Z+ C is bijective, h1 : C Z+ is also a bijection. C is countably innite by denition. Theorem 12.14. Let B be a nonempty set. Then the following are equivalent: (1) B is countable. (2) There is a surjection mapping Z+ onto B . (3) There is an injection mapping B into Z+ . Proof. (1) (2) Suppose B is countable. Then B is either countably innite or nite. If B is countably innite, then there exists a bijection f : B Z+ . Hence, f 1 : Z+ B is also a bijection. If B is nite, then there exists a bijection g : B {1, . . . , n} for some integer n 1. Hence, g 1 : {1, . . . , n} B is also a bijection. We can extend g 1 to a surjection h : Z+ B by dening h(i) = g 1 (i) g 1 (1) for 1 i n for i > n

In either case, if B is countable then a surjection mapping Z+ onto B . (2) (3) Let f : Z+ B be a surjection. Dene g : B Z+ by g (b) = min f 1 ({b}). Because f is surjective, b B the bre f 1 ({b}) is not empty. So g is a map from B into Z+ . If b = b , then the bres f 1 ({b}) and f 1 ({b }) are disjoint. So min f 1 ({b}) = min f 1 ({b }). That is, g (b) = g (b ). Hence, g : B Z+ is injective. (3) (1) Let g : B Z+ be injective. Notice g : B rng g dened by g (b) = g (b) is a bijection where rng g Z+ . If rng g is nite, then B is nite. If rng g is innite, then since every innite subset of Z+ is countable, there exists a bijection h mapping rng g into Z+ . So h g : B Z+ is a bijection. Thus, B is countably innite. Hence, in either case, B is countable. 60

Corollary 12.15. A subset of a countable set is countable. Proof. Let B be a countable set and let A B . Then since B is countable, there exists a bijection f : B Z+ . So the restriction of f to A, f |A : A Z+ is injective. Hence, by theorem 12.14, A is countable. Corollary 12.16. Arbitrary intersections of coutable sets are countable. Proof. Let {Ci : i I } be an indexed familiy of countable sets. Then the arbitrary intersection iI Ci is a subset of each Ci . Hence, by corollary 12.15, arbitrary intersections of countable sets are countable. Lemma 12.17. The set Z+ Z+ is countably innite. Proof. Dene a function f : Z+ Z+ Z+ by f (m, n) = 2m 3n . Suppose f (m, n) = f (p, q ) Then 2m 3n = 2p 3q . Assume m = p. Without loss of generality, we may further assume m < p. Then 3n = 2pm 3q , where p m > 0, which contradicts the fact that 3n is odd for all n Z+ . Hence, by contradiction, m = p. As a result, 3n = 3q . Hence n = q . That is, (m, n) = (p, q ). Thus, f : Z+ Z+ Z+ is an injection. So, by theorem 12.14, Z+ Z+ is countable. As it is not nite, it is countably innite. Theorem 12.18. Countable unions of countable sets are countable. Proof. Let {Ci : i I } be an indexed family of countable sets, where the indexing set I is either {1, . . . n} or Z+ . Without loss of generality, we may assume that each set Ci is not empty. Since each Ci is counatble, we can choose, for each i, a surjective function fi : Z+ Ci . Similarly, we can construct a surjective function g : Z+ I . Now dene h : Z+ Z+ iI Ci by h(m, n) = fg(m) (n). Let c iI Ci . Then c Ci for at least one i I . As g : Z+ I is a surjection, Ci = Cg(m) for at least one m Z+ . That is, c iI Ci , m Z+ such that c Cg(m) . Now Cg(m) , a surjection fg(m) : Z+ Cg(m) . So c = fg(m) (n) for at least one n Z+ . That is, c iI Ci , (m, n) Z+ Z+ such that c = h(m, n). Hence, h : Z+ Z+ iI Ci is surjective. Now since Z+ Z+ is countable, there exists a surjection k : Z+ Z+ Z+ . Hence, h k is a surjection mapping Z+ onto iI Ci . iI Ci is countable. Theorem 12.19. Finite products of countable sets are countable. Proof. Let A and B be countable sets. If either A or B is empty, then A B = is countable. Otherwise, surjections f : Z+ A and g : Z+ B . We dene h : Z+ Z+ A B by h(m, n) = (f (m), g (n)). Clearly, h is a surjection. As Z+ Z+ is countable, a surjection : Z+ Z+ Z+ . So h : Z+ A B is also surjective. Hence, A B is countable. That is, the product of any two countable sets is countable. Now we prove by induction that a nite product of countable sets is countable.

61

Let A1 , , An be countable sets. Clearly, A1 is countable. Now assume A1 An1 is countable. Then A1 An = (A1 An1 ) An is countable as it is the product of two countable sets. By induction, nite products of countable sets are countable.

12.3

Introduction to the Cocountable Topology

Denition 12.20. A subset A X is cocountable i X \ A is countable. That is, A contains all but countably many elements of X . Example 12.21. Let = {U X : U = or X \ U is countable}. Then is called the cocountable topology, or countable complement topology, on X . Exercise 12.22. Show that the cocountable topology on a set X satises the denition of a topology on X . Remark 12.23. Notice that if X is countable, then every subset of X is also countable. So the cocountable topology on a countable set X is the same as the discrete topology on X .

13
13.1

Lecture 13
Properties of the Conite and Cocountable Topologies

Example 13.1. Let X be a conite topological space where |X | 0 . Then X is not T2 (Hausdor). Proof. Let x, y X such that x = y . Then open neighborhood U of x, X \ U is nite. Similarly, open neighborhood V of y , X \ V is nite. So (X \ U ) (X \ V ) = X \ (U V ) is nite since nite unions of nite sets are nite. Hence U V is conite and as |X | 0 , U V contains innitely many points of X . In particular, U V = . X is not T2 (Hausdor). Example 13.2. Let X be a conite topological space where |X | 0 . Then every nite set F nowhere dense in X . Proof. As X is T1 (Fr echet), every nite point set is closed. So Cl(F ) = F . Since |X | 0 , every nonempty open set contains innitely many points of X . So Int(F ) = . Hence Int(Cl(F )) = Int(F ) = . F is nowhere dense in X . Example 13.3. Let = {U X : U = or X \ U is countable}. Then is a topology on X (called the cocountable topology). X is

62

Proof. (1) Clearly, . Since X \ X = is countable, X . (2) Let {Ui : i I } be an indexed family of nonempty sets in . Then X \ Ui is countable i I . So iI X \ Ui = X \ iI Ui is countable since arbitrary intersections of countable sets are countable. Hence iI Ui . Clearly, iI Ui . Thus arbitrary unions of sets in are also in . (3) Let U1 , . . . , Un be nonempty sets in . Then X \ Ui is countable i {1, . . . , n}. n So n i=1 X \ Ui = X \ i=1 Ui is countable since countable unions of countable sets are countable. n Hence i=1 Ui . Clearly, n i=1 Ui . Thus nite intersections of sets in are also in . Example 13.4. Every cocountable topological space X is T1 (Fr echet). Proof. Let x X . Since X \ (X \ {x}) = {x} is countable, X \ {x} is open. So {x} is closed. Hence every singleton set is closed. Thus X is T1 (Fr echet). Example 13.5. Let X be a cocountable topological space where |X | > 0 . Then X is not T2 (Hausdor). Proof. Let x, y X such that x = y . Then open neighborhood U of x, X \ U is countable. Similarly, open neighborhood V of y , X \ V is countable. So (X \ U ) (X \ V ) = X \ (U V ) is countable since countable unions of countable sets are countable. Hence U V is cocountable and as |X | > 0 , U V contains uncountably many elements of X . In particular, U V = . X is not T2 (Hausdor). Example 13.6. Let X be a cocountable topological space where |X | > 0 . Then every countable set C X is nowhere dense in X . Proof. Since C = X \ (X \ C ), X \ C is open and hence C is closed. So Cl(C ) = C . As |X | > 0 , every nonempty open set contains uncountably many elements of X . So Int(C ) = . Hence Int(Cl(C )) = Int(C ) = . C is nowhere dense in X . Example 13.7. Let X be a cocounatble topological space where |X | > 0 . Then X is not separable. Proof. Since every countable set C X is nowhere dense, X \ C is dense in X . That is, every nonempty open set is dense in X . However, every dense set is uncountable. So a countable dense subset of X . X is not separable.

63

Example 13.8. Let X be a cocountable topological space where |X | > 0 , let C and let U X be open. Subset C U X Interior U X Closure C X X

X be countable,

Since the only sets that are both opened and closed are and X , X is connected. Example 13.9. Let X be a cocountable topological space where |X | > 0 . Then every point x X is a cluster point of X . Proof. Let x X and let U be an open neighborhood of x. Then U contains all but countably many elements of X as it is cocountable. That is, every open neighborhood U of x contains uncountably many points of X . Since a cocountable topological space is T1 (Fr echet), by therorem 6.30 x is a cluster point of X . Hence every point x X is a cluster point of X . Remark 13.10. Since every point in a cocountable topological space on an uncountable set is a cluster point, there are no isolated points in the space. Remark 13.11. The topological invariants for the conite and cocountable topologies (each distinct from the discrete topology) which we have explored so far are summarized below: Conite Topology 0 Connected T1 Separable if |X | = 0 Cocountable Topology 0 Connected T1 Not Separable

# of Isolated Points Connectedness Separation Condition Countability Condition

13.2

More Countability Conditions

Denition 13.12. Let (X, ) be a topological space and let x X . A neighborhood basis at x is a collection Bx of neighborhoods of x such that neighborhood U of x, B Bx such that x B U . That is, Bx is a neighborhood basis at x i every neighborhood U of x contains at least one set B Bx . Denition 13.13. Let (X, ) be a topological space and let x X . A local basis at x is a neighborhood basis at x all of whose members are open. Remark 13.14. Clearly, every local basis at a point is a neigborhood basis at that point. Some sources make no distinction between a local basis and a neighborhood basis. For our purposes, a neighborhood basis is not a local basis unless all its members are open. Exercise 13.15. Let (X, ) be a topological space, let x X , and let Bx be a neighborhood basis at x. Show Cx = {Int(B ) : B Bx } is a local basis at x. Exercise 13.16. Let B be a basis for a topology on X . Show Bx = {B B : x B } is a local basis at x X . 64

Denition 13.17. Let (X, ) be a topological space. Then (1) X is rst-countable i it has a countable neighborhood basis at each point x X . (2) X is second-countable i it has a countable basis (for ). Remark 13.18. Notice that every II-countable space is I-countable since x X , a local basis Bx at the point x is a subset of the basis. That is, Bx B x X where |B | = 0 . Theorem 13.19. If a topological space is II-countable, then it is separable. Proof. Let X be II-countable. Then it has a countable basis B . We can construct a set D X so that it contains exactly one point from each basis element B B . Then |D| = |B | = 0 . Further, since every nonempty open set U X is a union of some basis elements, D U = . That is, D is a countable dense subset of X . X is separable. Remark 13.20. The converse of theorem 13.19 is not necessarily true. As we will see, the converse is true for a special class of topological spaces called metric spaces. However, the contrapositive is an equivalent statement: If a space is not separable, then it is not IIcountable.

13.3

The Order Topology on Q

a c Denition 13.21. The rational numbers is the set Q = { m n : m, n Z and n = 0} where b , d Q, c a b = d i ad = bc.

Remark 13.22. Notice that the rational numbers are those numbers which can be expressed as a ratio of integers. 2 2 m + Notice 3 = 3 since (2)(3) = 2(3). Indeed, q Q, q = n where m Z and n Z . m Further, Z = { n Q : m Z and n = 1}. So Z Q. Example 13.23. Z is countably innite. Proof. Let f : Z Z+ be dened by f (n) = 2n + 1 2n if n 0 if n < 0

There are two cases to consider: (1) m, n 0 suppose f (m) = f (n). Then 2m + 1 = 2n + 1. So 2m = 2n. Hence, m = n. Thus f is injective over nonnegative integers. (2) m, n < 0 suppose f (m) = f (n). Then 2m = 2n. Hence, m = n. Thus f is injective over negative integers. Therefore f is an injection. Hence, Z is countable. As Z is not nite, it is countably innite. Example 13.24. Q is countably innite.

65

+ Proof. Recall that q Q, q = m n where m Z and n Z . m + Let f : Z Z Q be dened by f (m, n) = n . Then q Q, (m, n) Z Z+ such that q = f (m, n). Hence f : Z Z+ Q is a surjection. As Z Z+ is a nite product of countable sets, it is coutable. So a surjection g : Z+ Z Z+ . Hence, f g : Z+ Q is a surjection. Thus Q is countable. As Q is not nite, it is countably innite.

Example 13.25. (Q, <) is a simply ordered set. Proof. (1) p, q Q such that p = q , either p < q or q < p. Hence < is a total ordering on Q. (2) q Q, q < q . Hence < is irreexive. (3) p, q, r Q, if p < q and q < r, then p < r. Hence < is transitive. Thus < is a simple ordering on Q. (Q, <) is a simply ordered set. Remark 13.26. Clearly, Q has neither a least element nor a greatest element. Further, Q is neither bounded below nor bounded above. So a basis for the order topology on Q is B = {(a, b) : a, b Q}. Example 13.27. The space Q, with the order topology, is II-countable. Proof. A basis for the order topology on Q is B = {(a, b) : a, b Q}. Let f : Q Q B be dened by f (a, b) = (a, b). That is, the ordered pair (a, b) Q Q is mapped onto the open interval (a, b) B . Clearly, f is a surjection. As Q Q is a a nite product of countable sets, it is countable. So a surjection g : Z+ Q Q. Hence f g : Z+ B is a surjection. Thus B is countable. Q is II-countable. Remark 13.28. Notice that the space Q with the order topology is distinct from the discrete space Q since discrete spaces are not separable and therefore not II-countable. Remark 13.29. Recall that every space X with the order topology is T2 (Hausdor). So Q is T2 (Hausdor) in the order topology. Exercise 13.30. Consider Q in the order topology. Determine whether Q is connected and determine the number of isolated points in Q. Remark 13.31. The standard topology on Q is the order topology. Similarly, the standard topology on R is the order topology. The space R with the order topology is called the real line. Our model of the real line comes from Euclidean geometry. Until irrational numbers were discovered, the Greeks originally though that every point on a line could be expressed as a rational number. 66

Lemma 13.32. Let a, b Z. (1) If a is even, then a2 is even. (2) If b is odd, then b2 is odd. Proof. (1) Suppose a is even. Then a = 2m for some m Z. So a2 = (2m)2 = 4m2 = 2(2m2 ) where 2m2 Z. Hence a2 is even. (2) Suppose b is odd. Then b = 2n + 1 for some n Z. So b2 = (2n + 1)2 = 4n2 + 4n + 1 = 2(2n2 + 2n) + 1 where 2n2 + 2n Z. Hence b2 is odd. Corollary 13.33. Let a Z, then a is even i a2 is even. Proof. Suppose a is even. Then by lemma 13.32, a2 is even. Conversely, suppose a2 is even and assume to the contrary that a is not even. Then a is odd. So by lemma 13.32, a2 is odd. Hence, by contradiction, a is even. Remark 13.34. Notice that q Q, q = m n for some m, n Z where m and n have no common factors. That is, every rational number can be uniquely represented by a reduced fraction. Theorem 13.35. q Q such that q 2 = 2. Proof. Suppose to the contrary that q Q such that q 2 = 2. Then q = m n for some m, n Z where m and n have no common factors. So ( m 2 ) = 2 n m 2 = 2 n2

Hence m2 is even. Thus m is even. That is, m = 2p for some p Z. So (2p)2 = 2n2 4p2 = 2n2 2p2 = n2 Hence n2 is even. Thus n is even. But then m and n have a common factor of 2. by contradiction, q Q such that q 2 = 2.

67

14
14.1

Lecture 14
More on Local Bases

Theorem 14.1. Let (X, ) be a topological space, let x X , and let Bx be a neighborhood basis at x. Then Cx = {Int(B ) : B Bx } is a local basis at x. Proof. As Bx is a neighborhood basis at x, neighborhood U of x B Bx such that x B U . Further, as each B Bx is a neighborhood of x, x Int(B ). That is, B Bx , x Int(B ) B . So neighborhood U of x, Int(B ) {Int(B ) : B Bx } such that x Int(B ) B U . Hence Cx = {Int(B ) : B Bx } is a local basis at x all of whose members are open. Thus Cx is a local basis at x. Theorem 14.2. Let B be a basis for a topology on X and let x X . Then Bx = {B B : x B } is a local basis at x. Proof. Let U be a neighborhood of x X . Then x Int(U ) U . As B is a basis for the topology on X , Int(U ) = j J Bj where Bj B . Further, Bj j J Bj , Bj Int(U ). As x Int(U ), x Bj for at least one j J . Let Bx = {B B : x B }. Then neighborhood U of x, B Bx such that x B Int(U ) U . Hence Bx is a neighborhood basis at x all of whose members are open. Thus Bx = {B B : x B } is a local basis at x. Remark 14.3. Theorem 14.2 gives us a way to construct a local basis at a point for any basis. We can also do the opposite. That is, given local bases for every point in a space X , we can construct a basis for the topology on X . Theorem 14.4. Let (X, ) be a topological space, let x X , and let Bx be a local basis at x. Then B = xX Bx is a basis for . Proof. Notice B = xX Bx = {B Bx : x X }. As Bx is a local basis at x, each B Bx is open. Hence B is a collection of open sets. Let U and let x U . Then B Bx such that x B U . That is, U and x U B B such that x B U . B is a basis for .

14.2

The Order Topology on Q and R

Example 14.5. Consider Q in the order topology. Let x Q and let U be an open neighborhood of x.

68

Then (a, b) B , where B is the basis for the order topology on Q, such that x (a, b) U . Notice that (a, b) Q \ {x} = (a, b) [(, x) (x, )] = [(a, b) (, x)] [(a, b) (x, )] = (a, x) (x, b)
a x+ a Since x, a Q, x+ 2 Q such that 2 (a, x). Hence (a, b) Q \ {x} = . So open neighborhood U of x, U Q \ {x} = . Hence x Q is a cluster point of Q. Thus there are no isolated points in Q. Now since Q is T2 (Hausdor), it is also T1 (Fr echet). As it is also true that x Q is a cluster point, by theorem 6.30 every nonempty open set contains innitely many points of Q. Further, every nite point set F is closed. So Int(Cl(F )) = Int(F ) = since there are no nonempty open sets that can be contained in a nite set. Hence F is nowhere dense in Q.

Theorem 14.6. Let a, b R where a < b. Then q Q such that a < q < b.
1 Proof. Since a < b, b a > 0. So b a > 0. 1 + As Z is not bounded above, n Z+ such that b a < n. Since b a > 0, 1 < bn an. Hence an + 1 < bn. Let m be the smallest integer such that an < m. If an Z, then an = m 1. If an / Z, then an > m 1. So an m 1. Hence an + 1 m. Thus, an < m an + 1 < bn. That is, an < m < bn. m Since n Z+ , a < m n < b where n Q. a, b R where a < b, q Q such that a < q < b.

Remark 14.7. Recall that our model for the real numbers is the set of all numbers which can be represented as points on the Euclidean line. In the previous lecture, we identied a point R where 2 = 2. So we can denote this number by = 2. By theorem 13.35, 2 is not a rational number. Example 14.8. Let A = {x Q : x < 2}. Then Q \ A = {x Q : x > 2}. Clearly, A Q \ A = Q and A Q \ A = . Hence {A, Q \ A} is a partition of Q. Let x A that a < x. . Then as Q is not bounded below, a Q such Since x, 2 R where x < 2, b Q such that x < b < 2. Hence x (a, b) A where (a, b) is a basic open set. That is, x A (a, b) B such that x (a, b) A. Hence A is open. Now let x Q \ A. Then since 2, x R where 2 < x, c Q such that 2 < c < x. As Q is not bounded above, d Q such that x < d. 69

Hence x (c, d) Q \ A where (c, d) is a basic open set. That is, x Q \ A (c, d) B such that x (c, d) Q \ A. Hence Q \ A is open. So {A, Q \ A} is a separation of Q. Q is not connected (i.e. Q is disconnected). Example 14.9. Notice Z Q \ Z = Q and Z Q \ Z = . Hence {Z, Q \ Z} is a partition of Z. Let x Z. Then as Z is unbounded, a, b Z such that a < x < b. Hence x (a, b) where (a, b) is a basic open set. However, since a, b R where a < b, q Q such that a < q < b. Further, since every nonempty open set contains innitely many points of Q, (a, b) contains innitely many rational numbers. So (a, b) B such that (a, b) Z. Hence the only open set that is contained in Z is . Thus Int(Z) = . Now let x Q \ Z and let m be the smallest integer such that x < m. Then since x / Z, x > m 1. So m 1 < x < m. Since m 1, x R such that m 1 < x, c Q such that m 1 < c < x and since x, m R such that x < m, d Q such that x < d < m. Hence x (c, d) Q \ Z where (c, d) is a basic open set. That is, x Q \ Z (c, d) B such that x (c, d) Q \ Z. Hence Q \ Z is open. So Z is closed. Thus Int(Cl(Z)) = Int(Z) = . Z is nowhere dense in Q. Remark 14.10. The topological invariants which we have explored so far are summarized below for the standard topology (i.e. order topology) on Q: # of Isolated Points Connectedness Separation Condition Countability Condition 0 Disconnected T2 II-Countable

Remark 14.11. Since Q is II-countable, it is also I-countable and separable. So the strongest countability condition satised by Q is the II-countable condition. We can demonstrate directly that Q is both I-countable and separable. Example 14.12. Recall Q is II-countable since the basis B = {(a, b) : a, b Q} is countable. Let x Q. Then Bx = {(a, b) B : x (a, b)} is a local basis at x. Further, x Q, Bx B . Hence x Q, Bx is countable. Thus Q is I-countable. Now since every nite point set F and Z are nowhere dense in Q, Q \ F and Q \ Z are dense in Q. Further, Q \ F and Q \ Z are countable since |Q| = 0 . Hence Q is separable. 70

Denition 14.13. A real number that is not rational is called irrational. We denote the set of all irrational numbers by R \ Q. Remark 14.14. In our construction of a separation of Q , { A, Q \ A } where A = { x Q : x < 2}, we can replace 2 by any R \ Q and {A, Q \ A} is also a separation of Q. Lemma 14.15. Let q Q \ {0} and let R \ Q. Then q R \ Q. Proof. Suppose to the contrary that q Q. Since q Q \ {0}, a Z \ {0} and b Z+ such that q = a b. Since q = 0 and = 0, q = 0. c So c Z \ {0} and d Z+ such that q = d . Hence a ( ) = b = where bc, ad Z \ {0}. That is, Q \ {0}. By contradiction, q R \ Q. Corollary 14.16. Let a, b R where a < b. Then R \ Q such that a < < b. b a Proof. Since a < b and 2 > 0, < . 2 2
b a <q< . By theorem 14.6, q Q such that 2 2 Without loss of generality, we may assume that q = 0. So a < q 2 < b where, by lemma 14.15, q 2 R \ Q. a, b R where a < b R \ Q such that a < < b.

c d bc ad

Remark 14.17. When we use the term the real line, or in the absence of a stated topology, R is the space with its standard topology (i.e. the order topology). Example 14.18. Recall that the standard basis for R is B = {(a, b) : a, b R}. Let U be a nonempty open set. Then x U (a, b) B such that x (a, b) U . Since a, x, b R where a < x < b, q Q such that a < q < x < b. Hence q (a, b) U . So nonempty open set U R, U Q = . Hence Q is dense in R. Further, as |Q| = 0 , R is separable. Now let x R. Then open neighborhood V of x, then q = x Q such that {x, q } V . Hence {x, q } R \ {x} V R \ {x} {q } V R \ {x} That is, x R and open neighborhood V of x, V R \ {x} = . Hence every x R is a cluster point of R Thus there are no isolated points in R. 71

Remark 14.19. Let R \ Q and let A = Q (, ). Notice that is an upper bound for A. In fact, the set of all upper bounds of A is [, ). So the least upper bound of A, Sup A, is . Notice that Sup A = / Q. If we consider the entire open ray (, ), then once again the set of all upper bounds for this open ray is [, ). In this case, Sup (, ) = R. That is, R has a property that Q does not: least upper bounds always occur in the set. While it is reasonable to assume that any nonempty subset of R that is bounded above has a least upper bound in R, we can not prove this. Instead, this condition is quaranteed axiomatically. Axiom 14.20. (Completeness Axiom) Every nonempty subset A R that is bounded above has a least upper bound in R. That is, Sup A exists and Sup A R. Theorem 14.21. R is connected. Proof. Suppose to the contrary that R is disconnected. Then there exists a separation {U, V } of R. That is, U and V are nonempty proper open subsets of R such that U V = R and U V = . Let a U and let b V . Then either a < b or b < a. Without loss of generality, we may assume a < b. Now let A = {x R : [a, x] U }. Notice a A and b is an upper bound for A since otherwise [a, b] U and hence U V = . That is, A is a nonempty subset of R that is bounded above. Hence, by the completeness axiom, Sup A exists and Sup A R. Now Sup A / U since otherwise (a, ) B such that Sup A (a, ) U where < b. That is, a < Sup A < < b and hence p Q such that a < Sup A < p < < b so that [a, p] U which contradicts the fact that Sup A is an upper bound of A. Similarly, Sup A / V since otherwise (, c) B such that Sup A (, c) V where c > b since Sup A b. That is, < Sup A b < c and hence q Q such that < q < Sup A b < c so that q is an upper bound of A which contradicts the fact that Sup A is the least upper bound of A. So Sup A R such that Sup A / U and Sup A / V. Hence U V = R. By contradiction, a separation of R and hence R is connected. Remark 14.22. So the real line is connected if the completeness axiom is true. You may well ask how it is that we know it is true. The answer is that we construct R from Q in such a way that the completeness axiom is satised. Such a construction is called a completion of Q. Example 14.23. The standard basis for R is B = {(a, b) : a, b R}. Let Br = {(p, q ) : p, q Q}. Notice (p, q ) is an open interval determined by rational numbers but that it contains all x R such that p < x < q . Notice further that Br contains the same number of intervals as the standard basis for Q. That is, Br is a countable collection of open sets. Let U be a nonempty open subset of R. 72

Then x U (a, b) B such that x (a, b) U . Since a < x < b where a, b R, p, q Q such that a < p < x < q < b. That is, open subset U R and x U (p, q ) Br such that x (p, q ) U . Hence Br = {(p, q ) : p, q Q} is a basis for R. Thus R is II-countable since Br is a countable basis for R. Remark 14.24. The topological invariants which we have explored so far are summarized below for the standard topology (i.e. order topology) on R: # of Isolated Points Connectedness Separation Condition Countability Condition 0 Connected T2 II-Countable

14.3

Introduction to the Subspace Topology

Denition 14.25. Let (X, X ) be a topological space and let Y X . Then Y = {Y U : U X } is called the subspace topology on Y . The space (Y, Y ) is called a subspace of X . Remark 14.26. Let Y be a subspace of X . We say that Y has the relative topology inherited from the topology on X . We also say that the topology on Y is induced from the ambient topology on X . Exercise 14.27. Show that the subspace topology saties the denition of a topology. Exercise 14.28. Show that the order topology on Q is the same as the topology Q inherits as a subspace of R. Exercise 14.29. Let Y = [0, 1) {2} and let Y R. Show that {2} is open in the subspace topology on Y but is not open in the order topology on Y .

15
15.1

Lecture 15
Properties of the Subspace Topology

Lemma 15.1. Let {Ui : i I } be an indexed family of sets and let Y be a set. Then (1) iI (Ui Y ) = ( iI Ui ) Y (2) iI (Ui Y ) = ( iI Ui ) Y Proof. (1) Let x iI (Ui Y ). Then i I such that x Ui Y . That is, x Ui and x Y . As x Ui , x iI Ui . So x ( iI Ui ) Y . Hence iI (Ui Y ) ( iI Ui ) Y . Conversely, let x ( iI Ui ) Y . Then x iI Ui and x Y . As x iI Ui , i I such that x Ui . So x Ui Y . 73

As x Ui Y , x iI (Ui Y ). Hence ( iI Ui ) Y iI (Ui Y ). iI (Ui Y ) = ( iI Ui ) Y . (2) Let x iI (Ui Y ). Then x Ui Y i I . That is, x Ui i I and x Y . As x Ui i I , x iI Ui . So x ( iI Ui ) Y . Hence iI (Ui Y ) ( iI Ui ) Y . Conversely, let x ( iI Ui ) Y . Then x iI Ui and x Y . As x iI Ui , x Ui i I . That is, i I , x Ui Y . So x iI (Ui Y ). Hence iI (Ui Y ) iI (Ui Y ). iI (Ui Y ) = ( iI Ui ) Y . Example 15.2. Let (X, X ) be a topological space and let Y Y . Then Y = {U Y : U X } is a topology on Y (called the subspace topology). Proof. (1) Since X , Y = Y . Similarly, as X X , X Y = Y Y . (2) Let {Vi : i I } be an indexed family of sets such that Vi Y i I . Then i I , Vi = Ui Y where Ui X . So ii Vi = iI (Ui Y ). Since iI (Ui Y ) = ( iI Ui ) Y and iI Ui X , iI Vi Y . (3) Let V1 , . . . , Vn Y . Then i {1, . . . , n}, Vi = Ui Y where Ui X . n So n 1 (Ui Y ). 1 Vi = n n n Since 1 (Ui Y ) = ( n 1 vi Y . 1 Ui ) Y and 1 Ui X , Y is a topology on Y by denition. Theorem 15.3. Let Y be a subspace of X and let B be a basis for the topology on X . Then BY = {B Y : B B } is a basis for the subspace topology on Y . Proof. Let V be open in Y and let y V . Then y V = U Y where U is open in X . That is, y U and y Y . As y U and U is open in X , B B such that y B U . As y Y , y B Y U Y = V where B Y BY . That is, open set V in Y and y V , BY BY such that y BY V . Hence BY is a basis for the subspace topology on Y . Example 15.4. Consider R in the order topology. A basis for R is B = {(p, q ) R : p, q Q}. So a basis for the subspace topology on Q is BQ = {B Q : B B } = {(p, q ) Q : p, q Q} 74

which is the standard basis for the order topology on Q. That is, the subspace topology and the order topology on Q are generated by the same basis. Hence the two topologies are equal. Example 15.5. Consider R in the order topology and let Y = [0, 1) {2} R. Since (1, 3) is open in R, (1, 3) Y = {2} is open in the subspace topology on Y . A basis B for the order topology on Y consists of all intervals of the form [0, b), (a, b), and (a, 2] where a, b Y . Notice 2 / [0, b) and 2 / (a, b) since 2 = max Y . Further, (a, 2] = {y Y : 0 a < y 2} = (a, 1) {2} where (a, 1) contains innitely many elements of Y . Since B B such that B {2}, {2} is not open in the order topology on Y . Lemma 15.6. Let Y be a subspace of X and let Z Y . Then (1) ClY (Z ) ClX (X ) (2) IntX (Z ) IntY (Z ) Proof. (1) Let y ClY (Z ) and let U be an open neighborhood of y in X . As y U and y Y , y U Y . Further, U Y is open in Y as U is open in X . That is, U Y is an open neighborhood of y in Y . So (U Y ) Z = since y Cly (Z ). Now (U Y ) Z = U (Y Z ) = U Z . So open neighborhood U of y in X , U Z = . Hence y ClX (Z ). Thus ClY (Z ) ClX (Z ). (2) Let x IntX (Z ). Then an open neighborhood U of x in X such that U Z . Since IntX (Z ) Z Y , U Y Z Y = Z . That is, U Y is an open neighborhoof of x in Y such that U Y Z . Hence x IntY (Z ). Thus IntX (Z ) IntY (Z ). Example 15.7. Let Y = [0, 1) {2} be a subspace of R (with the order topology). Since {2} is open in Y , Y \ {2} = [0, 1) is closed in Y . So ClY ([0, 1)) = [0, 1) while ClR ([0, 1)) = [0, 1]. Hence ClY ([0, 1)) ClR ([0, 1)). Now IntY ({2}) = {2} while IntR ({2}) = . Hence IntR ({2}) ClY ({2}). Theorem 15.8. Let Y be a subspace of X and let Z Y . Then (1) Z is open in Y i z = U Y with U open in X . (2) Z is closed in Y i Z = C Y with C closed in X . (3) ClX (Z ) Y = ClY (Z ). (4) IntX (Z ) Y IntY (Z ) with equality only if IntY (Z ) IntX (Z ). 75

Proof. (1) Trivial. This is just a restatement of the denition of the subspace topology. (2) Suppose Z is closed in Y . Then Y \ Z is open in Y . So Y \ Z = U Z with U open in X . Hence Z = Y \ (U Y ) = (Y \ U ) (Y \ Y ) = Y \U Now since Y X and U X , Y \ U = Y (X \ U ) where X \ U is closed in X . That is, Z = (X \ U ) Y with X \ U closed in X . Conversely, suppose Z = C Y with C closed in X . Then Y \ Z = Y \ (C Y ) = (Y \ C ) (Y \ Y ) = Y \C Now since Y X and C X , Y \ C = Y (X \ C ) where X \ C is open in X . Hence Y \ Z is open in Y . Thus Z is closed in Y . (3) Since ClY (Z ) ClX (Z ), ClY (Z ) Y ClX (Z ) Y ClY (Z ) ClX (Z ) Y Now let {Fi : i I } be the collection of all closed sets in Y containing Z . Then ClY (Z ) = iI Fi . Further, by (2) Fi Ci closed in X such that Fi = Ci Y . So ClY (Z ) = iI (Ci Y ) = ( iI Ci ) Y . Now i I , Z Ci Y Ci . So iI Ci is a closed set in X containing Z . Hence ClX (Z ) iI Ci . So ClX (Z ) Y ( iI Ci ) Y = ClY (Z ). ClX (Z ) Y = ClY (Z ). (4) Since IntX (Z ) IntY (Z ), IntX (Z ) Y IntX (Z ) Y Now suppose IntY (Z ) IntX (Z ). Then IntY (Z ) Y IntX (Z ) Y IntY (Z ) Y IntY (Z )

IntY (Z ) IntX (Z ) Y So in this case, IntX (Z ) Y = IntY (Z ). Exercise 15.9. Let Y be a subspace of X , let x Y , and let Bx be a neighborhood basis for x in the topology on X . Show {B Y : B Bx } is a neighborhood basis for x in the subspace topology on Y . 76

Remark 15.10. A set that is open in a subspace is not necessarily open in the ambient space. Similarly, a set that is closed in a subspace is not necessarily closed in the ambient space. Corollary 15.11. Let Y be a subspace of X and let Z Y . (1) If Z is open in Y and Y is open in X , then Z is open in X . (2) If Z is closed in Y and Y is closed in X , then Z is closed in X . Proof. (1) As Z is open in Y , Z = U Y where U is open in X . So if Y is open in X , then U Y = Z is open in X . (2) As Z is closed in Y , Z = C Y where C is closed in X . So if Y is closed in X , then C Y = Z is closed in X . Exercise 15.12. Let Y be a subspace of X and let Z Y . Show that the subspace topology on Z relative to Y is the same as the subspace topology on Z relative to X . In other words, show that if Y is a subspace of X and Z is a subspace of Y , then Z is a subspace of X .

15.2

Convex Subsets and Subbases

Remark 15.13. Let (X, <) be a simply ordered set and let Y X . The simple order relation on X , when restricted to Y , makes Y a simply ordered set. However, the resulting order topology on Y is not necesarily the same as the topology Y inherits as a subspace of X . Denition 15.14. Let (X, <) be a simply ordered set. A subset Y X is convex in X i for each pair of points a, b Y , with a < b, the entire interval (a, b) of points in X lies in Y . Example 15.15. Let Y = [0, 1) {2} R. Then Y is not a convex subset of R since 0, 2 Y but (0, 2) Y . Further, Y is a subset of R for which the order topology on Y is not the same as the topology Y inherits as a subspace of R. Remark 15.16. Notice that intervals and rays of a simply ordered set X are convex in X . Denition 15.17. Let (X, ) be a topological space and let B be a basis for . A subbasis S for is a collection of subsets of X such that each B B is a nite intersection of elements in S . That is, B B, B = n i=1 Si where Si S i {1, . . . , n}. Remark 15.18. A subbasis is sometimes called a subbase. Example 15.19. Let (X, <) be a simply ordered set. Then the collection S consisting of all open rays of the form (, b) and (a, ) where a, b X is a subbasis for the order topology on X . Proof. Let B be the standard basis for the order topology on X . Then (a, b) B , (a, b) = (, b) (a, ). If b0 = max X , then (a, ) = (a, b0 ]. Similarly, if a0 = min X , then (, b) = [a0 , b). That is, every B B is a nite intersection of elements in S . Hence S is a subbasis for the order topology on X . Denition 15.20. Let X be a set and let S = {Si : i I } be a collection of subsets of X such that X = iI Si . The topology generated by S is the collection of unions of nite intersections of elements in S . 77

Proof. Let B = { n j =1 Sj : Sj S}. Then x X = iI Si i = j I such that x Sj and hence B B such that x B . Now let B1 , B2 B . Then B1 = S1 Sm where Sj S and B2 = S1 Sn where Sj S . So B1 B2 = (S1 Sn ) (S1 Sn ) is a nite intersection of elements in S . Hence B1 B2 B . So if x B1 and x B2 , then B3 B such that x B3 B1 B2 . Thus B satises the conditions for a topology generated by a basis where the open sets in this topology are unions of some elements of B . This is precisely the topology generated by S . Corollary 15.21. Let Y be a subspace of X and let S be a subbasis for the topology on X . Then SY = {Si Y : Si S} is a subbasis for the subspace topology on Y . Proof. As S is a subbasis for the topology on X , B = { X . So BY
n j =1 Sj

: Sj S} is a basis for the topology on

= {B Y : B B }
n

= {(
j =1

SJ ) Y : Sj S}

is a basis for the subspace topology on Y . n n Since ( n j =1 (Sj Y ), BY = { j =1 (Sj Y ) : (Sj Y ) SY }. j =1 Sj ) Y = Hence SY is a subbasis for the subspace topology on Y . Example 15.22. Let (X, <) be a simply ordered set and let Y X be convex in X . Then for every open ray of the form (a, ) X , Y (a, ) = {x X : x Y and x > a}. If a Y , then Y (a, ) is an open ray in (Y, <). If a / Y , then a is either a lower bound of Y or an upper bound of Y since otherwise b, c Y such that (b, c) Y which contradicts the fact that Y is convex in X . If a is a lower bound of Y , then Y (a, ) = Y . If a is an upper bound of Y , then Y (a, ) = . Similarly, for every open ray of the form (, b) X , Y (, b) is exactly one of the following: an open ray in (Y, <), the entire set Y , or . Exercise 15.23. Let (X, <) be a simply ordered set with the order topology and let Y X be convex in X . Show that the order topology on Y is the same as the topology Y inherits as a subspace of X .

16
16.1

Lecture 16
Inherited Properties of Subspaces

Corollary 16.1. Let Y be a subspace of X , let x Y , and let Bx be a neighborhood basis at x in the topology on X . Then Cx = {B Y : B Bx } is a neighborhood basis at x in the subspace topology on Y. Proof. Let V be a neighborhood of x in Y . Then x IntY (V ) V . 78

As IntY (V ) is open in Y , IntY (V ) = Y U where U is open in X . That is, U is an open neighborhood of x in X . As Bx is a neighborhood basis at x in X , B Bx such that x B U . So x Y B Y U = IntY (V ) V . That is, neighborhood V of x in Y , B Y Cx such that x B Y V . Hence Cx is a neighborhood basis at x in Y . Theorem 16.2. Let Y be a subspace of X and let Z Y . Then the subspace topology on Z relative to Y is the same as the subspace topology on Z relative to X . Proof. Let Y be the subspace topology on Y and let X be the topology on X . Then the subspace topology on Z relative to Y is Z = {Z V : V Y }. Now V Y , V = Y U where U X . So Z = {Z (Y U ) : U X } = {(Z Y ) U ) : U X } = {Z U : U X } which is the subspace topology on Z relative to X . Theorem 16.3. Let (X, <) be a simply ordered set with the order topology and let Y X be convex in X . Then the order topology on Y is the same as the topology Y inherits as a subspace of X . Proof. A subbasis for the order topology on X is the collection S = {Si : i I } consisting of the open rays in (X, <). So SY = {Y Si : Si S} is a subbasis for the subspace topology on Y where each Y Si SY is exactly one of the following: an open ray in (Y, <), the entire set Y , or . As each Y Si SY is open in the order topology on Y , unions of nite intersections of elements in SY are open in the order topology on Y . That is, every open set in the subspace topology on Y is open in the order topology on Y . Now a subbasis for the order topology on Y consists of the open rays in (Y, <). Since each open ray in (Y, <) is in SY , unions of nite intersections of the open rays in (Y, <) are open in the subspace topology on Y . That is, every open set in the order topology on Y is open in the subspace topology on Y . The order topology on Y is the same as the subspace topology on Y . Remark 16.4. Notice that Y being convex in X is a sucient but not necessary condition for the subspace topology on Y to be the same as the order topology on Y . For example, Q is not convex in R but the subspace topology on Q is the same as the order topology on Q. Remark 16.5. In general, a subspace of a connected space need not be connected. For example, R is connected but the subspace Q is disconnected. Remark 16.6. In general, the number of isolated points in a subspace is not the same as the number of isolated points in the ambient space. For example, R has no isolated points but the subspace Y = [0, 1) {2} has one isolated point. However, if Y is a subspace of X and a point x Y is an isolated point of X , then x is an isolated point of the subspace Y .

79

Proof. Suppose x Y is an isolated point of X . Then {x} is open in X . So Y {x} = {x} is open in Y . Hence x is an isolated point of Y . Theorem 16.7. A subspace of a Hausdor space is Hausdor. Proof. Let Y be a subspace of a Hausdor space X and let x, y Y such that x = y . Then open neighborhoods U of x and V of y in X such that U V = . So U Y is an open neighborhood of x in Y and V Y is an open neighborhood of y in Y where (U Y ) (V Y ) = (U V ) Y = (U V ) Y = Y = Hence Y is Hausdor. Exercise 16.8. Let Y be a subspace of X . Prove (1) If X is T1 (Fr echet), then Y is T1 (Fr echet). (2) If X is T0 (Kolmogorov), then Y is T0 (Kolmogorov). Theorem 16.9. A subspace of a II-countbale space is II-countable. Proof. Suppose Y is a subspace of a II-countable subspace X . Then a countable basis B for the topology on X . So BY = {Y B : B B } is a countable basis for the subspace topology on Y . Hence Y is II-countable. Exercise 16.10. Prove that a subspace of a I-countable space is I-countable. Remark 16.11. In general, a subspace of a separable space need not be separable. However, every open subspace of a separable space is separable. Theorem 16.12. Let Y be an open subspace of a separable space X . Then Y is separable. Proof. As X is separable, a countable dense subset D X . So every nonempty open subset U in X intersects D. As Y is open in X , Y D = . Further, every open set V in Y is open in X . So V (Y D) = (V Y ) D = V D = Hence Y D is dense in Y . Further, Y D is countable since D is countable. Thus Y D is a countable dense subset of Y . Y is separable. 80

16.2

Functions Acting on Sets

Remark 16.13. Let f : X Y , let A X , and let B Y . (1) Recall f (A) = {f (x) Y : x A}. So y f (A) i x A such that y = f (x). (2) Recall f 1 (B ) = {x X : f (x) B }. So x f 1 (B ) i f (x) B . Theorem 16.14. Let f : X Y , let A B X , and let {Gi : i I } be an indexed family of subsets of X . Then (1) f () = . (2) f (A) f (B ). (3) f ( iI Gi ) = iI f (Gi ). (4) f ( iI Gi ) iI f (Gi ). Proof. (1) Trivial. There are no points in Y that are mapped onto from points in the empty set. (2) Let y f (A). Then x A such that y = f (x). Since A B , x B . That is, x B such that y = f (x). Hence y f (B ). Thus f (A) f (B ). (3) Let y f ( iI Gi ). Then x iI Gi such that y = f (x). So x Gi for at least one index i I . Hence y = f (x) f (Gi ) for at least one i I . So y iI f (Gi ). Thus f ( iI Gi ) iI f (Gi ). Conversely, let y iI f (Gi ). Then y f (Gi ) for at least one index i I . So x Gi such that y = f (x). As Gi iI Gi , x iI Gi . That is, x iI Gi such that y = f (x). Hence y f ( iI Gi ). Thus iI f (Gi ) f ( iI Gi ). (4) Let y f ( iI Gi ). Then x iI Gi such that y = f (x). As x iI Gi , x Gi i I . So y = f (x) f (Gi ) i I . Hence y iI f (Gi ). Thus f ( iI Gi ) iI f (Gi ). Example 16.15. Let f : R R be dened by f (x) = x2 . Then f ([0, )) = [0, ) and f ((, 0]) = [0, ). So f ((, 0]) f ([0, )) = [0, ). However, (, 0] [0, ) = {0} and so f ((, 0] [0, )) = {0}. That is, f ((, 0] [0, )) f ((, 0]) f ([0, )). Remark 16.16. In general, a direct image preserves inclusions and unions. 81

Theorem 16.17. Let f : X Y , let A B Y , and let {Gi : i I } be an indexed family of subsets of Y . Then (1) f 1 () = . (2) f 1 (A) f 1 (B ). (3) f 1 ( iI Gi ) = iI f 1 (Gi ). (4) f 1 ( iI Gi ) = iI f 1 (Gi ). (5) f 1 (Y \ A) = X \ f 1 (A). Proof. (1) Trivial. There are no points in X whose image lies in the emptyset. (2) Let x f 1 (A). Then f (x) A. As A B , f (x) B . Hence x f 1 (B ). Thus f 1 (A) f 1 (B ). (3) Let x f 1 ( iI Gi ). Then f (x) iI Gi . So f (x) Gi for at least one index i I . Hence x f 1 (Gi ) for at least one i I . So x iI f 1 (Gi ). Thus f 1 ( iI Gi ) iI f 1 (Gi ). Conversely, let x iI f 1 (Gi ). Then x f 1 (Gi ) for at least one index i I . So f (x) Gi for at least one i I . Hence f (x) iI Gi . So x f 1 ( iI Gi ). Thus iI f 1 (Gi ) f 1 ( iI Gi ). (4) Let x f 1 ( iI Gi ). Then f (x) iI Gi . So f (x) Gi i I . Hence x f 1 (Gi ) i I . So x iI f 1 (Gi ). Thus f 1 ( iI Gi ) iI f 1 (Gi ). Conversely, let x iI f 1 (Gi ). Then x f 1 (Gi ) i I . So f (x) Gi i I . Hence f (x) iI Gi . So x f 1 ( iI Gi ). Thus iI f 1 (Gi ) f 1 ( iI Gi ). (5) Let x f 1 (Y \ A). Then f (x) Y \ A. So f (x) / A. 1 Hence x / f (A). So x X \ f 1 (A). 1 Thus f (Y \ A) X \ f 1 (A). Conversely, let x X \ f 1 (A). Then x / f 1 (A). So f (x) / A. Hence f (x) Y \ A. So x f 1 (Y \ A). Thus X \ f 1 (A) f 1 (Y \ A). Remark 16.18. In general, an inverse image preserves inclusions, unions, intersections, and comple82

ments.

16.3

Introduction to Continuity

Denition 16.19. Let X and Y be topological spaces. Then f : X Y is continuous at a point x X i open neighborhood V of f (x) in Y an open neighborhood U of x in X such that U f 1 (V ). If f : X Y is is continuous at every point x X , then f is continuous. Theorem 16.20. Let X and Y be topological spaces. Then f : X Y is continuous i open set V in Y , f 1 (V ) is open in X . Proof. Let f be continuous, let V be an open set in Y , and let x f 1 (V ). Then f (x) V . That is, V is an open neighborhood of f (x) in Y . So an open neighborhood U of x in X such that U f 1 (V ). Hence x IntX (f 1 (V )). So f 1 (V ) IntX (f 1 (V )). Thus f 1 (V ) is open in X . Conversely, suppose that open set V in Y , f 1 (V ) is open in X . Let f (x) V . Then V is an open neighborhood of f (x). As f (x) V , x f 1 (V ) where f 1 (V ) is an open neighborhood of x. That is, open neighborhood V of f (x) in Y an open neighborhood U of x in X such that U f 1 (V ). Hence, f : X Y is continuous. Corollary 16.21. Let X and Y be topological spaces and let f : X Y be continuous. If Z is a subspace of X , then the restriction of f to Z is continuous. Proof. Let = f |Z and let V be open in Y . Notice Z f 1 (V ) = Z {x X : f (x) V } = {x Z : f (x) V } = 1 (V ) As f is continuous, f 1 (V ) is open in X . So Z f 1 (V ) = 1 (V ) is open in Z . f |Z : Z Y is continuous. Lemma 16.22. Let (X, X ) be a topological space, let Y X , and let f : Y X be dened by f (y ) = y . Then f 1 (X ) = {f 1 (U ) : U X } is the subspace topology on Y . Proof. Let U X . Then f 1 (U ) = {y Y : f (y ) U } = {y Y : y U } = {x X : x Y and x U } = Y U So f 1 (X ) = {Y U : U X } which is the subspace topology on Y . Remark 16.23. Notice that if Y is a subspace of X , then the function f : Y X dened by f (y ) = y is continuous. 83