A First Course
Textbook in Problems
O. Y. Viro, O. A. Ivanov,
N. Y. Netsvetaev, V. M. Kharlamov
This book includes basic material on general topology, introduces
algebraic topology via the fundamental group and covering spaces,
and provides a background on topological and smooth manifolds. It
is written mainly for students with a limited experience in mathe
matics, but determined to study the subject actively. The material
is presented in a concise form, proofs are omitted. Theorems, how
ever, are formulated in detail, and the reader is expected to treat
them as problems.
Foreword
Genre, Contents and Style of the Book
The core of the book is the material usually included in the Topol
ogy part of the two year Geometry lecture course at the Mathematical
Department of St. Petersburg University. It was composed by Vladimir
Abramovich Rokhlin in the sixties and has almost not changed since
then.
We believe this is the minimum topology that must be mastered by
any student who has decided to become a mathematician. Students
with research interests in topology and related ﬁelds will surely need
to go beyond this book, but it may serve as a starting point. The book
includes basic material on general topology, introduces algebraic topology
via its most classical and elementary part, the theory of the fundamental
group and covering spaces, and provides a background on topological
and smooth manifolds. It is written mainly for students with a limited
experience in mathematics, but who are determined to study the subject
actively.
The core material is presented in a concise form; proofs are omit
ted. Theorems, however, are formulated in detail. We present them as
problems and expect the reader to treat them as problems. Most of the
theorems are easy to ﬁnd elsewhere with complete proofs. We believe
that a serious attempt to prove a theorem must be the ﬁrst reaction to
its formulation. It should precede looking for a book where the theorem
is proved.
On the other hand, we want to emphasize the role of formulations.
In the early stages of studying mathematics it is especially important to
take each formulation seriously. We intentionally force a reader to think
about each simple statement. We hope that this will make the book
inconvenient for mere skimming.
The core material is enhanced by many problems of various sorts
and additional pieces of theory. Although they are closely related to the
main material, they can be (and usually are) kept outside of the standard
lecture course. These enhancements can be recognized by wider margins,
as the next paragraph.
iii
FOREWORD iv
The problems, which do not comprise separate topics and are intended
exclusively to be exercises, are typeset with small face. Some of them are
very easy and included just to provide additional examples. Few problems
are diﬃcult. They are to indicate relations with other parts of mathematics,
show possible directions of development of the subject, or just satisfy an
ambitious reader. Problems, whose solutions seem to be the most diﬃcult
(from the authors’ viewpoint), are marked with a star, as in many other
books.
Further, we want to deliver additional pieces of theory (with respect to the core
material) to more motivated and advanced students. Maybe, a mathematician, who
does not work in the ﬁelds geometric in ﬂavor, can aﬀord the luxury not to know
some of these things. Maybe, students studying topology can postpone this material
to their graduate study. We would like to include this in graduate lecture courses.
However, quite often it does not happen, because most of the topics of this sort are
rather isolated from the contents of traditional graduate courses. They are important,
but more related to the material of the very ﬁrst topology course. In the book these
topics are intertwined with the core material and exercises, but are distinguishable:
they are typeset, like these lines, with large face, theorems and problems in them are
numerated in a special manner described below.
Exercises and illustrative problems to the additional topics are typeset
with wider margins and marked in a diﬀerent way.
Thus, the whole book contains four layers:
• the core material,
• exercises and illustrative problems to the core material,
• additional topics,
• exercises and illustrative problems to additional topics.
The text of the core material is typeset with large face and smallest
margins.
The text of problems elaborating on the core material is typeset with
small face and larger margins.
The text of additional topics is typeset is typeset with large face as the problems
elaborating on the core material.
The text of problems illustrating additional topics is typeset with small
face and larger margins.
Therefore the book looks like a Russian folklore doll, matreshka com
posed of several dolls sitting inside each other. We apologize for being
nonconventional in this and hope that it may help some readers and does
not irritate the others too much.
The whole text of the book is divided into sections. Each section is
divided into subsections. Each of them is devoted to a single topic and
consists of deﬁnitions, commentaries, theorems, exercises, problems, and
riddles.
FOREWORD v
By a riddle we mean a problem of a special sort: its solution is not
contained in the formulation. One has to guess a solution, rather than
deduce it.
0.A. Theorems, exercises, problems and riddles belonging to the core
material are marked with pairs consisting of the number of section and
a letter separated with a dot. The letter identiﬁes the item inside the
section.
0.1. Exercises, problems, and riddles, which are not included in the core, but
are closely related to it (and typeset with small face) are marked with pairs
consisting of the number of the section and the number of the item inside the
section. The numbers in the pair are separated also by a dot.
Theorems, exercises, problems and riddles related to additional topics
are enumerated independently inside each section and denoted similarly.
0:A. The only diﬀerence is that the components of pairs marking the items are
separated by a colon (rather than dot).
We assume that the reader is familiar with naive set theory, but
anticipate that this familiarity may be superﬁcial. Therefore at points
where set theory is especially crucial we make settheoretic digressions
maintained in the same style as the rest of the book.
Advices to the Reader
Since the book contains a summary of elementary topology, you may
use the book while preparing for an examination (especially, if the exam
reduces to solving a collection of problems). However, if you attend
lectures on the subject, it would be much wiser to read the book prior
to the lectures and prove theorems before the lecturer gives the proofs.
We think that a reader who is able to prove statements of the core
of the book, does not need to solve all the other problems. It would be
reasonable instead to look through formulations and concentrate on the
most diﬃcult problems. The more diﬃcult the theorems of the main text
seem to you, the more carefully you should consider illustrative problems,
and the less time you should waste with problems marked with stars.
Keep in mind that sometimes a problem which seems to be diﬃcult is
followed by easier problems, which may suggest hints or serve as technical
lemmas. A chain of problems of this sort is often concluded with a
problem which suggests a return to the theorem, once you are armed
with the lemmas.
Most of our illustrative problems are easy to invent, and, moreover, if
you study the subject seriously, it is always worthwhile to invent problems
of this sort. To develop this style of studying mathematics while solving
FOREWORD vi
our problems one should attempt to invent one’s own problems and solve
them (it does not matter if they are similar to ours or not). Of course,
some problems presented in this book are not easy to invent.
Contents
Foreword iii
Genre, Contents and Style of the Book iii
Advices to the Reader v
Part 1. General Topology 1
Chapter 1. Generalities 3
'1. Digression on Sets 3
'1
◦
1 Sets and Elements 3
'1
◦
2 Equality of Sets 4
'1
◦
3 The Empty Set 4
'1
◦
4 Basic Sets of Numbers 5
'1
◦
5 Describing a Set by Listing of Its Elements 5
'1
◦
6 Subsets 6
'1
◦
7 Properties of Inclusion 6
'1
◦
8 To Prove Equality of Sets, Prove Inclusions 6
'1
◦
9 Inclusion Versus Belonging 6
'1
◦
10 Deﬁning a Set by a Condition 7
'1
◦
11 Intersection and Union 7
'1
◦
12 Diﬀerent Diﬀerences 9
Proofs and Comments 10
Hints, Comments, Advises, Solutions, and Answers 11
'2. Topology in a Set 12
'2
◦
1 Deﬁnition of Topological Space 12
'2
◦
2 The Simplest Examples 12
'2
◦
3 The Most Important Example: Real Line 13
'2
◦
4 Additional Examples 13
'2
◦
5 Using New Words: Points, Open and Closed Sets 13
'2
◦
6 SetTheoretic Digression. De Morgan Formulas 14
'2
◦
7 Properties of Closed Sets 14
'2
◦
8 Being Open or Closed 14
'2
◦
9 Cantor Set 15
'2
◦
10 Characterization of Topology in Terms of Closed Sets 15
'2
◦
11 Topology and Arithmetic Progressions 15
'2
◦
12 Neighborhoods 16
Proofs and Comments 16
Hints, Comments, Advises, Solutions, and Answers 17
vii
CONTENTS viii
'3. Bases 19
'3
◦
1 Deﬁnition of Base 19
'3
◦
2 When a Collection of Sets is a Base 19
'3
◦
3 Bases for Plane 19
'3
◦
4 Subbases 20
'3
◦
5 Inﬁniteness of the Set of Prime Numbers 20
'3
◦
6 Hierarchy of Topologies 20
Proofs and Comments 21
Hints, Comments, Advises, Solutions, and Answers 21
'4. Metric Spaces 23
'4
◦
1 Deﬁnition and First Examples 23
'4
◦
2 Further Examples 23
'4
◦
3 Balls and Spheres 24
'4
◦
4 Subspaces of a Metric Space 24
'4
◦
5 Surprising Balls 25
'4
◦
6 Segments (What Is Between) 25
'4
◦
7 Bounded Sets and Balls 25
'4
◦
8 Norms and Normed Spaces 25
'4
◦
9 Metric Topology 26
'4
◦
10 Openness and Closedness of Balls and Spheres 26
'4
◦
11 Metrizable Topological Spaces 27
'4
◦
12 Equivalent Metrics 27
'4
◦
13 Ultrametric 27
'4
◦
14 Operations with Metrics 28
'4
◦
15 Distance Between Point and Set 28
'4
◦
16 Distance Between Sets 28
'4
◦
17 Asymmetrics 29
Proofs and Comments 30
Hints, Comments, Advises, Solutions, and Answers 31
'5. Ordered Sets 34
'5
◦
1 Strict Orders 34
'5
◦
2 NonStrict Orders 34
'5
◦
3 Relation between Strict and NonStrict Orders 35
'5
◦
4 Cones 35
'5
◦
5 Position of an Element with Respect to a Set 36
'5
◦
6 Total Orders 37
'5
◦
7 Topologies Deﬁned by a Total Order 37
'5
◦
8 Poset Topology 38
'5
◦
9 How to Draw a Poset 39
'5
◦
10 Cyclic Orders in Finite Set 40
'5
◦
11 Cyclic Orders in Inﬁnite Sets 41
'5
◦
12 Topology of Cyclic Order 42
Proofs and Comments 42
Hints, Comments, Advises, Solutions, and Answers 43
CONTENTS ix
'6. Subspaces 45
'6
◦
1 Topology for a subset of a space 45
'6
◦
2 Relativity of Openness 45
'6
◦
3 Agreement on Notations of Topological Spaces 46
Proofs and Comments 46
Hints, Comments, Advises, Solutions, and Answers 47
'7. Position of a Point with Respect to a Set 48
'7
◦
1 Interior, Exterior and Boundary Points 48
'7
◦
2 Interior and Exterior 48
'7
◦
3 Closure 48
'7
◦
4 Closure in a Metric Space 49
'7
◦
5 Frontier 49
'7
◦
6 Closure and Interior with Respect to a Finer Topology 49
'7
◦
7 Properties of Interior and Closure 50
'7
◦
8 Compositions of Closure and Interior 50
'7
◦
9 Sets with Common Frontier 51
'7
◦
10 Convexity and Int, Cl, Fr 51
'7
◦
11 Characterization of Topology by Closure or Interior
Operations 51
'7
◦
12 Dense Sets 51
'7
◦
13 Nowhere Dense Sets 52
'7
◦
14 Limit Points and Isolated Points 52
'7
◦
15 Locally Closed Sets 53
Proofs and Comments 53
Hints, Comments, Advises, Solutions, and Answers 53
'8. SetTheoretic Digression. Maps 55
'8
◦
1 Maps and the Main Classes of Maps 55
'8
◦
2 Image and Preimage 55
'8
◦
3 Identity and Inclusion 56
'8
◦
4 Composition 56
'8
◦
5 Inverse and Invertible 57
'8
◦
6 Submappings 57
'9. Continuous Maps 58
'9
◦
1 Deﬁnition and Main Properties of Continuous Maps 58
'9
◦
2 Reformulations of Deﬁnition 59
'9
◦
3 More Examples 59
'9
◦
4 Behavior of Dense Sets 59
'9
◦
5 Local Continuity 60
'9
◦
6 Properties of Continuous Functions 60
'9
◦
7 Continuity of Distances 61
'9
◦
8 Isometry 61
'9
◦
9 GromovHausdorﬀ distance 61
'9
◦
10 Contractive maps 62
'9
◦
11 Monotone maps 62
CONTENTS x
'9
◦
12 Functions on Cantor Set and SquareFilling Curves 62
'9
◦
13 Sets Deﬁned by Systems of Equations and Inequalities 63
'9
◦
14 SetTheoretic Digression. Covers 64
'9
◦
15 Fundamental Covers 64
Hints, Comments, Advises, Solutions, and Answers 65
'10. Homeomorphisms 67
'10
◦
1 Deﬁnition and Main Properties of Homeomorphisms 67
'10
◦
2 Homeomorphic Spaces 67
'10
◦
3 Role of Homeomorphisms 67
'10
◦
4 More Examples of Homeomorphisms 68
'10
◦
5 Examples of Homeomorphic Spaces 69
'10
◦
6 Examples of Nonhomeomorphic Spaces 72
'10
◦
7 Homeomorphism Problem and Topological Properties 72
'10
◦
8 Information: Nonhomeomorphic Spaces 72
'10
◦
9 Embeddings 73
'10
◦
10 Equivalence of Embeddings 73
'10
◦
11 Information 74
Chapter 2. Topological Properties 75
'11. Connectedness 75
'11
◦
1 Deﬁnitions of Connectedness and First Examples 75
'11
◦
2 Connected Sets 75
'11
◦
3 Properties of Connected Sets 76
'11
◦
4 Connected Components 76
'11
◦
5 Totally Disconnected Spaces 77
'11
◦
6 Frontier and Connectedness 77
'11
◦
7 Connectedness and Continuous Maps 77
'11
◦
8 Connectedness on Line 78
'11
◦
9 Intermediate Value Theorem and Its Generalizations 79
'11
◦
10 Dividing Pancakes 79
'11
◦
11 Induction on Connectedness 79
'11
◦
12 Applications to Homeomorphism Problem 80
'12. PathConnectedness 81
'12
◦
1 Paths 81
'12
◦
2 PathConnected Spaces 81
'12
◦
3 PathConnected Sets 82
'12
◦
4 Properties of PathConnected Sets 82
'12
◦
5 PathConnected Components 82
'12
◦
6 PathConnectedness Versus Connectedness 83
'12
◦
7 PolygonConnectedness 83
'12
◦
8 Connectedness of Some Sets of Matrices 84
'13. Separation Axioms 85
'13
◦
1 The Hausdorﬀ Axiom 85
'13
◦
2 Limits of Sequence 85
CONTENTS xi
'13
◦
3 Coincidence Set and Fixed Point Set 86
'13
◦
4 Hereditary Properties 86
'13
◦
5 The First Separation Axiom 86
'13
◦
6 The Kolmogorov Axiom 87
'13
◦
7 The Third Separation Axiom 88
'13
◦
8 The Fourth Separation Axiom 88
'13
◦
9 Niemytski’s Space 89
'13
◦
10 Urysohn Lemma and Tietze Theorem 89
'14. Countability Axioms 90
'14
◦
1 SetTheoretic Digression. Countability 90
'14
◦
2 Second Countability and Separability 90
'14
◦
3 Embedding and Metrization Theorems 91
'14
◦
4 Bases at a Point 91
'14
◦
5 First Countability 92
'14
◦
6 Sequential Approach to Topology 92
'14
◦
7 Sequential Continuity 93
'15. Compactness 94
'15
◦
1 Deﬁnition of Compactness 94
'15
◦
2 Terminology Remarks 94
'15
◦
3 Compactness in Terms of Closed Sets 95
'15
◦
4 Compact Sets 95
'15
◦
5 Compact Sets Versus Closed Sets 95
'15
◦
6 Compactness and Separation Axioms 96
'15
◦
7 Compactness in Euclidean Space 96
'15
◦
8 Compactness and Continuous Maps 97
'15
◦
9 Closed Maps 97
'15
◦
10 Norms in R
n
98
'16. Local Compactness and Paracompactness 99
'16
◦
1 Local Compactness 99
'16
◦
2 OnePoint Compactiﬁcation 99
'16
◦
3 Proper Maps 100
'16
◦
4 Locally Finite Collections of Subsets 100
'16
◦
5 Paracompact Spaces 101
'16
◦
6 Paracompactness and Separation Axioms 101
'16
◦
7 Partitions of Unity 101
'16
◦
8 Application: Making Embeddings from Pieces 101
'17. Sequential Compactness 103
'17
◦
1 Sequential Compactness Versus Compactness 103
'17
◦
2 In Metric Space 103
'17
◦
3 Completeness and Compactness 104
'17
◦
4 NonCompact Balls in Inﬁnite Dimension 104
'17
◦
5 pAdic Numbers 104
'17
◦
6 Induction on Compactness 105
'17
◦
7 Spaces of Convex Figures 105
CONTENTS xii
Problems for Tests 107
Chapter 3. Topological Constructions 109
'18. Multiplication 109
'18
◦
1 SetTheoretic Digression. Product of Sets 109
'18
◦
2 Product of Topologies 110
'18
◦
3 Topological Properties of Projections and Fibers 110
'18
◦
4 Cartesian Products of Maps 111
'18
◦
5 Properties of Diagonal and Other Graphs 111
'18
◦
6 Topological Properties of Products 112
'18
◦
7 Representation of Special Spaces as Products 113
'19. Quotient Spaces 114
'19
◦
1 SetTheoretic Digression. Partitions and Equivalence
Relations 114
'19
◦
2 Quotient Topology 115
'19
◦
3 Topological Properties of Quotient Spaces 115
'19
◦
4 SetTheoretic Digression. Quotients and Maps 116
'19
◦
5 Continuity of Quotient Maps 116
'19
◦
6 Closed Partitions 117
'19
◦
7 Open Partitions 117
'19
◦
8 SetTheoretic Digression:
Splitting a transitive relation
into equivalence and partial order 117
'19
◦
9 Finite Topological Spaces 118
'19
◦
10 Simplicial schemes 119
'19
◦
11 Baricentric Subdivision of a Poset 119
'20. Zoo of Quotient Spaces 121
'20
◦
1 Tool for Identifying a Quotient Space with a Known
Space 121
'20
◦
2 Tools for Describing Partitions 121
'20
◦
3 Entrance to the Zoo 122
'20
◦
4 Transitivity of Factorization 123
'20
◦
5 M¨obius Strip 124
'20
◦
6 Contracting Subsets 124
'20
◦
7 Further Examples 125
'20
◦
8 Klein Bottle 125
'20
◦
9 Projective Plane 126
'20
◦
10 You May Have Been Provoked to Perform an Illegal
Operation 126
'20
◦
11 SetTheoretic Digression. Sums of Sets 126
'20
◦
12 Sums of Spaces 126
'20
◦
13 Attaching Space 127
'20
◦
14 Basic Surfaces 128
'21. Projective Spaces 130
CONTENTS xiii
'21
◦
1 Real Projective Space of Dimension n 130
'21
◦
2 Complex Projective Space of Dimension n 131
'21
◦
3 Quaternionic Projective Spaces 131
'22. Spaces of Continuous Maps 134
'22
◦
1 Sets of Continuous Mappings 134
'22
◦
2 Topologies on Set of Continuous Mappings 134
'22
◦
3 Topological Properties of Mapping Spaces 135
'22
◦
4 Metric Case 135
'22
◦
5 Interactions With Other Constructions 136
'22
◦
6 Mappings X Y →Z and X →((Y, Z) 136
Chapter 4. A Touch of Topological Algebra 138
'23. Digression. Generalities on Groups 139
'23
◦
1 The Notion of Group 139
'23
◦
2 Additive and Multiplicative Notations 140
'23
◦
3 Homomorphisms 141
'23
◦
4 Subgroups 141
'24. Topological Groups 143
'24
◦
1 The Notion of Topological Group 143
'24
◦
2 Examples of Topological Groups 143
'24
◦
3 SelfHomeomorphisms Making a Topological Group
Homogeneous 143
'24
◦
4 Neighborhoods 144
'24
◦
5 Separaion Axioms 145
'24
◦
6 Countability Axioms 145
'24
◦
7 Subgroups 145
'24
◦
8 Normal Subgroups 146
'24
◦
9 Homomorphisms 147
'24
◦
10 Local Isomorphisms 147
'24
◦
11 Direct Products 148
'25. Actions of Topological Groups 150
'25
◦
1 Actions of Group in Set 150
'25
◦
2 Continuous Actions 150
'25
◦
3 Orbit Spaces 150
'25
◦
4 Homogeneous Spaces 150
Part 2. Algebraic Topology 151
Chapter 5. Fundamental Group and Covering Spaces 153
'26. Homotopy 153
'26
◦
1 Continuous Deformations of Maps 153
'26
◦
2 Homotopy as Map and Family of Maps 153
'26
◦
3 Homotopy as Relation 154
'26
◦
4 StraightLine Homotopy 154
'26
◦
5 Maps to Star Convex Sets 154
CONTENTS xiv
'26
◦
6 Maps of Convex Sets 155
'26
◦
7 Easy Homotopies 155
'26
◦
8 Two Natural Properties of Homotopies 155
'26
◦
9 Stationary Homotopy 156
'26
◦
10 Homotopies and Paths 156
'26
◦
11 Homotopy of Paths 156
'27. Homotopy Properties of Path Multiplication 158
'27
◦
1 Multiplication of Homotopy Classes of Paths 158
'27
◦
2 Associativity 158
'27
◦
3 Unit 159
'27
◦
4 Inverse 159
'28. Fundamental Group 161
'28
◦
1 Deﬁnition of Fundamental Group 161
'28
◦
2 Why Index 1? 161
'28
◦
3 High Homotopy Groups 161
'28
◦
4 Circular loops 162
'28
◦
5 The Very First Calculations 163
'28
◦
6 Fundamental Group of Product 163
'28
◦
7 SimplyConnectedness 164
'28
◦
8 Fundamental Group of a Topological Group 165
'29. The Role of Base Point 166
'29
◦
1 Overview of the Role of Base Point 166
'29
◦
2 Deﬁnition of Translation Maps 166
'29
◦
3 Properties of T
s
166
'29
◦
4 Role of Path 167
'29
◦
5 High Homotopy Groups 167
'29
◦
6 In Topological Group 168
Chapter 6. Covering Spaces and Calculation of Fundamental
Groups 169
'30. Covering Spaces 169
'30
◦
1 Deﬁnition of Covering 169
'30
◦
2 More Examples 169
'30
◦
3 Local homeomorphisms versus coverings 170
'30
◦
4 Number of Sheets 171
'30
◦
5 Universal Coverings 171
'31. Theorems on Path Lifting 172
'31
◦
1 Lifting 172
'31
◦
2 Path Lifting 172
'31
◦
3 Homotopy Lifting 173
'31
◦
4 HighDimensional Homotopy Groups of Covering Space173
'32. Calculations of Fundamental Groups Using Universal
Coverings 174
'32
◦
1 Fundamental Group of Circle 174
CONTENTS xv
'32
◦
2 Fundamental Group of Projective Space 175
'32
◦
3 Fundamental Groups of Bouquet of Circles 175
'32
◦
4 Algebraic Digression. Free Groups 175
'32
◦
5 Universal Covering for Bouquet of Circles 177
'32
◦
6 Fundamental groups of some ﬁnite
topological spaces 178
Chapter 7. Fundamental Group and Mappings 179
'33. Induced Homomorphisms
and Their First Applications 179
'33
◦
1 Homomorphisms Induced by a Continuous Map 179
'33
◦
2 Fundamental Theorem of Algebra 180
'33
◦
3 Generalization of Intermediate Value Theorem 181
'33
◦
4 Winding Number 182
'33
◦
5 BorsukUlam Theorem 182
'34. Retractions and Fixed Points 184
'34
◦
1 Retractions and Retracts 184
'34
◦
2 Fundamental Group and Retractions 184
'34
◦
3 FixedPoint Property. 185
'35. Homotopy Equivalences 187
'35
◦
1 Homotopy Equivalence as Map 187
'35
◦
2 Homotopy Equivalence as Relation 187
'35
◦
3 Deformation Retraction 187
'35
◦
4 Examples 188
'35
◦
5 Deformation Retraction Versus Homotopy Equivalence188
'35
◦
6 Contractible Spaces 189
'35
◦
7 Fundamental Group and Homotopy Equivalences 189
'36. Covering Spaces via Fundamental Groups 191
'36
◦
1 Homomorphisms Induced by Covering Projections 191
'36
◦
2 Number of Sheets 191
'36
◦
3 Hierarchy of Coverings 192
'36
◦
4 Existence of subordinations 193
'36
◦
5 Micro Simply Connected Spaces 193
'36
◦
6 Existence of Coverings 194
'36
◦
7 Automorphisms of Covering 194
'36
◦
8 Regular Coverings 195
'36
◦
9 Lifting and Covering Maps 196
Chapter 8. Cellular Techniques 197
'37. Cellular Spaces 197
'37
◦
1 Deﬁnition of Cellular Spaces 197
'37
◦
2 First Examples 199
'37
◦
3 Further TwoDimensional Examples 200
'37
◦
4 Simplicial spaces 201
'37
◦
5 Topological Properties of Cellular Spaces 201
CONTENTS xvi
'37
◦
6 Embedding to Euclidean Space 202
'37
◦
7 Euler Characteristic 203
'37
◦
8 Collaps 203
'37
◦
9 Generalized collaps 204
'38. OneDimensional Cellular Spaces 206
'38
◦
1 Homotopy Classiﬁcation 206
'38
◦
2 Dividing Cells 206
'38
◦
3 Trees and Forests 207
'38
◦
4 Simple Paths 207
'38
◦
5 Maximal Trees 208
'39. Fundamental Group of a Cellular Space 209
'39
◦
1 OneDimensional Cellular Spaces 209
'39
◦
2 Generators 209
'39
◦
3 Relators 209
'39
◦
4 Writing Down Generators and Relators 210
'39
◦
5 Fundamental Groups of Basic Surfaces 211
'39
◦
6 Seifert  van Kampen Theorem 212
'40. OneDimensional Homology and Cohomology 213
'40
◦
1 Why and What for 213
'40
◦
2 OneDimensional Integer Homology 213
'40
◦
3 Zero Homologous Loops and Disks with Handles 214
'40
◦
4 Description of H
1
(X) in Terms of Free Circular Loops 214
'40
◦
5 Homology and Continuous Maps 215
'40
◦
6 OneDimensional Cohomology 216
'40
◦
7 Cohomology and Classiﬁcation of Regular Coverings 216
'40
◦
8 Integer Cohomology and Maps to S
1
216
'40
◦
9 OneDimensional Homology Modulo 2 217
Part 3. Manifolds 219
Chapter 9. Bare Manifolds 221
'41. Locally Euclidean Spaces 221
'41
◦
1 Deﬁnition of Locally Euclidean Space 221
'41
◦
2 Dimension 221
'41
◦
3 Interior and Boundary 222
'42. Manifolds 225
'42
◦
1 Deﬁnition of Manifold 225
'42
◦
2 Components of Manifold 225
'42
◦
3 Making New Manifolds out of Old Ones 225
'42
◦
4 Double 226
'42
◦
5 Collars and Bites 226
'43. Isotopy 228
'43
◦
1 Isotopy of Homeomorphisms 228
'43
◦
2 Isotopy of Embeddings and Sets 228
CONTENTS xvii
'43
◦
3 Isotopies and Attaching 229
'43
◦
4 Connected Sums 230
'44. OneDimensional Manifolds 231
'44
◦
1 ZeroDimensional Manifolds 231
'44
◦
2 Reduction to Connected Manifolds 231
'44
◦
3 Examples 231
'44
◦
4 Statements of Main Theorems 231
'44
◦
5 Lemma on 1Manifold Covered with Two Lines 232
'44
◦
6 Without Boundary 233
'44
◦
7 With Boundary 233
'44
◦
8 Consequences of Classiﬁcation 233
'44
◦
9 Mapping Class Groups 233
'45. TwoDimensional Manifolds: General Picture 234
'45
◦
1 Examples 234
'45
◦
2 Ends and Odds 234
'45
◦
3 Closed Surfaces 235
'46. Triangulations 237
'46
◦
1 Triangulations of Surfaces 237
'46
◦
2 Triangulation as cellular decomposition 237
'46
◦
3 Two Properties of Triangulations of Surfaces 237
'46
◦
4 Scheme of Triangulation 238
'46
◦
5 Examples 238
'46
◦
6 Subdivision of a Triangulation 239
'46
◦
7 Homotopy Type of Compact Surface with NonEmpty
Boundary 241
'46
◦
8 Triangulations in dimension one 241
'46
◦
9 Triangualtions in higher dimensions 242
'47. Handle Decomposition 243
'47
◦
1 Handles and Their Anatomy 243
'47
◦
2 Handle Decomposition of Manifold 243
'47
◦
3 Handle Decomposition and Triangulation 244
'47
◦
4 Regular Neighborhoods 245
'47
◦
5 Cutting 2Manifold Along a Curve 245
'48. Orientations 248
'48
◦
1 Orientations of Edges and Triangles 248
'48
◦
2 Orientation of Triangulation 248
'49. Classical Approach to Topological Classiﬁcation
of Compact surfaces 249
'49
◦
1 Families of Polygons 249
'49
◦
2 Operations on Family of Polygons 250
'49
◦
3 Topological and Homotopy Classiﬁcation of Closed
Surfaces 250
'49
◦
4 Recognizing Closed Surfaces 251
'49
◦
5 Orientations 252
CONTENTS xviii
'49
◦
6 More About Recognizing Closed Surfaces 252
'49
◦
7 Compact Surfaces with Boundary 253
'49
◦
8 Simply Connected Surfaces 253
'50. OneDimensional mod2Homology of Surfaces 254
'50
◦
1 Polygonal Paths on Surface 254
'50
◦
2 Subdivisions of Triangulation 254
'50
◦
3 Bringing Loops to General Position 255
'50
◦
4 Cutting Surface Along Curve 256
'50
◦
5 Curves on Surfaces and TwoFold Coverings 257
'50
◦
6 OneDimensional Z
2
Cohomology of Surface 257
'50
◦
7 OneDimensional Z
2
Homology of Surface 258
'50
◦
8 Poincar´e Duality 258
'50
◦
9 OneSided and TwoSided Simple Closed Curves on
Surfaces 258
'50
◦
10 Orientation Covering and First StiefelWhitney Class 258
'50
◦
11 Relative Homology 258
'51. Surfaces Beyond Classiﬁcation 259
'51
◦
1 Genus of Surface 259
'51
◦
2 Systems of disjoint curves on a surface 259
'51
◦
3 Polygonal Jordan and Sch¨onﬂies Theorems 259
'51
◦
4 Polygonal Annulus Theorem 259
'51
◦
5 Dehn Twists 259
'51
◦
6 Coverings of Surfaces 259
'51
◦
7 Branched Coverings 259
'51
◦
8 Mapping Class Group of Torus 259
'51
◦
9 Braid Groups 259
'52. ThreeDimensional Manifolds 260
'52
◦
1 Poincar´e Conjecture 260
'52
◦
2 Lens Spaces 260
'52
◦
3 Seifert Manifolds 260
'52
◦
4 Fibrations over Circle 260
'52
◦
5 Heegaard Splitting and Diagrams 260
Proofs and Comments 261
Part 1
General Topology
The goal of this part of the book is to teach the language of math
ematics. More speciﬁcally, one of its most important components: the
language of settheoretic topology, which treats the basic notions related
to continuity. The term general topology means: this is the topology that
is needed and used by most mathematicians.
As a research ﬁeld, it was completed a long time ago. A permanent
usage in the capacity of a common mathematical language has polished
its system of deﬁnitions and theorems. Nowadays studying general topol
ogy really resembles studying a language rather than mathematics: one
needs to learn a lot of new words, while proofs of all theorems are ex
tremely simple. On the other hand, the theorems are numerous, for they
play the role of rules regulating usage of words.
We have to warn students, for whom this is one of the ﬁrst mathe
matical subjects. Do not hurry to fall in love with it too seriously, do
not let an imprinting happen. This ﬁeld may seam to be charming, but
it is not very active. It hardly provides as much room for exciting new
research as most of other ﬁelds.
CHAPTER 1
Generalities
'1 Digression on Sets
We begin with a digression, which we would like to consider unnec
essary. Its subject is the ﬁrst basic notions of the naive set theory. This
is a part of the common mathematical language, too, but even more
profound than general topology. We would not be able to say anything
about topology without this part (see the next section to make sure that
this is not an exaggeration). Naturally, one may expect that naive set
theory becomes familiar to a student when she or he studies Calculus
or Algebra, the subjects which usually precede topology. If this is what
really happened to you, please, glance through this section and move to
the next one.
'1
◦
1 Sets and Elements
In any intellectual activity, one of the most profound action is gath
ering objects into groups. The gathering is performed in minds and is
not accompanied with any action in the physical world. As soon as the
group has been created and assigned with a name, it may be subject of
thoughts and arguments and, in particular, may be included into other
groups. In Mathematics there is an elaborated system of notions which
organizes and regulate creating of those groups and manipulating them.
This system is called the naive set theory, a slightly misleading name,
because this is rather a language, than a theory.
The ﬁrst words in this language are set and element. y a set we
understand an arbitrary collection of various objects. An object included
into the collection is called an element of the set. A set consists of its
elements. It is formed by them. To diversify wording, the word set is
replaced by the word collection. Sometimes other words, such as class,
family and group, are used in the same sense, but it is not quite safe,
since each of these words is associated in the modern mathematics with
a more special meaning, and hence should be used instead of the word
set cautiously.
If x is an element of a set A, we write x ∈ A and say x belongs to A
and A contains x. The sign ∈ is a version of Greek letter epsilon, which
is the ﬁrst letter of the Latin word element. To make formulas more
3
§1. DIGRESSION ON SETS 4
ﬂexible, the formula x ∈ A is allowed to be written also as A ÷ x. So
the origin of notation is ignored, but a more meaningful similarity to the
inequality symbols < and > is emphasized. To state that x is not an
element of A, we write x ∈ A or A ÷ x.
'1
◦
2 Equality of Sets
A set is deﬁned by its elements. It is nothing but a collection of its
elements. This manifests most sharply in the following principle: two
sets are considered equal, if and only if they have the same elements. In
this sense the word set has slightly disparaging meaning. When one calls
something a set, this shows, maybe unintentionally, a lack of interest to
whatever organization of the elements of this set.
For example, when we say that a line is a set of points, we indicate
that two lines coincide if and only if they consist of the same points. On
the other hand, we commit ourselves to consider all the relations between
points on a line (e.g. the distance between points, the order of points on
the line) separately from the notion of line.
We may think of sets as boxes, which can be built eﬀortlessly around
elements, just to distinguish them from the rest of the world. The cost of
this lightness is that such a box is not more than the collection of elements
placed inside. It is a little more than just a name: it is a declaration of
our wish to think about this collection of things as of entity and not to
go into details about the nature of its memberselements. Elements, in
turn, may also be sets, but as long as we consider them elements, they
play the role of atoms with their own original nature ignored.
In modern Mathematics the words set and element are very common
and appear in most of texts. They are even overused. There are instances
when it is not appropriate to use them. For example, it is not good
to use the word element as a replacement for other, more meaningful
words. When you call something an element, the set, whose element
is this one, should be clear. The word element makes sense only in a
combination with the word set, unless we deal with nonmathematical
term (like chemical element), or a rare oldfashioned exception from the
common mathematical terminology (sometimes the expression under the
sign of integral is called an inﬁnitesimal element, in old texts lines, planes
and other geometric images are called elements). Euclid’s famous book
on Geometry is called Elements.
'1
◦
3 The Empty Set
Thus, an element may not be without a set. However a set may
be without elements. There is a set which has no element. This set is
unique, because a set is deﬁned completely by its elements. It is called
§1. DIGRESSION ON SETS 5
the empty set and denoted by ∅. Other notations, like Λ, also were in
use, but ∅ has become common.
'1
◦
4 Basic Sets of Numbers
Besides ∅, there are few other sets, which are so important that have
their own unique names and notation. The set of all natural numbers,
i.e., 1, 2, 3, 4, 5, . . . , etc., is denoted by N. The set of all integer
numbers, both positive (that is natural numbers) and negative and the
zero, is denoted by Z. The set of all the rational numbers (add to the
integers those numbers which can be presented by fractions, like
2
3
, −
7
5
)
is denoted by Q. The set of all the real numbers (obtained by adjoining
to rational numbers the numbers like
√
2 and π = 3.14 . . . ) is denoted
by R. The set of complex numbers is denoted by C.
'1
◦
5 Describing a Set by Listing of Its Elements
The set presented by the list a, b, . . . , x of its elements is denoted
by symbol ¦a, b, . . . , x¦. In other words, the list of objects enclosed in a
curly brackets denotes the set, whose elements are listed. For example,
¦1, 2, 123¦ denotes the set which consists of numbers 1, 2 and 123. No
tation ¦a, x, A¦ means the set which consists of three elements, a, x and
A, whatever these three letters denote.
1.1. What is ¦∅¦? How many elements does it contain?
1.2. Which of the following formulas are correct:
1) ∅ ∈ ¦∅, ¦∅¦¦; 2) ¦∅¦ ∈ ¦¦∅¦¦; 3) ∅ ∈ ¦¦∅¦¦?
A set consisting of a single element is called a singleton. This is any
set which can be presented as ¦a¦ for some a.
1.3. Is ¦¦∅¦¦ a singleton?
Notice that sets ¦1, 2, 3¦ and ¦3, 2, 1, 2¦ are equal, since they consists
of the same elements. At ﬁrst glance, a list with repetition of elements
is never needed. There arises even a temptation to prohibit usage of
lists with repetitions in such a notation. However, as it often happens
to temptations to prohibit something, this would not be wise. In fact,
quite often one cannot say a priori if there are repetitions or not. For
example, the elements of the list may depend on parameter, and under
certain values of the parameters some entries of the list coincide, while
for other values, they don’t.
1.4. How many elements do the following sets contain?
1) ¦1, 2, 1¦; 2) ¦1, 2, ¦1, 2¦¦; 3) ¦¦2¦¦;
4) ¦¦1¦, 1¦; 5) ¦1, ∅¦; 6) ¦¦∅¦, ∅¦;
7) ¦¦∅¦, ¦∅¦¦; 8) ¦x, 3x −1¦ for x ∈ R.
§1. DIGRESSION ON SETS 6
'1
◦
6 Subsets
If A and B are sets and every element of A belongs also to B, we say
that A is a subset of B, or B includes A, and write A ⊂ B or B ⊃ A.
The inclusion signs ⊂ and ⊃ recall the inequality signs < and
> for a good reason: in the world of sets the inclusion signs are obvious
counterparts for the signs of inequalities.
1.A. Let a set A consists of a elements, and a set B of b elements. Prove
that if A ⊂ B then a ≤ b.
Thus, the inclusion signs are not completely true counterparts of the
inequality signs < and >. They are closer to ≤ and ≥.
'1
◦
7 Properties of Inclusion
1.B Reﬂexivity of Inclusion. Any set includes itself: A ⊂ A holds
true for any A.
Notice that there is no number a satisfying inequality a < a.
1.C The Empty Set Is Everywhere. ∅ ⊂ A for any set A. In other
words, the empty set is present in each set as a subset.
Thus, each set A has two obvious subsets: the empty set ∅ and A
itself. A subset of A diﬀerent from ∅ and A is called a proper subset of
A. This word is used when one does not want to consider the obvious
subsets (which are called improper).
1.D Transitivity of Inclusion. If A, B and C are sets, A ⊂ B and
B ⊂ C, then A ⊂ C.
'1
◦
8 To Prove Equality of Sets, Prove Inclusions
Working with sets, we need from time to time to prove that two sets,
say A and B, which may have emerged in quite diﬀerent ways, are equal.
The most common way to do this is provided by the following theorem.
1.E Criterium of Equality for Sets.
A = B, if and only if A ⊂ B and B ⊂ A.
'1
◦
9 Inclusion Versus Belonging
1.F. x ∈ A, if and only if ¦x¦ ⊂ A.
Despite this obvious relation between the notions of belonging ∈ and
inclusion ⊂ and similarity of the symbols ∈ and ⊂, the concepts are
very diﬀerent. Indeed, A ∈ B means that A is one of the elements of B
(that is one of indivisible pieces comprising B), while A ⊂ B means that
A is made of some of the elements of B.
§1. DIGRESSION ON SETS 7
In particular, A ⊂ A, while A ∈ A for any reasonable A. Thus,
belonging is not reﬂexive. Yet another diﬀerence: belonging is not tran
sitive, while inclusion is.
1.G NonReﬂexivity of Belonging. Construct sets A and B such
that A ∈ A, while B ∈ B. Cf. 1.B.
1.H NonTransitivity of Belonging. Construct sets A, B and C
such that A ∈ B and B ∈ C, but A ∈ C. Cf. 1.D.
'1
◦
10 Deﬁning a Set by a Condition
As we know (see Section '1
◦
5), a set can be described by presenting
a list of its elements. This simplest way may be not available or, at least,
be not the easiest one. For example, it is easy to say: “the set of all the
solutions of the following equation” and write down the equation. This is
a reasonable description of the set. At least, it is unambiguous. Having
accepted it, we may start speaking on the set, studying its properties,
and eventually may be lucky to solve the equation and get the list of its
solutions. However the latter may be diﬃcult and should not prevent us
from discussing the set.
Thus we see another way for description of a set: to formulate the
properties which distinguish the elements of the set among elements of
some wider and already known set. Here is the corresponding notation:
the subset of a set A consisting of elements x which satisfy condition
P(x) is denoted by ¦x ∈ A [ P(x)¦.
1.5. Present the following sets by lists of their elements (i.e., in the form
¦a, b, . . . ¦)
(a) ¦x ∈ N [ x < 5¦, (b) ¦x ∈ N [ x < 0¦, (c) ¦x ∈ Z [ x < 0¦.
'1
◦
11 Intersection and Union
The intersection of sets A and B is the set consisting of their common
elements, that is elements belonging both to A and B. It is denoted by
A∩ B and can be described by formula
A∩ B = ¦x [ x ∈ A x ∈ B¦.
Sets A and B are said to be disjoint, if their intersection is empty, i.e.,
A∪ B = ∅.
The union of sets A and B is the set consisting of those elements each
of which belongs to at least one of these sets. The union of A and B is
denoted by A∪ B. It can be described by formula
A ∪ B = ¦x [ x ∈ A or x ∈ B¦.
§1. DIGRESSION ON SETS 8
Here the conjunction or should be understood in the inclusive way: the
statement “x ∈ A or x ∈ B” means that x belongs to at least one of the
sets A and B, but, maybe, to both of them.
A B A B A B
A∩ B A∪ B
Figure 1. Disks A and B, their intersection A ∩ B and
union A∪ B.
1.I Commutativity of ∩ and ∪. For any sets A and B
A∩ B = B ∩ A A∪ B = B ∪ A.
1.6. Prove that for any set A
A∩ A = A, A ∪ A = A, A ∪ ∅ = A and A∩ ∅ = ∅.
1.7. Prove that for any sets A and B
A ⊂ B, iﬀ A ∩ B = A, iﬀ A ∪ B = B.
1.J Associativity of ∩ and ∪. For any sets A, B and C
(A∩ B) ∩ C = A∩ (B ∩ C) and (A ∪ B) ∪ C = A∪ (B ∪ C).
Associativity allows us do not care about brackets and sometimes
even omit them. One deﬁnes A ∩ B ∩ C = (A ∩ B) ∩ C = A ∩ (B ∩ C)
and A∪B ∪C = (A∪B) ∪C = A∪(B ∪C). However, intersection and
union of arbitrarily large (in particular, inﬁnite) collection of sets can be
deﬁned directly, without reference to intersection or union of two sets.
Indeed, let Γ be a collection of sets. The intersection of the sets belonging
to Γ is the set formed by elements which belong to every set, belonging
to Γ. This set is denoted by ∩
A∈Γ
A or
¸
A∈Γ
A. Similarly, the union of
the sets belonging to Γ is the set formed by elements which belong to at
least one of the sets belonging to Γ. This set is denoted by ∪
A∈Γ
A or
¸
A∈Γ
A .
1.K. The notions of intersection and union of arbitrary collection of
sets generalize the notions of intersection and union of two sets: for
Γ = ¦A, B¦
¸
C∈Γ
C = A∩ B and
¸
C∈Γ
C = A∪ B.
1.8. Enigma. How are related to each other the notions of system of equa
tions and intersection of sets?
1.L Two Distributivities. For any sets A, B and C
(A∩ B) ∪ C = (A∪ C) ∩ (B ∪ C). (1)
(A∪ B) ∩ C = (A∩ C) ∪ (B ∩ C) (2)
§1. DIGRESSION ON SETS 9
A A B B
C C C
(A ∩ B) ∪ C (A∪ C) (B ∪ C) = ∩
= ∩
Figure 2. The lefthand side (A∩B) ∪C of the equality
(1) and the sets A ∪ C B ∪ C, whose intersection is the
righthand side of the equation (1). A∪ B.
In Figure 2 the ﬁrst of two equalities of Theorem 1.L is illustrated by
a sort of comics. Such comics are called Venn diagrams. They are very
useful and we strongly recommend to draw them for each formula about
sets.
1.M. Draw a Venn diagram illustrating (2). Prove (1) and (2) tracing
all the details of the proofs in Venn diagrams. Draw Venn diagrams
illustrating all formulas below in this section.
1.9. Enigma. Generalize Theorem 1.L to the case of arbitrary collection of
sets.
1.N Yet Another Pair of Distributivities. Let A be a set and Γ be
a set consisting of sets. Then
A ∩
¸
B∈Γ
B =
¸
B∈Γ
(A∩ B) and A∪
¸
B∈Γ
B =
¸
B∈Γ
(A∪ B).
'1
◦
12 Diﬀerent Diﬀerences
A diﬀerence AB of sets A and B is the set of those elements of A
which do not belong to B. Here it is not assumed that A ⊃ B.
If A ⊃ B, the set AB is called also the complement of B in A.
1.10. Prove that for any sets A and B their union A∪B can be represented
as the union of the following three sets: A B, B A and A ∩ B, and that
these sets are pairwise disjoint.
1.11. Prove that A(A B) = A∩ B for any sets A and B.
1.12. Prove that A ⊂ B, if and only if A B = ∅.
1.13. Prove that A∩(B C) = (A∩B) (A∩C) for any sets A, B and C.
The set (A B) ∪ (B A) is called the symmetric diﬀerence of sets
A and B. It is denoted by A B.
1.14. Prove that for any sets A and B
A B = (A ∪ B) (A ∩ B)
§1. DIGRESSION ON SETS 10
A B A B A B
B A A B A B
Figure 3. Diﬀerences of disks A and B.
1.15 Associativity of Symmetric Diﬀerence. Prove that for any sets
A, B and C
(A B) C = A (B C).
1.16. Enigma. Find a symmetric deﬁnition of symmetric diﬀerence (A
B) C of three sets and generalize it to any ﬁnite collection of sets.
1.17 Distributivity. Prove that (A B) ∩ C = (A∩ C) (B ∩ C) for any
sets A, B and C.
1.18. Does the following equality hold true for any sets A, B C
(A B) ∪ C = (A ∪ C) (B ∪ C)?
Proofs and Comments
1.A The question is so elementary that it is diﬃcult to ﬁnd more
elementary facts, which a proof can be based on. What does it mean that
A consists of a elements? It means, say, that we can count elements of A
one by one assigning to them numbers 1, 2, 3, and the last element will
get number a. It is known that the result does not depend on the order
in which we count. (In fact, one can develop the set theory, which would
include a theory of counting, and in which this is a theorem. But since
we have no doubts in this fact, let us use it without proof.) Therefore
we can start counting of elements of B with counting the elements of A.
The counting of elements of A will be done, ﬁrst, and then, if there are
some elements of B which are not in A, counting will continue. Thus the
number of elements in A is less than or equal to the number of elements
in B.
1.B Recall that, by the deﬁnition of inclusion, A ⊂ B means that
each element of A is an element of B. Therefore the statement that
we have to prove can be rephrased as follows: each element of A is an
element of A. This is a tautology.
1.C Recall that, by the deﬁnition of inclusion, A ⊂ B means that
each element of A is an element of B. Thus we need to prove that any
element of ∅ belongs to A. This is correct, because there is no elements
in ∅. If you are not satisﬁed with this argument (since it sounds too
crazy), let us resort to a question, whether this can be wrong. How can
it happen that ∅ is not a subset of A? It could happen, only if there was
an element of ∅ which would not be an element of A. But there is no
such an element in ∅, because ∅ has no elements at all.
§1. DIGRESSION ON SETS 11
1.D We have to prove that each element of A is an element of C.
Let x ∈ A. Since A ⊂ B, it follows that x ∈ B. Since b ⊂ C, the latter
(i.e., x ∈ B) implies x ∈ C. This is what we had to prove.
1.E We have already seen that A ⊂ A. Hence if A = B then A ⊂ B
and B ⊂ A. On the other hand, A ⊂ B means that each element of
A belongs to B and B ⊂ A means that each element of B belongs to
A. Hence A and B have the same elements, which means that they are
equal.
1.G It is easy to construct a set A with A ∈ A. Take A = ∅, or
A = N, or A = ¦1¦,. . . A set B such that B ∈ B is a strange creature.
It would not appear in real problems, unless you think really globally.
Take for B the set of all sets. Mathematicians avoid such sets. There
are good reasons for this. If we consider the set of all sets, why not to
consider the set Y of all the sets X such that X ∈ X? Does Y belongs to
itself? If Y ∈ Y then Y ∈ Y , since each element X of Y has the property
that X ∈ X. If Y ∈ Y then Y ∈ Y since Y is the set of ALL the sets
X such that X ∈ X. This contradiction shows that our deﬁnition of Y
does not make sense. An easy way to avoid this paradox is to prohibit
consideration of sets with the property X ∈ X. The the set of all sets is
not a legitimate set.
1.H Take A = ¦1¦, B = ¦¦1¦¦ and C = ¦¦¦1¦¦¦. It is more diﬃcult
to construct sets A, B and C such that A ∈ B, B ∈ C, and A ∈ C. Take
A = ¦1¦, B = ¦¦1¦¦, C = ¦¦1¦, ¦¦1¦¦¦.
Hints, Comments, Advises, Solutions, and Answers
1.1 The set ¦∅¦ consists of one element, which is the empty set ∅. Of
course, this element itself is the empty set and contains no element, but the
set ¦∅¦ consists of a single element ∅.
1.2 1) and 2) are correct, 3) is not.
1.3 Yes, ¦¦∅¦¦ is a singleton.
1.4 2, 3, 1, 2, 2, 2, 1, 2 for x =
1
2
and 1 if x =
1
2
.
1.5 (a) ¦1, 2, 3, 4¦; (b) ¦¦; (c) ¦−1, −2, −3, −4, −5, −6, . . . ¦
1.8 The set of solutions for a system of equations is equal to the in
tersection of the sets of solutions of individual equations belonging to the
system.
'2 Topology in a Set
'2
◦
1 Deﬁnition of Topological Space
Let X be a set. Let Ω be a collection of its subsets such that:
(a) the union of a collection of sets, which are elements of Ω, belongs to
Ω;
(b) the intersection of a ﬁnite collection of sets, which are elements of
Ω, belongs to Ω;
(c) the empty set ∅ and the whole X belong to Ω.
Then
• Ω is called a topological structure or just a topology
1
in X;
• the pair (X, Ω) is called a topological space;
• an element of X is called a point of this topological space;
• an element of Ω is called an open set of the topological space (X, Ω).
The conditions in the deﬁnition above are called the axioms of topological
structure.
'2
◦
2 The Simplest Examples
A discrete topological space is a set with the topological structure
which consists of all the subsets.
2.A. Check that this is a topological space, i.e., all axioms of topological
structure hold true.
An indiscrete topological space is the opposite example, in which the
topological structure is the most meager. It consists only of X and ∅.
2.B. This is a topological structure, is it not?
Here are slightly less trivial examples.
2.1. Let X be the ray [0, +∞), and Ω consists of ∅, X, and all the rays
(a, +∞) with a ≥ 0. Prove that Ω is a topological structure.
2.2. Let X be a plane. Let Σ consist of ∅, X, and all open disks with center
at the origin. Is this a topological structure?
2.3. Let X consist of four elements: X = ¦a, b, c, d¦. Which of the follow
ing collections of its subsets are topological structures in X, i.e., satisfy the
axioms of topological structure:
(a) ∅, X, ¦a¦, ¦b¦, ¦a, c¦, ¦a, b, c¦, ¦a, b¦;
(b) ∅, X, ¦a¦, ¦b¦, ¦a, b¦, ¦b, d¦;
(c) ∅, X, ¦a, c, d¦, ¦b, c, d¦?
The space of 2.1 is called an arrow. We denote the space of 2.3 (a) by .
It is a sort of toy space made of 4 points. Both of these spaces, as well as the
the arrow:
∞
space of 2.2, are not important, but provide good simple examples.
1
Thus Ω is important: it is called by the same word as the whole branch of
mathematics. Of course, this does not mean that Ω coincides with the subject of
topology, but everything in this subject is related to Ω.
12
§2. TOPOLOGY IN A SET 13
'2
◦
3 The Most Important Example: Real Line
Let X be the set R of all real numbers, Ω be the set of unions of all
intervals (a, b) with a, b ∈ R.
2.C. Check if Ω satisﬁes the axioms of topological structure.
This is the topological structure which is always meant when R is
considered as a topological space (unless other topological structure is
explicitly speciﬁed). This space is called usually the real line and the
structure is referred to as the canonical or standard topology in R.
'2
◦
4 Additional Examples
2.4. Let X be R, and Ω consists of empty set and all the inﬁnite subsets of
R. Is Ω a topological structure?
2.5. Let X be R, and Ω consists of empty set and complements of all ﬁnite
subsets of R. Is Ω a topological structure?
The space of 2.5 is denoted by R
T1
and called the line with T
1
topology.
2.6. Let (X, Ω) be a topological space and Y be the set obtained from X by
adding a single element a. Is
¦¦a¦ ∪ U : U ∈ Ω¦ ∪ ¦∅¦
a topological structure in Y ?
2.7. Is the set ¦∅, ¦0¦, ¦0, 1¦¦ a topological structure in ¦0, 1¦?
In Problem 2.6, if topology Ω discrete, the topology in Y is called a
particular point topology or topology of everywhere dense point. The topology
in Problem 2.7 is a particular point topology; it is called also the topology of
connected pair of points or Sierpinski topology.
2.8. List all the topological structures in a twoelement set, say, in ¦0, 1¦.
'2
◦
5 Using New Words: Points, Open and Closed Sets
Recall that, for a topological space (X, Ω), elements of X are called
points, and elements of Ω are called open sets.
2
2.D. Reformulate the axioms of topological structure using the words
open set wherever possible.
A set F ⊂ X is said to be closed in the space (X, Ω) if its complement
X F is open (i.e., X F ∈ Ω).
2
The letter Ω stands for the letter O which is the initial of the words with the
same meaning: Open in English, Otkrytyj in Russian, Oﬀen in German, Ouvert in
French.
§2. TOPOLOGY IN A SET 14
'2
◦
6 SetTheoretic Digression. De Morgan Formulas
2.E. Let Γ be an arbitrary collection of subsets of a set X. Then
(3) X
¸
A∈Γ
A =
¸
A∈Γ
(X A)
(4) X
¸
A∈Γ
A =
¸
A∈Γ
(X A).
Formula (4) is deduced from (3) in one step, is it not? These formulas are
nonsymmetric cases of a single formulation, which contains in a symmetric
way sets and their complements, unions and intersections.
2.9. Enigma. Find such a formulation.
'2
◦
7 Properties of Closed Sets
2.F. Prove that:
(a) the intersection of any collection of closed sets is closed;
(b) union of any ﬁnite number of closed sets is closed;
(c) empty set and the whole space (i.e., the underlying set of the topo
logical structure) are closed.
'2
◦
8 Being Open or Closed
Notice that the property of being closed is not a negation of the
property of being open.
(They are not exact antonyms in everyday usage, too).
2.G. Find examples of sets, which
(a) are both open, and closed simultaneously;
(b) are neither open, nor closed.
2.10. Give an explicit description of closed sets in
(a) a discrete space; (b) an indiscrete space;
(c) the arrow; (d) ;
(e) R
T1
.
2.H. Is a closed segment [a, b] closed in R.
Concepts of closed and open sets are similar in a number of ways.
The main diﬀerence is that the intersection of an inﬁnite collection of
open sets does not have to be necessarily open, while the intersection of
any collection of closed sets is closed. Along the same lines, the union
of an inﬁnite collection of closed sets is not necessarily closed, while the
union of any collection of open sets is open.
§2. TOPOLOGY IN A SET 15
2.11. Prove that the halfopen interval [0, 1) is neither open nor closed in R,
but can be presented as either the union of closed sets or intersection of open
sets.
2.12. Prove that every open set of the real line is a union of disjoint open
intervals.
2.13. Prove that the set A = ¦0¦ ∪
1
n
∞
n=1
is closed in R.
'2
◦
9 Cantor Set
Let K be the set of real numbers which can be presented as sums of series of the
form
¸
∞
k=1
a
k
3
k
with a
k
= 0 or 2. In other words, K is the set of real numbers which
in the positional system with base 3 are presented as 0.a
1
a
2
. . . a
k
. . . without digit 1.
2:A. Find a geometric description of K.
2:A.1. Prove that
(a) K is contained in [0, 1],
(b) K does not intersect
1
3
,
2
3
,
(c) K does not intersect
3s+1
3
k
,
3s+2
3
k
for any integers k and s.
2:A.2. Present K as [0, 1] with an inﬁnite family of open intervals removed.
2:A.3. Try to draw K.
The set K is called the Cantor set. It has a lot of remarkable properties and is
involved in numerous problems below.
2:B. Prove that K is a closed set in the real line.
'2
◦
10 Characterization of Topology in Terms of Closed Sets
2.14. Prove that if a collection T of subsets of X satisﬁes the following
conditions:
(a) the intersection of any family of sets from T belongs to T;
(b) the union of any ﬁnite number sets from T belongs to T;
(c) ∅ and X belong to T,
then T is the set of all closed sets of a topological space (which one?).
2.15. List all collections of subsets of a threeelement set such that there
exist topologies, in which these collections are complete sets of closed sets.
'2
◦
11 Topology and Arithmetic Progressions
2.16*. Consider the following property of a subset F of the set N of natural
numbers: there exists N ∈ N such that F does not contain an arithmetic
progression of length greater than N. Prove, that subsets with this property
together with the whole N form a collection of closed subsets in some topology
in N.
Solving this problem, you probably are not able to avoid the following
combinatorial theorem.
§2. TOPOLOGY IN A SET 16
2.17 Van der Waerden’s Theorem*. For every n ∈ N there exists
N ∈ N such that for any A ⊂ ¦1, 2, . . . , N¦, either A or ¦1, 2, . . . , N¦ A
contains an arithmetic progression of length n.
'2
◦
12 Neighborhoods
By a neighborhood of a point one means any open set containing
this point. Analysts and French mathematicians (following N. Bourbaki)
prefer a wider notion of neighborhood: they use this word for any set
containing a neighborhood in the sense above.
2.18. Give an explicit description of all neighborhoods of a point in
(a) a discrete space; (b) an indiscrete space;
(c) the arrow; (d) ;
(e) connected pair of points; (f) particular point topology.
Proofs and Comments
2.A What should we check? The ﬁrst axiom reads here that the
union of any collection of subsets of X is a subset of X? Well, this is
right. If A ⊂ X for each A ∈ Γ then ∪
A∈Γ
A ⊂ X. Indeed, take arbitrary
point b ∈ ∪
A∈Γ
A. Since it belongs to the union, it belongs to at least one
of A ∈ Γ, and since A ⊂ X, it belongs to X. Exactly in the same way
one checks the second axiom. Finally, of course, ∅ ⊂ X and X ⊂ X.
2.B Yes, it is. Here we can list all the collections of sets that we
need to consider. If one of the united sets is X then the union is X.
What if it is not there? Then what is there? Empty set, at most. Then
the union is also empty. With intersections the situation is simialr. If
one of the sets to intersect is ∅ the the intersection is ∅. If it is not
there, then what is? Only the whole X. Then the intersection equals X.
2.C First, show that
¸
A∈Γ
A∩
¸
B∈Σ
B =
¸
A∈Γ,B∈Σ
(A∩ B). Therefore if
A and B are intervals then the righthand side is a union of intervals.
If you think that a set which is a union of intervals is too simple,
please, try to answer the following question (which has nothing to do with
the problem under consideration, though). Let ¦r
n
¦
∞
n=1
= Q (i. e., we
numbered all the rational numbers). Prove that
¸
(r −2
−n
; r +2
−n
) = R,
although this is a union of some intervals, which contains all (!) the
rational numbers.
2.D The union of any collection of open sets is open. The intersec
tion of any ﬁnite collection of open sets is open. The empty set and the
whole space are open.
§2. TOPOLOGY IN A SET 17
2.E
(a)
x ∈ ∩
A∈Γ
(X A) ⇐⇒ ∀ A x ∈ X A ⇐⇒
∀ λ x / ∈ A
λ
⇐⇒ x / ∈ ∪
A∈Γ
A ⇐⇒
x ∈ X ∪
A∈Γ
A
(b) Replace both sides of the formula by their complements in X and
put B = X A.
2.G In any topological space the empty set and the whole space
are both open and closed. In a discrete space any set is both open and
closed. Semiopen interval is neither open nor closed on the line. Cf. also
the next problem.
2.H Yes, it is, because R [a; b] = (−∞; a) ∪ (b; +∞) is open.
Hints, Comments, Advises, Solutions, and Answers
2.1 The solution is based on the equality ∪(a
α
; +∞) = (inf a
α
; +∞).
Prove it. By the way the collection of closed rays [a; +∞) is not a topological
structure, since it may happen that ∪[a
α
; +∞) = (a
0
; +∞) (ﬁnd an example).
2.2 Yes, it is. A proof coincides almost literally with the solution of the
preceding problem.
2.3 The main point here is to realize that the axioms of topological
structure are conditions on the collection of subsets and if these conditions
are satisﬁed then the collection is called a topological structure. The second
collection is not a topological structure, because the sets ¦a¦, ¦b, d¦ are con
tained in it, while ¦a, b, d¦ = ¦a¦ ∪ ¦b, d¦ is not. Find two elements of the
third collection such that there intersection does not belong to it. By this
you would prove that this is not a topology. Finally, it is not diﬃcult to see
that all the unions and intersections of elements of the ﬁrst collection still
belong to the ﬁrst collection.
2.10 The following sets are closed
(a) in a disctrete space: all sets;
(b) in an indiscrete: only those which are also open, that is the empty set
and the whole space;
(c) in the arrow: ∅, the whole space and segments of the form [0; a];
(d) in : sets X, ∅, ¦b, c, d¦, ¦a, c, d¦, ¦b, d¦, ¦d¦, ¦c, d¦;
(e) in R
T1
: all ﬁnite sets and the whole R.
2.11 Here it is important to overcome the feeling that the question is
completely obvious. Why is not (0, 1] open? If (0; 1] = ∪(a
α
; b
α
) then 1 ∈
(a
α0
; b
α0
) for some α
0
, hence b
α0
> 1, and it follows that ∪(a
α
; b
α
) = (0; 1].
Similarly
R (0; 1] = (−∞; 0] ∪ (1; +∞)
§2. TOPOLOGY IN A SET 18
is not open. On the other hand,
(0; 1] =
∞
¸
n=1
¸
1
n
; 1
=
∞
¸
n=1
0;
n + 1
n
.
2.14 Check that Ω = ¦U [ X U ∈ T¦ is a topological structure.
2.15 Control indication: there number of such collections is 14.
2.16 The conditions (a) and (c) from 2.14 are obviously satisﬁed. To
prove (b), let us use 2.17. Let sets A and B do not contain arithmetic
progression of length ≥ n. If the set A ∪ B contained a suﬃciently long
progression, in one of the original sets there would be a progression of length
n.
2.18 By this point you have to learn already everything needed for
solving this problem, and must solve it on your own. Please, don’t be lazy.
'3 Bases
'3
◦
1 Deﬁnition of Base
Usually the topological structure is presented by describing its part,
which is suﬃcient to recover the whole structure. A collection Σ of open
sets is called a base for a topology, if each nonempty open set is a union
of sets belonging to Σ. For instance, all intervals form a base for the real
line.
3.1. Are there diﬀerent topological structures with the same base?
3.2. Find some bases of topology of
(a) a discrete space; (b) ;
(c) an indiscrete space; (d) the arrow.
Try to choose the bases as small as possible.
3.3. Describe all topological structures having exactly one base.
3.4. Prove that any base of the canonical topology in R can be diminished.
'3
◦
2 When a Collection of Sets is a Base
3.A. A collection Σ of open sets is a base for the topology, iﬀ for any open
set U and any point x ∈ U there is a set V ∈ Σ such that x ∈ V ⊂ U.
3.B. A collection Σ of subsets of a set X is a base for some topology in
X, iﬀ X is a union of sets of Σ and intersection of any two sets of Σ is
a union of sets in Σ.
3.C. Show that the second condition in 3.B (on intersection) is equiva
lent to the following: the intersection of any two sets of Σ contains, to
gether with any of its points, some set of Σ containing this point (cf. 3.A).
'3
◦
3 Bases for Plane
Consider the following three collections of subsets of R
2
:
• Σ
2
which consists of all possible open disks (i.e., disks without its
boundary circles);
• Σ
∞
which consists of all possible open squares (i.e., squares without
their sides and vertices) with sides parallel to the coordinate axis;
• Σ
1
which consists of all possible open squares with sides parallel to the
bisectors of the coordinate angles.
(Squares of Σ
∞
and Σ
1
are deﬁned by inequalities max¦[x−a[, [y−b[¦ < ρ
and [x −a[ +[y −b[ < ρ, respectively.)
3.5. Prove that every element of Σ
2
is a union of elements of Σ
∞
.
3.6. Prove that intersection of any two elements of Σ
1
is a union of elements
of Σ
1
.
3.7. Prove that each of the collections Σ
2
, Σ
∞
, Σ
1
is a base for some topo
logical structure in R
2
, and that the structures deﬁned by these collections
coincide.
19
§3. BASES 20
Figure 1. Elements of Σ
∞
(left) and Σ
1
(right).
'3
◦
4 Subbases
Let (X, Ω) be a topological space. A collection ∆ of its open subsets is
called a subbase for Ω, provided the collection
Σ = ¦V [ V = ∩
k
i=1
W
i
, W
i
∈ ∆, k ∈ N¦
of all ﬁnite intersections of sets belonging to ∆ is a base for Ω.
3.8. Prove that for any set X a collection ∆ of its subsets is a subbase of a
topology in X, iﬀ ∆ = ∅ and X = ∪
W∈∆
W.
'3
◦
5 Inﬁniteness of the Set of Prime Numbers
3.9. Prove that all inﬁnite arithmetic progressions consisting of natural num
bers form a base for some topology in N.
3.10. Using this topology prove that the set of all prime numbers is inﬁnite.
'3
◦
6 Hierarchy of Topologies
If Ω
1
and Ω
2
are topological structures in a set X such that Ω
1
⊂ Ω
2
then Ω
2
is said to be ﬁner than Ω
1
, and Ω
1
coarser than Ω
2
. For instance,
among all topological structures in the same set the indiscrete topology
is the coarsest topology, and the discrete topology is the ﬁnest one, is it
not?
3.11. Show that T
1
topology (see Section '2) is coarser than the canonical
topology in the real line.
3.12. Enigma. Let Σ
1
and Σ
2
be bases for topological structures Ω
1
and Ω
2
in a set X. Find necessary and suﬃcient condition for Ω
1
⊂ Ω
2
in terms of
the bases Σ
1
and Σ
2
without explicit referring to Ω
1
and Ω
2
(cf. 3.7).
Bases deﬁning the same topological structure are said to be equivalent.
3.D. Enigma. Formulate a necessary and suﬃcient condition for two
bases to be equivalent without explicit mentioning of topological struc
tures deﬁned by the bases. (Cf. 3.7: bases Σ
2
, Σ
∞
, and Σ
1
must satisfy
the condition you are looking for.)
§3. BASES 21
Proofs and Comments
3.A Let Σ be a base of Ω and U ∈ Ω. Present U as a union of
elements of Σ. Each point x ∈ U is contained in some of these sets. Such
a set can be chosen as V . It is contained in U, since it participates in a
union which is equal to U.
Vice versa, assume that for any U ∈ Ω and any point x ∈ U there
exists a set V ∈ Σ such that x ∈ V ⊂ U, and show that Σ is a base of
Ω. For this we need to prove that any U ∈ Ω can be represented as a
union of elements of Σ. For each point x ∈ U choose according to the
assumption a set V
x
∈ Σ such that x ∈ V
x
⊂ U and consider ∪
x∈U
V
x
.
Notice that ∪
x∈U
V
x
⊂ U, since V
x
⊂ U for each x ∈ U. On the other
hand, each point x ∈ U is contained in its V
x
and hence in ∪
x∈U
V
x
.
Therefore U ⊂ ∪
x∈U
V
x
. Thus, U = ∪
x∈U
V
x
.
3.B Assume that Σ is a base of a topology. Then X, being an open
set in any topology, can be presented as a union of some sets belonging to
[GS. The intersection of any two sets belonging to Σ is open, therefore
it also can be presented as a union of base sets.
Vice versa, assume that Σ is a collection of subsets of X such that
X is a union of sets belonging to Σ and the intersection of any two sets
belonging to Σ is a union of sets belonging to Σ. Let us prove that the
set of unions of all the collections of elements of Σ satisﬁes the axioms
of topological structure. The ﬁrst axiom is obviously satisﬁed, since the
union of some unions is a union. Let us prove the second axiom (the
intersection of two open sets is open). Let U = ∪
α
A
α
V = ∪
β
B
β
,
A
α
, B
β
∈ Σ. Then U ∩ V = (∪
α
A
α
) ∩ (∪
β
B
β
) = ∪
α,β
(A
α
∩ B
β
), and
since, by the assumpiton, A
α
∩B
β
can be presented as union of elements
of Σ, the intersection U ∩ V can be presented in this form, too. In the
third axiom, we need to check only the part concerning the whole X. By
the assumption, X is a union of sets belonging to Σ.
3.D Let Σ
1
and Σ
2
be bases of topological structures Ω
1
and Ω
2
in a
set X. Obviously, Ω
1
⊂ Ω
2
, iﬀ ∀ U ∈ Σ
1
∀ x ∈ U ∃ V ∈ Σ
2
: x ∈ V ⊂ U.
Now recall that Ω
1
= Ω
2
⇐⇒ Ω
1
⊂ Ω
2
and Ω
2
⊂ Ω
1
.
Hints, Comments, Advises, Solutions, and Answers
3.1 Of course, not! A topological structure is recovered from its base
as the set of unions of all collections of sets which belong to the base.
3.2
(a) A discrete space admits a base consisting of all onepoint subsets of the
space and this base is minimal. (why?)
(b) For a base in one can take, say, ¦a¦, ¦b¦, ¦a, c¦, ¦a, b, c, d¦.
§3. BASES 22
(c) In indiscrete space the minimal base is formed by a single set, the whole
space.
(d) In the arrow ¦[0, +∞), (r, +∞)¦
r∈Q+
is a base.
3.3 The whole topological structure is its own base. So, the question is
when this is the only base. In such a space any open set cannot be represented
as a union of two open sets distinct from it. Hence open sets are linearly
ordered by inclusion. Moreover, the space should contain only ﬁnite number
of open sets, since otherwisean open set could be obtained as a union of
inﬁnite increasing sequence of open sets.
3.4 We will show that removing of any element from any base of the
standard topology of the line gives a base of the same topology! Let U be an
arbitrary element of a base. It can be presented as a union of open intervals,
which are shorter than distance between some two points of U. We would need
at least two such intervals. Each of the intervals, in turn, can be presented as
a union of sets of the base under consideration. U is not involved into these
unions, since it is not contained in so short intervals. Hence U is a union of
elements of the base distinct from U and it can be replaced by this union in
a presentation of an open set as a union of elements of the base.
3.5, 3.6 In solution of each of these problems the following easy lemma
may help: A =
¸
B
α
, where B
α
∈ B, iﬀ ∀ x ∈ A ∃ B
x
∈ B : x ∈ B
x
⊂ A.
3.7 The statement: “B is a base of a topological structure” is equivalent
to the following: the set of unions of all collections of sets belonging to B is a
topological structure. Σ
1
is a base of some topology by 3.B and 3.6. So, you
need to prove analogues of 3.6 for Σ
2
and Σ
∞
. To prove that the structures
deﬁned, say, by bases Σ
1
and Σ
2
, you need to prove that a union of disks can
be presented as a union of squares and vice versa. Is it enough to prove that
a disk is a union of squares? What is the simplest way to do this (cf. our
advice concerning 3.5 and 3.6)?
3.9 Observe that intersection of arithmetic progressions is an arithmetic
progression.
3.10 Since the sets ¦i, i +d, i +2d, . . .¦, i = 1, . . . , d, are open, pairwise
disjoint and cover the whole N, it follows that each of them is closed. In
particular, for each prime number p the set ¦p, 2p, 3p, . . .¦ is closed. All
together the sets of the form ¦p, 2p, 3p, . . .¦ cover N¦1¦. Hence if the set of
prime numbers was ﬁnite, the set ¦1¦ would be open. But it cannot presented
as union of arithmetic progressions.
3.11 Inclusion Ω
1
⊂ Ω
2
means that a set open in the ﬁrst topology
(i.e., belonging to Ω
1
) belongs also to Ω
2
. Therefore, you should prove that
R ¦x
i
¦
n
i=1
is open in the canonical topology of the line.
3.12 Ω
1
⊂ Ω
2
, iﬀ ∀ U ∈ Σ
1
∀ x ∈ U ∃ V ∈ Σ
2
: x ∈ V ⊂ U.
'4 Metric Spaces
'4
◦
1 Deﬁnition and First Examples
A function ρ : X X → R
+
= ¦ x ∈ R [ x ≥ 0 ¦ is called a metric
(or distance) in X, if
(a) ρ(x, y) = 0, iﬀ x = y;
(b) ρ(x, y) = ρ(y, x) for every x, y ∈ X;
(c) ρ(x, y) ≤ ρ(x, z) +ρ(z, y) for every x, y, z ∈ X.
The pair (X, ρ), where ρ is a metric in X, is called a metric space.
The condition (c) is triangle inequality.
4.A. Prove that for any set X
ρ : X X →R
+
: (x, y) →
0, if x = y;
1, if x = y
is a metric.
4.B. Prove that R R →R
+
: (x, y) →[x −y[ is a metric.
4.C. Prove that R
n
R
n
→R
+
: (x, y) →
¸
n
i=1
(x
i
−y
i
)
2
is a metric.
Metrics 4.B and 4.C are always meant when R and R
n
are considered
as metric spaces unless another metric is speciﬁed explicitly. Metric 4.B
is a special case of metric 4.C. These metrics are called Euclidean.
'4
◦
2 Further Examples
4.1. Prove that R
n
R
n
→R
+
: (x, y) →max
i=1,...,n
[x
i
−y
i
[ is a metric.
4.2. Prove that R
n
R
n
→R
+
: (x, y) →
¸
n
i=1
[x
i
−y
i
[ is a metric.
Metrics in R
n
introduced in 4.C–4.2 are included in inﬁnite series of the
metrics
ρ
(p)
: (x, y) →
n
¸
i=1
[x
i
−y
i
[
p
1
p
, p ≥ 1.
4.3. Prove that ρ
(p)
is a metric for any p ≥ 1.
4.3.1 H¨ older Inequality. Prove that
n
¸
i=1
x
i
y
i
≤
n
¸
i=1
x
p
i
1/p
n
¸
i=1
y
q
i
1/q
if x
i
, y
i
≥ 0, p, q > 0 and
1
p
+
1
q
= 1.
Metric of 4.C is ρ
(2)
, metric of 4.2 is ρ
(1)
, and metric of 4.1 can be denoted
by ρ
(∞)
and adjoined to the series since
lim
p→+∞
n
¸
i=1
a
p
i
1
p
= max a
i
,
for any positive a
1
, a
2
, . . . , a
n
.
23
§4. METRIC SPACES 24
4.4. Enigma. How is this related to Σ
2
, Σ
∞
, and Σ
1
from Section '3?
For a real number p ≥ 1 denote by l
(p)
the set of sequences x = ¦x
i
¦
i=1,2,...
such that the series
¸
∞
i=1
[x[
p
converges.
4.5. Prove that for any two elements x, y ∈ l
(p)
the series
¸
∞
i=1
[x
i
− y
i
[
p
converges and that
(x, y) →
∞
¸
i=1
[x
i
−y
i
[
p
1
p
, p ≥ 1
is a metric in l
(p)
.
'4
◦
3 Balls and Spheres
Let (X, ρ) be a metric space, let a be its point, and let r be a positive
real number. The sets
B
r
(a) = ¦ x ∈ X [ ρ(a, x) < r ¦, (5)
D
r
(a) = ¦ x ∈ X [ ρ(a, x) ≤ r ¦, (6)
S
r
(a) = ¦ x ∈ X [ ρ(a, x) = r ¦ (7)
are called, respectively, open ball, closed ball, and sphere of the space
(X, ρ) with center at a and radius r.
'4
◦
4 Subspaces of a Metric Space
If (X, ρ) is a metric space and A ⊂ X, then the restriction of metric
ρ to AA is a metric in A, and (A, ρ
A×A
) is a metric space. It is called
a subspace of (X, ρ).
The ball D
1
(0) and sphere S
1
(0) in R
n
(with Euclidean metric, see
4.C) are denoted by symbols D
n
and S
n−1
and called ndimensional ball
and (n − 1)dimensional sphere. They are considered as metric spaces
(with the metric restricted from R
n
).
4.D. Check that D
1
is the segment [−1, 1]; D
2
is a disk; S
0
is the pair
of points ¦−1, 1¦; S
1
is a circle; S
2
is a sphere; D
3
is a ball.
The last two statements clarify the origin of terms sphere and ball (in
the context of metric spaces).
Some properties of balls and spheres in arbitrary metric space resem
ble familiar properties of planar disks and circles and spatial balls and
spheres.
4.E. Prove that for points x and a of any metric space and any r >
ρ(a, x)
D
r−ρ(a,x)
(x) ⊂ D
r
(a).
4.6. Enigma. What if r < ρ(x, a)? What is an analogue for the statement
of Problem 4.E in this case?
§4. METRIC SPACES 25
'4
◦
5 Surprising Balls
However in other metric spaces balls and spheres may have rather sur
prising properties.
4.7. What are balls and spheres in R
2
with metrics of 4.1 and 4.2 (cf. 4.4)?
4.8. Find D
1
(a), D1
2
(a), and S1
2
(a) in the space of 4.A.
4.9. Find a metric space and two balls in it such that the ball with the
smaller radius contains the ball with the bigger one and does not coincide
with it.
4.10. What is the minimal number of points in the space which is required
to be constructed in 4.9.
4.11. Prove that in 4.9 the big radius does not exceed double the smaller
radius.
'4
◦
6 Segments (What Is Between)
4.12. Prove that the segment with end points a, b ∈ R
n
can be described as
¦ x ∈ R
n
[ ρ(a, x) +ρ(x, b) = ρ(a, b) ¦,
where ρ is the Euclidean metric.
4.13. How do the sets deﬁned as in 4.12 look like with ρ of 4.1 and 4.2?
(Consider the case n = 2 if it appears to be easier.)
'4
◦
7 Bounded Sets and Balls
A subset A of a metric space (X, ρ) is said to be bounded, if there is a
number d > 0 such that ρ(x, y) < d for any x, y ∈ A. The greatest lower
bound of such d is called the diameter of A and denoted by diam(A).
4.F. Prove that a set A is bounded, iﬀ it is contained in a ball.
4.14. What is the relation between the minimal radius of such a ball and
diam(A)?
'4
◦
8 Norms and Normed Spaces
Let X be a vector space (over R). Function X → R
+
: x →[[x[[ is called
a norm if
(a) [[x[[ = 0, iﬀ x = 0;
(b) [[λx[[ = [λ[[[x[[ for any λ ∈ R and x ∈ X;
(c) [[x +y[[ ≤ [[x[[ +[[y[[ for any x, y ∈ X.
4.15. Prove that if x →[[x[[ is a norm then
ρ : X X →R
+
: (x, y) →[[x −y[[
is a metric.
The vector space equipped with a norm is called a normed space. The
metric deﬁned by the norm as in 4.15 turns the normed space into the metric
one in a canonical way.
§4. METRIC SPACES 26
4.16. Look through the problems of this section and ﬁgure out which of the
metric spaces involved are, in fact, normed vector spaces.
4.17. Prove that every ball in the normed space is a convex
3
set symmetric
with respect to the center of the ball.
4.18*. Prove that every convex closed bounded set in R
n
, which is symmet
ric with respect to its center and is not contained in any aﬃne space except
R
n
itself, is the unit ball with respect to some norm, and that this norm is
uniquely deﬁned by this ball.
'4
◦
9 Metric Topology
4.G. The collection of all open balls in the metric space is a base for
some topology (cf. 3.A, 3.B and 4.E).
4.G.1 Lemma. In any metric space, B
r
(a) ⊃ B
r−ρ(a,x)
(x) for any point a,
real number r > 0 and point x ∈ B
r
(a).
This topology is called metric topology. It is said to be induced by the
metric. This topological structure is always meant whenever the metric
space is considered as a topological one (for instance, when one says
about open and closed sets, neighborhoods, etc. in this space).
4.H. Prove that the standard topological structure in R introduced in
Section '2 is induced by metric (x, y) →[x −y[.
4.19. What topological structure is induced by the metric of 4.A?
4.I. A set is open in a metric space, iﬀ it contains together with any its
point a ball with center at this point.
'4
◦
10 Openness and Closedness of Balls and Spheres
4.20. Prove that a closed ball is closed (with respect to the metric topology).
4.21. Find a closed ball, which is open (with respect to the metric topology).
4.22. Find an open ball, which is closed (with respect to the metric topology).
4.23. Prove that a sphere is closed.
4.24. Find a sphere, which is open.
3
Recall that a set A is said to be convex if for any x, y ∈ A the segment connecting
x, y is contained in A. Of course, this deﬁnition is based on the notion of segment,
so it makes sense only for subsets of spaces, where the notion of segment connecting
two point is deﬁned. This is the case in vector and aﬃne spaces over R
§4. METRIC SPACES 27
'4
◦
11 Metrizable Topological Spaces
A topological space is said to be metrizable if its topological structure
is induced by some metric.
4.J. An indiscrete space is not metrizable unless it consists of a single
point (it has too few open sets).
4.K. A ﬁnite space is metrizable iﬀ it is discrete.
4.25. Which topological spaces described in Section '2 are metrizable?
'4
◦
12 Equivalent Metrics
Two metrics in the same set are said to be equivalent if they induce
the same topology.
4.26. Are the metrics of 4.C, 4.1, and 4.2 equivalent?
4.27. Prove that metrics ρ
1
, ρ
2
in X are equivalent if there are numbers
c, C > 0 such that
cρ
1
(x, y) ≤ ρ
2
(x, y) ≤ Cρ
1
(x, y)
for any x, y ∈ X.
4.28. Generally speaking the inverse is not true.
4.29. Enigma. Hence the condition of the equivalence of metrics formulated
in 4.27 can be weakened. How?
4.30. Metrics ρ
(p)
in R
n
deﬁned right above Problem 4.3 are equivalent.
4.31*. Prove that the following two metrics ρ
1
, ρ
C
in the set of all contin
uous functions [0, 1] →R are not equivalent:
ρ
1
(f, g) =
1
0
f(x) −g(x)
dx; ρ
C
(f, g) = max
x∈[0,1]
f(x) −g(x)
.
Is it true that topological structure deﬁned by one of them is ﬁner than
another?
'4
◦
13 Ultrametric
A metric ρ is called an ultrametric if it satisﬁes to ultrametric triangle inequality:
ρ(x, y) ≤ max¦ρ(x, z), ρ(z, y)¦
for any x, y, z.
A metric space (X, ρ) with ultrametric ρ is called an ultrametric space.
4:A. Check that only one metric in 4.A–4.2 is ultrametric. Which one?
4:B. Prove that in an ultrametric space all triangles are isosceles (i.e., for any three
points a, b, c two of the three distances ρ(a, b), ρ(b, c), ρ(a, c) are equal).
4:C. Prove that in a ultrametric space spheres are not only closed (cf. 4.23) but also
open.
§4. METRIC SPACES 28
The most important example of ultrametric is padic metric in the set Q of all
rational numbers. Let p be a prime number. For x, y ∈ Q, present the diﬀerence
x−y as
r
s
p
α
, where r, s, and α are integers, and r, s are relatively prime with p. Put
ρ(x, y) = p
−α
.
4:D. Prove that this is an ultrametric.
'4
◦
14 Operations with Metrics
4.32. Prove that if ρ
1
, ρ
2
are metrics in X then ρ
1
+ρ
2
and max¦ρ
1
, ρ
2
¦ are
also metrics. Are the functions min¦ρ
1
, ρ
2
¦,
ρ
1
ρ
2
, and ρ
1
ρ
2
metrics?
4.33. Prove that if ρ : X X →R
+
is a metric then
(a) function
(x, y) →
ρ(x, y)
1 +ρ(x, y)
is a metric;
(b) function
(x, y) →f
ρ(x, y)
is a metric, if f satisﬁes the following conditions:
(1) f(0) = 0,
(2) f is a monotone increasing function, and
(3) f(x +y) ≤ f(x) +f(y) for any x, y ∈ R.
4.34. Prove that metrics ρ and
ρ
1 +ρ
are equivalent.
'4
◦
15 Distance Between Point and Set
Let (X, ρ) be a metric space, A ⊂ X, b ∈ X. The inf¦ ρ(b, a) [ a ∈ A¦
is called a distance from the point b to the set A and denoted by ρ(b, A).
4.L. Let A be a closed set. Prove that ρ(b, A) = 0, iﬀ b ∈ A.
4.35. Prove that [ρ(x, A) − ρ(y, A)[ ≤ ρ(x, y) for any set A and points x, y
of the same metric space.
'4
◦
16 Distance Between Sets
Let A and B be bounded subsets in the metric space (X, ρ). Put
d
ρ
(A, B) = max
sup
a∈A
ρ(a, B), sup
b∈B
ρ(b, A)
¸
.
This number is called the Hausdorﬀ distance between A and B.
4:E. Prove that the Hausdorﬀ distance in the set of all bounded subsets of a metric
space satisﬁes the conditions (b) and (c) of the deﬁnition of metric.
4:F. Prove that for every metric space the Hausdorﬀ distance is a metric in the set
of its closed bounded subsets.
§4. METRIC SPACES 29
Let A and B be bounded polygons in the plane
4
. Put
d
∆
(A, B) = S(A) +S(B) −2S(A∩ B),
where S(C) is the area of polygon C.
4:G. Prove that d
∆
is a metric in the set of all plane bounded polygons.
We will call d
∆
the area metric.
4:H. Prove that in the set of all bounded plane polygons the area metric is not
equivalent to the Hausdorﬀ metric.
4:I. Prove that in the set of convex bounded plane polygons the area metric is equiv
alent to the Hausdorﬀ metric.
'4
◦
17 Asymmetrics
A function ρ : X X →R
+
is called an asymmetric in set X, if
(a) ρ(x, y) = 0 and ρ(y, x) = 0, iﬀ x = y;
(b) ρ(x, y) ≤ ρ(x, z) +ρ(z, y) for any x, y, z ∈ X.
Thus, an asymmetric satisﬁes the conditions a and c of the deﬁnition of metric,
but does not satisfy condition b.
An example of asymmetric taken from “the real life”: the shortest length of path
from one point to another by a car in a city in which there exist one way streets.
4:J. Prove that if ρ : X X →R
+
is an asymmetric then the function
(x, y) →ρ(x, y) +ρ(y, x)
is a metric in X.
Let A and B be bounded subsets of a metric space (X, ρ). The number a
ρ
(A, B) =
sup
b∈B
ρ(b, A) is called the asymmetric distance from A to B.
4:K. a
ρ
in the set of nounded subsets of a metric space satisﬁes the triangle inequality
from the deﬁnition of asymmetric.
4:L. In a metric space (X, ρ), a set B is contained in all the closed sets containing
A, iﬀ a
ρ
(A, B) = 0.
4:M. Prove that a
ρ
is an asymmetric in the set of all bounded closed subsets of a
metric space (X, ρ) .
Let A and B be polygons on the plane. Put
a
∆
(A, B) = S(B) −S(A ∩ B) = S(B A),
where S(C) is the area of polygon C.
4:1. Prove that a
∆
is an asymmetric in the set of all planar polygons.
4
Although we assume that the notion of bounded polygon is wellknown from
elementary geometry, recall the deﬁnition. A bounded plane polygon is a set of the
points of a simple closed polygonal line and the points surrounded by this line. By
a simple closed polygonal line we mean a cyclic sequence of segments such that each
of them starts at the point where the previous one ﬁnishes and these are the only
pairwise intersections of the segments.
§4. METRIC SPACES 30
A pair (X, ρ), where ρ is an asymmetric in X, is called an asymmetric space. Of
course, any metric space is an asymmetric space, too. In an asymmetric space, balls
(open and closed) and spheres are deﬁned like in a metric space, see '4
◦
3.
4:N. The set of all open balls of an asymmetric space is a base of some topology.
This topology is said to be generated by the asymmetric.
4:2. Prove that formula a(x, y) = min(x − y, 0) deﬁnes an asymmetric in
[0, ∞), and that the topology generated by this asymmetric coincides with
the arrow topology, see '2
◦
2.
Proofs and Comments
4.A Indeed, it makes sense to check that all the conditions of the
deﬁnition of metric is satisﬁed for each combination of points x, y z.
4.B Triangle inequality in this case looks as follows [x − y[ ≤ [x −
z[ +[z −y[. Put a = x −z, b = z −y. This turns the triangle ineguality
to a wellknown inequality [a +b[ ≤ [a[ +[b[.
4.C As in the solution of Problem 4.B, the triangle inequality can be
rewritten as follows:
¸
n
i=1
(a
i
+b
i
)
2
≤
¸
n
i=1
a
2
i
+
¸
n
i=1
b
2
i
. By two
squaring followed by an obvious simpliﬁcation, this inequality is reduced
to the wellknown Cauchy inequality (
¸
a
i
b
i
)
2
≤
¸
a
2
i
¸
b
2
i
.
4.F Show that if d = diamA and a ∈ A then A ⊂ D
d
(a). Vice
versa: diamD
d
(a) ≤ 2d (cf. 4.11).
4.G.1 We have to prove that any point y ∈ B
r−ρ(a,x)
(x) belongs to
B
r
(a). In terms of distances, this means that ρ(y, a) < r, if ρ(y, x) <
r −ρ(a, x) and ρ(a, x) < r. By the triangle inequality, ρ(y, a) ≤ ρ(y, x) +
ρ(x, a). Replacing in the righthand side of the latter inequality the ﬁrst
summand by a greater number r −ρ(a, x), we get the desired inequality.
4.G It is claimed that Ω = ¦∪B
r
(x) [ r > 0, x ∈ X¦ is a topological
structure. This follows from Lemma 4.G.1 and Theorems 3.B and 3.C.
4.H For this metric, the balls are open intervals. Each open interval
in R appears as a ball. The standard topology in R is deﬁned by the base
consisting of all open intervals.
4.I If a set contains together with any of its points a ball with
center at this point, this set is the union of those balls. Thus, it is open
in the metric topology. If a ∈ U, where U is open, then a ∈ B
r
(x) and
B
r−ρ(a,x)
(a) ∈ B
r
(x) ⊂ U, see Lemma 4.G.1.
4.J An indiscrete space does not have enough open sets. For x, y ∈
X and r = ρ(x, y) > 0, the ball D
r
(x) is not empty and does not coincide
with the whole space.
§4. METRIC SPACES 31
4:A Clearly, the metric in 4.A is an ultrametric. The other metrics
are not: for each of them you can ﬁnd points x, y, z such that ρ(x, y) =
ρ(x, z) +ρ(z, y).
4:B The deﬁnition of ultrametric implies that no one of pairwise
distances between points a, b, c can be greater than each of the other
two.
4:C By 4:B, if y ∈ S
r
(x) and r > s > 0 then B
s
(y) ⊂ S
r
(x).
4:D Let x − z =
r
1
s
1
p
α
1
, z − y =
r
2
s
2
p
α
2
and α
1
≤ α
2
. Then: x −
y = p
α
1
r
1
s
1
+
r
2
s
2
p
α
2
−α
1
= p
α
1
r
1
s
2
+r
2
s
1
p
α
2
−α
1
s
1
s
2
, hence p(x, y) ≤ p
−α
1
=
max¦ρ(x, z), ρ(z, y)¦.
4.L Condition ρ(b, A) = 0 means that each ball centered at b meets
A. In turn, this means that b does not belong to the complement of A
(since A is closed).
Hints, Comments, Advises, Solutions, and Answers
4.2 Cf. 4.B.
4.4 Look for an answer in 4.7.
4.7 Squares with sides parallel to the coordinate axes and bisectors of
the coordinate angles, respectively.
4.8 D
1
(a) = X, D
1/2
(a) = ¦a¦, S
1/2
(a) = ∅.
4.9 For example, X = D
1
(0) ⊂ R
1
, and D
3/2
(5/6) ⊂ D
1
(0).
4.10 Three points suﬃce.
4.11 Let R > r and D
R
(b) ⊂ D
r
(a). Take c ∈ D
R
(b) and use the
triangle inequality ρ(b, c) ≤ ρ(b, a) +ρ(a, c).
4.12 Put u = b − x and t = x − a. The Cauchy inequality becomes
equality, iﬀ the vectors u and t has the same direction, i.e., x lies on the
segment connecting a and b.
4.13 For metric ρ
(p)
with p > 1 this set coincides with the segment
connecting a and b, and for metric ρ
(1)
it is a rectangular parallelipiped
whose opposite vertices are those two points.
4.14 See the proof of 4.F.
4.19 The discrete one.
4.20 Let us just remind you that you need to prove that X D
r
(a) =
¦x [ ρ(x, a) > r¦ is open.
4.23 Use the obvious equality X S
r
(a) = B
r
(a) ∪ (X D
r
(a)) and
the result of 4.20.
4.K For x ∈ X put r = min¦p(x, y)[y ∈ X x¦. Which points are in
D
r
(x)?
§4. METRIC SPACES 32
4.25 Only line and discrete spaces.
4.26 According to 3.7, for n = 2 metrics ρ
(2)
, ρ
(1)
ρ
(∞)
are equivalent;
similar arguments work for n > 2, too. However in this case it is more
convenient to use the result of the next problem: to show that for any pair of
metrics ρ
(p)
(1 ≤ p ≤ ∞) there exist appropriate constants c and C, required
in 4.27.
4.27 First, let us prove that Ω
2
⊂ Ω
1
, provided ρ
2
(x, y) ≤ Cρ
1
(x, y). In
deed, inequality ρ
2
≤ Cρ
1
implies B
(ρ1)
r
(a) ⊂ B
(ρ2)
Cr
. Now let us use Theorem
4.I. Inequality cρ
1
(x, y) ≤ ρ
2
(x, y) can be represented as ρ
1
(x, y) ≤
1
c
ρ
2
(x, y).
Hence Ω
1
⊂ Ω
2
.
4.28 Metrics ρ
1
(x, y) = [x −y[ ρ
2
(x, y) = arctan[x −y[ on the line are
equivalent, but obviously there is no number C such that ρ
1
≤ Cρ
2
.
4.29 Metrics ρ
1
, ρ
2
are equivalent, if there exist c, C, d > 0, such that
ρ
1
(x, y) ≤ d implies cρ
1
(x, y) ≤ ρ
2
(x, y) ≤ Cρ
1
(x, y).
4.31 Ω
1
⊂ Ω
2
, because ρ
1
(f, g) ≤ ρ
C
(f, g). On the other hand, there is
no ball centered at the origin for metric ρ
1
, which would not ﬁt to B
(ρC)
1
(0),
since ∀ ε > 0 ∃ f
[f[ < ε, max
[0,1]
[f[ ≥ 1.
4.32 Clearly in all the cases the only thing which is to be proved and
is not completely obvious is the triangle inequality. Forρ
1
+ ρ
2
it is obvious
either. Furthermore
ρ
1
(x, y) ≤ ρ
1
(x, z) +ρ
1
(z, y) ≤
max¦ρ
1
(x, z), ρ
2
(x, z)¦ + max¦ρ
1
(y, z), ρ
2
(y, z)¦.
A similar inequality holds true for ρ
2
(x, y), therefore max¦ρ
1
, ρ
2
¦ is a met
ric. Construct examples which would prove that neither min¦ρ
1
, ρ
2
¦, nor
ρ1
ρ2
,
nor ρ
1
ρ
2
is a metric (for this it would be enough to ﬁnd three points with
appropriate pairwise distances).
4.33 The latter statement is quite obvious. The ﬁrst and the second
ones follow from the last one for f(t) =
t
1+t
and f(t) = min 1, t, respectively.
Thus it suﬃes to check that these function satisfy the conditions of the last
statement.
4.34 Since
ρ
1+ρ
≤ ρ, and for ρ(x, y) ≤ 1 inequality
1
2
ρ(x, y) ≤
ρ(x,y)
1+ρ(x,y)
holds true, the statement follows from 4.29.
4:E Condition (b) is obviously satisﬁed. Put r(A, B) = sup
a∈A
ρ(a, B), so
that d
ρ
(A, B) = max¦r(A, B), r(B, A)¦. To prove that (c) is also satisﬁed, it
suﬁcies to prove that r(A, C) ≤ r(A, B) +r(B, C) for any sets A, B, C ⊂ X.
One can easily see that ρ(a, C) ≤ ρ(a, b) +ρ(b, C) for all a ∈ A, b ∈ B. Hence
ρ(a, C) ≤ ρ(a, b)+r(B, C), and therefore ρ(a, C) ≤ inf
b∈B
ρ(a, b)+r(B, C) =
ρ(a, B) +r(B, C) ≤ r(A, B) +r(B, C), which implies the desired inequality.
4:F By 4:E, d
ρ
satisﬁes (b) (c) from the deﬁnition of metric. From 4.L
it follows that if the Hausdorﬀ metric between closed sets A and B equals
zero then A ⊂ B and B ⊂ A, i. e. A = B. Thus d
ρ
satisﬁes (a).
4:G d
∆
(A, B) is the area of the symmetric diﬀerence of A and B, i.
e. the area of A∆B = (A B) ∪ (B A). The ﬁrst two axioms of metric
§4. METRIC SPACES 33
are obviously satisﬁed. Prove the triangle inequality using inclusion AB ⊂
(C B) ∪ (A C).
'5 Ordered Sets
This section is devoted to orders. They are structures in sets and
occupy in Mathematics a position almost as profound as topological
structures. After a short general introduction, we will focus on rela
tions between structures of these two types. Like metric spaces, partially
ordered sets give rise to natural topological structures. This is a source
of interesting and important examples of topological spaces. As we will
see later (in Section '19), virtually all ﬁnite topological spaces appear in
this way.
'5
◦
1 Strict Orders
A binary relation in a set X is a set of ordered pairs of elements of
X, that is a subset R ⊂ X X. Many relations are denoted by special
symbols, like ≺, or ¬, or ≡, or ∼. In the case, if such a notation is used,
there is a tradition to write xRy instead of writing (x, y) ∈ R. So, we
write x ¬ y, or x ∼ y, or x ≺ y, etc. This generalizes the notation for
classical binary relations =, <, >, ≤, ⊂.
A binary relation ≺ in a set X is called a strict partial order, or just
strict order, if it satisﬁes the following two conditions:
• Irreﬂexivity: There is no a ∈ X such that a ≺ a holds.
• Transitivity: a ≺ b and b ≺ c imply a ≺ c for any a, b, c ∈ X.
5.A Antisymmetry. Let ≺ be a strict partial order in a set X. There
exist no x, y ∈ X such that both x ≺ y and y ≺ x hold true.
5.B. Relation < in the set R of real numbers is a strict order.
Formula a ≺ b is read sometimes as “a is less than b” or “b is greater
than a”, but often it is read as “a is followed by a” or “a precedes b”.
The advantage of the latter two ways of reading is that then the relation
≺ is not associated too closely to the inequality between real numbers.
'5
◦
2 NonStrict Orders
Binary relation _ in a set X is called a nonstrict partial order, or just
nonstrict order, if it satisﬁes the following three conditions:
• Transitivity: If a _ b and b _ c then a _ c for any a, b, c ∈ X.
• Antisymmetry: If a _ b and b _ a then a = b for any a, b ∈ X.
• Reﬂexivity: a _ a for any a ∈ X.
5.C. Relation ≤ in R is a nonstrict order.
5.D. In the set N of natural numbers the relation a[b (a divides b) is a
nonstrict partial order.
5.1. Is the relation a[b a nonstrict partial order in the set Z of integers?
34
§5. ORDERED SETS 35
5.E. In the set of subsets of a set X inclusion is a nonstrict partial
order.
'5
◦
3 Relation between Strict and NonStrict Orders
5.F. For each strict order ≺, there is a relation _ deﬁned in the same
set as follows: a _ b, if either a ≺ b or a = b. This relation is a nonstrict
order.
One says that the nonstrict order _ of 5.F is associated to the
original strict order ≺.
5.G. For each nonstrict order _, there is a relation ≺ deﬁned in the
same set as follows: a ≺ b, if a _ b and a = b. This relation is a strict
order.
One says that the strict order ≺ of 5.G is associated to the original
nonstrict order _.
5.H. The construction of the two preceding problems are inverse to each
other: applied one after another in any order, they give the initial rela
tion.
Thus, strict and nonstrict orders are deﬁned by each other. They
are just diﬀerent incarnations of the same structure of order. We have
already met with a similar phenomenon in topology: open and closed
sets in a topological space deﬁne each other and provide diﬀerent ways
of presenting a topological structure.
A set equipped with a partial order (either strict or nonstrict) is
called a partially ordered set or poset. More formally speaking, a partially
ordered set is a pair (X, ≺) formed by a set X and a strict partial order
≺ in X. Certainly, instead of a strict partial order ≺ one can use the
corresponding nonstrict one _.
Which of the orders, strict or nonstrict, prevails in each concrete
case is a matter of convenience, taste and tradition. Although it would
be handy to keep both of them available, nonstrict orders conquer situ
ation by situation. For instance, nobody introduces notation for a strict
divisibility. Another example: symbol ⊆, which used to denote nonstrict
inclusion, is replaced by symbol ⊂, which is almost never understood as
notation solely for strict inclusion.
In abstract considerations we will use both kinds of orders, a strict
partial order is denoted by symbol ≺, a nonstrict one by symbol _.
'5
◦
4 Cones
Let (X, ≺) be a poset and a ∈ X. The set ¦x ∈ X [ a ≺ x¦ is called
the upper cone of a, and the set ¦x ∈ X [ x ≺ a¦ the — lower cone of
§5. ORDERED SETS 36
a. The element a does not belong to its cones. By adding it to them, we
get completed cones: the upper completed cone or star C
+
X
(a) = ¦x ∈ X [
a _ x¦ and the lower completed cone C
−
X
(a) = ¦x ∈ X [ x _ a¦.
5.I Properties of Cones. Let (X, ≺) be a poset.
(a) C
+
X
(b) ⊂ C
+
X
(a), provided b ∈ C
+
X
(a);
(b) a ∈ C
+
X
(a) for each a ∈ X.
(c) C
+
X
(a) = C
+
X
(b) implies a = b;
5.J Cones Determine an Order. Let X be an arbitrary set. Suppose
for any a ∈ X one ﬁxes a subset C
a
⊂ X. If
(a) b ∈ C
a
implies C
b
⊂ C
a
;
(b) a ∈ C
a
for each a ∈ X.
(c) C
a
= C
b
implies a = b;
Let us write a ≺ b, if b ∈ C
a
. Then the relation ≺ is a nonstrict order
in X and for this order C
+
X
(a) = C
a
.
5.2. Let C ⊂ R
3
be a set. Consider relation
C
in R
3
, which is deﬁned as
follows: a
C
b, if b − a ∈ C. What properties of C would imply that
C
would be a partial order in R
3
? In the poset (R
3
,
C
), what are the upper
and lower cones?
5.3. Prove that any convex cone C in R
3
with vertex (0, 0, 0) such that
P ∩ C = (0, 0, 0) for some plane P satisﬁes the conditions found in solution
of the preceding problem.
5.4. In the spacetime R
4
of special relativity theory (where points rep
resent moment point events, the ﬁrst three coordinates, x
1
, x
2
, x
3
, are the
spatial coordinates and the fourth one, t, is the time) there is a relation event
(x
1
, x
2
, x
3
, t) precedes (and can inﬂuence) event (¯ x
1
, ¯ x
2
, ¯ x
3
,
¯
t). This relation
is deﬁned by inequality
c(
˜
t −t) ≥
(¯ x
1
−x
1
)
2
+ (¯ x
2
−x
2
)
2
+ (¯ x
3
−x
3
)
2
.
Is this a partual ordr? If yes, what are the upper and lower cones of an event?
5.5. Answer the versions of questions of the preceding problem concerning
twodimensional and threedimensional versions of this space, in which the
number of spatial coordinates is 1 and 2 respectively.
'5
◦
5 Position of an Element with Respect to a Set
Let (X, ≺) be a poset, A be its subset. We say that b is the greatest
element of A, if b ∈ A and c _ b for every c ∈ A. Similarly, b is said to
be the smallest element of A, if b ∈ A and b _ c for every c ∈ A.
5.K. b is the smallest element of A, iﬀ A ⊂ C
+
X
(b).
b is the greatest element of A, iﬀ A ⊂ C
−
X
(b).
5.L. Each set has at most one greatest and at most one smallest element.
An element b of a set A is called its maximal element, if A does not
contain an element c such that b ≺ c. An element b of a set A is called
its minimal element, if A does not contain an element c such that c ≺ b.
§5. ORDERED SETS 37
5.M. An element b of A is maximal, iﬀ A∩ C
−
X
(b) = b.
An element b of A is minimal, iﬀ A∩ C
+
X
(b) = b.
5.6. Enigma. How are the notions of maximal and greatest elements re
lated? What can you say about a poset, in which for any subset these notions
coincide?
'5
◦
6 Total Orders
Please, notice: the deﬁnition of a strict order does not require that
for any a, b ∈ X either a ≺ b, or b ≺ a, or a = b. This condition is called
a trichotomy. In terms of the corresponding nonstrict order, it can be
reformulated as follows: any two elements a, b ∈ X are comparable: either
a _ b, or a _ b.
A strict order which satisﬁes trichotomy is said to be total or linear.
The corresponding poset is said to be linearly or totally ordered. It is
called also just an ordered set.
5
Some orders do satisfy trichotomy.
5.N. The order < in the set R of real numbers is linear.
This is the most important example of totally ordered set. The words
and images rooted in it are often extended to all totally ordered set. For
example, cones are called rays, the upper cones turn to right rays, while
lower cones to left rays.
5.7. A poset (X, ≺) is linearly ordered, iﬀ X = C
+
X
(a) ∪ C
−
X
(a) for each
a ∈ X.
5.8. In the set N of natural numbers the order a[b is not linear.
5.9. For which X the relation of inclusion in the set of all subsets of X is a
linear order?
'5
◦
7 Topologies Deﬁned by a Total Order
5.O. Let (X, ≺) be a totally ordered set. The set of all its subsets con
sisting of all the right rays, i.e., sets of the form ¦x ∈ X [ a ≺ x¦, where
a runs over X, and the set X itself is a base of a topological structure in
X.
5
Quite a bit of confusion was brought into the terminology by Bourbaki. Then
total orders were called orders, nontotal orders were called partial orders, and in
occasions when it was not known if the order under consideration was total, the fact
that this was unknown was explicitly stated. Bourbaki suggested to withdraw the
word partial. The motivation for this was that a partial order, as a phenomenon more
general than a total order, deserves a shorter and simpler name. In French literature
this suggestion was commonly accepted, but in English it would imply abolishing of
a nice short word poset that seems to be absolutely impossible.
§5. ORDERED SETS 38
The topological structure deﬁned by this base is called the right ray
topology of linearly ordered set (X, ≺). Left ray topology is deﬁned sim
ilarly: it is generated by the base consisting of X and sets of the form
¦x ∈ X [ x ≺ a¦ with a ∈ X.
5.10. The topology of the arrow (see. '2) coincides with the right ray topol
ogy of halﬂine [0, ∞) equipped with the order <.
5.11. Enigma. To what extent is the assumption that the order is linear
necessary in Theorem 5.O? Find a weaker condition which would imply
the conclusion of Theorem 5.O and allow one to speak about the topological
structure described in Problem 2.2 as the right ray topology of an appropriate
partial order on the plane.
5.P. Let (X, ≺) be a totally ordered set. The set of its subsets consisting
of X and all sets of the form ¦x ∈ X [ a ≺ x ≺ b¦, where a and b run
over the whole X is a base of a topological structure in X.
The topological structure deﬁned by this base is called the interval
topology of a linearly ordered set.
5.12. Prove that the interval topology is the smallest topological structure
containing the right ray and left ray topological structures.
5.Q. The canonical topology of the line coincides with the interval topol
ogy of (R, <).
'5
◦
8 Poset Topology
5.R. Let (X, _) be a poset. The set of subsets of X consisting of all the
sets of form ¦x ∈ X [ a _ x¦, where a runs over the whole X, is a base
of a topological structure in X.
The topological strucuter generated by this base is called a poset
topology.
5.S. In a poset topology each point a ∈ X has the smallest (with respect
to inclusion) neighborhood. This is ¦x ∈ X [ a _ x¦.
5.T. The following properties of a topological space are equivalent:
• each point has a smallest neighborhood,
• the intersection of any collection of open sets is open,
• the union of any collection of closed sets is closed.
A space satisfying the conditions of Theorem 5.T is called a smallest
neighborhood space.
6
In a smallest neighborhood space open and closed
sets satisfy the same conditions. In particular, the set of all closed sets of
6
This class of topological spaces was introduced and studied by P. S. Aleksandrov
in 1935. Aleksandrov called them discrete. Nowadays the term discrete space is used
for a much narrower class of topological spaces (see Section '2). The term smallest
neighborhood space was introduced by Christer Kiselman.
§5. ORDERED SETS 39
a smallest neighborhood space is a topological structure. This structure
is said to be dual to the original one. It corresponds to the opposite
partial order.
5.13. How to characterize points open in a poset topology in terms of the
partial order? The same question about closed points.
5.14. Describe directly open sets in the poset topology of R with order <.
5.15. In set ¦a, b, c, d¦ consider a partial order in which the strict inequalities
are: c ≺ a, d ≺ c, d ≺ a, d ≺ b. Check that this is a partial order and that
the poset topology coincides with the topology of , described in Problem
2.3 (a).
'5
◦
9 How to Draw a Poset
Now we can explain a pictogram , which we use to denote the space
introduced in Problem 2.3 (a). It describes the partial order in ¦a, b, c, d¦,
which deﬁnes the topology of this space by 5.15. Indeed, if we place the
elements of the set under consideration at vertices of the
graph of the pictogram, as shown in the picture, the ver
tices corresponding to comparable elements occur to be
connected by a segment or ascending broken line and the
greater element correspond to the higher vertex.
d
c
a
b
In this way one can draw a scheme representing any ﬁnite poset. El
ements of the poset are represented by points and a ≺ b, iﬀ the point
repsenting b is above the point representing a and those points are con
nected either by a segment or broken line consisting of segments which
connect points representing intermediate elements of a chain a ≺ c
1
≺
c
2
≺ ≺ c
n
≺ b. One could connect by a segment any two points
corresponding to comparable elements, but this would make the diagram
excessively cumbersome. This is why segments which could be recovered
from the others by transitivity are not drawn. A diagram of this kind
representing a poset is called its Hasse diagram.
5.U. Prove that any ﬁnite poset can be described by a Hasse diagram.
5.V. Discribe the topological structure in set Z of integers which is the
poset topology deﬁned by the following Hasse diagram
0
−1
−2
−3
−4
−5 1
2
3
4
5
6
The space of Problem 5.V is called the digital line or Khalimsky line.
In this space each even number is closed and each odd open.
§5. ORDERED SETS 40
5.16. Associate to each even integer 2k the interval (2k −1, 2k +1) of length
2 centered at this point, and to each odd integer 2k−1, the singleton ¦2k−1¦.
Prove that a set of integers is open in the Khalimsky topology, iﬀ the union
of sets associated to its elements is open in R with the standard topology.
5.17. Among the topological spaces described in Section '2, ﬁnd all that
can be obtained as posets with the poset topology. In the cases of ﬁnite sets,
draw Hasse diagrams describing the corresponding partial orders.
'5
◦
10 Cyclic Orders in Finite Set
Recall that a cyclic order in a ﬁnite set X is a linear order considered
up to cyclic permutatuion. A linear order allows one to enumerate el
ements of the set X by natural numbers, so that X = ¦x
1
, x
2
, . . . , x
n
¦.
A cyclic permutation transpose the ﬁrst k elements with the last n − k
elements without change of the ordering inside of each of these two parts
of the set:
(x
1
, x
2
, . . . , x
k
, x
k+1
, x
k+2
, . . . , x
n
) →(x
x+1
, x
k+2
, . . . , x
n
, x
1
, x
2
, . . . , x
k
).
When we consider a cyclic order, it does not make sense to say that
one of its elements is greater than other one, since an appropriate cyclic
permutation put these two elements in the opposite order. However, it
makes sense to say that an element is followed immediately by other
one. Certainly, the very last element is followed immediately by the very
ﬁrst: indeed, any nonidentity cyclic permutation puts the ﬁrst element
immediately after the last one.
In a cyclicly ordered ﬁnite set for each element a there is a unique
element b next to a, that is which follows a immediatly. This deﬁnes a
map of the set onto itself, the simplest cyclic permutation
x
i
→
x
i+1
, if i < n
x
1
, if i = n.
This permutation acts transitively (i.e., any element is mapped to any
other one by an appropriate iteration of it).
5.W. Any map T : X → X, which acts transitively in X, deﬁnes a
cyclic order in X such that each a ∈ X is followed by T(a).
5.X. In a set consisting of n elements possesses exactly (n−1)! pairwise
distinct cyclic orders.
In particular, in a two element set there is only one cyclic order (which
so uninteresting that sometimes it is said to make no sense), and a set
consisting of three elements possesses two cyclic orders.
§5. ORDERED SETS 41
'5
◦
11 Cyclic Orders in Inﬁnite Sets
One can consider cyclic orders in an inﬁnite set. However most of what was said
above cannot apply to cyclic orders in inﬁnite sets without adjustment. In particular,
most of them cannot be described by showing pairs of elements which are next to each
other. For example, points of a circle can be cyclically ordered clockwise (or counter
clockwise), but with respect to this cyclic order no point follows is immediately by
other point.
Such “continuous” cyclic orders can be deﬁned almost in the same way as cyclic
orders in ﬁnite sets were deﬁned above. The diﬀerence is that sometimes one cannot
deﬁne cyclic permutations of the set in necessary quantity, and they have to be re
placed by cyclic tranformations of the linear orders. Namely, a cyclic order is deﬁned
as a linear order considered up to cyclic transformations, where by cyclic transforma
tion of a linear order ≺ in a set X we mean a passage from ≺ to a linear order ≺
such that X splits into subsets A and B such that the restrictions of ≺ ≺
to each
of them coincide, while a ≺ b and b ≺
a for any a ∈ A and b ∈ B.
5:A. Existense of a cyclic transformation turning linear orders to each other is an
equivalence relation on the set of all linear orders in a set.
A cyclic order in a set is an equivalence class of linear orders under the relation
of existence of a cyclic transformation.
5:B. Prove that for a ﬁnite set this deﬁnition is equivalent to the deﬁnition of the
preceding Section.
5:C. Prove that the cyclic “counterclockwise” order on a circle can be deﬁned along
the deﬁniton of this Section, but cannot be deﬁned as a linear order modulo cyclic
transformations of the set for whatever deﬁnition of cyclic transformations of cir
cle. Describe the linear orders on circle which deﬁne this cyclic order up to cyclic
transfomrations of orders.
5:D. Let A be a subset of a set X. If linear orders ≺
and ≺ on X can be obtained
from each other by a cyclic transformation, than their restrictions to A can also be
obtained from each other by a cyclic transformation.
5:E Corollary. A cyclic order in a set induces a welldeﬁned cyclic order in every
subset of this set.
5:F. A cyclic order in a set can be recovered from cyclic orders induced by it on all
its subsets consisting of three elements.
5:F.1. A cyclic order in a set X can be recovered from cyclic orders induced
by it all its subsets consisting of three elements and containing a ﬁxed
element a ∈ X.
Theorem 5:F allows one to describe acyclic order as a ternary relation. Namely,
for a set ¦a, b, c¦ denote by [a ≺ b ≺ c] a cyclic order deﬁned by the linear order in
which the inequalities presented in the brackets hold true (i.e., b follows by a and c
follows by b).
5:G. Cyclic orders induced in three element subsets of X induced by the same cyclic
order on X, satisfy the following requirements:
(a) [a ≺ a ≺ b] is not true for any a, b ∈ X;
(b) for any pairwise distinct a, b, c ∈ X either [a ≺ b ≺ c], or [b ≺ a ≺ c] is true,
but both cannot be true simultaneously;
§5. ORDERED SETS 42
(c) [a ≺ b ≺ c], iﬀ [b ≺ c ≺ a], iﬀ [c ≺ a ≺ b] a, b, c ∈ X;
(d) if [a ≺ b ≺ c] and [a ≺ c ≺ d], then [a ≺ b ≺ d].
Vice versa, a ternary relation having these four properties in a set X deﬁnes a cyclic
order in X.
'5
◦
12 Topology of Cyclic Order
5:H. Let X be a set with a cyclic order. The collection of sets which belong to
the interval topology of every linear order deﬁning this cyclic order is a topological
structure in X.
The topology deﬁned in 5:H is called a cyclic order topology.
5:I. A cyclic order topology deﬁned by cyclic counterclockwise order of S
1
coincides
with the topology deﬁned by the metric ρ(x, y) = [x −y[ on S
1
⊂ C.
Proofs and Comments
5.F We need to check that the relation “a ≺ b, or a = b” satisﬁes the
three conditions from the deﬁnition of a nonstrict order. Doing this, we
can use only the fact that ≺ satisﬁes the conditions from the deﬁnition
of a strict order. Let us check transitivity. Let a _ b and b _ c. This
meens that either a ≺ b and b ≺ c, or a = b and b ≺ c, or a ≺ b and
b = c, or a = b and b = c. In the ﬁrst case a ≺ c by transitivity of ≺,
hence a _ c; in the second case a = b ≺ c, hence a ≺ c and a _ c; in
the third case a ≺ b = c, hence a ≺ c and a _ c; ﬁnally, at the fourth
case a = b = c, hence a = c and a _ c. Other conditions are checked
similarly.
5.I The ﬁrst assertion follows from transitivity of the order. Indeed,
consider arbitrary c ∈ C
+
X
(b). By the deﬁnition of cone, b _ c, while
the condition b ∈ C
+
X
(a) means that a _ b. By transitivity this implies
a _ c, that is c ∈ C
+
X
(a). Thereby we proved that each element of C
+
X
(b)
belongs to C
+
X
(a). Hence C
+
X
(b) ⊂ C
+
X
(a), which is what we had to prove.
The second assertion follows from the deﬁnition of cone and reﬂexivity
of order. Indeed, by the deﬁnition, C
+
X
(a) consists of b such that a _ b,
and, by reﬂexivity of order, a _ a.
The third assertion follows similarly from antisymmetricity: the as
sumption C
+
X
(a) = C
+
X
(b) together with the second assertion implies
a _ b and b _ a, and this together with antisymmetricity implies that
a = b.
5.J By the preceding theorem 5.I, in a poset cones have the prop
erties which form the hypothesis of the theorem under consideration.
Proving Theorem 5.I, we have shown that these properties follows from
§5. ORDERED SETS 43
the corresponding conditions from the deﬁnition of partial nonstrict or
der. In fact, they are equivalent to these conditions. Permute words in
the proof of Theorem 5.I, to get a proof of Theorem 5.J.
5.O By Theorem 3.B, it suﬃces to prove that the intersection of
any two right rays is a union of right rays. Consider the intersection of
¦x ∈ X [ a ≺ x¦ and ¦x ∈ X [ b ≺ x¦. The order is total, therefore
either a ≺ b, or b ≺ a. Let a ≺ b. Then ¦x ∈ X [ a ≺ x¦ ∩ ¦x ∈ X [ b ≺
x¦ = ¦x ∈ X [ b ≺ x¦.
5.R By Theorem 3.C, it suﬃces to prove that any element of the
intersection of two conus, say, C
+
X
(a) and C
+
X
(b), is contained in the
intersection together with a whole cone of the same kind. Let c ∈ C
+
X
(a)∩
C
+
X
(b) and d ∈ C
+
X
(c). Then a _ c _ d and b _ c _ d, hence a _ d and
b _ d. Therefore d ∈ C
+
X
(a) ∩ C
+
X
(b). Hence, C
+
X
(c) ⊂ C
+
X
(a) ∩ C
+
X
(b).
5.T Equivalence of the second and third properties is proved using de
Morgan formulas, as 2.F. Let us prove that the ﬁrst property implies the
second one. Consider the intersection of an arbitrary collection of open
sets. For each of its points, every set of this collection is a neighborhood.
Therefore its smallest neighborhood is contained in every of the sets
which are to be intersected. Hence, the smallest neighborhood of the
point is contained in the intersection. Thus, each point of the intersection
is contained in the intersection together with its neighborhood. The
intersection is the union of these neighborhoods. Therefore it is open.
Now let us prove that if the intersection of any collection of open sets
is open then any point has a smallest neighborhood. Where can one get
such a neighborhood from? How to construct it? Take all the neigh
borhoods of a point and consider their intersection. By the assumption,
this intersection is open. It contains the point. Therefore this is a neigh
borhood of the point. This neighborhood, being the intersection of all
neighborhoods, is contained in each of the neighborhoods. Thus, this is
the smallest neighborhood.
5.V The minimal base of this topology consists of singletons of the
form¦2k−1¦ with k ∈ Z and threepoint sets of the form¦2k−1, 2k, 2k+
1¦, where again k ∈ Z.
Hints, Comments, Advises, Solutions, and Answers
5.1 No, for it is not antisymmetric. Indeed, −1[1 and 1[ −1, but −1 = 1.
5.2 The hypothesis of Theorem 5.J turn into the following restrictions
on C: it should be closed with respect to addition, contain the zero and no
nonidentity translation can map it bijectively onto itself.
§5. ORDERED SETS 44
5.6 Obviously, the greatest element is maximal and the smallest one is
minimal, but the opposite statements are not true. For any subset of a poset
these notions coincide, iﬀ any two elements of the poset are comparable (i. e.
one of them is greater than the other). Indeed, consider a twoelement subset.
If the two elements were incomparable, each of them would be maximal, and
hence the greates. But the greatest element is unique. This contradiction
proves that the elements are comparable. Vice versa, comparability of any
two elements implies obiously that in any subset any maximal element is the
greatest one, and any minimal, the smallest.
5.9 The relation of inclusion in the set of all subsets of X is a linear
order, iﬀ X is either empty, or singletone.
5.11 Consider, say, the following condition: for arbitrary a, b and c
such that a ≺ c and b ≺ c, there exists an element d such that a ≺ d, b ≺ d
and d ≺ c. Show that this condition holds true in any totally ordered set
and it implies that right rays form a base of a topology. Show also that this
condition holds true, if right rays form a base of a topology.
5.13 A point open in a poset topology is a point maximal in the whole
poset. Similarly, a point closed in a poset topology is a point miniimal in the
whole poset.
5.14 Rays of the forms (a, ∞) and [a, ∞), the empty set and the whole
line.
'6 Subspaces
'6
◦
1 Topology for a subset of a space
Let (X, Ω) be a topological space, and A ⊂ X. Denote by Ω
A
the
collection of sets A ∩ V , where V ∈ Ω.
6.A. Ω
A
is a topological structure in A.
The pair (A, Ω
A
) is called a subspace of the space (X, Ω). The col
lection Ω
A
is called the subspace topology or the relative topology or the
topology induced on A by Ω, and its elements are called open sets in A.
6.B. The canonical topology in R
1
and the topology induced on R
1
as
a subspace of R
2
coincide.
6.1. Enigma. How to construct a base for the topology induced on A using
the base for the topology in X?
6.2. Describe the topological structures induced
(a) on the set N of natural numbers by the topology of the real line;
(b) on N by the topology of the arrow;
(c) on the twopoint set ¦1, 2¦ by the topology of R
T1
;
(d) on the same set by the topology of the arrow.
6.3. Is the halfopen interval [0, 1) open in the segment [0, 2] considered as
a subspace of the real line?
6.C. A set is closed in a subspace, iﬀ it is the intersection of the subspace
and a closed subset of the ambient space.
'6
◦
2 Relativity of Openness
Sets, which are open in the subspace, are not necessarily open in the
ambient space.
6.D. The unique set open in R
1
, which is also open in R
2
, is ∅.
However:
6.E. An open set of an open subspace is open in the ambient space, i.
e., if A ∈ Ω then Ω
A
⊂ Ω.
The same relation holds true for closed sets. Sets, which are closed in
the subspace, are not necessarily closed in the ambient space. However:
6.F. Closed sets of the closed subspace are closed in the ambient space.
6.4. Prove that a set U is open in X, iﬀ every its point has a neighborhood
V in X such that U ∩ V is open in V .
It allows one to say that the property of being open is local. Indeed, we
can reformulate 6.4 as follows: a set is open, iﬀ it is open in a neighborhood
of each of its points.
45
§6. SUBSPACES 46
6.5. Show that the property of being closed is not local.
6.G Transitivity of Induced Topology. Let (X, Ω) be a topological
space, and X ⊃ A ⊃ B. Then (Ω
A
)
B
= Ω
B
, i.e., the topology induced on
B by the topology induced on A coincides with the topology induced on B
directly.
6.6. Let (X, ρ) be a metric space, and A ⊂ X. Then the topology in A
generated by metric ρ
A×A
coincides with the topology induced on A by the
topology in X generated by metric ρ.
6.7. Enigma. The statement 6.6 is equivalent to a couple of inclusions.
Which of them is less obvious?
'6
◦
3 Agreement on Notations of Topological Spaces
Diﬀerent topological structures in the same set are not considered
simultaneously very often. That is why a topological space is usually
denoted by the same symbol as the set of its points, i.e., instead of
(X, Ω) one writes just X. The same is applied for metric spaces: instead
of (X, ρ) one writes just X.
Proofs and Comments
6.A We need to check that Ω
A
satisﬁes the axioms of topological
structure. Consider the ﬁrst axiom. Let Γ ⊂ Ω
A
be a collection of
sets belonging to Ω
A
. We have to prove that ∪
U∈Γ
U ∈ Ω
A
. For each
U ∈ Γ ﬁnd U
X
∈ Ω such that U = A ∩ U
X
. This is possible due to the
deﬁnition of Ω
A
. Transform the union under consideration: ∪
U∈Γ
U =
∪
U∈Γ
(A ∩ U
X
) = A ∩ (∪
U∈Γ
U
X
). The union ∪
U∈Γ
U
X
belongs to Ω (i.
e. is open in X) as the union of sets open in X. (Here we use the fact
that Ω, being a topology in X, satisﬁes the ﬁrst axiom of topological
structure.) Therefore A ∩ (∪
U∈Γ
U
X
) belongs to Ω
A
. Similarly one can
check the second axiom. The third axiom: A = A∩ X, and ∅ = A∩ ∅.
6.B The intersection of an open disk with a line is either an open
interval or empty. Any open set in the plane is a union of open disks.
Therefore the intersection of any open set of the plane with a line is a
union of open intervals. Thus, it is open in the line.
6.C If a set F is closed in A then its complement AF is open in A,
i. e. AF = A∩U, where U is open in X. What closed set cuts F on A?
It is cut by XU. Indeed, A∩(XU) = A(A∩U) = A(AF) = F.
Similarly one can prove that the intersection with A of the set closed in
X is closed in A.
6.D No disk of R
2
ﬁts into R.
§6. SUBSPACES 47
6.E If A ∈ Ω and B ∈ Ω
A
then B = A∩U, where U ∈ Ω. Therefore
B ∈ Ω, for this is the intersection of two sets, A and U, which belong to
Ω.
6.F Act as in the solution of the preceding problem 6.E, but use
6.C instead of deﬁnition of induced topology.
6.G The core of the proof is equality (U ∩A) ∩B = U ∩B. It takes
place, for B ⊂ A, and is applied to U ∈ Ω. As U runs over Ω, the right
hand side of the equality (U ∩ A) ∩ B = U ∩ B runs over Ω
B
, while the
left hand side runs over (Ω
A
)
B
. Indeed, elements of Ω
B
are intersections
U ∩ B with U ∈ Ω, and elements of (Ω
A
)
B
are intersections V ∩ B with
V ∈ Ω
A
, but V , in turn, being an element of Ω
A
, is intersection U ∩ A
with U ∈ Ω.
Hints, Comments, Advises, Solutions, and Answers
6.1 In the same way as the induced topology: if Σ is a base in X then
Σ
A
= ¦A∩ V [ V ∈ Σ¦ is a base of the induced topology in A.
6.2
(a) Discrete, for (n −1, n + 1) ∩ N = ¦n¦;
(b) Ω
N
= ¦(k, k + 1, k + 2...)¦
k∈N
;
(c) discrete;
(d) Ω = ¦∅, ¦2¦, ¦1, 2¦¦.
6.3 Yes, it is open, since [0, 1) = (−1, 1) ∩ [0, 2], and (−1, 1) is open on
the line.
6.4 For V one can take U itself, if U is open. To prove the opposite
implication, use problem 6.E.
6.5 Consider interval (−1, 1) ⊂ R and open disk with radius 1 and
center at (0, 0) on plane R
2
.
6.7 The topology induced by the metric in A, is deﬁned by base Σ
1
=
¦B
A
r
(a) [ a ∈ A¦, where B
A
r
(a) = ¦x ∈ A [ ρ(x, a) < r¦ is an open ball
in A with center a and radius r. The other topology is deﬁned by base
Σ
2
= ¦A ∩ B
r
(x) [ x ∈ X¦. Here B
r
(x) is an open ball in X. Obviously
B
A
r
(a) = A ∩ B
r
(a) for a ∈ A. Therefore Σ
1
⊂ Σ
2
. However, it may happen
that Σ
1
= Σ
2
.
6.6 It is left to prove that elements of Σ
2
are open in the topology
deﬁned by Σ
1
. For a point x of an element U of Σ
2
, ﬁnd V ∈ Σ
1
such that
x ∈ V ⊂ U.
'7 Position of a Point with Respect to a Set
This section is devoted to a further expansion of the vocabulary
needed when one speaks of phenomena in a topological space.
'7
◦
1 Interior, Exterior and Boundary Points
Let X be a topological space, A ⊂ X, and b ∈ X. The point b is
called
• an interior point of the set A, if it has a neighborhood contained
in A;
• an exterior point of the set A, if it has a neighborhood disjoint with
A;
• a boundary point of the set A, if any its neighborhood intersects
both A and the complement of A.
'7
◦
2 Interior and Exterior
The interior of a set A in a topological space X is the greatest (with
respect to inclusion) open in X set contained in A, i.e., an open set,
which contains any other open subset of A. It is denoted Int A or, going
into details, Int
X
A.
7.A. Every subset of a topological space has interior. It is the union of
all open sets contained in this set.
7.B. The interior of a set is the union of its interior points.
7.C. A set is open, iﬀ it coincides with its interior.
7.D. Prove that in R:
(a) Int[0, 1) = (0, 1),
(b) Int Q = ∅ and
(c) Int(R Q) = ∅.
7.1. Find the interior of ¦a, b, d¦ in space .
7.2. Find the interior of the interval (0; 1) on the line with the Zariski topol
ogy.
The exterior of a set is the maximal open set disjoint from A. It is
obvious that the exterior of A is Int(X A).
'7
◦
3 Closure
The closure of a set A is the smallest closed set containing A. It is
denoted Cl A or, going into details, Cl
X
A.
7.E. Every subset of topological space has closure. It is the intersection
of all closed sets containing this set.
48
§7. POSITION OF A POINT WITH RESPECT TO A SET 49
7.3. Prove that if A is a subspace of X, and B ⊂ A, then Cl
A
B = (Cl
X
B) ∩
A. Is it true that Int
A
B = (Int
X
B) ∩ A?
A point b is called an adherent point for a set A if all of its neighbor
hood intersect A.
7.F. The closure of a set is the set of its adherent points.
7.G. A set A is closed, iﬀ A = Cl A.
7.H. The closure of a set is the complement of its exterior. In formulas:
Cl A = X Int(X A), where X is the space and A ⊂ X.
7.I. Prove that in R:
(a) Cl[0, 1) = [0, 1],
(b) Cl Q = R,
(c) Cl(R Q) = R.
7.4. Find the closure of ¦a¦ in .
7.5. Describe the closure of a point in a poset topology.
'7
◦
4 Closure in a Metric Space
Let A be a subset, and b be a point of a metric space (X, ρ). Recall
(see Section '4) that the distance ρ(b, A) from the point b to the set A
is the inf¦ ρ(b, a) [ a ∈ A¦.
7.J. Prove that b ∈ Cl A, iﬀ ρ(b, A) = 0.
'7
◦
5 Frontier
The frontier of a set A is the set Cl A Int A. It is denoted by Fr A
or, more precisely, Fr
X
A.
7.6. In ﬁnd the frontier of ¦a¦.
7.K. The frontier of a set is the set of its boundary points.
7.L. Prove that a set A is closed, iﬀ Fr A ⊂ A.
7.7. Prove that Fr A = Fr(X A). Find a formula for Fr A, which is sym
metric with respect to A and X A.
7.8. The frontier of a set A equals the intersection of the closure of A and
the closure of the complement of A:
Fr A = Cl A ∩ Cl(X A).
'7
◦
6 Closure and Interior with Respect to a Finer Topology
7.9. Let Ω
1
, Ω
2
be topological structure in X, and Ω
1
⊂ Ω
2
. Let Cl
i
denote
the closure with respect to Ω
i
. Prove that Cl
1
A ⊃ Cl
2
A for any A ⊂ X.
7.10. Formulate and prove an analogous statement about interior.
§7. POSITION OF A POINT WITH RESPECT TO A SET 50
'7
◦
7 Properties of Interior and Closure
7.11. Prove that if A ⊂ B then Int A ⊂ Int B.
7.12. Prove that Int Int A = Int A.
7.13. Is it true that for any sets A and B the following equalities hold true:
Int(A ∩ B) = Int A ∩ Int B, (8)
Int(A ∪ B) = Int A ∪ Int B? (9)
7.14. Give an example in which one of that equalities does not hold true.
7.15. In the example that you have found solving the previous problem an
inclusion of one hand side into another one holds true. Does this inclusion
hold true for any A and B?
7.16. Study the operator Cl in a way suggested by the investigation of Int
undertaken in 7.11–7.15.
7.17. Find Cl¦1¦, Int[0, 1], and Fr(2, +∞) in the arrow.
7.18. Find Int
(0, 1] ∪ ¦2¦
, Cl
¦
1
n
[ n ∈ N¦
, and Fr Q in R.
7.19. Find Cl N, Int(0, 1), and Fr[0, 1] in R
T1
. How to ﬁnd the closure and
interior of a set in this space?
7.20. Does a sphere contain the frontier of the open ball with the same center
and radius?
7.21. Does a sphere contain the frontier of the closed ball with the same
center and radius?
7.22. Find an example in which a sphere is disjoint from the closure of the
open ball with the same center and radius.
'7
◦
8 Compositions of Closure and Interior
7.23 The Kuratowski Problem. How many pairwise distinct sets can one
obtain out of a single set using operators Cl and Int?
The following problems will help you to solve problem 7.23.
7.23.1. Find a set A ⊂ R such that the sets A, Cl A, and Int A
would be pairwise distinct.
7.23.2. Is there a set A ⊂ R such that
(a) A, Cl A, Int A, Cl Int A are pairwise distinct;
(b) A, Cl A, Int A, Int Cl A are pairwise distinct;
(c) A, Cl A, Int A, Cl Int A, Int Cl A are pairwise distinct?
If you ﬁnd such sets, keep on going in the same way, and when
fail, try to formulate a theorem explaining the failure.
7.23.3. Prove that Cl Int Cl Int A = Cl Int A.
§7. POSITION OF A POINT WITH RESPECT TO A SET 51
'7
◦
9 Sets with Common Frontier
7.24*. Find three sets in the real line, which have the same frontier. Is it
possible to increase the number of such sets?
'7
◦
10 Convexity and Int, Cl, Fr
Recall that a set A ⊂ R
n
is said to be convex if together with any two
points it contains the whole interval connecting them (i.e., for any x, y ∈ A
any point z belonging to the segment [x, y] belongs to A).
Let A be a convex set in R
n
.
7.25. Prove that Cl A and Int A are convex.
7.26. Prove that A contains a ball, unless A is not contained in an (n −1)
dimensional aﬃne subspace of R
n
.
7.27. When is Fr A convex?
'7
◦
11 Characterization of Topology by Closure or Interior
Operations
7.28*. Let in the set of all subset of a set X exist an operator Cl
∗
which
has the following properties:
(a) Cl
∗
∅ = ∅;
(b) Cl
∗
A ⊃ A;
(c) Cl
∗
(A ∪ B) = Cl
∗
A ∪ Cl
∗
B;
(d) Cl
∗
Cl
∗
A = Cl
∗
A.
Prove that Ω = ¦ U ⊂ X [ Cl
∗
(X U) = X U ¦ is a topological
structure, and Cl
∗
A is the closure of a set A in the space (X, Ω).
7.29. Find an analogous system of axioms for Int.
'7
◦
12 Dense Sets
Let A and B be sets in a topological space X. A is said to be dense
in B if Cl A ⊃ B, and everywhere dense if Cl A = X.
7.M. A set is everywhere dense, iﬀ it intersects any nonempty open set.
7.N. The set Q is everywhere dense in R.
7.30. Give a characterization of everywhere dense sets in an indiscrete space,
in the arrow and in R
T1
.
7.31. Prove that a topological space is a discrete space, iﬀ it has a unique
everywhere dense set (by the way, which one).
7.32. Formulate a necessary and suﬃcient condition on the topology of a
space which has an everywhere dense point. Find spaces satisfying the con
dition in Section '2.
7.33. Which singletons are dense in a poset topology?
7.34. Is it true that the union of everywhere dense sets is everywhere dense,
and that the intersection of everywhere dense sets is everywhere dense?
§7. POSITION OF A POINT WITH RESPECT TO A SET 52
7.35. Prove that the intersection of two open everywhere dense sets is ev
erywhere dense.
7.36. Which condition in the previous problem is redundant?
7.37*. Prove that in R a countable intersection of open everywhere dense
sets is everywhere dense. Is it possible to replace R here by an arbitrary
topological space?
7.38*. Prove that Q cannot be presented as a countable intersection of
open sets dense in R.
'7
◦
13 Nowhere Dense Sets
A set is called nowhere dense if its exterior is everywhere dense.
7.39. Can a set be everywhere dense and nowhere dense simultaneously?
7.O. A set A is nowhere dense in X, iﬀ any neighborhood of any point
x ∈ X contains a point y such that the complement of A contains y
together with one of its neighborhoods.
7.40. Enigma. What can you say about the interior of a nowhere dense
set?
7.41. Is R nowhere dense in R
2
?
7.42. Prove that if A is nowhere dense then Int Cl A = ∅.
7.43. Prove that the frontier of a closed set is nowhere dense. Is this true
for the boundary of an open set; boundary of an arbitrary set?
7.44. Prove that a ﬁnite union of nowhere dense sets is nowhere dense.
7.45. Prove that in R
n
(n ≥ 1) every proper algebraic set (i.e., a set deﬁned
by algebraic equations) is nowhere dense.
7.46. Prove that for every set A there exists a maximal open set B in which
A is dense. The extreme cases B = X and B = ∅ mean that A is either
everywhere dense or nowhere dense respectively.
'7
◦
14 Limit Points and Isolated Points
A point b is called a limit point of a set A if any neighborhood of b
intersects A¦b¦.
7.P. Every limit point of a set is its adherent point.
7.47. Give an example proving that an adherent point may be not a limit
one.
A point b is called an isolated point of a set A if b ∈ A and there exists
a neighborhood of b disjoint with A ¦b¦.
7.Q. A set A is closed, iﬀ it contains all its limit points.
7.48. Find limit and isolated points of the sets (0, 1] ∪ ¦2¦, ¦
1
n
[ n ∈ N¦
in Q and in R.
7.49. Find limit and isolated points of the set N in R
T1
.
§7. POSITION OF A POINT WITH RESPECT TO A SET 53
'7
◦
15 Locally Closed Sets
A subset A of a topological space X is called locally closed if each of its
points has a neighborhood U such that A∩ U is closed in U (cf. 6.4–6.5).
7.50. Prove that the following conditions are equivalent:
(a) A is locally closed in X;
(b) A is an open subset of its closure Cl
X
A;
(c) A is the intersection of open and closed subsets of X.
Proofs and Comments
7.A The union of all open sets contained in A, ﬁrstly, is open (as
a union of open sets), and, secondly, contains every open set which is
contained in A (i. e., it is the greatest one among those sets).
7.B Let x be an interior point of A (i. e., there exists an open set
U
x
such that x ∈ U
x
⊂ A). Then U
x
⊂ Int A (for Int A is the greatest
among all open sets contained in A), and hence x ∈ Int A. Vice versa, if
x ∈ Int A, the set Int A itself is a neighborhood of x contained in A.
7.C A set is the greatest one among all of its subsets, therefore,
if it is open, it is the greatest among all of its open subsets, and hence
coincides with its interior. Vice versa, a set, which coincides with its
interior, is open, since the interior is open.
7.D
(a) [0; 1) is not open on the line, while (0; 1) is, therefore Int[0; 1) =
(0; 1).
(b) Since any interval contains irrational points, Q does not contain a
nonempty set open in the classical topology of the line. Therefore,
Int Q = ∅.
(c) Since any interval contains rational points, RQ does not contain a
nonempty set open in the classical topology of the line. Therefore,
Int(R Q) = ∅.
7.E The intersection of all closed sets containing A, ﬁrstly, is closed
(as an intersection of closed sets), and, secondly, is contained in every
closed set which contains A (i. e., it is the smallest one among those sets).
Cf. the proof of Theorem 7.A. In general, properties of closure can be
obtained from properties of interior by replacing unions with intersections
and vice versa.
Hints, Comments, Advises, Solutions, and Answers
7.1 Int¦a, b, d¦ = ¦a, b¦, since this is really the greatest among all sets
open in and contained in ¦a, b, d¦.
§7. POSITION OF A POINT WITH RESPECT TO A SET 54
7.2 The interior of the interval (0; 1) on the line with the Zariski topology
is empty, because no nonempty open set of this space is contained in (0; 1).
7.3 Indeed,
Cl
A
B =
¸
F⊃B,
AF∈ΩA
F =
¸
H⊃B,
XH∈Ω
(H ∩ A) = A∩
¸
H⊃B,
XH∈Ω
H = A ∩ Cl
X
B.
The second equality may be obviously wrong. Really, let X = R
2
, A = R
1
,
B = A. Then Int
A
B = R
1
= ∅ = (Int
X
B) ∩ A.
'8 SetTheoretic Digression. Maps
'8
◦
1 Maps and the Main Classes of Maps
A mapping f of a set X to a set Y is a triple consisting of X, Y , and
a rule,
7
which assigns to every element of X exactly one element of Y .
There are other words with the same meaning: map, function.
If f is a mapping of X to Y then one writes f : X →Y , or X
f
→Y .
The element b of Y assigned by f to an element a of X is denoted by
f(a) and called the image of a under f. One writes b = f(a), or a
f
→ b,
or f : a →b.
A mapping f : X →Y is called a surjective map, or just a surjection if
every element of Y is an image of at least one element of X. A mapping
f : X → Y is called an injective map, injection, or onetoone map if
every element of Y is an image of not more than one element of X. A
mapping is called a bijective map, bijection, or invertible if it is surjective
and injective.
'8
◦
2 Image and Preimage
The image of a set A ⊂ X under a map f : X → Y is the set of
images of all points of A. It is denoted by f(A). Thus
f(A) = ¦f(x) : x ∈ A¦.
The image of the entire set X (i.e., f(X)) is called the image of f. The
preimage of a set B ⊂ Y under a map f : X → Y is the set of elements
of X whose images belong to B. It is denoted by f
−1
(B). Thus
f
−1
(B) = ¦a ∈ X : f(a) ∈ B¦.
Be careful with these terms: their etymology can be misleading. For
example, the image of the preimage of a set B can diﬀer from B. And
even if it does not diﬀer, It may happen that the preimage is not the
only set with this property. Hence, the preimage cannot be deﬁned as a
set whose image is a given set.
8.A. f
f
−1
(B)
= B, iﬀ B is contained in the image of f.
8.B. f
f
−1
(B)
⊂ B for any map f : X →Y and B ⊂ Y .
8.C. Let f : X → Y and B ⊂ Y such that f
f
−1
(B)
= B. Then the
following statements are equivalent:
7
Of course, the rule (as everything in the set theory) may be thought of as a set.
Namely, one considers a set of ordered pairs (x, y) with x ∈ X, y ∈ Y such that the
rule assigns y to x. This set is called the graph of f. It is a subset of the set X Y
of all ordered pairs (x, y).
55
§8. SETTHEORETIC DIGRESSION. MAPS 56
(a) f
−1
(B) is the unique subset of X whose image equals B;
(b) for any a
1
, a
2
∈ f
−1
(B) the equality f(a
1
) = f(a
2
) implies a
1
= a
2
.
8.D. A map f : X → Y is an injection, iﬀ for any B ⊂ Y such that
f
f
−1
(B)
= B the preimage f
−1
(B) is the unique subset of X whose
image equals B.
8.E. f
−1
f(A)
⊃ A for any map f : X →Y and A ⊂ X.
8.F. f
−1
f(A)
= A, iﬀ f(A) ∩ f(X A) = ∅.
8.1. Do the following equalities hold true for any A, B ⊂ Y and any f : X →
Y :
f
−1
(A ∪ B) = f
−1
(A) ∪ f
−1
(B), (10)
f
−1
(A ∩ B) = f
−1
(A) ∩ f
−1
(B), (11)
f
−1
(Y A) = X f
−1
(A)? (12)
8.2. Do the following equalities hold true for any A, B ⊂ X and any f : X →
Y :
f(A ∪ B) = f(A) ∪ f(B), (13)
f(A ∩ B) = f(A) ∩ f(B), (14)
f(X A) = Y f(A)? (15)
8.3. Give examples in which two of the equalities above are false.
8.4. Replace the false equalities of 8.2 by correct inclusions.
8.5. What simple condition on f : X → Y should be imposed in order to
make correct all the equalities of 8.2 for any A, B ⊂ X ?
8.6. Prove that for any map f : X →Y , and subsets A ⊂ X, B ⊂ Y :
B ∩ f(A) = f
f
−1
(B) ∩ A
.
'8
◦
3 Identity and Inclusion
The identity map of a set X is the map X → X deﬁned by formula
x → x. It is denoted by id
X
, or just id, when there is no ambiguity. If
A is a subset of X then the map A → X deﬁned by formula x → x is
called an inclusion map, or just inclusion, of A into X and denoted by
in : A →X, or just in, when A and X are clear.
8.G. The preimage of a set B under an inclusion in : A →X is B ∩ A.
'8
◦
4 Composition
The composition of mappings f : X → Y and g : Y → Z is the
mapping g ◦ f : X →Z deﬁned by formula x →g
f(x)
.
§8. SETTHEORETIC DIGRESSION. MAPS 57
8.H. h ◦ (g ◦ f) = (h ◦ g) ◦ f for any maps f : X → Y , g : Y →Z, and
h : Z →U.
8.I. f ◦ (id
X
) = f = (id
X
) ◦ f for any f : X →Y .
8.J. The composition of injections is injective.
8.K. If the composition g ◦ f is injective then f is injective.
8.L. The composition of surjections is surjective.
8.M. If the composition g ◦ f is surjective then g is surjective.
8.N. The composition of bijections is a bijection.
8.7. Let a composition g ◦f be bijective. Is then f or g necessarily bijective?
'8
◦
5 Inverse and Invertible
A map g : Y → X is said to be inverse to a map f : X → Y if
g ◦ f = id
X
and f ◦ g = id
Y
. A map, for which an inverse map exists, is
said to be invertible.
8.O. A mapping is invertible, iﬀ it is a bijection.
8.P. If an inverse map exists then it is unique.
'8
◦
6 Submappings
If A ⊂ X and B ⊂ Y then for every f : X →Y such that f(A) ⊂ B
there is mapping ab(f) : A →B deﬁned by formula x →f(x) and called
an abbreviation of the mapping f to A, B, or submapping, or submap. If
B = Y then ab f : A →Y is denoted by f
A
and called the restriction of
f to A. If B = Y then ab f : A →B is denoted by f
A,B
or even simply
f [.
8.Q. The restriction of a map f : X → Y to A ⊂ X is the composition
of inclusion in A :→X and f. In other words, f
A
= f ◦ in.
8.R. Any abbreviation (including any restriction) of injections is injec
tive.
8.S. If a restriction of a mapping is surjective then the original mapping
is surjective.
'9 Continuous Maps
'9
◦
1 Deﬁnition and Main Properties of Continuous Maps
Let X, Y be topological spaces. A map f : X → Y is said to be
continuous if the preimage of any open subset of Y is an open subset of
X.
9.A. A map is continuous, iﬀ the preimage of any closed set is closed.
9.B. The identity map of any topological space is continuous.
9.1. Let Ω
1
, Ω
2
be topological structures in X. Prove that the identity
mapping of X
id : (X, Ω
1
) →(X, Ω
2
)
is continuous, iﬀ Ω
2
⊂ Ω
1
.
9.2. Let f : X →Y be a continuous map. Is it continuous with respect to
(a) a ﬁner topology in X and the same topology in Y ,
(b) a coarser topology in X and the same topology in Y ,
(c) a ﬁner topology in Y and the same topology in X,
(d) a coarser topology in Y and the same topology in X?
9.3. Let X be a discrete space and Y an arbitrary space. Which maps
X →Y and Y →X are continuous?
9.4. Let X be an indiscrete space and Y an arbitrary space. Which maps
X →Y and Y →X are continuous?
9.C. Let A be a subspace of X. The inclusion in : A →X is continuous.
9.D. The topology Ω
A
induced on A ⊂ X by the topology of X is the
coarsest topology in A such that the inclusion mapping in : A → X is
continuous with respect to it.
9.5. Enigma. The statement 9.D admits a natural generalization with the
inclusion map replaced by an arbitrary map f : A → X of an arbitrary set
A. Find this generalization.
9.E. A composition of continuous maps is continuous.
9.F. A submap of a continuous map is continuous.
9.G. A map f : X → Y is continuous, iﬀ abf : X → f(X) is continu
ous.
9.H. Any constant map (i.e., a map with image consisting of a single
point) is continuous.
58
§9. CONTINUOUS MAPS 59
'9
◦
2 Reformulations of Deﬁnition
9.6. Prove that a mapping f : X →Y is continuous, iﬀ
Cl f
−1
(A) ⊂ f
−1
(Cl A)
for any A ⊂ Y .
9.7. Formulate and prove similar criteria of continuity in terms of Int f
−1
(A)
and f
−1
(Int A). Do the same for Cl f(A) and f(Cl A).
9.8. Let Σ be a base for topology in Y . Prove that a map f : X → Y is
continuous, iﬀ f
−1
(U) is open for any U ∈ Σ.
'9
◦
3 More Examples
9.9. Is the mapping f : [0, 2] →[0, 2] deﬁned by formula
f(x) =
x, if x ∈ [0, 1);
3 −x, if x ∈ [1, 2]
continuous (with respect to the topology induced from the real line)?
9.10. Is the map f of segment [0, 2] (with the topology induced by the
topology of the real line) into the arrow (see Section '2) deﬁned by formula
f(x) =
x, if x ∈ [0, 1];
x + 1, if x ∈ (1, 2]
continuous?
9.11. Give an explicit characterization of continuous mappings of R
T1
(see
Section '2) to R.
9.12. Which maps R
T1
→R
T1
are continuous?
9.13. Give an explicit characterization of continuous mappings of the arrow
to itself.
9.14. Let f be a mapping of the set Z
+
of nonnegative numbers onto R
deﬁned by formula
f(x) =
1
x
, if x = 0;
0, if x = 0.
Let g : Z
+
→f(Z
+
) be its submap. Induce topology on Z
+
and f(Z
+
) from
R. Are f and the map g
−1
, inverse to g, continuous?
'9
◦
4 Behavior of Dense Sets
9.15. Prove that the image of an everywhere dense set under a surjective
continuous map is everywhere dense.
9.16. Is it true that the image of nowhere dense set under a continuous map
is nowhere dense.
9.17*. Does there exist a nowhere dense set A of [0, 1] (with the topology
induced out of the real line) and a continuous map f : [0, 1] →[0, 1] such that
f(A) = [0, 1]?
§9. CONTINUOUS MAPS 60
'9
◦
5 Local Continuity
A map f of a topological space X to a topological space Y is said to
be continuous at a point a ∈ X if for every neighborhood U of f(a) there
exists a neighborhood V of a such that f(V ) ⊂ U.
9.I. A map f : X → Y is continuous, iﬀ it is continuous at each point
of X.
9.J. Let X, Y be metric spaces, and a ∈ X. A map f : X → Y is
continuous at a, iﬀ for every ball with center at f(a) there exists a ball
with center at a whose image is contained in the ﬁrst ball.
9.K. Let X, Y be metric spaces, and a ∈ X. A mapping f : X → Y is
continuous at the point a, iﬀ for every ε > 0 there exists δ > 0 such that
for every point x ∈ X inequality ρ(x, a) < δ implies ρ
f(x), f(a)
< ε.
Theorem 9.K means that continuity introduced above coincides with
the one that is usually studied in Calculus.
'9
◦
6 Properties of Continuous Functions
9.18. Let f, g : X → R be continuous. Prove that the mappings X → R
deﬁned by formulas
x →f(x) +g(x), (16)
x →f(x)g(x), (17)
x →f(x) −g(x), (18)
x →
f(x)
, (19)
x →max¦f(x), g(x)¦, (20)
x →min¦f(x), g(x)¦ (21)
are continuous.
9.19. Prove that if 0 / ∈ g(X) then a mapping X →R deﬁned by formula
x →
f(x)
g(x)
is continuous.
9.20. Find a sequence of continuous functions f
i
: R →R, (i ∈ N) such that
the formula
x →sup¦ f
i
(x) [ i ∈ N¦
deﬁnes a function R →R which is not continuous.
9.21. Let X be any topological space. Prove that a function f : X → R
n
:
x → (f
1
(x), . . . , f
n
(x)) is continuous, iﬀ all the functions f
i
: X → R with
i = 1, . . . , n are continuous.
§9. CONTINUOUS MAPS 61
Real pqmatrices comprise a space Mat(pq, R), which diﬀers fromR
pq
only in the way of numeration of its natural coordinates (they are numerated
by pairs of indices).
9.22. Let f : X → Mat(p q, R) and g : X →Mat(q r, R) be continuous
maps. Prove that then
X →Mat(p r, R) : x →g(x)f(x)
is a continuous map.
Recall that GL(n; R) is the subspace of Mat(n n, R) consisting of all
the invertible matrices.
9.23. Let f : X → GL(n; R) be a continuous map. Prove that X →
GL(n; R) : x →(f(x))
−1
is continuous.
'9
◦
7 Continuity of Distances
9.L. For every subset A of a metric space X the function deﬁned by
formula x →ρ(x, A) (see Section '4) is continuous.
9.24. Prove that a topology of a metric space is the coarsest topology, with
respect to which for every A ⊂ X the function X → R deﬁned by formula
x →ρ(x, A) is continuous.
'9
◦
8 Isometry
A mapping f of a metric space X into a metric space Y is called
an isometric embedding if ρ
f(a), f(b)
= ρ(a, b) for every a, b ∈ X. A
bijection which is an isometric embedding is called an isometry.
9.M. Every isometric embedding is injective.
9.N. Every isometric embedding is continuous.
'9
◦
9 GromovHausdorﬀ distance
9:A. For any metric spaces X and Y there exists a metric space Z such that X and
Y can be isometrically embedded into Z.
Having embedded isometrically two metric space in a single one, we can consider
the Hausdorﬀ distance between their images (see. '4
◦
16). The inﬁnum of such
Hausdorﬀ distances over all pairs of isometric embeddings of metric spaces X and Y
to metric spaces is called the GromovHausdorﬀ distance between X and Y .
9:B. Does there exist metric spaces with inﬁnite GromovHausdorﬀ distance?
9:C. Prove that the GromovHausdorﬀ distance is symmetric and satisﬁes the triangle
inequality.
9:D. Enigma. In what sense the GromovHausdorﬀ distance can satisfy the ﬁrst
axiom of metric?
§9. CONTINUOUS MAPS 62
'9
◦
10 Contractive maps
A mapping f : X → X of a metric space X is called contractive if there
exists α ∈ (0, 1) such that ρ
f(a), f(b)
≤ αρ(a, b) for every a, b ∈ X.
9.25. Prove that every contractive mapping is continuous.
Let X, Y be metric spaces. A mapping f : X → Y is said to be H¨older
if there exist C > 0 and α > 0 such that ρ
f(a), f(b)
≤ Cρ(a, b)
α
for every
a, b ∈ X.
9.26. Prove that every H¨older mapping is continuous.
'9
◦
11 Monotone maps
Let (X, ≺) and (Y, ≺) be posets. A map f : X →Y is said to be
• (nonstrictly) monotonically increasing or just monotone, if
f(a) _ f(b) for any a, b ∈ X with a _ b ;
• (nonstrictly) monotonically decreasing or antimotone, if
f(b) _ f(a) for any a, b ∈ X with a _ b ;
• strictly monotonically increasing or just strictly monotone, if
f(a) ≺ f(b) for any a, b ∈ X with a ≺ b ;
• strictly monotonically decreasing or strictly antimonotone, if
f(b) ≺ f(a) for any a, b ∈ X with a ≺ b.
9.O. Let X and Y be linearly ordered sets. With respect to the interval
topology in X and Y any surjective monotone or antimonotone mapping
X →Y is continuous.
9.27. Show that the surjectivity condition in 9.O is needed.
9.28. Under conditions of Theorem 9.O, is f continuous with respect to right
ray or left ray topologies?
9.P. A mapping of a poset to a poset is monotone, iﬀ it is continuous
with respect to the poset topologies.
'9
◦
12 Functions on Cantor Set and SquareFilling Curves
Recall that Cantor set K is the set of real numbers which can be presented as
sums of series of the form
¸
∞
k=1
a
k
3
k
with a
k
= 0 or 2.
9:E. Let γ
1
be a map K →I deﬁned by
∞
¸
k=1
a
k
3
k
→
∞
¸
k=1
a
k
2
k+1
.
Prove that γ
1
: K →I is a continuous surjection. Draw the graph of ϕ.
§9. CONTINUOUS MAPS 63
9:F. Prove that the function K →K deﬁned by
∞
¸
k=1
a
k
3
k
→
∞
¸
k=1
a
2k
3
k
is continuous.
Denote by K
2
the set ¦(x, y) ∈ R
2
: x ∈ K, y ∈ K¦.
9:G. Prove that the map γ
2
: K →K
2
deﬁned by
∞
¸
k=1
a
k
3
k
→
∞
¸
k=1
a
2k−1
3
k
,
∞
¸
k=1
a
2k
3
k
is a continuous surjection.
9:H. Prove that the map γ
3
: K → I
2
deﬁned as the composition of γ
2
: K → K
2
and K
2
→I
2
: (x, y) →(γ
1
(x), γ
1
(y)) is a continuous surjection.
9:I. Prove that the map γ
3
: K → I
2
is a restriction of a continuous map. (Cf.
2:A.2.)
The latter map is a continuous surjection I →I
2
. Thus, this is a curve ﬁlling the
square. A curve with this property was ﬁrst constructed by G. Peano in 1890. Though
the construction sketched above is based on the same ideas as the original Peano’s
construction, they are slightly diﬀerent. Since then a lot of other similar examples
have been found. You may ﬁnd a nice survey of them in a book by Hans Sagan, Space
Filling Curves, SpringerVerlag 1994. Here is a sketch of Hilbert’s construction.
9:J. Prove that there exists a sequence of polygonal maps f
k
: I →I
2
such that
(a) f
k
connects all centers of the squares forming the obvious subdivision of I
2
into
4
k
equal squares with side 1/2
k
;
(b) dist(f
k
(x), f
k−1
(x)) ≤
√
2/2
k+1
for any x ∈ I (here dist means the metric
induced on I
2
from the standard Euclidean metric of R
2
).
9:K. Prove that any sequence of paths f
k
: I → I
2
satisfying the conditions of 9:J
converges to a map f : I → I
2
(i.e. for any x ∈ I there exists a limit f(x) =
lim
k→∞
f
k
(x)) and this map is continuous and its image is dense in I
2
.
9:L.
8
Prove that any continuous map I → I
2
with dense image is surjective.
9:M. Generalize 9:G – 9:I 9:J – 9:L to obtain a continuous surjection of I onto I
n
.
'9
◦
13 Sets Deﬁned by Systems of Equations and Inequalities
9.Q. Let f
i
(i = 1, . . . , n) be continuous mappings X → R. Then the
subset of X consisting of solutions of the system of equations
f
1
(x) = 0, . . . , f
n
(x) = 0
is closed.
8
Although this problem can be solved using theorems wellknown from Calculus,
we have to mention that it would be more appropriate after Section '15. Cf. Problems
15.O, 15.T, 15.K.
§9. CONTINUOUS MAPS 64
9.R. Let f
i
(i = 1, . . . , n) be continuous mappings X → R. Then the
subset of X consisting of solutions of the system of inequalities
f
1
(x) ≥ 0, . . . , f
n
(x) ≥ 0
is closed, while the set consisting of solutions of the system of inequalities
f
1
(x) > 0, . . . , f
n
(x) > 0
is open.
9.29. Where in 9.Q and 9.R a ﬁnite system can be replaced by an inﬁnite
one.
'9
◦
14 SetTheoretic Digression. Covers
A collection Γ of subsets of a set X is called a cover or a covering
of X if X is a union of sets of belonging to Γ, i.e., X =
¸
A∈Γ
A. In this
case elements of Γ are said to cover X.
There is also a more general meaning of these words. A collection Γ
of subsets of a set Y is called a cover or a covering of a set X ⊂ Y if X
is contained in the union of the sets belonging to Γ, i.e., X ⊂
¸
A∈Γ
A.
In this case, sets belonging to Γ are also said to cover X.
'9
◦
15 Fundamental Covers
Consider a cover Γ of a topological space X. Each element of Γ inher
its from X a topological structure. When are these structures suﬃcient
for recovering the topology of X? In particular, under what conditions
on Γ does continuity of a map f : X → Y follow from continuity of
its restrictions to elements of Γ. To answer these questions, solve the
problems 9.30–9.31 and 9.S–9.X.
9.30. Is this true for the following coverings:
(a) X = [0, 2], Γ = ¦[0, 1], (1, 2]¦;
(b) X = [0, 2], Γ = ¦[0, 1], [1, 2]¦;
(c) X = R, Γ = ¦Q, R Q¦;
(d) X = R, Γ is a set of all onepoint subsets of R?
A cover Γ of a space X is said to be fundamental if a set U ⊂ X is
open, iﬀ for every A ∈ Γ the set U ∩ A is open in A.
9.S. A covering Γ of a space X is fundamental, iﬀ a set U ⊂ X is open,
provided U ∩ A is open in A for every A ∈ Γ.
9.T. A covering Γ of a space X is fundamental, iﬀ a set F ⊂ X is closed,
provided F ∩ A is closed A for every A ∈ Γ.
§9. CONTINUOUS MAPS 65
9.31. A cover of a topological space consisting of singletons is fundamental,
iﬀ the space is discrete.
A cover of a topological space is said to be open if it consists of open
sets, and closed if it consists of closed sets. A cover of a topological space
is said to be locally ﬁnite if every point of the space has a neighborhood
intersecting only a ﬁnite number of elements of the cover.
9.U. Every open cover is fundamental.
9.V. Every ﬁnite closed cover is fundamental.
9.W. Every locally ﬁnite closed cover is fundamental.
9.X. Let Γ be a fundamental cover of a topological space X. If the
restriction of a mapping f : X → Y to each element of Γ is continuous
then f is continuous.
A cover Γ
is said to be a reﬁnement of a cover Γ if every element of Γ
is
contained in some element of Γ.
9.32. Prove that if a cover Γ
is a reﬁnement of a cover Γ, and Γ
is funda
mental then Γ is also fundamental.
9.33. Prove that if Γ is a fundamental cover, and A is a subspace of a space
X, then Γ
A
= ¦ U ∩ A [ U ∈ Γ¦ is a fundamental cover of A.
9.34. Let ∆ be a fundamental cover of a topological space X, and Γ be a
cover of X such that Γ
A
= ¦ U ∩ A [ U ∈ Γ¦ is a fundamental cover for
subspace A ⊂ X for every A ∈ ∆. Prove that Γ is a fundamental cover.
9.35. Prove that the property of being fundamental is local, i.e., if every
point of a space X has a neighborhood V such that Γ
V
= ¦ U ∩ V [ U ∈ Γ¦
is fundamental, then Γ is fundamental.
Hints, Comments, Advises, Solutions, and Answers
9:A Construct Z as the union of disjoint copies of X and Y . In the
union, put the distance between points, which belong to a copy of one of
the spaces, to be equal to the distance between the corresponding points of
the original space. To deﬁne the distance bewteen points of diﬀerent copies,
choose points x
0
∈ X and y
0
∈ Y , and put ρ(a, b) = ρ
X
(a, x
0
) +ρ
Y
(y
0
, b) +1
for a ∈ X and b ∈ Y . Check (this is easy, really), that this deﬁnes a metric.
9:B Yes. For example a singleton and any nonbounded space.
9:C Although, as we have seen solving the previous problem, the Gromov
Hausdorﬀ distance can be inﬁnite, while symmetricity and triangle inequality
were formulated above only for functions with ﬁnite values, these two prop
erties make sense if inﬁnite values are admitted. (The triangle inequality
should be considered satisﬁed, if two or three of the quantities involved are
inﬁnite, and not satisﬁed, if only one of them is inﬁnite.) The following con
struction helps to prove the triangle inequality. Let metric spaces X and Y
are isometrically embedded into a metric space A, and metric spaces Y and
§9. CONTINUOUS MAPS 66
Z are isometrically embedded into a metric space B. Construct a new met
ric space in which A and B would be isometrically embedded meeting in Y .
For this, add to A all points of B A. Put distances between these points
to be equal to the distances between them in B. Put the distance between
x ∈ A B and z ∈ B A equal to inf¦ρ
A
(x, y) + ρ
B
(y, z) [ y ∈ A ∩ B¦.
Compare this construction to the construction from the solution of Problem
9:A. Prove that this gives a metric space and use the triangle inequality for
the Hausdorﬀ distance between X, Y and Z in this space.
9:D Partially, the answer is obvious. Certainly, the GromovHausdorﬀ
distance is nonnegative! But what if it is zero, in what sense the spaces
should be equal then? First, the most optimistic idea is that then there
should exist an isometric bijection between the spaces. But this is not true,
as one can see looking at spaces Q and R with standard distances in them.
We promise ahead of time that for compact metric spaces this is true.
'10 Homeomorphisms
'10
◦
1 Deﬁnition and Main Properties of Homeomorphisms
An invertible mapping is called a homeomorphism, if both this map
ping and its inverse are continuous.
10.A. Find an example of a continuous bijection, which is not a home
omorphism.
10.B. Find a continuous bijection [0, 1) → S
1
, which is not a homeo
morphism.
10.C. The identity map of a topological space is a homeomorphism.
10.D. A composition of homeomorphisms is a homeomorphism.
10.E. The inverse of a homeomorphism is a homeomorphism.
'10
◦
2 Homeomorphic Spaces
A topological space X is said to be homeomorphic to space Y , if there
exists a homeomorphism X →Y .
10.F. Being homeomorphic is an equivalence relation.
10.1. Enigma. How is Theorem 10.F related to 10.C–10.E?
'10
◦
3 Role of Homeomorphisms
10.G. Let f : X → Y be a homeomorphism. Then U ⊂ X is open (in
X), iﬀ f(U) is open (in Y ).
10.H. f : X →Y is a homeomorphism, iﬀ f is a bijection and deﬁnes a
bijection between the topological structures of X and Y .
10.I. Let f : X →Y be a homeomorphism. Then for every A ⊂ X
(a) A is closed in X, iﬀ f(A) is closed in Y ;
(b) f(Cl A) = Cl f(A);
(c) f(Int A) = Int f(A);
(d) f(Fr A) = Fr f(A);
(e) A is a neighborhood of a point x ∈ X, iﬀ f(A) is a neighborhood of
the point f(x);
(f) etc.
Therefore from the topological point of view homeomorphic spaces
are completely identical: a homeomorphism X → Y establishes oneto
one correspondence between all phenomena in X and Y which can be
expressed in terms of topological structures.
This phenomenon was used as a basis for a deﬁnition of the subject
of topology in the ﬁrst stages of its development, when the notion of
67
§10. HOMEOMORPHISMS 68
topological space had not been developed yet. Then mathematicians
studied only subspaces of Euclidean spaces, their continuous mappings
and homeomorphisms. Felix Klein in his famous Erlangen Program,
9
where he classiﬁed various geometries that had emerged up to that time,
like Euclidean, Lobachevsky, aﬃne, and projective geometries, deﬁned
topology as a part of geometry which deals with the properties preserved
by homeomorphisms.
'10
◦
4 More Examples of Homeomorphisms
10.J. Let f : X →Y be a homeomorphism. Prove that for every A ⊂ X
the reduction ab(f) : A →f(A) is also a homeomorphism.
10.K. Prove that every isometry (see Section '9) is a homeomorphism.
10.L. Prove that every nondegenerate aﬃne transformation of R
n
is a
homeomorphism.
10.M. Let X and Y be totally ordered sets. Any surjective strictly mono
tone map X →Y is a homeomorphism with respect to the interval topo
logical structures in X and Y .
10.N Corollary. Any surjective stricly monotone function f : [a, b] →
[c, d] is a homeomorphism.
10.2. Prove that inversion
x →
Rx
[x[
2
: R
n
¦0¦ →R
n
¦0¦
is a homeomorphism.
10.3. Let H = ¦ z ∈ C [ Imz > 0 ¦ be the upper halfplane. Prove that
mapping f : H → H deﬁned by f(z) =
az +b
cz +d
, where a, b, c, d ∈ R, is a
homeomorphism if
a b
c d
> 0.
10.4. Prove that a bijection R →R is a homeomorphism, iﬀ it is a monotone
function.
10.5. Prove that every bijection of an indiscrete space onto itself is a home
omorphism. Prove that the same holds true for a discrete space and R
T1
.
10.6. Find all homeomorphisms of the space (see Section '2) to itself.
10.7. Prove that every continuous bijection of the arrow onto itself is a
homeomorphism.
10.8. Find two homeomorphic spaces X and Y and a continuous bijection
X →Y , which is not a homeomorphism.
9
In fact it was not assumed to be a program in the sense of being planned,
although it became a kind of program. It was a sort of dissertation presented by
Klein for getting the position as a professor at Erlangen University.
§10. HOMEOMORPHISMS 69
10.9. Is γ
2
: K →K
2
considered in Problem 9:G a homeomorphism? Recall
that K is the Cantor set, K
2
= ¦(x, y) ∈ R
2
: x ∈ K, y ∈ K¦ and γ
2
is
deﬁned by
∞
¸
k=1
a
k
3
k
→
∞
¸
k=1
a
2k−1
3
k
,
∞
¸
k=1
a
2k
3
k
'10
◦
5 Examples of Homeomorphic Spaces
Below the homeomorphism relation is denoted by
∼
=. It is not a
commonly accepted notation. In other textbooks any sign close to, but
distinct from =, e. g. ∼, ·, ≈, is used.
10.O. Prove that
(a) [0, 1]
∼
= [a, b] for any a < b;
(b) [0, 1)
∼
= [a, b)
∼
= (0, 1]
∼
= (a, b] for any a < b;
(c) (0, 1)
∼
= (a, b) for any a < b;
(d) (−1, 1)
∼
= R;
(e) [0, 1)
∼
= [0, +∞) and (0, 1)
∼
= (0, +∞).
10.P. S
1
¦(0, 1)¦
∼
= R
1
.
10.Q. S
n
¦point¦
∼
= R
n
.
10.10. Prove that the following plane ﬁgures are homeomorphic:
(a) the whole plane R
2
;
(b) open square ¦ (x, y) ∈ R
2
[ x, y ∈ (0, 1) ¦;
(c) open strip ¦ (x, y) ∈ R
2
[ x ∈ (0, 1) ¦;
(d) halfplane ¦ (x, y) ∈ R
2
[ y > 0 ¦;
(e) open halfstrip ¦ (x, y) ∈ R
2
[ x > 0, y ∈ (0, 1) ¦;
(f) open disk ¦ (x, y) ∈ R
2
[ x
2
+y
2
< 1 ¦;
(g) open rectangle ¦ (x, y) ∈ R
2
[ a < x < b, c < y < d ¦;
(h) open quadrant ¦ (x, y) ∈ R
2
[ x, y > 0 ¦;
(i) ¦ (x, y) ∈ R
2
[ y
2
+ [x[ > x¦, i.e., plane cut along the ray ¦ y = 0, x ≥
0 ¦.
10.R. Prove that
(a) closed disk D
2
is homeomorphic to square I
2
= ¦ (x, y) ∈ R
2
[ x, y ∈
[0, 1] ¦;
(b) open disc Int D
2
= ¦ (x, y) ∈ R
2
[ x
2
+ y
2
< 1 ¦ is homeomorphic to
open square Int I
2
= ¦ (x, y) ∈ R
2
[ x, y ∈ (0, 1) ¦;
(c) circle S
1
is homeomorphic to the boundary of square ∂I
2
= I
2
Int I
2
.
10.S. Prove that
(a) every bounded closed convex set in the plane with nonempty interior
is homeomorphic to D
2
;
(b) every bounded open convex nonempty set in the plane is homeomor
phic to the plane;
(c) boundary of every bounded convex set in the plane with nonempty
interior is homeomorphic to S
1
.
§10. HOMEOMORPHISMS 70
10.11. In which of the situations considered in 10.S can the assumption that
the set is bounded be omitted?
10.12. Classify up to homeomorphism all closed convex sets in the plane.
(Make a list without repeats; prove that every such set is homeomorphic
to one in the list; postpone a proof of nonexistence of homeomorphisms till
Section '11.)
10.13*. Generalize the previous three problems to the case of sets in R
n
with arbitrary n.
The latter four problems show that angles are not essential in topol
ogy, i.e., for a line or boundary of a domain the property of having angles
is not preserved by homeomorphism. And now two more problems on
this.
10.14. Prove that every closed simple (i.e., without selfintersections) poly
gon in R
2
(and in R
n
with n > 2) is homeomorphic to the circle S
1
.
10.15. Prove that every nonclosed simple ﬁnite unit polyline in R
2
(and
in R
n
with n > 2) is homeomorphic to the segment [0, 1].
10.16. Prove that R
2
¦ [x[, [y[ > 1 ¦
∼
= I
2
¦(±1, ±1), (±1, ±1)¦.
10.17. Prove that the following plane ﬁgures are homeomorphic to each
other:
(a) ¦ (x, y) [ 0 ≤ x, y < 1 ¦;
(b) ¦ (x, y) [ 0 < x < 1, 0 ≤ y < 1 ¦;
(c) ¦ (x, y) [ 0 ≤ x ≤ 1, 0 ≤ y < 1 ¦;
(d) ¦ (x, y) [ x, y ≥ 0 ¦;
(e) ¦ (x, y) [ x ≥ 0 ¦;
(f) ¦ (x, y) [ x ≥ y ≥ 0 ¦;
(g) ¦ (x, y) [ x
2
+ y
2
≤ 1, x = 1 ¦.
10.18. Prove that the following plane ﬁgures are homeomorphic to each
other:
(a) punctured plane R
2
¦(0, 0)¦;
(b) punctured disc ¦ (x, y) [ 0 < x
2
+y
2
< 1 ¦;
(c) annulus ¦ (x, y) [ a < x
2
+y
2
< b ¦ where 0 < a < b;
(d) plane without disc ¦ (x, y) [ x
2
+y
2
> 1 ¦;
(e) plane without square ¦ (x, y) [ 0 ≤ x, y ≤ 1 ¦;
(f) plane without segment R
2
[0, 1].
10.19. Let X ⊂ R
2
be an union of several segments with a common end
point. Prove that the complement R
2
X is homeomorphic to the punctured
plane.
10.20. Let X ⊂ R
2
simple nonclosed ﬁnite polyline. Prove that its comple
ment R
2
X is homeomorphic to the punctured plane.
10.21. Let D
1
, . . . , D
n
⊂ R
2
be pairwise disjoint closed discs. The com
plement of the union of its interior is said to be plane with n holes. Prove
that any two planes with n holes are homeomorphic, i.e., dislocation of discs
D
1
, . . . , D
n
does not aﬀect on the topological type of R
2
∪
n
i=1
Int D
i
.
10.22. Prove that for continuous functions f, g : R →R such that f < g, the
space between their graphs ¦ (x, y) ∈ R
2
[ f(x) ≤ y ≤ g(x) ¦ is homeomorphic
to a closed strip ¦ (x, y) [ y ∈ [0, 1] ¦.
§10. HOMEOMORPHISMS 71
10.23. Prove that a mug (with handle) is homeomorphic to a doughnut.
10.24. Arrange the following items to homeomorphism classes: a cup, a
saucer, a glass, a spoon, a fork, a knife, a plate, a coin, a nail, a screw, a bolt,
a nut, a wedding ring, a drill, a ﬂower pot (with hole in the bottom), a key.
10.25. In a spherical shell (the space between two concentric spheres) one
drilled out a cylindrical hole connecting the boundary spheres. Prove that
the rest is homeomorphic to D
3
.
10.26. In a spherical shell one made a hole connecting the boundary spheres
and having the shape of a knotted tube (see Figure 1.). Prove that the rest
of the shell is homeomorphic to D
3
.
Figure 1.
10.27. Prove that surfaces shown in Figure 2 are homeomorphic (they are
called handles).
Figure 2.
10.28. Prove that surfaces shown in the Figure 3 are homeomorphic. (They
are homeomorphic to Klein bottle with two holes. More details about this is
given in Section '20.)
10.29*. Prove that R
3
S
1 ∼
= R
3
R
1
∪ ¦(1, 1, 1)¦
.
10.30. Prove that subset of the sphere S
n
deﬁned in standard coordinates
in R
n+1
by inequality x
2
1
+x
2
2
+ +x
2
k
< x
2
k+1
+ +x
2
n
is homeomorphic
to R
n
R
n−k
.
§10. HOMEOMORPHISMS 72
Figure 3.
'10
◦
6 Examples of Nonhomeomorphic Spaces
10.T. Spaces consisting of diﬀerent number of points are not homeo
morphic.
10.U. A discrete space and an indiscrete space (which have more than
one point) are not homeomorphic.
10.31. Prove that the spaces Z, Q (with topology induced from R), R, R
T1
and the arrow are pairwise nonhomeomorphic.
10.32. Find two nonhomeomorphic spaces X and Y for which there exist
continuous bijections X → Y and Y →X.
'10
◦
7 Homeomorphism Problem and Topological Properties
One of the classic problems of topology is the homeomorphism problem:
to ﬁnd out whether two given topological spaces are homeomorphic. In
each special case the character of solution depends mainly on the answer.
To prove that spaces are homeomorphic, it is enough to present a home
omorphism between them. Essentially this is what one usually does in
this case. To prove that spaces are not homeomorphic, it does not suﬃce
to consider any special mapping, and usually it is impossible to review all
the mappings. Therefore for proving nonexistence of a homeomorphism
one uses indirect arguments. In particular, one ﬁnds a property or a
characteristic shared by homeomorphic spaces and such that one of the
spaces has it, while the other does not. Properties and characteristics
which are shared by homeomorphic spaces are called topological proper
ties and invariants. Obvious examples of them are the cardinality (i.e.,
the number of elements) of the set of points and the set of open sets (cf.
Problems 10.30 and 10.T). Less obvious examples are the main object
of the next chapter.
'10
◦
8 Information: Nonhomeomorphic Spaces
Euclidean spaces of diﬀerent dimensions are not homeomorphic. The
balls D
p
, D
q
with p = q are not homeomorphic. The spheres S
p
, S
q
with p = q are not homeomorphic. Euclidean spaces are homeomorphic
neither to balls, nor to spheres (of any dimension). Letters A and B are
not homeomorphic (if the lines are absolutely thin!). Punctured plane
§10. HOMEOMORPHISMS 73
R
2
¦point¦ is not homeomorphic to the plane with hole R
2
¦ x
2
+y
2
<
1 ¦.
These statements are of diﬀerent degrees of diﬃculty. Some of them
will be considered in the next section. However some of them can not be
proven by techniques of this course. (See, e.g., D. B. Fuchs, V. A. Rokhlin.
Beginner’s course in topology: Geometric chapters. Berlin; New York:
SpringerVerlag, 1984.)
'10
◦
9 Embeddings
Continuous mapping f : X → Y is called a (topological ) embedding
if the submapping ab(f) : X →f(X) is a homeomorphism.
10.V. The inclusion of a subspace into a space is an embedding.
10.W. Composition of embeddings is an embedding.
10.X. Give an example of continuous injective map, which is not a topo
logical embedding. (Find such an example above and create a new one.)
10.33. Find topological spaces X and Y such that X can be embedded into
Y , Y can be embedded into X, but X
∼
= Y .
10.34. Prove that Q cannot be embedded into Z.
10.35. Can a discrete space be embedded into an indiscrete space? How
about vice versa?
10.36. Prove that spaces R, R
T1
, and the arrow cannot be embedded into
each other.
10.37 Corollary of Inverse Function Theorem. Deduce from the Inver
se Function Theorem (see, e.g., any course of advanced calculus) the following
statement:
For any diﬀerentiable function f : R
n
→ R
n
whose Jacobian det(
∂fi
∂xj
)
does not vanish at the origin 0 ∈ R
n
there exists a neighborhood U of the
origin such that f[
U
: U →R
n
is an embedding and f(U) is open.
'10
◦
10 Equivalence of Embeddings
Embeddings f
1
, f
2
: X → Y are said to be equivalent, if there exist
homeomorphisms h
X
: X → X and h
Y
: Y → Y such that f
2
◦ h
X
=
h
Y
◦ f
1
(the latter equality may be stated as follows: the diagram
X
f
1
−−−→ Y
h
X
h
Y
X
f
2
−−−→ Y
is commutative).
§10. HOMEOMORPHISMS 74
An embedding of the circle S
1
into R
3
is called a knot.
10.38. Prove that knots f
1
, f
2
: S
1
→ R
3
with f
1
(S
1
) = f
2
(S
1
) are equiva
lent.
10.39. Prove that knots are equivalent.
'10
◦
11 Information
There are nonequivalent knots. For instance, and .
CHAPTER 2
Topological Properties
'11 Connectedness
'11
◦
1 Deﬁnitions of Connectedness and First Examples
A topological space X is said to be connected if it has only two subsets
which are both open and closed: ∅ and the entire X.
A partition of a set is a cover of this set with pairwise disjoint sets.
To partition a set means to construct such a cover.
11.A. A topological space is connected, iﬀ it cannot be partitioned into
two nonempty open sets, iﬀ it cannot be partitioned into two nonempty
closed sets.
11.1. Is an indiscrete space connected? The same for the arrow and R
T1
.
11.2. Describe explicitly all connected discrete spaces.
11.3. Is the set Q of rational numbers (with the topology induced from R)
connected? The same about the set of irrational numbers.
11.4. Let Ω
1
, Ω
2
be topological structures in a set X, and Ω
2
be ﬁner than
Ω
1
(i.e., Ω
1
⊂ Ω
2
). If (X, Ω
1
) is connected, is (X, Ω
2
) connected? If (X, Ω
2
)
is connected, is (X, Ω
1
) connected?
'11
◦
2 Connected Sets
When one says that a set is connected, it means that this set lies in
some topological space (which should be clear from the context), and,
with the induced topology, is a connected topological space.
11.5. Give a deﬁnition of disconnected subset without relying on the induced
topology.
11.6. Is the set ¦0, 1¦ connected in R, in the arrow, in R
T1
?
11.7. Describe explicitly all connected subsets of the arrow, of R
T1
.
11.8. Show that the set [0, 1] ∪ (2, 3] is disconnected in R.
11.9. Prove that every nonconvex subset of the real line is disconnected.
11.10. Let A be a subset of a topological space X. Prove that A is dis
connected, iﬀ there exist nonempty sets B and C such that A = B ∪ C,
B ∩ Cl
X
C = ∅, and C ∩ Cl
X
B = ∅.
75
§11. CONNECTEDNESS 76
11.11. Find a topological space X and disconnected subset A ⊂ X such that
for any disjoint open sets U and V , which form a cover of X, either U ⊃ A,
or V ⊃ A.
11.12. Prove that for every disconnected set A in R
n
there exist disjoint
open sets U and V such that A ⊂ U ∪ V , U ∩ A = ∅, and V ∩ A = ∅.
Compare 11.10–11.12 with 11.5.
'11
◦
3 Properties of Connected Sets
11.B. The closure of a connected set is connected.
11.13. Prove that if a set A is connected and A ⊂ B ⊂ Cl A, then B is
connected.
11.C. Let ¦A
λ
¦
λ∈Λ
be a family of connected subsets of a space X. As
sume that any two sets of this family intersect. Then
¸
λ∈Λ
A
λ
is con
nected. (In other words: the union of pairwise intersecting connected sets
is connected.)
11.D. Let ¦A
k
¦
k∈Z
be a family of connected sets such that A
k
∩A
k+1
= ∅
for any k ∈ Z. Prove that
¸
k∈Z
A
k
is connected.
11.14. Let A, B be connected sets, and A∩ Cl B = ∅. Prove that A ∪ B is
connected.
11.15. Let A be a connected subset of a connected space X, and B ⊂ XA
be an open and closed set in the topology of the subspace X A of the
space X. Prove that A ∪ B is connected.
11.16. Does connectedness of A ∪ B and A ∩ B imply connectedness of A
and B?
11.17. Prove that if A and B are either both closed or both open sets, and
their union and intersection are connected then A and B are connected, too.
11.18. Let A
1
⊃ A
2
⊃ be an inﬁnite descending sequence of connected
spaces. Is
¸
∞
k=1
A
k
a connected set?
'11
◦
4 Connected Components
A connected component of a space X is its maximal connected subset,
that is a connected subset, which is not contained in any other (strictly)
larger connected subset of X.
11.E. Every point belongs to some connected component. Moreover, this
component is unique. It is the union of all connected sets containing this
point.
11.F. Connected components are closed.
11.G. Two connected components either are disjoint or coincide.
§11. CONNECTEDNESS 77
A connected component of a space X is called just a component of X.
Theorems 11.E and 11.G mean that connected components comprise a
partition of the whole space. The next theorem describes the correspond
ing equivalence relation.
11.H. Prove that two points are in the same component, iﬀ they belong
to the same connected set.
11.19. Let x and y belong to the same component. Prove that any set,
which is closed and open, either contains both x and y or does not contain
either of them (cf. 11.29).
11.20. Let a space X has a group structure, and the multiplication by an
element of the group is a continuous map. Prove that the component of unity
is a normal subgroup.
'11
◦
5 Totally Disconnected Spaces
A topological space is called totally disconnected if each of its compo
nents consists of a single point.
11.I Obvious Example. Any discrete space is totally disconnected.
11.J. The space Q (with the topology induced from R) is totally dis
connected.
Note that Q is not discrete.
11.21. Give an example of an uncountable closed totally disconnected subset
of the line.
11.22. Prove that Cantor set (see 2:A) is totally disconnected.
'11
◦
6 Frontier and Connectedness
11.23. Prove that if A is a proper nonempty subset of a connected topological
space then Fr A = ∅.
11.24. Let F be a connected subset of X. Prove that if A ⊂ X, F ∩ A, and
F ∩ (X A) = ∅ then F ∩ Fr A = ∅.
11.25. Let A be a subset of connected topological space. Prove that if Fr A
is a connected set then Cl A is also connected.
'11
◦
7 Connectedness and Continuous Maps
A continuous image of a space is its image under a continuous map
ping.
11.K. A continuous image of a connected space is connected. (In other
words if f : X →Y is a continuous map, and X is connected then f(X)
is also connected.)
11.L Corollary. Connectedness is a topological property.
§11. CONNECTEDNESS 78
11.M Corollary. The number of connected components is a topological
invariant.
11.N. A space X is not connected, iﬀ there is a continuous surjection
X →S
0
.
'11
◦
8 Connectedness on Line
11.O. The segment I = [0, 1] is connected.
There are several ways to prove 11.O. One is suggested by 11.N, but
refers to a famous Intermediate Value Theorem from calculus, see 11.T.
Basically the same proof as a combination of 11.N with a traditional proof
of Intermidiate Value Theorem is sketched in the following two problems.
Cf. also 11.26 below.
11.O.1. Let U, V be subsets of I with V = U V . Let a ∈ U, b ∈ V
and a > b. Prove that there exists a descending sequence a
n
with a
1
= a,
a
n
∈ U and an ascending sequence b
n
with b
1
= b, b
n
∈ V such that both
a
n
and b
n
have the same limit c.
11.O.2. If under assumptions of 11.O.1 U and V are open, then in which
of them can be c?
11.26. Prove that every open subset of the real line is a union of disjoint
open intervals (do not use 11.O). Deduce 11.O from this.
11.P. Prove that the set of connected components of an open subset of
R is countable.
11.Q. Prove that R
1
is connected.
11.R. Describe explicitly all connected subsets of the line.
11.S. Prove that every convex set in R
n
is connected.
11.27. Consider the union of spiral
r = exp
1
1 +ϕ
2
, with ϕ ≥ 0
(r, ϕ are the polar coordinates) and circle S
1
. Is this set connected? Would
the answer change, if the entire circle was replaced by some its subset?
(Cf. 11.13)
11.28. Consider the subset of the plane R
2
consisting of points with both
coordinates rational or both coordinates irrational. Is it connected?
11.29. Find a space and two points belonging to its diﬀerent components
such that each simultaneously open and closed set contains either both of the
points, or neither of them (cf. 11.19).
§11. CONNECTEDNESS 79
'11
◦
9 Intermediate Value Theorem and Its Generalizations
The following theorem is usually included in Calculus. You can easily
deduce it from the matterial of this section. In fact, in a sense it is
equivalent to connectedness of interval.
11.T Intermediate Value Theorem. A continuous function
f : [a, b] →R
takes every value between f(a) and f(b).
Many problems which can be solved using Intermediate Value Theorem
can be found in Calculus textbooks. Here are few of them.
11.30. Prove that any polynomial of odd degree in one variable with real
coeﬃcients has at least one real root.
11.U Generalization. Let X be a connected space and f : X → R a
continuous function. Then f(X) is a convex subset of R.
'11
◦
10 Dividing Pancakes
11.31. Any irregularly shaped pancake can be cut in half by one stroke of
the knife made in any prescribed direction. In other words, if A is a bounded
open set in the plane and l is a line in the plane, then there exists a line L
parallel to l which divides A in half by area.
11.32. If, under the conditions of 11.31, A is connected then L is unique.
11.33. Suppose two irregularly shaped pancakes lie on the same platter;
show that it is possible to cut both exactly in half by one stroke of the knife.
In other words: if A and B are two bounded regions in the plane, then there
exists a line in the plane which divides each region in half by area.
11.34 Dividing Pancake. Prove that a plane pancake of any shape can be
divided to four pieces of equal area by two straight cuts orthogonal to each
other. In other words, if A is a bounded connected open set in the plane,
then there are two perpendicular lines which divide A into four parts having
equal areas.
11.35. Enigma. What if the knife is not makes cuts of a shape diﬀerent
from straight line? For which shapes of the blade you can formulate and solve
problems similar to 11.31 – 11.34?
11.36. Enigma. Formulate and solve counterparts of Problems 11.31 –
11.34 for regions in the threedimensional space. Can you increase the number
of regions in the counterpart of 11.31 and 11.33?
11.37. Enigma. What about pancakes in R
n
?
'11
◦
11 Induction on Connectedness
A function is said to be locally constant if each point of its source
space has a neighborhood such that the restriction of the function to this
neighborhood is constant.
11.V. A locally constant function on a connected set is constant.
§11. CONNECTEDNESS 80
11.38. Enigma. How are 11.24 and 11.V related?
11.39. Let G be a group equipped with a topology such that for any g ∈
G the map G → G deﬁned by x → xgx
−1
is continuous, and let G with
this topology be connected. Prove that if the topology induced in a normal
subgroup H of G is discrete, then H is contained in the center of G (i.e.,
hg = gh for any h ∈ H and g ∈ G).
11.40 Induction on Connectedness. Let c be a property of subsets of a
topological space such that the union of sets with nonempty pairwise inter
sections inherits this property from the sets involved. Prove that if the space
is connected and each its point has a neighborhood with property c, then the
space has property c.
11.41. Prove 11.V and solve 11.39 using 11.40.
For more applications of induction on connectedness see 12.R, 12.14,
12.16 and 12.18.
'11
◦
12 Applications to Homeomorphism Problem
Connectedness is a topological property, and the number of connected
components is a topological invariant (see Section '10).
11.W. [0, 2] and [0, 1] ∪ [2, 3] are not homeomorphic.
Simple constructions, which assign homeomorphic spaces to home
omorphic ones (e.g. deleting one or several points), allow one to use
connectedness for proving that some connected spaces are not homeo
morphic.
11.X. I, R
1
, S
1
and [0, ∞) are pairwise nonhomeomorphic.
11.42. Prove that a circle is not homeomorphic to any subspace of R
1
.
11.43. Give a topological classiﬁcation of the letters: A, B, C, D,. . . , consid
ered as subsets of the plane (the arcs comprising the letters are assumed to
have zero thickness).
11.44. Prove that square and segment are not homeomorphic.
Recall that there exist continuous surjections of the segment onto
square and these maps are called Peano curves, see Section '9.
11.Y. R
1
and R
n
are not homeomorphic if n > 1.
Information. R
p
and R
q
are not homeomorphic unless p = q. It
follows, for instance, from the LebesgueBrower Theorem on invariance
of dimension (see, e.g., W. Hurewicz and H. Wallman, Dimension Theory
Princeton, NJ, 1941).
11.45. The statement “R
p
is not homeomorphic to R
q
unless p = q” implies
that S
p
is not homeomorphic to S
q
unless p = q.
'12 PathConnectedness
'12
◦
1 Paths
A path in a topological space X is a continuous mapping of the in
terval I = [0, 1] to X. The point s(0) is called the initial point of a path
s : I →X, while s(1) is called its ﬁnal point. One says that path s con
nects s(0) with s(1). This terminology is inspired by an image of moving
point: at the moment t ∈ [0, 1] it is in s(t). To tell the truth, this is
more than what is usually called a path, since besides an information on
trajectory of the point it contains a complete account on the movement:
the schedule saying when the point goes through each point.
A constant map s : I → X is called a stationary path and denoted
by e
a
where a = s(I). For a path s the inverse path is the path deﬁned
by t → s(1 − t). It is denoted by s
−1
. Although, strictly speaking,
this notation is already used (for the inverse mapping), the ambiguity
of notations does not lead to confusion: in the context involving paths,
inverse mappings, as a rule, do not appear.
Let u : I →X, v : I →X be paths such that u(1) = v(0). Set
(22) uv(t) =
u(2t), if t ∈ [0,
1
2
]
v(2t −1), if t ∈ [
1
2
, 1].
12.A. Prove that the map uv : I →X deﬁned by ('12
◦
1) is continuous
(i.e., it is a path). Cf. 9.V and 9.X.
Path uv is called the product of paths u and v. Recall that it is
deﬁned only if the ﬁnal point u(1) of u coincides with the initial point
v(0) of v.
'12
◦
2 PathConnected Spaces
A topological space is said to be pathconnected or pathwise con
nected, if any two points can be connected in it by a path.
12.B. Prove that I is pathwise connected.
12.C. Prove that the Euclidean space of any dimension is pathwise con
nected.
12.D. Prove that sphere of dimension n > 0 is pathconnected.
12.E. Prove that the zerodimensional sphere S
0
is not pathconnected.
12.1. Which of the following topological spaces are pathconnected:
(a) a discrete space; (b) an indiscrete space;
(c) the arrow; (d) R
T1
;
(e) ?
81
§12. PATHCONNECTEDNESS 82
'12
◦
3 PathConnected Sets
By a pathconnected set or pathwise connected set one calls a subset
of a topological space (which should be clear from the context) path
connected as a space with the topology induced from the ambient space.
12.2. Prove that a subset A of a topological space X is pathconnected, iﬀ
any two points in it can be connected by a path s : I →X with s(I) ⊂ A.
12.3. Prove that a convex subset of Euclidean space is pathconnected.
12.4. Prove that the set of plane convex polygons with topology deﬁned by
the Hausdorﬀ metric is pathconnected.
'12
◦
4 Properties of PathConnected Sets
Pathconnectedness is very similar to connectedness. Further, in some
important situations it is even equivalent to connectedness. However,
some properties of connectedness do not carry over pathconnectedness
(see 12.O, 12.P). For properties, which carry over, proofs are usually
easier in the case of pathconnectedness.
12.F. The union of a family of pairwise intersecting pathconnected sets
is pathconnected.
12.5. Prove that if sets A and B are both closed or both open and their union
and intersection are pathconnected, then A and B are also pathconnected.
12.6. Prove that interior and frontier of a pathconnected set may not be
pathconnected and that connectedness shares this property.
12.7. Let A be a subset of Euclidean space. Prove that if Fr A is connected
then Cl A is also connected.
12.8. Prove that the same holds true for a subset of an arbitrary path
connected space.
'12
◦
5 PathConnected Components
A pathconnected component or pathwise connected component of a
topological space X is a pathconnected subset of X such that no other
pathconnected subset of X contains it.
12.G. Every point belongs to a pathconnected component.
12.H. Two pathconnected components either coincide or are disjoint.
12.I. Prove that two points belong to the same pathconnected compo
nent, iﬀ they can be connected by a path.
Unlike to the case of connectedness, pathconnected components may
be nonclosed. (See 12.O, cf. 12.N, 12.P.)
12.J. A continuous image of a pathwise connected space is pathwise con
nected.
§12. PATHCONNECTEDNESS 83
12.9. Let s : I →X be a path connecting a point of a set A with a point of
X A. Prove that s(I) ∩ Fr(A) = ∅.
'12
◦
6 PathConnectedness Versus Connectedness
12.K. Any pathconnected space is connected.
Put
A =
(x, y) ∈ R
2
: x > 0, y = sin
1
x
and X = A∪ ¦(0, 0)¦.
12.10. Draw A.
12.L. Prove that A is pathconnected and X is connected.
12.M. Prove that deleting any point from A makes A and X discon
nected (and hence, not pathconnected).
12.N. X is not pathconnected.
12.O. Find an example of a pathconnected set, whose closure is not
pathconnected.
12.P. Find an example of a pathconnected component that is not
closed.
12.Q. If each point of a space has a pathconnected neighborhood, then
each pathconnected component is open.
12.R. If each point of a space has a pathconnected neighborhood, then
the space is pathconnected, iﬀ it is connected.
12.S. For an open subset of Euclidean space connectedness is equivalent
to pathconnectedness.
12.11. For subsets of the real line pathconnectedness and connectedness are
equivalent.
12.12. Prove that for any ε > 0 an εneighborhood of a connected subset of
Euclidean space is pathconnected.
12.13. Prove that any neighborhood of a connected subset of Euclidean
space contains a pathconnected neighborhood of the same set.
'12
◦
7 PolygonConnectedness
A subset A of Euclidean space is said to be polygonconnected if any two
points of A can be connected by a ﬁnite polygonal line contained in A.
12.14. Prove that for open subsets of Euclidean space connectedness is equiv
alent to polygonconnectedness.
§12. PATHCONNECTEDNESS 84
12.15. Construct a pathconnected subset A of Euclidean space such that A
consists of more than one point and no two distinct points can be connected
with a polygon in A.
12.16. Let X ⊂ R
2
be a countable set. Prove that then R
2
X is polygon
connected.
12.17. Let X ⊂ R
n
be a union of a countable collection of aﬃne subspaces
with dimensions not greater than n −2. Prove that then R
n
X is polygon
connected.
12.18. Let X ⊂ C
n
be a union of a countable collection of algebraic sub
sets (i.e., subsets deﬁned by systems of algebraic equations in the standard
coordinates of C
n
) Prove that then C
n
X is polygonconnected.
'12
◦
8 Connectedness of Some Sets of Matrices
Recall, that real nnmatrices comprise a space, which diﬀers from R
n
2
only in the way of enumeration of its natural coordinates (they are numerated
by pairs of indices). The same relation holds between the set of complex nn
matrix and C
n
2
(homeomorphic to R
2n
2
).
12.19. Find connected and pathconnected components of the following sub
spaces of the space of real n nmatrices:
• GL(n; R) = ¦A : det A = 0¦;
• O(n; R) = ¦A : A (
t
A) = ¦;
• Symm(n; R) = ¦A :
t
A = A¦;
• Symm(n; R) ∩ GL(n; R);
• ¦A : A
2
= ¦.
12.20. Find connected and pathconnected components of the following sub
spaces of the space of complex n nmatrices:
• GL(n; C) = ¦A : det A = 0¦;
• U(n; C) = ¦A : A (
t
¯
A) = ¦;
• Herm(n; C) = ¦A :
t
A =
¯
A¦;
• Herm(n; C) ∩ GL(n; C).
'13 Separation Axioms
The aim of this section is to consider natural restrictions on topolog
ical structure making the structure closer to being metrizable. A lot of
separation axioms are known. We restrict ourselves to the most impor
tant ﬁve of them. They are numerated and denoted by T
0
, T
1
, T
2
, T
3
,
and T
4
, respectively.
1
'13
◦
1 The Hausdorﬀ Axiom
Let us start with the most important second axiom. Besides the
notation T
2
it has a name, the Hausdorﬀ axiom. A topological space
satisfying it is called a Hausdorﬀ space. This axiom is stated as follows:
any two distinct points possess disjoint neighborhoods.
13.A. Any metric space is Hausdorﬀ.
13.1. Which of the following spaces are Hausdorﬀ:
(a) a discrete space;
(b) an indiscrete space;
(c) the arrow;
(d) R
T1
;
(e) ?
If the next problem holds you up even for a minute, we advise you to
think over all deﬁnitions and solve all simple problems.
13.B. Is the segment [0, 1] with the topology induced fromR a Hausdorﬀ
space? Do the points 0 and 1 possess disjoint neighborhoods? Which if
any?
'13
◦
2 Limits of Sequence
Let ¦a
n
¦ be a sequence of points of a topological space X. A point
b ∈ X is called its limit, if for any neighborhood U of b there exists a
number N such that a
n
∈ U for any n > N. The sequence is said to
converge or tend to b as n tends to inﬁnity.
13.2. Explain the meaning of the statement “ b is not a limit of sequence a
n
”
avoiding as much as you can negations (i.e., the words no, not, none, etc..)
13.C. In a Hausdorﬀ space any sequence has at most one limit.
13.D. Prove that in the space R
T
1
each point is a limit of the sequence
¦a
n
= n¦.
1
Letter T in these notations originates from a German word Trennungsaxiom,
which means separation axiom.
85
§13. SEPARATION AXIOMS 86
'13
◦
3 Coincidence Set and Fixed Point Set
Let f, g : X →Y be maps. Then the set ¦x ∈ X : f(x) = g(x)¦ is called
the coincidence set of f and g.
13.3. Prove that the coincidence set for two continuous maps of an arbitrary
topological space to a Hausdorﬀ space is closed.
13.4. Construct an example proving that the Hausdorﬀ condition in 13.3 is
essential.
A point x ∈ X is called a ﬁxed point of a map f : X → X if f(x) = x.
The set of all ﬁxed points of a map f is called the ﬁxed point set of f.
13.5. Prove that the ﬁxed point set of a continuous map of a Hausdorﬀ space
to itself is closed.
13.6. Construct an example proving that the Hausdorﬀ condition in 13.5 is
essential.
13.7. Prove that if f, g : X → Y are continuous maps, Y is Hausdorﬀ, A is
everywhere dense in X, and f[
A
= g[
A
then f = g.
13.8. Enigma. How are problems 13.3, 13.5, and 13.7 related?
'13
◦
4 Hereditary Properties
A topological property is called hereditary if it is carried over from a
space to its subspaces, i.e. if a space X possesses this property then any
subspace of X possesses it.
13.9. Which of the following topological properties are hereditary:
• ﬁniteness of the set of points;
• ﬁniteness of the topological structure;
• inﬁniteness of the set of points;
• connectedness;
• pathconnectedness?
13.E. The property of being Hausdorﬀ space is hereditary.
'13
◦
5 The First Separation Axiom
A topological space is said to satisfy the ﬁrst separation axiom T
1
if
each of any two points of the space has a neighborhood which does not
contain the other point.
13.F. A topological space X satisﬁes the ﬁrst separation axiom,
• iﬀ all onepoint sets in X are closed,
• iﬀ all ﬁnite sets in X are closed.
13.10. Prove that a space X satisﬁes the ﬁrst separation axiom, iﬀ any point
of X coincides with the intersection of all its neighborhoods.
13.11. Any Hausdorﬀ space satisﬁes the ﬁrst separation axiom.
13.G. In a Hausdorﬀ space any ﬁnite set is closed.
13.H. A metric space satisﬁes the ﬁrst separation axiom.
§13. SEPARATION AXIOMS 87
13.12. Find an example showing that the ﬁrst separation axiom does not
imply the Hausdorﬀ axiom.
13.I. Show that R
T
1
meets the ﬁrst separation axiom, but is not a Haus
dorﬀ space (cf. 13.12).
13.J. The ﬁrst separation axiom is hereditary.
13.13. Prove that if for any two distinct points a and b of a topological
space X there exists a continuous map f of X to a space with the ﬁrst
separation axiom such that f(a) = f(b) then X possesses the ﬁrst separation
axiom.
13.14. Prove that a continuous mapping of an indiscrete space to a space
satisfying axiom T
1
is constant.
13.15. Prove that in every set there exists a coarsest topological structure
satisfying the ﬁrst separation axiom. Describe this structure.
'13
◦
6 The Kolmogorov Axiom
The ﬁrst separation axiom emerges as a weakened Hausdorﬀ axiom.
13.K. Enigma. How can the ﬁrst separation axiom be weakened?
A topological space is said to satisfy the Kolmogorov axiom or the
zeroth separation axiom T
0
, if at least one of any two distinct points of
this space has a neighborhood which does not contain the other of these
points.
13.L. An indiscrete space containing at least two points does not satisfy
T
0
.
13.M. The following properties of a topological space X are equivalent:
• X satisﬁes the Kolmogorov axiom;
• any two diﬀerent points of X has diﬀerent closures;
• X contains no indiscrete subspace consisting of more than one point;
• X contains no indiscrete subspace consisting of two points.
13.N. A topology is a poset topology, iﬀ this is a smallest neighborhood
topology satisfying the Kolmogorov axiom.
13.N.1. Express axioms of nonstrict order in terms of poset topology.
13.N.2. Prove that if (X, Ω) is a smallest neighborhood topological space
satisfying the Kolmogorov axiom then the relation “a belongs to any closed
set which contain b” is a nonstrict partial order in X and the poset topol
ogy deﬁned by this partial order coincides with Ω.
Thus, on one hand, posets give rise to numerous examples of topolog
ical spaces, among which we see the most important spaces, like the line
with the standard topology, on the other hand, all posets are obtained
from topological spaces of a special kind, which is quite far away from
the class of metric spaces.
§13. SEPARATION AXIOMS 88
'13
◦
7 The Third Separation Axiom
A topological space X is said to satisfy the third separation axiom if
any closed set and a point of its complement have disjoint neighborhoods,
i.e., for any closed set F ⊂ X and point b ∈ X ` F there exist open sets
U, V ⊂ X such that U ∩ V = ∅, F ⊂ U, and b ∈ V .
A topological space is called regular if it satisﬁes the ﬁrst and third
separation axioms.
13.O. A regular space is Hausdorﬀ space.
13.P. A space is regular, iﬀ it satisﬁes the second and third separation
axioms.
13.16. Find a Hausdorﬀ space which is not regular.
13.17. Find a space satisfying the third, but not the second separation ax
iom.
13.18. Prove that a space satisﬁes the third separation axiom, iﬀ any neigh
borhood of any point contains the closure of some neighborhood of the same
point.
13.19. Prove that the third separation axiom is hereditary.
13.Q. Any metric space is regular.
'13
◦
8 The Fourth Separation Axiom
A topological space X is said to satisfy the fourth separation axiom
if any two disjoint closed sets have disjoint neighborhoods, i.e., for any
closed sets A, B ⊂ X such that A∩B = ∅ there exist open sets U, V ⊂ X
such that U ∩ V = ∅, A ⊂ U, and B ⊂ V .
A topological space is called normal if it satisﬁes the ﬁrst and fourth
separation axioms.
13.R. A normal space is regular (and hence Hausdorﬀ).
13.S. A space is normal, iﬀ it satisﬁes the second and fourth separation
axioms.
13.20. Find a space which satisﬁes the fourth, but not second separation
axiom.
13.21. Prove that a space satisﬁes the fourth separation axiom, iﬀ in any
neighborhood of any closed set contains the closure of some neighborhood of
the same set.
13.22. Prove that any closed subspace of a normal space is normal.
13.23. Find closed disjoint subsets A and B of some metric space such that
inf¦ρ(a, b) [ a ∈ A, b ∈ B¦ = 0.
13.T. Any metric space is normal.
13.24. Let f : X →Y be a continuous surjection such that the image of any
closed set is closed. Prove that if X is normal then Y is normal.
§13. SEPARATION AXIOMS 89
'13
◦
9 Niemytski’s Space
Denote by H the open upper halfplane ¦(x, y) ∈ R
2
: y > 0¦ equipped
with the topology induced by the Euclidean metric. Denote by X the union
of H and its boundary line L = ¦(x, y) ∈ R
2
: y = 0¦, but equip it with the
topology, which is obtained by adjoining to the Euclidean topology the sets
of the form x∪D, where x ∈ R
1
and D is an open disc in H which is tangent
to L at the point x. This is the Niemytski space. It can be used to clarify
properties of the fourth separation axiom.
13.25. Prove that the Niemytski space is Hausdorﬀ.
13.26. Prove that the Niemytski space is regular.
13.27. What topological structure is induced on L from X?
13.28. Prove that the Niemytski space is not normal.
13.29 Corollary. There exists a regular space, which is not normal.
13.30. Embed the Niemytski space into a normal space in such a way that
the complement of the image would be a single point.
13.31 Corollary. Theorem 13.22 does not extend to nonclosed subspaces,
i.e., the property of being normal is not hereditary?
'13
◦
10 Urysohn Lemma and Tietze Theorem
13:A*. Let Y be a topological space satisfying the ﬁrst separation axiom. Let T
be a subbase
2
of the topology of Y . Let Σ be an open cover of a space X. Prove that
if there exists a bijection φ : Σ → T which preserves inclusions then there exists a
continuous map f : X →Y such that f
−1
(V ) = φ
−1
(V ) for any V ∈ T.
13:B. Prove that intervals [0, r) and (r, 1] where r =
n
2
q
, n, q ∈ N form a subbase for
[0, 1], i.e., a collection of open sets in [0, 1], whose ﬁnite intersections form a base of
the standard topology in [0, 1].
13:C Urysohn Lemma. Let A and B be disjoint closed subsets of a normal space X.
Then there exists a continuous function f : X →I such that f(A) = 0 and f(B) = 1.
13:D. Let A be a closed subset of a normal space X. Let f : A → [−1, 1] be a
continuous function. Prove that there exists a continuous function g : X →
−
1
3
,
1
3
such that [f(x) −g(x)[ ≤
2
3
for any x ∈ A.
13:E. Prove that under the conditions of 13:D for any ε > 0 there exists a continuous
function φ : X →[−1, 1] such that [f(x) −φ(x)[ ≤ ε for any x ∈ A.
13:F Tietze Extension Theorem. Prove that under the conditions of 13:D there
exists a continuous function F : X →[−1, 1] such that F
A
= f.
13:G. Would the statement of Tietze Theorem remain true if in the hypothesis the
segment [−1, 1] was replaced by R, R
n
, S
1
, or S
2
?
2
Recall that a subbase of the topology of Y is a collection T of open sets of Y such
that all ﬁnite intersections of sets from T form a base of topology of Y , see Section
'3.
'14 Countability Axioms
In this section we continue to study topological properties which are
imposed additionally on a topological structure to make the abstract
situation under consideration closer to special situations and hence richer
in contents. Restrictions studied in this section bound a topological
structure from above: they require something to be countable.
'14
◦
1 SetTheoretic Digression. Countability
Recall that two sets are said to be of equal cardinality if there exists
a bijection of one of them onto the other. A set of the same cardinality
as a subset of the set N of natural numbers is said to be countable.
Sometimes this term is used only for inﬁnite countable sets, i.e. for
set of the cardinality of the whole set N of natural numbers, while a
set countable in the sense above is called at most countable. This is
less convenient. In particular, if we adopted this terminology, then this
section would have to be called “At Most Countability Axioms”. This
would lead to other more serious inconveniences as well. Our terminology
has the following advantageous properties.
14.A. Any subset of a countable set is countable.
14.B. The image of a countable set under any mapping is countable.
14.C. The union of a countable family of countable sets is countable.
'14
◦
2 Second Countability and Separability
In this section we study three restrictions on topological structure.
Two of them have numbers (one and two), the third one has no number.
As in the previous section, we start from the restriction having number
two.
A topological space is said to satisfy the second axiom of countability
or to be second countable if it has a countable base. A space is called
separable if it contains a countable dense set. (This is the countability
axiom without a number mentioned above.)
14.D. The second axiom of countability implies separability.
14.E. The second axiom of countability is hereditary.
14.1. Are the arrow and R
T1
second countable?
14.2. Are the arrow and R
T1
separable?
14.3. Construct an example proving that separability is not hereditary.
14.F. A metric separable space is second countable.
14.G Corollary. For metric spaces, separability is equivalent to the
second axiom of countability.
90
§14. COUNTABILITY AXIOMS 91
14.H. (Cf. 14.3.) Prove that for metric spaces separability is hereditary.
14.I. Prove that Euclidean spaces and all their subspaces are separable
and second countable.
14.4. Construct a metric space which is not second countable.
14.J. A continuous image of a separable space is separable.
14.5. Construct an example proving that a continuous image of a second
countable space may be not second countable.
14.K Lindel¨ of Theorem. Any open cover of a second countable space
contains a countable part, which also covers the space.
14.6. Prove that any base of a second countable space contains a countable
part which is also a base.
14.7. Prove that in a separable space any collection of pairwise disjoint open
sets is countable.
14.8. Prove that the set of components of an open set A ⊂ R
n
is countable.
14.9. Prove that any set of disjoint ﬁgure eight curves in the plane is count
able.
14.10 Brower Theorem*. Let ¦K
λ
¦ be a family of closed sets of a second
countable space and let for any descending sequence K
1
⊃ K
2
⊃ . . . of sets
belonging to this family the intersection ∩
∞
1
K
n
also belongs to the family.
Then the family contains a minimal set, i.e., a set such that no proper its
subset belongs to the family.
'14
◦
3 Embedding and Metrization Theorems
14:A. Prove that the space l
2
is separable and second countable.
14:B. Prove that a regular second countable space is normal.
14:C. Prove that a normal second countable space can be embedded into l
2
. (Use
Urysohn Lemma 13:C.)
14:D. Prove that a second countable space is metrizable, iﬀ it is regular.
'14
◦
4 Bases at a Point
Let X be a topological space, and a its point. A neighborhood base
at a or just base of X at a is a collection of neighborhoods of a such that
any neighborhood of a contains a neighborhood from this collection.
14.L. If Σ is a base of a space X then ¦U ∈ Σ : a ∈ U¦ is a base of X
at a.
14.11. In a metric space the following collections of balls are neighborhood
bases at a point a:
• the set of all open balls of center a;
• the set of all open balls of center a and rational radii;
§14. COUNTABILITY AXIOMS 92
• the set of all open balls of center a and radii r
n
, where ¦r
n
¦ is any
sequence of positive numbers converging to zero.
14.12. What are the minimal bases at a point in the discrete and indiscrete
spaces?
'14
◦
5 First Countability
A topological space X is says to satisfy the ﬁrst axiom of countability
or to be a ﬁrst countable space if it has a countable neighborhood base
at each point.
14.M. Any metric space is ﬁrst countable.
14.N. The second axiom of countability implies the ﬁrst one.
14.O. Find a ﬁrst countable space which is not second countable. (Cf.
14.4.)
14.13. Which of the following spaces are ﬁrst countable:
(a) the arrow; (b) R
T1
;
(c) a discrete space; (d) an indiscrete space?
14.14. Find a ﬁrst countable separable space which is not second countable.
'14
◦
6 Sequential Approach to Topology
Specialists in Mathematical Analysis love sequences and their limits.
Moreover they like to talk about all topological notions relying on the
notions of sequence and its limit. This tradition has almost no mathe
matical justiﬁcation, except for a long history descending from the XIX
century studies on the foundations of analysis. In fact, almost always
3
it
is more convenient to avoid sequences, provided you deal with topolog
ical notions, except summing of series, where sequences are involved in
the underlying deﬁnitions. Paying a tribute to this tradition we explain
here how and in what situations topological notions can be described in
terms of sequences.
Let A be a subset of a topological space X. The set of limits of all
sequences a
n
with a
n
∈ A is called a sequential closure of A and denoted
by SCl A.
14.P. Prove that SCl A ⊂ Cl A.
14.Q. If a space X is ﬁrst countable then the for any A ⊂ X the opposite
inclusion Cl A ⊂ SCl A holds also true, and hence SCl A = Cl A.
3
The exceptions which one may ﬁnd in the standard curriculum of a mathematical
department can be counted on two hands.
§14. COUNTABILITY AXIOMS 93
Therefore, in a second countable space (in particular, any metric
spaces) one can recover (hence, deﬁne) the closure of a set provided it is
known which sequences are convergent and what the limits are. In turn,
knowledge of closures allows one to recover which sets are closed. As a
consequence, knowledge of closed sets allows one to recover open sets and
all other topological notions.
14.15. Let X be the set of real numbers equipped with the topology con
sisting of ∅ and complements of all countable subsets. Describe convergent
sequences, sequential closure and closure in X. Prove that in X there exists
a set A with SCl A = Cl A.
'14
◦
7 Sequential Continuity
Consider now continuity of maps along the same lines. A map f :
X → Y is said to be sequentially continuous if for any b ∈ X and a
sequence a
n
∈ X, which converges to b, the sequence f(a
n
) converges to
f(b).
14.R. Any continuous map is sequentially continuous.
14.S. The preimage of a sequentially closed set under a sequentially con
tinuous map is sequentially closed.
14.T. If X is a ﬁrst countable space then any sequentially continuous
map f : X →Y is continuous.
Thus for mappings of a ﬁrst countable space continuity and sequential
continuity are equivalent.
14.16. Construct a sequentially continuous, but discontinuous map. (Cf.
14.15)
'15 Compactness
'15
◦
1 Deﬁnition of Compactness
This section is devoted to a topological property, which plays a very
special role in topology and its applications. It is sort of topological
counterpart for the property of being ﬁnite in the context of set theory.
(It seems though that this analogy has never been formalized.)
Topological space is said to be compact if any of its open covers
contains a ﬁnite part which covers the space.
If Γ is a cover of X and Σ ⊂ Γ is a cover of X then GS is called a
subcover (or subcovering) of Γ. Thus, a topological space is compact if
every open cover has a ﬁnite subcover.
15.A. Any ﬁnite topological space and indiscrete space are compact.
15.B. Which discrete topological spaces are compact?
15.1. Let Ω
1
⊂ Ω
2
be topological structures in X. Does compactness of
(X, Ω
2
) imply compactness of (X, Ω
1
)? And vice versa?
15.C. The line R is not compact.
15.D. A topological space X is not compact, iﬀ there exists an open
covering which contains no ﬁnite subcovering.
15.2. Is the arrow compact? Is R
T1
compact?
'15
◦
2 Terminology Remarks
Originally the word compactness was used for the following weaker
property: any countable open cover contains a ﬁnite subcover.
15.E. For a second countable space the original deﬁnition of compactness
is equivalent to the modern one.
The modern notion of compactness was introduced by P. S. Alexan
droﬀ (1896–1982) and P. S. Urysohn (1898–1924). They suggested for it
the term bicompactness. This notion appeared to be so successful that
it has displaced the original one and even took its name, i.e. compact
ness. The term bicompactness is sometimes used (mainly by topologists
of Alexandroﬀ school).
Another deviation from the terminology used here comes from Bour
baki: we do not include the Hausdorﬀ property into the deﬁnition of
compactness, which Bourbaki includes. According to our deﬁnition, R
T
1
is compact, according to Bourbaki it is not.
94
§15. COMPACTNESS 95
'15
◦
3 Compactness in Terms of Closed Sets
A collection of subsets of a set is said to be centered if the intersection
of any ﬁnite subcollection is not empty.
15.F. A collection Σ of subsets of a set X is centered, iﬀ there exists
no ﬁnite Σ
1
⊂ Σ such that the complements of the sets belonging to Σ
1
cover X.
15.G. A topological space is compact, iﬀ any centered collection of its
closed sets has nonempty intersection.
'15
◦
4 Compact Sets
By a compact set one means a subset of a topological space (the latter
must be clear from the context) provided it is compact as a space with
the topology induced from the ambient space.
15.H. A subset A of a topological space X is compact, iﬀ any cover
which consists of sets open in X contains a ﬁnite subcover.
15.3. Is [1, 2) ⊂ R compact?
15.4. Is the same set [1, 2) compact in the arrow?
15.5. Find a necessary and suﬃcient condition (formulated not in topological
terms) for a subset of the arrow to be compact?
15.6. Prove that any subset of R
T1
is compact.
15.7. Let A and B be compact subsets of a topological space X. Does it
follow that A∪ B is compact? Does it follow that A ∩ B is compact?
15.8. Prove that the set A = ¦0¦ ∪ ¦
1
n
¦
∞
n=1
in R is compact.
'15
◦
5 Compact Sets Versus Closed Sets
15.I. Is compactness hereditary?
15.J. Any closed subset of a compact space is compact.
15.K. Any compact subset of a Hausdorﬀ space is closed.
15.L Lemma to 15.K, but not only . . . . Let A be a compact subset
of a Hausdorﬀ space X and b a point of X which does not belong to A.
Then there exists open sets U, V ⊂ X such that b ∈ V , A ⊂ U and
U ∩ V = ∅.
15.9. Construct a nonclosed compact subset of some topological space. What
is the minimal number of points needed?
§15. COMPACTNESS 96
'15
◦
6 Compactness and Separation Axioms
15.M. A compact Hausdorﬀ space is regular.
15.N. Prove that a compact Hausdorﬀ space is normal.
15.10. Prove that the intersection of any family of compact subsets of a
Hausdorﬀ space is compact. (Cf. 15.7.)
15.11. Let X be a Hausdorﬀ space, let ¦K
α
¦
α∈Λ
be a family of its compact
subsets, and let U be an open set containing ∩
α∈Λ
K
α
. Prove that U ⊃
∩
α∈A
K
α
for some ﬁnite A ⊂ Λ.
15.12. Let ¦K
n
¦ be a decreasing sequence of compact nonempty connected
subset of a Hausdorﬀ space. Prove that the intersection ∩
∞
n=1
K
n
is nonempty
and connected.
15.13. Construct a decreasing sequence of connected subsets of the plane
with nonconnected intersection.
15.14. Let K be a connected component of a compact Hausdorﬀ space X
and let U be an open set containing K. Prove that there exists an open and
closed set V such that K ⊂ V ⊂ U.
'15
◦
7 Compactness in Euclidean Space
15.O. The interval I is compact.
Recall that ndimensional cube is the set
I
n
= ¦x ∈ R
n
[ x
i
∈ [0, 1] for i = 1, . . . , n¦.
15.P. The cube I
n
is compact.
15.Q. Any compact subset of a metric space is bounded.
Therefore, any compact subset of a metric space is closed and bounded,
see 15.K and 15.Q.
15.R. Construct a closed and bounded, but noncompact set of a metric
space.
15.15. Are the metric spaces of Problem 4.A compact?
15.S. A subset of a Euclidean space is compact, iﬀ it is closed and
bounded.
15.16. Which of the following sets are compact:
(a) [0, 1); (b) ray R
+
= ¦x ∈ R[ x ≥ 0¦; (c) S
1
;
(d) S
n
; (e) onesheeted hyperboloid; (f) ellipsoid;
(g) [0, 1] ∩ Q?
Matrix (a
ij
) with 1 ≤ i ≤ n, 1 ≤ j ≤ k with real a
i
j can be considered as
a point of R
nk
. For this, one needs to enumerate somehow (e.g, lexicograph
ically) its elements by numbers from 1 till nk. This identiﬁes the set L(nk)
of all matrices like that with R
nk
and endows it with a topological structure.
(Cf. Section '12.)
§15. COMPACTNESS 97
15.17. Which of the following subsets of L(n, n) are compact:
(a) GL(n) = ¦A ∈ L(n, n) [ det A = 0¦;
(b) SL(n) = ¦A ∈ L(n, n) [ det A = 1¦;
(c) O(n) = ¦A ∈ L(n, n) : [ A is an orthogonal matrix¦;
(d) ¦A ∈ L(n, n) [ A
2
= E¦, here E is the unit matrix?
'15
◦
8 Compactness and Continuous Maps
15.T. A continuous image of a compact set is compact. (In other words,
if X is a compact space and f : X →Y is a continuous map then f(X)
is compact.)
15.U. On a compact set any continuous function is bounded and attains
its maximal and minimal values. (In other words, if X is a compact space
and f : X → R is a continuous function, then there exist a, b ∈ X such
that f(a) ≤ f(x) ≤ f(b) for any x ∈ X.) Cf. 15.T and 15.S.
15.18. Prove that if f : I → R is a continuous function then f(I) is an
interval.
15.19. Prove that if F and G are disjoint subsets of a metric space, F is
closed and G compact then ρ(F, G) = inf ¦ρ(x, y) [ x ∈ F, y ∈ G¦ > 0.
15.20. Prove that any open set containing a compact set A of a metric space
X contains an εneighborhood of A. (i.e., the set ¦x ∈ X[ ρ(x, A) < ε¦ for
some ε > 0).
15.21. Let A be a closed connected subset of R
n
and let V be its closed
εneighborhood (i.e., V = ¦x ∈ R
n
[ ρ(x, A) < ε¦). Prove that V is path
connected.
15.22. Prove that if in a compact metric space the closure of any open ball
is the closed ball with the same center and radius then any ball of this space
is connected.
15.23. Let X be a compact metric space and f : X → X be a map such
that ρ(f(x), f(y)) < ρ(x, y) for any x, y ∈ X with x = y. Prove that f has
a unique ﬁxed point. (Recall that a ﬁxed point of f is a point x such that
f(x) = x.)
15.24. Prove that for any open cover of a compact metric space there exists
a number r > 0 such that any open ball of radius r is contained in some
element of the cover.
15.V Lebesgue Lemma. Let f : X → Y be a continuous map of a
compact metric space X to a topological space Y , and let Γ be an open
cover of Y . Then there exists a number δ > 0 such that for any set
A ⊂ X with diameter diam(A) < δ the image f(A) is contained in some
element of Γ.
'15
◦
9 Closed Maps
A continuous map is said to be closed if the image of any closed set
under this map is closed.
§15. COMPACTNESS 98
15.W. A continuous map of a compact space to a Hausdorﬀ space is
closed.
Here are two important corollaries of this theorem.
15.X. A continuous injection of a compact space to a Hausdorﬀ space
is a topological embedding.
15.Y. A continuous bijection of a compact space to a Hausdorﬀ space is
a homeomorphism.
15.25. Show that none of the hypothesis in 15.Y can be omitted without
making the statement false.
15.26. Does there exist a noncompact subspace of Euclidian space such that
any its map to a Hausdorﬀ space is closed? (Cf. 15.U and 15.W.)
'15
◦
10 Norms in R
n
15.27. Prove that any normR
n
→R (see Section '4) is a continuous function
(with respect to the standard topology of R
n
).
15.28. Prove that any two norms in R
n
are equivalent (i.e. deﬁne the same
topological structure). See 4.27, cf. 4.31.
15.29. Does the same hold true for metrics in R
n
?
'16 Local Compactness and Paracompactness
'16
◦
1 Local Compactness
A topological space X is called locally compact if each of its points has a neigh
borhood with compact closure.
16:A. Local compactness is a local property, i.e., a space is locally compact, iﬀ each
of its points has a locally compact neighborhood.
16:B. Is local compactness hereditary?
16:C. A closed subset of a locally compact space is locally compact.
16:D. An open subset of a locally compact Hausdorﬀ space is locally compact.
16:1. Which of the following spaces are locally compact:
(a) R;
(b) Q;
(c) R
n
;
(d) a discrete space?
16:2. Find two locally compact sets on the line such that their union is not
locally compact.
'16
◦
2 OnePoint Compactiﬁcation
Let X be a Hausdorﬀ topological space. Let X
∗
be the set obtained by adding a
point to X (of course, the point does not belong to X). Let Ω
∗
be the collection of
subsets of X
∗
consisting of
• sets open in X and
• sets of the form X
∗
C, where C ⊂ X is a compact set.
16:E. Prove that Ω
∗
is a topological structure.
16:F. Prove that the space (X
∗
, Ω
∗
) is compact.
16:G. Prove that the inclusion X → X
∗
is a topological embedding (with respect to
the original topology of X and Ω
∗
).
16:H. Prove that if X is locally compact then the space (X
∗
, Ω
∗
) is Hausdorﬀ. (Recall
that X is assumed to be Hausdorﬀ.)
A topological embedding of a space X into a compact space Y is called a com
pactiﬁcation of X if the image of X is dense in Y . In this situation Y is also called a
compactiﬁcation of X.
16:I. Prove that if X is a locally compact Hausdorﬀ space and Y is its compactiﬁ
cation with Y X consisting of a single point then there exists a homeomorphism
Y →X
∗
which is the identity on X.
The space Y of Problem 16:I is called a onepoint compactiﬁcation or Alexandroﬀ
compactiﬁcation of X.
16:J. Prove that the onepoint compactiﬁcation of the plane is homeomorphic to S
2
.
16:3. Prove that the onepoint compactiﬁcation of R
n
is homeomorphic to
S
n
.
16:4. Give explicit descriptions of onepoint compactiﬁcations of the follow
ing spaces:
99
§16. LOCAL COMPACTNESS AND PARACOMPACTNESS 100
(a) annulus ¦(x, y) ∈ R
2
[ 1 < x
2
+y
2
< 2¦;
(b) square without vertices ¦(x, y) ∈ R
2
[ x, y ∈ [−1, 1], [xy[ < 1¦;
(c) strip ¦(x, y) ∈ R
2
[ x ∈ [0, 1]¦;
(d) a compact space.
16:K. Prove that a locally compact Hausdorﬀ space is regular.
'16
◦
3 Proper Maps
A continuous map f : X →Y is said to be proper if the preimage of any compact
subset of Y is compact.
Let X, Y be Hausdorﬀ spaces. Any continuous map f : X → Y is naturally
extended to a map X
∗
→ Y
∗
deﬁned by the following formula:
f
∗
(x) =
f(x), if x ∈ X
Y
∗
Y, otherwise, i.e., if x = X
∗
X.
16:L. Prove that f
∗
is continuous, iﬀ f is proper.
16:M. Prove that any proper map of a Hausdorﬀ space to a Hausdorﬀ locally compact
space is closed.
Problem 16:M is related to Theorem 15.W.
16:N. Extend this analogy: formulate and prove statements corresponding to theo
rems 15.X and 15.Y.
'16
◦
4 Locally Finite Collections of Subsets
A collection Γ of subsets of a space X is said to be locally ﬁnite if each point
b ∈ X has a neighborhood U such that A∩ U = ∅ for all but ﬁnite number of A ∈ Γ.
16:O. Any locally ﬁnite cover of a compact space is ﬁnite.
16:5. If a collection Γ of subsets of a space X is locally ﬁnite then so is
¦Cl A [ A ∈ Γ¦.
16:6. If a collection Γ of subsets of a space X is locally ﬁnite and Cl A is
compact for each A ∈ Γ then each A ∈ Γ intersects only ﬁnite number of
elements of Γ.
16:7. Any locally ﬁnite cover of a sequentially compact space is ﬁnite.
16:P. Find an example of an open cover of R
n
which does not possess a locally ﬁnite
subcover.
Let Γ and ∆ be covers of a set X. Then ∆ is said to be a reﬁnement of Γ if for
each A ∈ ∆ there exists B ∈ Γ such that A ⊂ B.
16:Q. Prove that any open cover of R
n
has a locally ﬁnite open reﬁnement.
16:R. Let ¦U
i
¦
i∈N
be a locally ﬁnite open cover of R
n
. Prove that there exist an
open cover ¦V
i
¦
i∈N
such that Cl V
i
⊂ U
i
for each i ∈ N.
§16. LOCAL COMPACTNESS AND PARACOMPACTNESS 101
'16
◦
5 Paracompact Spaces
A space X is said to be paracompact if any its open cover has a locally ﬁnite open
reﬁnement.
16:S. Any compact space is paracompact.
16:T. R
n
is paracompact.
16:U. Let X = ∪
∞
i=1
X
i
and X
i
are compact sets. Then X is paracompact.
16:V. Any closed subspace of a paracompact space is paracompact.
16:8. A disjoint union of paracompact spaces is paracompact.
'16
◦
6 Paracompactness and Separation Axioms
16:9. Any Hausdorﬀ paracompact space is regular.
16:10. Any Hausdorﬀ paracompact space is normal.
16:11. Let X be a normal space and Γ its locally ﬁnite open cover. Then
there exists a locally ﬁnite open cover ∆ such that ¦Cl V [ V ∈ ∆¦ is a
reﬁnement of Γ.
Information. Any metrizable space is paracompact.
'16
◦
7 Partitions of Unity
For a function f : X → R, the set Cl¦x ∈ X [ f(x) = 0¦ is called the support of
f and denoted by suppf.
16:W. Let ¦f
α
¦
α∈Λ
be a family of continuous functions X → R such that the sets
supp(f
α
) comprise a locally ﬁnite cover of the space X. Prove that the relation
f(x) =
¸
α∈Λ
f
α
(x)
deﬁnes a continuous function f : X →R.
A family of nonnegative functions f
α
X →R
+
is called a partition of unity if the
sets supp(f
α
) comprise a locally ﬁnite cover of the space X and
¸
α∈Λ
f
α
(x) = 1.
A partition of unity ¦f
α
¦ is said to be subordinate to a cover Γ if each supp(f
α
)
is contained in an element of Γ.
16:X. For every normal space X there exists a partition of unity which is subordinate
to a given locally ﬁnite open cover of X.
16:Y. A Hausdorﬀ space is paracompact, iﬀ any its open cover admits a partition of
unity which is subordinate to this cover.
'16
◦
8 Application: Making Embeddings from Pieces
16:Z. Let h
i
U
i
→R
n
, i = 1, . . . , k, be embeddings, where U
i
comprise an open cover
of a space X. Then X can be embedded in R
k(n+1)
.
§16. LOCAL COMPACTNESS AND PARACOMPACTNESS 102
16:Z.1. Show that the map x → (f
i
(x)
ˆ
h
i
(x)), where f
i
X → R comrise
a partition of unity, which is subordinate to the given cover and
ˆ
h
i
(x) =
(h
i
(x), 1) ∈ R
n+1
, is an embedding.
'17 Sequential Compactness
'17
◦
1 Sequential Compactness Versus Compactness
A topological space is said to be sequentially compact if every sequence
of its points contains a convergent subsequence.
17.A. Any compact ﬁrst countable space is sequentially compact.
A point b is called an accumulation point of a set A if every neighborhood
of b contains inﬁnitely many points of A.
17.A.1. Prove that in a space satisfying the ﬁrst separation axiom the
notions of accumulation point and limit point coincide.
17.A.2. In a compact space any inﬁnite set has an accumulation point.
17.A.3. The space, in which any inﬁnite set has an accumulation point, is
sequentially compact.
17.B. A sequentially compact second countable space is compact.
17.B.1. In a sequentially compact space a decreasing sequence of nonempty
closed sets has a nonempty intersection.
17.B.2. Prove that in a topological space every decreasing sequence of
nonempty closed sets has nonempty intersection, iﬀ any centered countable
collection of closed sets has nonempty intersection.
17.C. For second countable spaces compactness and sequential compact
ness are equivalent.
'17
◦
2 In Metric Space
A subset A of a metric space X is called an εnet (where ε is a positive
number) if ρ(x, A) < ε for each point x ∈ X.
17.D. Prove that in any compact metric space for any ε > 0 there exists
a ﬁnite εnet.
17.E. Prove that in any sequentially compact metric space for any ε > 0
there exists a ﬁnite εnet.
17.F. Prove that a subset of a metric space is everywhere dense, iﬀ it is
an εnet for any ε > 0.
17.G. Any sequentially compact metric space is separable.
17.H. Any sequentially compact metric space is second countable.
17.I. For metric spaces compactness and sequential compactness are
equivalent.
103
§17. SEQUENTIAL COMPACTNESS 104
17.1. Prove that a sequentially compact metric space is bounded. (Cf. 17.E
and 17.I.)
17.2. Prove that in any metric space for any ε > 0 there exists
(a) a discrete εnet and even
(b) an εnet such that the distance between any two of its points is greater
than ε.
'17
◦
3 Completeness and Compactness
A sequence ¦x
n
¦
n∈N
of points of a metric space is called a Cauchy sequence if for
any ε > 0 there exists a number N such that ρ(x
n
, x
m
), ε for any n, m > N. A metric
space is said to be complete if each Cauchy sequence in it is convergent.
17:A. A Cauchy sequence, which contains a convergent subsequence, converges.
17:B. Prove that a metric space is complete, iﬀ any decreasing sequence of its closed
balls with radii tending to 0 has nonempty intersection.
17:C. Prove that a compact metric space is complete?
17:D. Is any locally compact, but not compact metric space complete?
17:E. Prove that a complete metric space is compact, iﬀ for any ε > 0 it contains a
ﬁnite εnet.
17:F. Prove that a complete metric space is compact iﬀ for any ε > 0 it contains a
compact εnet.
'17
◦
4 NonCompact Balls in Inﬁnite Dimension
By l
∞
denote the set of all bounded sequences of real numbers. This is
a vector space with respect to the componentwise operations. There is a
natural norm in it: [[x[[ = sup¦[x
n
[ : n ∈ N¦.
17.3. Are closed balls of l
∞
compact? What about spheres?
17.4. Is the set ¦x ∈ l
∞
: [x
n
[ ≤ 2
−n
, n ∈ N¦ compact?
17.5. Prove that the set ¦x ∈ l
∞
: [x
n
[ = 2
−n
, n ∈ N¦ is homeomorphic to
the Cantor set K introduced in Section '2.
17.6*. Does there exist an inﬁnitely dimensional normed space, in which
closed balls are compact?
'17
◦
5 pAdic Numbers
Fix a prime integer p. By Z
p
denote the set of series of the form a
0
+
a
1
p + +a
n
p
n
+. . . with 0 ≤ a
n
< p, a
n
∈ N. For x, y ∈ Z
p
put ρ(x, y) = 0
if x = y and ρ(x, y) = p
−m
, if m is the smallest number such that the mth
coeﬃcients in the series x and y diﬀer.
17.7. Prove that ρ is a metric in Z
p
.
This metric space is called the space of integer padic numbers. There is
an injection Z → Z
p
assigning to a
0
+a
1
p + +a
n
p
n
∈ Z with 0 ≤ a
k
< p
the series
a
0
+ a
1
p + +a
n
p
n
+ 0p
n+1
+ 0p
n+2
+ ∈ Z
p
§17. SEQUENTIAL COMPACTNESS 105
and to −(a
0
+a
1
p + +a
n
p
n
) ∈ Z with 0 ≤ a
k
< p the series
b
0
+b
1
p + +b
n
p
n
+ (p −1)p
n+1
+ (p −1)p
n+2
+. . . ,
where
b
0
+b
1
p + +b
n
p
n
= p
n+1
−(a
0
+a
1
p + +a
n
p
n
).
Cf. 4:D.
17.8. Prove that the image of the injection Z →Z
p
is dense in Z
p
.
17.9. Is Z
p
a complete space?
17.10. Is Z
p
compact?
'17
◦
6 Induction on Compactness
A function f : X → R is locally bounded if for any point a ∈ X there
exists a neighborhood U and a number M > 0 such that [f(x)[ ≤ M for
x ∈ U (i.e., each point has a neighborhood such that the restriction of f to
this neighborhood is bounded).
17.11. Prove that if a space X is compact and a function f : X → R is
locally bounded then f is bounded.
This statement is one of the simplest applications of a general princi
ple formulated below in 17.12. This principle may be called induction on
compactness (cf. induction on connectedness discussed in Section '11).
Let X be a topological space, ( a property of subsets of X. We say that
( is additive if the union of any ﬁnite family of sets having ( also has (. The
space X is said to possess ( locally if each point of X has a neighborhood
with property (.
17.12. Prove that a compact space which possesses locally an additive prop
erty has this property itself.
17.13. Deduce from this principle the statements of problems 15.Q, 17:E,
and 17:F.
'17
◦
7 Spaces of Convex Figures
Let D ⊂ R
2
be a closed disc of radius p. Consider the set of all convex
polygons P with the following properties:
• the perimeter of P is at most p;
• P is contained in D;
• P has ≤ n vertices (the cases of one and two vertices are not excluded).
See 4:E, cf. 4:G.
17.14. Equip this set with a natural topological structure. For instance,
deﬁne a natural metric.
17.15. Prove that this space is compact.
17.16. Prove that there exists a polygon belonging to this set and having
the maximal area.
17.17. Prove that this is a regular ngon.
§17. SEQUENTIAL COMPACTNESS 106
Consider now the set of all convex polygons of perimeter ≤ p contained
in D. In other words, consider the union of the sets of ≤ ngons considered
above.
17.18. Construct a topological structure in this set such that it induces the
structures introduced above in the spaces of ≤ npolygons.
17.19. Prove that the space provided by the solution of Problem 17.18 is
not compact.
Consider now the set of all convex subsets of the plane of perimeter ≤ p
contained in D.
17.20. Construct a topological structure in this set such that it induces the
structure introduced above in the spaces of polygons.
17.21. Prove that the space provided by the solution of Problem 17.20 is
compact.
17.22. Prove that there exists a convex plane set with perimeter ≤ p having
a maximal area.
17.23. Prove that this is a disc of radius
p
2π
.
17.24. Consider the set of all bounded subsets of a compact metric space.
Prove that this set endowed with the Hausdorﬀ metric (see 4:F) is a compact
space.
Problems for Tests
Test.1. Let X be a topological space. Fill Table 1 with pluses and minuses
according to your answers to the corresponding questions.
If X is: non non second
connected Hausdorﬀ Hausdorﬀ separable compact compact countable
Has Y
the same
property, if:
Y ⊂ X
Y is open
subset of X
Y is closed
subset of X
X is dense
in Y
Y is quotient
space of X
Y = X as sets,
Ω
X
⊂ Ω
Y
Y is open
subset of R
n
Y is anti
discrete
Table 1.
107
PROBLEMS FOR TESTS 108
Test.2. Let X be a topological space. Fill Table 2 with pluses and minuses
according to your answers to the corresponding questions.
If X is: non non second
connected Hausdorﬀ Hausdorﬀ separable compact compact countable
Has Y
the same
property, if:
X = Y Z
Y = X Z
Y is open
dense in X
X is open
dense in Y
X is quotient
space of Y
Y = X as sets,
Ω
X
⊃ Ω
Y
Y is closed
and bounded
subset of R
n
Y is discrete
Table 2.
Test.3. Give as many proves as you can for nonexistence of a homeomor
phism between
(a) S
1
and R
1
,
(b) I and I
2
,
(c) R and R
T1
(d) R and R
+
= ¦x ∈ R : x ≥ 0¦.
CHAPTER 3
Topological Constructions
'18 Multiplication
'18
◦
1 SetTheoretic Digression. Product of Sets
Let X and Y be sets. The set of ordered pairs (x, y) with x ∈ X and
y ∈ Y is called a direct product or Cartesian product or just product of X
and Y and denoted by XY . If A ⊂ X and B ⊂ Y then AB ⊂ XY .
Sets X ¦b¦ with b ∈ Y and ¦a¦ Y with a ∈ X are called ﬁbers of the
product X Y .
18.A. Prove that for any A
1
, A
2
⊂ X and B
1
, B
2
⊂ Y
(A
1
∪A
2
) (B
1
∪B
2
) = (A
1
B
1
) ∪(A
1
B
2
) ∪(A
2
B
1
) ∪(A
2
B
2
),
(A
1
B
1
) ∩ (A
2
B
2
) = (A
1
∩ A
2
) (B
1
∩ B
2
).
There are natural maps of X Y onto X and Y deﬁned by formulas
(x, y) → x and (x, y) → y. They are denoted by pr
X
and pr
Y
and are
called (natural) projections.
18.B. Prove that pr
−1
X
(A) = AY for A ⊂ X.
18.1. Write down the corresponding formula for B ⊂ Y .
To a map f : X →Y there corresponds a subset Γ
f
of XY deﬁned
by Γ
f
= ¦(x, f(x)) : x ∈ X¦ and called the graph of f.
18.C. A set Γ ⊂ XY is the graph of a map X →Y , iﬀ for each a ∈ X
the intersection Γ ∩ (a Y ) contains exactly one point.
18.2. Prove that for any map f : X →Y and any set A ⊂ X,
f(A) = pr
Y
(Γ
f
∩ (A Y )) = pr
Y
(Γ
f
∩ pr
−1
X
(A))
and f
−1
(B) = pr
X
(Γ ∩ (X B)) for any B ⊂ Y .
18.3. Let A and B be subsets of X and ∆ = ¦(x, y) ∈ X X : x = y¦.
Prove that (A B) ∩ ∆ = ∅, iﬀ A ∩ B = ∅
18.4. Prove that the map pr
X
Γ
f
is bijective.
18.5. Prove that f is injective, iﬀ pr
Y
Γ
f
is injective.
18.6. Let T : X Y →Y X be the map deﬁned by (x, y) →(y, x). Prove
that Γ
f
−1 = T(Γ
f
) for any invertible map f : X →Y .
109
§18. MULTIPLICATION 110
'18
◦
2 Product of Topologies
Let X and Y be topological spaces. If U is an open set of X and B is
an open set of Y , then we say that U V is an elementary set of X Y .
18.D. The set of elementary sets of X Y is a base of a topological
structure in X Y .
The product of topological spaces X and Y is the set XY with the
topological structure deﬁned by the base consisting of elementary sets.
18.7. Prove that for any subspaces A and B of spaces X and Y the topology
of the product A B coincides with the topology induced from X Y via
the natural inclusion A B ⊂ X Y .
18.E. Y X is canonically homeomorphic to X Y .
The word canonically means here that a homeomorphism between
XY and Y X which exists according to the statement can be chosen
in a nice special (or even obvious?) way, so that one may expect that it
has additional pleasant properties.
18.F. X (Y Z) is canonically homeomorphic to (X Y ) Z.
18.8. Prove that if A is closed in X and B is closed in Y then AB is closed
in X Y .
18.9. Prove that Cl(A B) = Cl A Cl B for any A ⊂ X and B ⊂ Y .
18.10. Is it true that Int(A B) = Int AInt B?
18.11. Is it true that Fr(A B) = Fr A Fr B?
18.12. Is it true that Fr(A B) = (Fr A B) ∪ (A Fr B)?
18.13. Prove that Fr(AB) = (Fr AB) ∪(AFr B) for closed A and B?
18.14. Find a formula for Fr(A B) in terms of A, Fr A, B and Fr B.
'18
◦
3 Topological Properties of Projections and Fibers
18.G. The natural projections pr
X
: X Y →X and pr
Y
: X Y →Y
are continuous.
18.H. The topology of product is the coarsest topology with respect to
which pr
X
and pr
Y
are continuous.
18.I. A ﬁber of a product is canonically homeomorphic to the correspond
ing factor. The canonical homeomorphism is the restriction to the ﬁber
of the natural projection of the product onto the factor.
18.J. Prove that R
1
R
1
= R
2
, (R
1
)
n
= R
n
, (I)
n
= I
n
(recall that
I
n
is the ndimensional cube).
§18. MULTIPLICATION 111
18.15. Let Σ
X
and Σ
Y
be bases of topological spaces X and Y . Prove that
sets U V with U ∈ Σ
X
and V ∈ Σ
Y
constitute a base for X Y .
18.16. Prove that a map f : X → Y is continuous, iﬀ pr
X
[
Γ
f
is a homeo
morphism.
18.17. Prove that if W is open in X Y then pr
X
(W) is open in X.
A map of a topological space X to a topological space Y is said to be
open, if the image of any open set under this map is open. Therefore 18.17
states that pr
X
: X Y →X is an open map.
18.18. Is pr
X
a closed map?
18.19. Prove that for each topological space X and each compact topological
space Y the map pr
X
: X Y →X is closed.
'18
◦
4 Cartesian Products of Maps
Let X, Y , and Z be sets. To a map f : Z → X Y one assigns the
compositions f
1
= pr
X
◦ f : Z → X and f
2
= pr
Y
◦ f : Z → Y . These
compositions are called factors of f. Indeed, f can be recovered from them
as a sort of product.
18.20. Prove that for any maps f
1
: Z → X and f
2
: Z → Y there exists a
unique map f : Z →X Y with pr
X
◦ f = f
1
and pr
Y
◦ f = f
2
18.21. Let X, Y , and Z be topological spaces. Prove that f is continuous,
iﬀ f
1
and f
2
are continuous.
For any maps g
1
: X
1
→Y
1
and g
2
: X
2
→Y
2
there is a map X
1
X
2
→
Y
1
Y
2
deﬁned by formula (x
1
, x
2
) → (g
1
(x
1
), g
2
(x
2
)). This map is called a
(Cartesian) product of g
1
and g
2
and denoted by g
1
g
2
.
18.22. Prove that the Cartesian product of continuous maps is continuous.
18.23. Prove that the Cartesian product of open maps is open.
18.24. Prove that a metric ρ : XX →R is continuous with respect to the
topology deﬁned by the metric.
'18
◦
5 Properties of Diagonal and Other Graphs
18.25. Prove that a topological space is Hausdorﬀ, iﬀ the set ∆ = ¦(x, x) :
x ∈ X¦ (which is called the diagonal of X X) is closed.
18.26. Prove that if Y is a Hausdorﬀ space and a map f : X → Y is
continuous then the graph Γ
f
is closed in X Y .
18.27. Let Y be a compact space and Γ
f
be closed. Prove that then f is
continuous.
18.28. Prove that in 18.27 the hypothesis on compactness is necessary.
18.29. Let f R →R be a continuous function. Prove that its graph is:
(a) closed;
(b) connected;
(c) path connected;
(d) locally connected;
§18. MULTIPLICATION 112
(e) locally compact.
18.30. Does any of properties of the graph of a function mentioned in 18.29
imply its continuity?
18.31. Let Γ
f
be closed. Then the following assertions are equivalent:
(a) f is continuous;
(b) f is locally bounded;
(c) the graph Γ
f
of f is connected.
18.32. Prove that if Γ
f
is connected and locally connected then f is contin
uous.
18.33. Prove that if Γ
f
is connected and locally compact then f is continu
ous.
18.34. Are some of assertions in problems 18.31 – 18.33 true for mappings
f : R
2
→R?
'18
◦
6 Topological Properties of Products
18.K. The product of Hausdorﬀ spaces is Hausdorﬀ.
18.35. Prove that the product of regular spaces is regular.
18.36. The product of normal spaces is not necessarily normal.
18.36.1. Prove that the set of real numbers with the topology de
ﬁned by the base which consists of all semiopen intervals [a, b) is
normal.
18.36.2. Prove that in the Cartesian square of the space introduced
in 18.36.1 the subspace ¦(x, y) : x = −y¦ is closed and discrete.
18.36.3. Find two disjoint subsets of ¦(x, y) : x = −y¦ which have
no disjoint neighborhoods in the Cartesian square of the space of
18.36.1.
18.L. The product of separable spaces is separable.
18.M. First countability of factors implies ﬁrst countability of the prod
uct.
18.N. The product of second countable spaces is second countable.
18.O. The product of metrizable spaces is metrizable.
18.P. The product of connected spaces is connected.
18.37. Prove that for connected spaces X and Y and any proper subsets
A ⊂ X, B ⊂ Y the set X Y A B is connected.
18.Q. The product of pathconnected spaces is pathconnected.
18.R. The product of compact spaces is compact.
18.38. Prove that the product of locally compact spaces is locally compact.
18.39. If X is a paracompact space and Y compact then XY is paracom
pact.
18.40. For which of the topological properties studied above, if X Y has
the property then X also has?
§18. MULTIPLICATION 113
'18
◦
7 Representation of Special Spaces as Products
18.S. Prove that R
2
¦0¦ is homeomorphic to S
1
R.
18.41. Prove that R
n
R
k
is homeomorphic to S
n−k−1
R
k+1
.
18.42. Prove that S
n
∩ ¦x ∈ R
n+1
: x
2
1
+ + x
2
k
≤ x
2
k+1
+ + x
2
n+1
¦ is
homeomorphic to S
k−1
D
n−k+1
.
18.43. Prove that O(n) is homeomorphic to SO(n) O(1).
18.44. Prove that GL(n) is homeomorphic to SL(n) GL(1).
18.45. Prove that GL
+
(n) is homeomorphic to SO(n) R
n(n+1)
2
, where
GL
+
(n) = ¦A ∈ L(n, n) : det A > 0¦.
18.46. Prove that SO(4) is homeomorphic to S
3
SO(3).
The space S
1
S
1
is called a torus.
18.T. Construct a topological embedding of the torus to R
3
.
The product S
1
S
1
of k factors is called the kdimensional torus.
18.U. Prove that the kdimensional torus can be topologically embedded
into R
k+1
.
18.V. Find topological embeddings of S
1
D
2
, S
1
S
1
I, and S
2
I
into R
3
.
'19 Quotient Spaces
'19
◦
1 SetTheoretic Digression. Partitions and Equivalence
Relations
Recall that a partition of a set is its cover consisting of pairwise disjoint
sets.
Each partition of a set X gives rise to an equivalence relation (i.e., a
relation, which is reﬂexive, symmetric and transitive): two elements of
X are said to be equivalent if they belong to the same element of the
partition. Vice versa, each equivalence relation in X gives rise to the
partition of X to classes of equivalent elements. Thus partitions of a set
into nonempty subsets and equivalence relations in the set are essentially
the same. More precisely, they are two ways of describing the same
phenomenon.
Let X be a set, and S be a partition. The set whose elements are
members of the partition S (which are subsets of X) is called the quotient
set or factor set of X by S and denoted by X/
S
.
19.1. Enigma. How is this operation related to division of numbers? Why
is there a similarity in terminology and notations?
At ﬁrst glance, the deﬁnition of quotient set contradicts one of the
very profound principles of the set theory which states that a set is deﬁned
by its elements. Indeed, according to this principle, X/
S
= S, since S
and X/
S
have the same elements. Hence, there seems to be no need to
introduce X/
S
.
The real sense of the notion of quotient set is not in its literal set
theoretic meaning, but in our way of thinking of elements of partitions.
If we remember that they are subsets of the original set and want to keep
track of their internal structure (at least, of their elements), we speak of
a partition. If we think of them as atoms, getting rid of their possible
internal structure then we speak on the quotient set.
The set X/
S
is called also the set of equivalence classes for the equiv
alence relation corresponding to the partition S.
The mapping X → X/
S
that maps x ∈ X to the element of S
containing this point is called a (canonical) projection and denoted by pr.
A subset of X which is a union of elements of a partition is said to be
saturated. The smallest saturated set containing a subset A of X is called
the saturation of A.
19.2. Prove that A ⊂ X is an element of a partition S of X, iﬀ A =
pr
−1
(point) where pr : X →X/
S
is the natural projection.
19.A. Prove that the saturation of a set A equals pr
−1
pr(A)
.
114
§19. QUOTIENT SPACES 115
19.B. Prove that a set is saturated, iﬀ it is equal to its saturation.
'19
◦
2 Quotient Topology
A quotient set X/
S
of a topological space X with respect to a parti
tion S into nonempty subsets is provided with a natural topology: a set
U ⊂ X/
S
is said to be open in X/
S
if its preimage pr
−1
(U) under the
canonical projection pr : X →X/
S
is open.
19.C. The collection of these sets is a topological structure in the quo
tient set X/
S
.
This topological structure is called the quotient topology. The set X/
S
with this topology is called the quotient space of the space X by parti
tion S.
19.3. Give an explicit description of the quotient space of the segment [0, 1]
by the partition consisting of [0,
1
3
], (
1
3
,
2
3
], (
2
3
, 1].
19.4. What can you say about a partition S of a topological space X if the
quotient space X/
S
is known to be discrete?
19.D. A subset of a quotient space X/
S
is open, iﬀ it is the image of an
open saturated set under the canonical projection pr.
19.E. A subset of a quotient space X/
S
is closed, iﬀ its preimage under
pr is closed in X, iﬀ it is the image of a closed saturated set.
19.F. The canonical projection pr : X →X/
S
is continuous.
19.G. Prove that the quotient topology is the ﬁnest topology in X/
S
such that the canonical projection pr is continuous with respect to it.
'19
◦
3 Topological Properties of Quotient Spaces
19.H. A quotient space of a connected space is connected.
19.I. A quotient space of a pathconnected space is pathconnected.
19.J. A quotient space of a separable space is separable.
19.K. A quotient space of a compact space is compact.
19.L. The quotient space of the real line by partition R
+
, R R
+
is
not Hausdorﬀ.
19.M. The quotient space of a topological space X by a partition S is
Hausdorﬀ, iﬀ any two elements of S possess disjoint saturated neighbor
hoods.
19.5. Formulate similar necessary and suﬃcient conditions for a quotient
space to satisfy other separation axioms and countability axioms.
19.6. Give an example showing that second countability may get lost when
we go over to a quotient space.
§19. QUOTIENT SPACES 116
'19
◦
4 SetTheoretic Digression. Quotients and Maps
Let S be a partition of a set X into nonempty subsets. Let f : X →Y
be a map which is constant on each element of S. Then there is a map
X/
S
→Y which assigns to each element A of S the element f(A). This
map is denoted by f/
S
and called the quotient map or factor map of f
(by partition S).
19.N. Prove that a map f : X → Y is constant on each element of
a partition S of X iﬀ there exists a map g : X/
S
→ Y such that the
following diagram is commutative:
X
f
−−−→ Y
pr
g
X/
S
Prove that such a map g coincides with f/
S
.
More generally, if S and T are partitions of sets X and Y then every
map f : X → Y , which maps each element of S into an element of
T, gives rise to a map X/
S
→ Y/
T
which assigns to an element A of
partition S the element of partition T containing f(A). This map is
denoted by f/
S, T
and called the quotient map or factor map of f (with
respect to S and T).
19.O. Formulate and prove for f/
S, T
a statement which generalizes
19.N.
A map f : X → Y deﬁnes a partition of the set X into nonempty
preimages of the elements of Y . This partition is denoted by S(f).
19.P. The map f/
S(f)
: X/
S(f)
→Y is injective.
This map is called injective factor (or injective quotient) of the map f.
'19
◦
5 Continuity of Quotient Maps
19.Q. Let X, Y be topological spaces, S be a partition of X into nonempty
sets, and f : X → Y be a continuous map, which is constant on each
element of S. Then the factor f/
S
of f is continuous.
19.7. Let X, Y be topological spaces, S be a partition of X into nonempty
sets. Prove that the formula f → f/
S
deﬁnes a bijection of the set of all
continuous maps X →Y , which are constant on each element of the partition
S, onto the set of all continuous maps X/
S
→Y .
19.R. Let X, Y be topological spaces, S and T partitions of X and Y ,
and f : X → Y a continuous map, which maps each element of S into
an element of T. Then the map f/
S, T
: X/
S
→Y/
T
is continuous.
§19. QUOTIENT SPACES 117
'19
◦
6 Closed Partitions
A partition S of a topological space X is called closed, if the saturation of each
closed set is closed.
19:1. Prove that a partition is closed, iﬀ the canonical projection X →X/
S
is a closed map.
19:2. Prove that a partition, which contains only one element consisting of
more than one point, is closed if this element is a closed set.
19:A. The quotient space of a topological space satisfying the ﬁrst separation axiom
with respect to a closed partition satisﬁes the ﬁrst separation axiom.
19:B. The quotient space of a normal topological space with respect to a closed par
tition is normal.
'19
◦
7 Open Partitions
A partition S of a topological space X is called open, if the saturation of each
open set is open.
19:3. Prove that a partition is open, iﬀ the canonical projection X → X/
S
is an open map.
19:4. Prove that if a set A is saturated with respect to an open partition,
then Int A and Cl A are also saturated.
19:C. The quotient space of a second countable space with respect to an open partition
is second countable.
19:D. The quotient space of a ﬁrst countable space with respect to an open partition
is ﬁrst countable.
19:E. Let S be an open partition of a topological space X and T be an open partition
of a topological space Y . Denote by S T the partition of X Y consisting of AB
with A ∈ S and B ∈ T. Then the injective factor X Y/
S T
→ X/
S
Y/
T
of
pr pr X Y →X/
S
Y/
T
is a homeomorphism.
'19
◦
8 SetTheoretic Digression:
Splitting a transitive relation
into equivalence and partial order
In the deﬁnitions of equivalence and partial order relations the condition of tran
sitivity seems to be the most important. Below we supply a formal justiﬁcation of this
feeling by showing that the other conditions are natural companions of transitivity,
although are not its consequences.
19:F. Let ≺ be a transitive relation in a set X. Then the relation deﬁned as follows
a b, if a ≺ b or a = b,
is also transitive (and furthermore it is certainly reﬂexive, i. e. a a for each a ∈ X).
A binary relation in a set X is called a preorder if it is transitive and reﬂective,
that is it satisﬁes the following conditions:
• Transitivity. If a b and b c, then a c.
• Reﬂexivity. a a for any a.
A set X equipped with a preorder is called preordered.
If a preorder is antisymmetric then this is a nonstrict order.
§19. QUOTIENT SPACES 118
19:5. Is the relation a[b a preorder in the set Z of integers?
19:G. If (X, ) is a preordered set then the relation ∼ deﬁned by
a ∼ b, if a b and b a
is an equivalence relation (i. e., it is symmetric, reﬂexive and transitive) in X.
19:6. What equivalence relation is deﬁned in Z by the preorder a[b?
19:H. Let (X, ) be a preordered set and ∼ be an equivalence relation deﬁned in X
by according to 19:G. Then a
∼ a, a b and b ∼ b
imply a
b
and in this
way deﬁnes a relation in the set of equivalence classes X/
∼
. This relation is a
nonstrict partial order.
Thus any transitive relation generates an equivalence relation and a partial order
in the set of equivalence classes.
19:I. How this chain of constructions would degenerate, if the original relation was
(a) an equivalence relation, or
(b) nonstric partial order.
19:J. In any topological space the relation deﬁned by
a b, if a ∈ Cl¦b¦,
is a preorder.
19:7. In the set of all subsets of an arbitrary topological space the relation
A B, if A ⊂ Cl¦B¦,
is a preorder. This preorder deﬁnes the following equivalence relation: sets
are equivalent iﬀ they have the same closure.
19:K. The equivalence relation deﬁned by the preorder of Theorem 19:J deﬁnes the
partition of the space to maximal (with respect to inclusion) indiscrete subspaces. The
quotient space satisﬁes the Kolmogorov separation axiom T
0
.
The quotient space of Theorem 19:K is called the maximal T
0
quotient of X.
19:L. A continuous image of an indiscrete topological space is indiscrete.
19:M. Prove that any continuous map X → Y induces a continuous map of the
maximal T
0
quotient of X to the maximal T
0
quotient of Y .
'19
◦
9 Finite Topological Spaces
The results of the preceding subsection provide a key to understanding of struc
ture of ﬁnite topological spaces. Let X be a ﬁnite topological space. By Theorem
19:K, it is partitioned to indiscrete clusters of points. By 19:L, continuous maps
between ﬁnite spaces respect these clusters, and by 19:M induce continuous maps
between the maximal T
0
quotient spaces.
This means that we can consider a ﬁnite topological space as its maximal T
0

qoutient whose points are equipped with multiplicities, which are natural numbers,
the numbers of points in the corresponding clusters of the original space.
The maximal T
0
quotient of a ﬁnite space is a smallest neighborhood space (as
a ﬁnite space). By Theorem 13.N, its topology is deﬁned by a partial order. By
Theorem 9.P, homeomorphisms between spaces with poset topologies are monotone
bijections.
§19. QUOTIENT SPACES 119
Thus a ﬁnite topological space is characterised up to a homeomorphism by a
ﬁnite poset whose elements are equipped with multiplicities (natural numbers). Two
such spaces are homeomorphic, iﬀ there exists a monotone bijection between the
corresponding posets which preserve the multiplicities. To recover the topological
space from the poset with the multiplicities, one has to equip the poset with the
poset topology and then replace each of its elements with an indiscrete cluster of
points, the number points in which is the multiplicity of the element.
'19
◦
10 Simplicial schemes
Let V be a set and Σ be a set of some of its subsets. A pair (V, Σ) is called a
simplicial scheme with set of vertices V and set of simplices Σ, if
• each subset of any element of Σ belongs to Σ,
• the intersection of any collection of elements of Σ belongs to Σ,
• each oneelement subset of V belongs to Σ.
The set Σ is partially ordered by inclusion. When equipped the poset topology of this
partial order, it is called the space of simplices of the simplicial scheme (X, Σ).
Each simplicial scheme gives rise also to another topological space. Namely, for a
simplicial scheme (V, Σ) consider the set S(V, Σ) of all functions c : V → I (= [0, 1])
such that
Supp(c) = ¦v ∈ V : c(v) = 0¦ ∈ Σ
and
¸
v∈V
c(v) = 1. Equip S(V, Σ) with the topology generated by metric
ρ(c
1
, c
2
) = sup
v∈V
[c
1
(v) −c
2
(v)[.
Space S(V, Σ) is called simplicial or triangulated space. It is covered by sets
¦c ∈ S [ Supp(c) = σ¦, where σ ∈ Σ, which are called its (open) simplices.
19:8. Which open simplices of a simplicial space are open sets, which are
closed, and which are neither closed nor open?
19:N. Find for each σ ∈ Σ a homeomorphism of the space
¦c ∈ S [ Supp(c) = σ¦ ⊂ S(V, Σ)
onto an open simplex of dimension equal to the number of vertices belonging to σ
minus one (recall that ndimensional open simplex is the set ¦(x
1
, . . . , x
n+1
) ∈ R
n+1
[
x
j
> 0 for j = 1, . . . , n + 1 and
¸
n+1
i=1
x
i
= 1, ¦).
19:O. Prove that for any simplicial scheme (V, Σ) the quotient space of the simplicial
space S(V, Σ) by its partition to open simplices is homeomorphic to the space Σ of
simplices of the simplicial scheme (V, Σ).
'19
◦
11 Baricentric Subdivision of a Poset
19:P. Find a poset, which is not isomorphic to the set of simplices (ordered by
inclusion) of whatever simplicial scheme.
Let (X, ≺) be a poset. Consider the set of all nonempty ﬁnite stricly increasing
sequences a
1
≺ a
2
≺ ≺ a
n
of elements of X. It can be described also as the set of
all nonempty ﬁnite subsets of X in each of which ≺ deﬁnes a linear order. Denote
this set by X
. It is naturally ordered by inclusion.
Poset (X
, ⊂) is called the baricentric subdivision of (X, ≺).
§19. QUOTIENT SPACES 120
19:Q. For any poset (X, ≺), pair (X, X
) is a simplicial scheme.
There is a natural map X
→ X which maps an element of X
(i. e., nonempty
ﬁnite linearly ordered subset of X) to its greatest element.
19:R. Is this map monotone? Strictly monotone? The same questions concerning a
similar map which maps a nonempty ﬁnite linearly ordered subset of X to its smallest
element.
Let (V, Σ) be a simplicial scheme and Σ
be the baricentric subdivision of Σ
(ordered by inclusion). Simplicial scheme (Σ, Σ
) is called the baricentric subdivision
of the simplicial scheme (V, Σ)
There is a natural mapping Σ → S(V, Σ) which maps a simplex σ ∈ Σ (i. e., a
subset ¦v
0
, v
1
, . . . , v
n
¦ of V ) to function b
σ
: V →R with b
σ
(v
i
) =
1
n+1
and b
σ
(v) = 0
for any v ∈ σ.
Deﬁne a map β : S(Σ, Σ
) → S(V, Σ), which maps a function ϕ : Σ → R to
function
V →R : v →
¸
σ∈Σ
ϕ(σ)b
σ
(v).
19:S. Prove that the map β : S(Σ, Σ
) → S(V, Σ) is a homeomorphism and consti
tutes, together with projections S(V, Σ) →Σ and S(Σ, Σ
) →Σ
and the natural map
Σ
→Σ a commutative diagram
S(Σ, Σ
)
β
−−−−→ S(V, Σ)
Σ
−−−−→ Σ
'20 Zoo of Quotient Spaces
'20
◦
1 Tool for Identifying a Quotient Space with a Known
Space
20.A. If f : X → Y is a continuous map of a compact space X onto a
Hausdorﬀ space Y then the injective factor f/
S(f)
: X/
S(f)
→ Y is a
homeomorphism.
20.B. The injective factor of a continuous map of a compact space to a
Hausdorﬀ one is a topological embedding.
20.1. Describe explicitly partitions of a segment such that the corresponding
quotient spaces are all the connected letters of the alphabet.
20.2. Prove that there exists a partition of a segment I with the quotient
space homeomorphic to square I I.
'20
◦
2 Tools for Describing Partitions
Usually an accurate literal description of a partition is cumbersome,
but can be shortened and made more understandable. Of course, this re
quires a more ﬂexible vocabulary with lots of words with almost the same
meanings. For instance, the words factorize and pass to a quotient can be
replaced by attach, glue, identify, contract, and other words accompanying
these ones in everyday life.
Some elements of this language are easy to formalize. For instance,
factorization of a space X with respect to a partition consisting of a set
A and onepoint subsets of the complement of A is called a contraction
(of the subset A to a point), and the result is denoted by X/
A
.
20.3. Let A, B ⊂ X form a fundamental cover of a topological space X.
Prove that the quotient map A/
A∩ B
→ X/
B
of the inclusion A → X is a
homeomorphism.
If A and B are disjoint subspaces of a space X, and f : A → B is
a homeomorphism then passing to the quotient of the space X by the
partition into onepoint subsets of the set X (A ∪ B) and twopoint
sets ¦x, f(x)¦, where x ∈ A, is called gluing or identifying (of sets A and
B by homeomorphism f).
Rather convenient and ﬂexible way for describing partitions is to de
scribe the corresponding equivalence relations. The main advantage of
this approach is that, due to transitivity, it suﬃces to specify only some
pairs of equivalent elements: if one states that x ∼ y and y ∼ z then it
is not needed to state x ∼ z, since this follows.
Hence, a partition is represented by a list of statements of the form
x ∼ y, which are suﬃcient to recover the equivalence relation. By such
121
§20. ZOO OF QUOTIENT SPACES 122
a list enclosed into square brackets, we denote the corresponding par
tition. For example, the quotient of a space X obtained by identify
ing subsets A and B by a homeomorphism f : A → B is denoted by
X/
[a ∼ f(a) for any a ∈ A]
or just X/
[a ∼ f(a)]
.
Some partitions are easy to describe by a picture, especially if the
original space can be embedded into plane. In such a case, as in the
pictures below, one draws arrows on segments to be identiﬁed to show
directions which are to be identiﬁed.
Below we introduce all these kinds of descriptions for partitions and
give examples of their usage, providing simultaneously literal descrip
tions. The latter are not nice, but they may help to keep the reader
conﬁdent about the meaning of the new words and, on the other hand,
appreciating the improvement the new words bring in.
'20
◦
3 Entrance to the Zoo
20.C. Prove that I/
[0 ∼ 1]
is homeomorphic to S
1
.
In other words, the quotient space of segment I by the partition
consisting of ¦0, 1¦ and ¦a¦ with a ∈ (0, 1) is homeomorphic to a circle.
20.C.1. Find a surjective continuous map I → S
1
such that the corre
sponding partition into preimages of points consists of onepoint subsets of
the interior of the segment and the pair of boundary points of the segment.
20.D. Prove that D
n
/
S
n−1
is homeomorphic to S
n
.
In 20.D we deal with the quotient space of ball D
n
by the partition
into S
n−1
and onepoint subsets of its interior.
Reformulation of 20.D: Contracting the boundary of an ndimensional
ball to a point gives rise to an ndimensional sphere.
20.D.1. Find a continuous map of ball D
n
to the sphere S
n
that maps the
boundary of the ball to a single point, and maps the interior of the ball
bijectively onto the complement of this point.
20.E. Prove that I
2
/
[(0, t) ∼ (1, t) for t ∈I]
is homeomorphic to S
1
I.
Here the partition consisits of pairs of points ¦(0, t), (1, t)¦ where
t ∈ I, and onepoint subsets of (0, 1) I.
Reformulation of 20.E: If we glue the side edges of a square identifying
points on the same hight, we get a cylinder.
§20. ZOO OF QUOTIENT SPACES 123
20.F. Let X and Y be topological spaces, S a partition of X. Denote by
T the partition of X Y into sets A y with A ∈ S, y ∈ Y . Then the
natural bijection X/
S
Y →X Y/
T
is a homeomorphism.
20.G. Enigma. How are the problems 20.C, 20.E and 20.F related?
20.H. S
1
I/
[(z, 0) ∼ (z, 1) for z ∈ S
1
]
is homeomorphic to S
1
S
1
.
Here the partition consists of onepoint subsets of S
1
(0, 1), and
pairs of points of the basis circles lying on the same generatrix of the
cylinder.
Reformulation of 20.H: If we glue the basis circles of a cylinder iden
tifying points on the same generatrix, then we get a torus.
20.I. I
2
/
[(0, t) ∼ (1, t), (t, 0) ∼ (t, 1)]
is homeomorphic to S
1
S
1
.
In 20.I the partition consists of
• onepoint subsets of the interior (0, 1) (0, 1) of the square,
• pairs of points on the vertical sides, which are the same distance
from the bottom side (i.e., pairs ¦(0, t), (1, t)¦ with t ∈ (0, 1)),
• pairs of points on the horizontal sides which lie on the same vertical
line (i.e., pairs ¦(t, 0), (t, 1)¦ with t ∈ (0, 1)),
• the four vertices of the square
Reformulation of 20.I: Identifying the sides of a square according to
the picture , we get a torus .
'20
◦
4 Transitivity of Factorization
A solution of Problem 20.I can be based on Problems 20.E and 20.H
and the following general theorem.
20.J Transitivity of Factorization. Let S be a partition of a space
X, and let S
be a partition of the space X/
S
. Then the quotient space
(X/
S
)/
S
is canonically homeomorphic to X/
T
, where T is the parti
tion of the space X into preimages of elements of the partition S
under
projection X →X/
S
.
§20. ZOO OF QUOTIENT SPACES 124
'20
◦
5 M¨obius Strip
M¨obius strip or M¨obius band is I
2
/
[(0, t) ∼ (1, 1 −t)]
. In other words,
this is the quotient space of square I
2
by the partition into pairs of
points symmetric with respect to the center of the square and lying on
the vertical edges and onepoint set which do not lie on the vertical
edges. Figuratively speaking, the M¨obius strip is obtained by identifying
the vertical sides of a square in such a way that the directions shown on
them by arrows are superimposed .
20.K. Prove that the M¨obius strip is homeomorphic to the surface swept
in R
3
by an interval, which rotates in a halfplane around the middle point
while the halfplane rotates around its boundary line. The ratio of the
angular velocities of these rotations is such that rotation of the halfplane
by 360
◦
takes the same time as rotation of the interval by 180
◦
. See
Figure 1.
Figure 1.
'20
◦
6 Contracting Subsets
20.4. Prove that [0, 1]/
[
1
3
,
2
3
]
is homeomorphic to [0, 1], and [0, 1]/
¦
1
3
, 1¦
is
homeomorphic to letter P.
20.5. Prove that the following spaces are homeomorphic: (a) R
2
; (b) R
2
/
I
; (c)
(e) R
2
/
A
, where A is a union of several segments with a common end
point;
(f) R
2
/
B
, where B is a simple ﬁnite polygonal line, i.e., a union of a
ﬁnite sequence of segments I
1
, . . . , I
n
such that the initial point of
I
i+1
coincides with the ﬁnal point of I
i
.
20.6. Prove that if f : X → Y is a homeomorphism then the quotient spaces
X/
A
and Y/
f(A)
are homeomorphic.
20.7. Prove that R
2
/
[0, +∞)
is homeomorphic to Int D
2
∪ ¦(0, 1)¦.
§20. ZOO OF QUOTIENT SPACES 125
'20
◦
7 Further Examples
20.8. Prove that S
1
/
[z ∼ e
2πi/3
z]
is homeomorphic to S
1
.
In 20.8 the partition consists of triples of points which are vertices of
equilateral inscribed triangles.
20.9. Prove that the following quotient spaces of disk D
2
are homeomorphic
to D
2
:
(a) D
2
/
[(x, y) ∼ (−x, −y)]
,
(b) D
2
/
[(x, y) ∼ (x, −y)]
,
(c) D
2
/
[(x, y) ∼ (−y, x)]
.
20.10. Find a generalization of 20.9 with D
n
substituted for D
2
.
20.11. Describe explicitly the quotient space of line R
1
by equivalence rela
tion x ∼ y ⇔x −y ∈ Z.
20.12. Represent the M¨obius strip as a quotient space of cylinder S
1
I.
'20
◦
8 Klein Bottle
Klein bottle is I
2
/
[(t, 0) ∼ (t, 1), (0, t) ∼ (1, 1 −t)]
. In other words,
this is the quotient space of square I
2
by the partition into
• onepoint subsets of its interior,
• pairs of points (t, 0), (t, 1) on horizontal edges which lie on the same
vertical line,
• pairs of points (0, t), (1, 1 −t) symmetric with respect to the center
of the square which lie on the vertical edges, and
• the quadruple of vertices.
20.13. Present the Klein bottle as a quotient space of
(a) a cylinder;
(b) the M¨obius strip.
20.14. Prove that S
1
S
1
/
[(z, w) ∼ (−z, ¯ w)]
is homeomorphic to the Klein
bottle. (Here ¯ w denotes the complex number conjugate to w.)
20.15. Embed the Klein bottle into R
4
(cf. 20.K and 18.T).
20.16. Embed the Klein bottle into R
4
so that the image of this embedding
under the orthogonal projection R
4
→ R
3
would look as follows:
§20. ZOO OF QUOTIENT SPACES 126
'20
◦
9 Projective Plane
Let us identify each boundary point of the disk D
2
with the antipodal
point, i.e., factorize the disk by the partition consisting of onepoint
subsets of the interior of the disk and pairs of points on the boundary
circle symmetric with respect to the center of the disk. The result is
called the projective plane. This space cannot be embedded into R
3
, too.
Thus we are not able to draw it. Instead, we present it in other way.
20.L. A projective plane is the result of gluing of a disk and the M¨obius
strip by homeomorphism between boundary circle of the disk and bound
ary circle of the M¨obius strip.
'20
◦
10 You May Have Been Provoked to Perform an Illegal
Operation
Solving the previous problem you did something which does not ﬁt
into the theory presented above. Indeed, the operation with two spaces
called gluing in 20.L has not appeared yet. It is a combination of two
operations: ﬁrst we must make a single space consisting of disjoint copies
of the original spaces, and then we factorize this space identifying points
of one copy with points of another. Let us consider the ﬁrst operation in
details.
'20
◦
11 SetTheoretic Digression. Sums of Sets
A sum of a family of sets ¦X
α
¦
α∈A
is the set of pairs (x
α
, α) such that
x
α
∈ X
α
. The sum is denoted by
¸
α∈A
X
α
. The map of X
β
(β ∈ A)
to
¸
α∈A
X
α
deﬁned by formula x → (x, β) is an injection and denoted
by in
β
. If only sets X and Y are involved and they are distinct, we can
avoid indices and deﬁne the sum by setting
X HY = ¦(x, X) [ x ∈ X¦ ∪ ¦(y, Y ) [ y ∈ Y ¦.
'20
◦
12 Sums of Spaces
20.M. If ¦X
α
¦
α∈A
is a collection of topological spaces then the collection
of subsets of
¸
α∈A
X
α
whose preimages under all inclusions in
α
(α ∈ A)
are open, is a topological structure.
The sum
¸
α∈A
X
α
with this topology is called the (disjoint) sum of
topological spaces X
α
, (α ∈ A).
20.N. Topology described in 20.M is the ﬁnest topology with respect to
which all inclusions in
α
are continuous.
20.17. The maps in
β
: X
β
→
¸
α∈A
X
α
are topological embedding, and their
images are both open and closed in
¸
α∈A
X
α
.
§20. ZOO OF QUOTIENT SPACES 127
20.18. Which topological properties are inherited from summands X
α
by
the sum
¸
α∈A
X
α
? Which are not?
'20
◦
13 Attaching Space
Let X, Y be topological spaces, A a subset of Y , and f : A → X
a continuous map. The quotient space (X HY )/
[a ∼ f(a) for a ∈ A]
is
denoted by X∪
f
Y , and is said to be the result of attaching or gluing the
space Y to the space X by f. The latter is called the attaching map.
Here the partition of XHY consists of onepoint subsets of in
2
(Y A)
and in
1
(X f(A)), and sets in
1
(x) ∪ in
2
f
−1
(x)
with x ∈ f(A).
20.19. Prove that the composition of inclusion X → X H Y and projection
X HY →X ∪
f
Y is a topological embedding.
20.20. Prove that if X is a point then X ∪
f
Y is Y/
A
.
20.O. Prove that attaching a ball D
n
to its copy by the identity map of
the boundary sphere S
n−1
gives rise to a space homeomorphic to S
n
.
20.21. Prove that the Klein bottle can be obtained as a result of gluing two
copies of the M¨obius strip by the identity map of the boundary circle.
20.22. Prove that the result of gluing two copies of a cylinder by the identity
map of the boundary circles (of one copy to the boundary circles of the other)
is homeomorphic to S
1
S
1
.
20.23. Prove that the result of gluing two copies of solid torus S
1
D
2
by
the identity map of the boundary torus S
1
S
1
is homeomorphic to S
1
S
2
.
20.24. Obtain the Klein bottle by gluing two copies of the cylinder S
1
I
to each other.
20.25. Prove that the result of gluing two copies of solid torus S
1
D
2
by
the map
S
1
S
1
→S
1
S
1
: (x, y) →(y, x)
of the boundary torus to its copy is homeomorphic to S
3
.
20.P. Let X, Y be topological spaces, A a subset of Y , and f, g : A →X
continuous maps. Prove that if there exists a homeomorphism h : X →X
such that h ◦ f = g then X ∪
f
Y and X ∪
g
Y are homeomorphic.
20.Q. Prove that D
n
∪
h
D
n
is homeomorphic to S
n
for any homeomor
phism h : S
n−1
→S
n−1
.
20.26. Classify up to homeomorphism topological spaces, which can be ob
tained from a square by identifying a pair of opposite sides by a homeomor
phism.
20.27. Classify up to homeomorphism the spaces which can be obtained
from two copies of S
1
I by identifying of the copies of S
1
¦0, 1¦ by a
homeomorphism.
§20. ZOO OF QUOTIENT SPACES 128
20.28. Prove that the topological type of the space resulting in gluing two
copies of the M¨obius strip by a homeomorphism of the boundary circle does
not depend on the homeomorphism.
20.29. Classify up to homeomorphism topological spaces, which can be ob
tained from S
1
I by identifying S
1
0 with S
1
1 by a homeomorphism.
'20
◦
14 Basic Surfaces
A torus S
1
S
1
with the interior of an embedded disk deleted is
called a handle. A twodimensional sphere with the interior of n disjoint
embedded disks deleted is called a sphere with n holes.
20.R. A sphere with a hole is homeomorphic to disk D
2
.
20.S. A sphere with two holes is homeomorphic to cylinder S
1
I.
A sphere with three holes has a special name. It is called pantaloons.
The result of attaching p copies of a handle to a sphere with p holes
by embeddings of the boundary circles of handles onto the boundary
circles of the holes (the boundaries of the holes) is called a sphere with
p handles, or, more ceremonial (and less understandable, for a while),
orientable connected closed surface of genus p.
20.30. Prove that a sphere with p handles is welldeﬁned up to homeomor
phism (i.e., the topological type of the result of gluing does not depend on
the attaching embeddings).
20.T. A sphere with one handle is homeomorphic to torus S
1
S
1
.
20.U. A sphere with two handles is homeomorphic to the result of gluing
two copies of a handle by the identity map of the boundary circle.
A sphere with two handles is called a pretzel. Sometimes this word
denotes also a sphere with more handles.
The space obtained from a sphere with q holes by attaching q copies
of the M¨obius strip by embeddings of the boundary circles of the M¨obius
strips onto the boundary circles of the holes (the boundaries of the holes)
is called a sphere with q crosscaps, or nonorientable connected closed sur
face of genus q.
20.31. Prove that a sphere with q crosscaps is welldeﬁned up to homeomor
phism (i.e., the topological type of the result of gluing does not depend on
the attaching embeddings).
20.V. A sphere with one crosscap is homeomorphic to the projective
plane.
20.W. A sphere with two crosscaps is homeomorphic to the Klein bottle.
A sphere, spheres with handles, and spheres with crosscaps are called
basic surfaces.
§20. ZOO OF QUOTIENT SPACES 129
20.X. Prove that a sphere with p handles and q crosscaps is homeomor
phic to a sphere with 2p +q crosscaps (here q > 0).
20.32. Classify up to homeomorphisms topological spaces, which can be ob
tained by attaching to a sphere with 2p holes p copies of S
1
I by embeddings
of the boundary circles of the cylinders onto the boundary circles of the sphere
with holes.
'21 Projective Spaces
This section can be considered as a continuation of the previous one.
The quotient spaces described here are of too great importance to con
sider them just as examples of quotient spaces.
'21
◦
1 Real Projective Space of Dimension n
This space is deﬁned as the quotient space of the sphere S
n
by the
partition into pairs of antipodal points, and denoted by RP
n
.
21.A. The space RP
n
is homeomorphic to the quotient space of the
ball D
n
by the partition into onepoint subsets of the interior of D
n
,
and pairs of antipodal point of the boundary sphere S
n−1
.
21.B. RP
0
is a point.
21.C. The space RP
1
is homeomorphic to the circle S
1
.
21.D. The space RP
2
is homeomorphic to the projective plane deﬁned
in the previous section.
21.E. The space RP
n
is canonically homeomorphic to the quotient space
of R
n+1
¦0¦ by the partition into onedimensional vector subspaces of
R
n+1
punctured at 0.
A point of the space R
n+1
¦0¦ is a sequence of real numbers which are
not all zeros. These numbers are called homogeneous coordinates of the
corresponding point of RP
n
. The point with homogeneous coordinates
x
0
, x
1
, . . . , x
n
is denoted by (x
0
: x
1
: : x
n
). Homogeneous coordinates
deﬁne a point of RP
n
, but are not deﬁned by this point: proportional
vectors of coordinates (x
0
, x
1
, . . . , x
n
) and (λx
0
, λx
1
, . . . , λx
n
) deﬁne the
same point of RP
n
.
21.F. The space RP
n
is canonically homeomorphic to the metric space,
whose points are lines of R
n+1
passing through the origin 0 = (0, . . . , 0)
and the metric is deﬁned as the angle between lines (which takes values
in [0,
π
2
]). Prove that this is really a metric.
21.G. Prove that the map
i : R
n
→RP
n
: (x
1
, . . . , x
n
) →(1 : x
1
: : x
n
)
is a topological embedding. What is its image? What is the inverse map
of its image onto R
n
?
21.H. Construct a topological embedding RP
n−1
→ RP
n
such that its
image is RP
n
i(R
n
), where i is the embedding from Problem 21.G.
Therefore the projective space RP
n
can be considered as the result
of extending of R
n
by adjoining “nonproper” or “inﬁnite” points, which
constitute a projective space RP
n−1
.
130
§21. PROJECTIVE SPACES 131
21.1. Introduce a natural topological structure in the set of all lines on
the plane and prove that the resulting space is homeomorphic to a) RP
2
¦ point ¦; b) open M¨obius strip (i. e. M¨obius strip with the boundary circle
removed).
21.2. Prove that the set of all rotations of the space R
3
around lines passing
through the origin equipped with the natural topology is homeomorphic to
RP
3
.
'21
◦
2 Complex Projective Space of Dimension n
This space is deﬁned as the quotient space of unit sphere S
2n+1
of the space C
n+1
by the partition into circles which cut by (complex) lines of C
n+1
passing through the
point 0. It is denoted by CP
n
.
21:A. CP
n
is homeomorphic to the quotient space of the unit ball D
2n
of the space C
n
by the partition whose elements are onepoint subsets of the interior of D
2n
and circles
cut on the boundary sphere S
2n−1
by (complex) lines of the space C
n
passing through
the origin 0 ∈ C
n
.
21:B. CP
0
is a point.
The space CP
1
is called a complex projective line.
21:C. The complex projective line CP
1
is homeomorphic to S
2
.
21:D. The space CP
n
is canonically homeomorphic to the quotient space of the space
C
n+1
¦0¦ by the partition into complex lines of C
n+1
punctured at 0.
Hence, CP
n
can be viewed as the space of complexproportional nonzero complex
sequences (x
0
, x
1
, . . . , x
n
). Notation (x
0
: x
1
: : x
n
) and term homogeneous
coordinates introduced for the real case are used in the same way for the complex
case.
21:E. The space CP
n
is canonically homeomorphic to the metric space, whose points
are the (complex) lines of the space C
n+1
passing through the origin 0 and the metric
is deﬁned to be the angle between lines (which takes values in [0,
π
2
]).
'21
◦
3 Quaternionic Projective Spaces
Recall that in R
4
there is a remarkable multiplication, which was discovered by
R.W.Hamilton in 1843. It can be deﬁned by the formula
(x
1
, x
1
, x
3
, x
4
) (y
1
, y
2
, y
3
, y
4
) =
(x
1
y
1
−x
2
y
2
−x
3
y
3
−x
4
y
4
, x
1
y
2
+x
2
y
1
+x
3
y
4
−x
4
y
3
,
x
1
y
3
−x
2
y
4
+x
3
y
1
+x
4
y
2
, x
1
y
4
+x
2
y
3
−x
3
y
2
+x
4
y
1
).
It is bilinear and to describe it in a shorter way it suﬃces to specify the products of the
basis vectors. The latter are traditionally denoted in this case, following Hamilton,
as follows:
1 = (1, 0, 0, 0), i = (0, 1, 0, 0), j = (0, 0, 1, 0) k = (0, 0, 0, 1).
In this notation, 1 is really a unity: (1, 0, 0, 0) x = x for any x ∈ R
4
. The rest of
multiplication table looks as follows:
ij = k, jk = i, ki = j, ji = −k, kj = −i ik = −j.
§21. PROJECTIVE SPACES 132
Together with coordinatewise addition, this multiplication deﬁnes a structure of al
gebra in R
4
. Its elements are called quaternions.
21:F. Check that the quaternion multiplication is associative.
It is not commutative (e. g., ij = k = −k = ji). Otherwise, quaternions are very
similar to complex numbers. As in C, there is a transformation called conjugation
acting in the set of quaternions. It is denoted, as the conjugation of complex numbers,
by a bar: x → x. It is deﬁned by the formula (x
1
, x
2
, x
3
, x
4
) → (x
1
, −x
2
, −x
3
, −x
4
)
and has the following two remarkable properties:
21:G. ab = ba for any quaternions a and b.
21:H. aa = [a[
2
, i. e. the product of any quaternion a by the conjugate quaternion
a equals ([a[
2
, 0, 0, 0).
The latter property allows one to deﬁne, for any a ∈ R
4
, the inverse quaternion
a
−1
= [a[
−2
a
such that aa
−1
= 1.
Hence, the quaternion algebra is a division algebra or a skew ﬁeld. It is denoted
by H after Hamilton, who discovered it.
In the space H
n
= R
4n
, there are right quaternionic lines, i. e. subsets ¦(a
1
ξ, . . . , a
n
ξ) [
ξ ∈ H¦, and similar left quaternionic lines ¦(ξa
1
, . . . , ξa
n
) [ ξ ∈ H¦. Each of them is
a real 4dimensional subspace of H
n
= R
4n
.
21:I. Find a right quaternionic line, that is not a left quaternionic line.
21:J. Prove that two right quaternionic lines in H
n
either meet only at 0, or coincide.
The quotient space of the unit sphere S
4n+3
of the space H
n+1
= R
4n+4
by
the partition into its intersections with right quaternionic lines is called the (right)
quaternionic projective space of dimension n. Similarly, but with left quaternionic lines,
the (left) quaternionic projective space of dimension n.
21:K. Are the right and left quaternionic projective space of the same dimension
homeomorphic?
The left quaternionic projective space of dimension n is denoted by HP
n
.
21:L. HP
0
consists of a single point.
21:M. HP
n
is homeomorphic to the quotient space of the unit closed ball D
4n
of the
space H
n
by the partition into points of the interior of D
4n
and 3sphere which are
intersections of the boundary sphere S
4n−1
with (left quaternionic) lines of H
n
.
Space HP
1
is called the quaternionic projective line.
21:N. Quaternionic projective line HP
1
is homeomorphic to S
4
.
21:O. HP
n
is canonically homeomorphic to the quotient space of H
n+1
¦0¦ by the
partition to left quaternionic lines of H
n+1
passing through the origin and punctured
at it.
Hence, HP
n
can be presented as the space of classes of left proportional (in the
quaternionic sense) nonzero sequences (x
0
, . . . , x
n
) of quaternions. Notation (x
0
:
x
1
: . . . : x
n
) and term homogenious coordinates introduced above in the real case are
used in the same way in the quaternionic situation.
§21. PROJECTIVE SPACES 133
21:P. HP
n
is canonically homeomorphic to the set of (left quaternionic) lines of
H
n+1
equipped with the topology induced by the angular metric (which takes values
in [0,
π
2
]).
'22 Spaces of Continuous Maps
'22
◦
1 Sets of Continuous Mappings
By ((X, Y ) we denote the set of all continuous mappings of a topologival space
X to a topological space Y .
22:1. Prove that ((X, Y ) consists of a single element iﬀ so does Y .
22:2. Prove that there exists an injection Y →((X, Y ). In other words, the
cardinality card((X, Y ) of ((X, Y ) is greater than or equal to cardY .
22:3. Enigma. Find natural conditions implying ((X, Y ) = Y .
22:4. Let Y = ¦0, 1¦ equipped with topology ¦∅, ¦0¦, Y ¦. Prove that there
exists a bijection between ((X, Y ) and the topological structure of X.
22:5. Let X be a set of n points with discrete topology. Prove that ((X, Y )
can be identiﬁed with Y . . . Y (n times).
22:6. Let Y be a set of k points with discrete topology. Find necessary and
suﬃcient condition for the set ((X, Y ) contain k
2
elements.
'22
◦
2 Topologies on Set of Continuous Mappings
Let X, Y be topological spaces, A ⊂ X, B ⊂ Y . Denote by W(A, B) the set
¦f ∈ ((X, Y ) [ f(A) ⊂ B¦. Denote by ∆
(pw)
the set
¦W(a, U) [ a ∈ X, U is open in Y ¦
and by ∆
(co)
the set
¦W(C, U) [ C ⊂ X is compact, U is open in Y ¦
22:A. ∆
(pw)
is a subbase of a topological structure on ((X, Y ).
The topological structure generated by ∆
(pw)
is called the topology of pointwise
convergency. The set ((X, Y ) equipped with this structure is denoted by (
(pw)
(X, Y ).
22:B. ∆
(co)
is a subbase of a topological structures on ((X, Y ).
The topological structure deﬁned by ∆
(co)
is called the compactopen topology.
Hereafter we denote by ((X, Y ) the space of all continuous mappings X → Y with
the compactopen topology, unless the contrary is speciﬁed explicitly.
22:C CompactOpen Versus Pointwise. The compactopen topology is ﬁner than
the topology of pointwise convergence.
22:7. Prove that ((I, I) is not homeomorphic to (
(pw)
(I, I).
Denote by Const(X, Y ) the set of all constant mappings f : X →Y .
22:8. Prove that the topology of pointwise convergence and compactopen
topology of ((X, Y ) induce the same topological structure on Const(X, Y ),
which, with this topology, is homeomorphic Y .
22:9. Let X be a discrete space of n points. Prove that (
(pw)
(X, Y ) is
homeomorphic Y . . . Y (n times). Is this true for ((X, Y )?
134
§22. SPACES OF CONTINUOUS MAPS 135
'22
◦
3 Topological Properties of Mapping Spaces
22:D. Prove that if Y is Hausdorﬀ, then (
(pw)
(X, Y ) is Hausdorﬀ for any topological
space X. Is this true for ((X, Y )?
22:10. Prove that ((I, X) is path connected iﬀ X is path connected.
22:11. Prove that (
(pw)
(I, I) is not compact. Is the space ((I, I) compact?
'22
◦
4 Metric Case
22:E. If Y is metrizable and X is compact then ((X, Y ) is metrizable.
Let (Y, ρ) be a metric space and X a compact space. For continuous maps f, g :
X →Y put
d(f, g) = max¦ρ(f(x), g(x)) [ x ∈ X¦.
22:F This is a Metric. If X is a compact space and Y a metric space, then d is a
metric on the set ((X, Y ).
Let X be a topological space and Y a metric space with metric ρ. A sequence
f
n
of maps X → Y is said to uniformly converge to f : X → Y if for any ε > 0 there
exists a natural N such that ρ(f
n
(x), f(x)) < ε for any n > N and x ∈ X. This is a
straightforward generalization of the notion of uniform convergence which is known
from Calculus.
22:G Metric of Uniform Convergence. Let X be a compact space and Y a
metric space. A sequence f
n
of maps X → Y converges to f : X →Y in the topology
deﬁned by d, iﬀ f
n
uniformly converges to f.
22:H Uniform Convergence Versus CompactOpen. Let X be a compact space
and Y a metric space. Then the topology deﬁned by d on ((X, Y ) coincides with the
compactopen topology.
22:12. Prove that the space ((R, I) is metrizable.
22:13. Let Y be a bounded metric space and X a topological space which
admits presentation X =
¸
∞
i=1
X
i
, where X
i
is compact and X
i
⊂ Int X
i+1
for i = 1, 2, . . .. Prove that ((X, Y ) is metrizable.
Denote by (
b
(X, Y ) the set of all continuous bounded maps from a topological
space X to a metric space Y . For maps f, g ∈ (
b
(X, Y ), put
d
∞
(f, g) = sup¦ρ(f(x), g(x)) [ x ∈ X¦.
22:I Metric on Bounded Mappings. This is a metric in (
b
(X, Y ).
22:J d
∞
and Uniform Convergence. Let X be a topological space and Y a
metric space. A sequence f
n
of bounded maps X → Y converges to f : X → Y in
the topology deﬁned by d
∞
, iﬀ f
n
uniformly converges to f.
22:K When Uniform Is Not CompactOpen. Find X and Y such that the topol
ogy deﬁned by d
∞
on (
b
(X, Y ) does not coincide with the compactopen topology.
§22. SPACES OF CONTINUOUS MAPS 136
'22
◦
5 Interactions With Other Constructions
22:L Continuity of Restricting. Let X, Y be topological spaces and A ⊂ X.
Prove that the map ((X, Y ) →((A, Y ) : f →f[
A
is continuous.
22:M Continuity of Composing. Let X be a topological space and Y a locally
compact Hausdorﬀ space. Prove that the map
((X, Y ) ((Y, Z) →((X, Z) : (f, g) →g ◦ f
is continuous.
22:14. Is local compactness of Y necessary in 22:M?
22:N Extending Target. For any topological spaces X, Y and B ⊂ Y the map
((X, B) →((X, Y ) : f →i
B
◦ f is a topological embedding.
22:O Maps to Product. For any topological spaces X, Y and Z the space ((X, Y
Z) is canonically homeomorphic to ((X, Y ) ((X, Z).
22:P Restricting to Sets Covering Source. Let ¦X
1
, . . . , X
n
¦ be a fundumental
cover of X. Prove that for any topological space Y ,
((X, Y ) →
n
¸
i=1
((X
i
, Y ) : f →(f[
X1
, . . . , f[
Xn
)
is a topological embedding. What if the cover is not fundamental?
22:Q Factorizing Source. Let S be a closed partition
1
of a Hausdorﬀ compact
space X. Prove that for any topological space Y the mapping
((X/S, Y ) →((X, Y )
is a topological embedding.
22:15. Are the conditions imposed on S and X in 22:Q necessary?
22:R The Evaluation Map. Let X, Y be topological spaces. Prove that if X is
locally compact and Hausdorﬀ then the map
((X, Y ) X →Y : (f, x) →f(x)
is continuous.
22:16. Are the conditions imposed on X in 22:R necessary?
'22
◦
6 Mappings X Y →Z and X →((Y, Z)
22:S. Let X, Y and Z be topological spaces and f XY →Z be a continuous map.
Then the map
F : X → ((Y, Z) : F(x) : y →f(x, y),
is continuous.
22:T. Let X, Z be topological spaces and Y a Hausdorﬀ locally compact space. Let
F : X → ((Y, Z) be a continuous mapping. Then the mapping f : X Y → Z :
(x, y) →F(x)(y) is continuous.
1
Recall that a partition is called closed, if the saturation of each closed set is
closed.
§22. SPACES OF CONTINUOUS MAPS 137
22:U. Let X, Y and Z be topological spaces. Let the mapping
Φ : ((X Y, Z) → ((X, ((Y, Z))
be deﬁned by the relation
Φ(f)(x) : y →f(x, y).
Then
(a) Φ is continuous;
(b) if Y is locally compact and Hausdorﬀ then Φ is a homeomorphism.
CHAPTER 4
A Touch of Topological Algebra
In this chapter we are going to study topological spaces strongly related to groups:
they either themselves are groups in a nice way (so that all the maps coming from
the group theory would be continuous), or groups act on topological spaces and can
be thought of as consisting of homeomorphisms.
This is a sort of interdisciplinary material. Although it plays important roles
in many areas of Mathematics, it is not that important in the frameworks of general
topology. Quite often, this material can be postponed till the introductory chapters of
the mathematical courses (functional analysis, Lie groups, etc.) which really require
it. In the frameworks of general topology, this material provides a greate collection
of exercises.
In the second part of the book, which is devoted to algebraic topology, groups
appear in a more profound way. So, sooner or later, the reader will meet groups. At
latest by the next chapter, for study of fundamental groups.
Groups are attributed to Algebra. In the Mathematics built on sets, the main
objects are sets with additional structure. Above we met a few of the most fundamen
tal of these structures: topology, metric, partial order. Topology and metric evolved
from geometric considerations. Algebra studied algebraic operations with numbers
and similar objects and contributed to the settheoretic Mathematics various struc
tures based on operations. One of the simplest (and most versatile) of these structures
is the structure of group. It emerges in overwhelming majority of mathematical envi
ronments. Often it appears together with topology and in a nice interaction with it.
This interaction is a subject of Topological Algebra.
The second part of this book is called Algebraic Topology. It is also about an
interaction of Topology and Algebra, spaces and groups. But this is a completely
diﬀerent interaction. There structures of topological space and group do not live on
he same set, but the group is built to encode topological properties of the space.
138
'23 Digression. Generalities on Groups
This section is included mainly to recall the most elementary deﬁnitions and
statements concerning groups. We do not mean to provide a selfcontained outline of
the group theory. The reader is assumed to be familiar with groups, homomorphisms,
subgroups, quotient groups, etc.
If this is not yet so, we recommend to read one of numerous algebraic textbooks
covering the elementary group theory. The mathematical culture, which hadto be
acquired for mastering of the material presented above in this book, would make this
an easy and pleasant exercise.
As a temporary solution, the reader can read few deﬁnitions and prove few the
orems gathered in this section. They provide a suﬃcient bases for most of what
follows.
'23
◦
1 The Notion of Group
Recall that a group is a set G equipped with a group operation. A group operation
in a set G is a map ω : G G → G satisfying the following three conditions (known
as group axioms):
• Associativity. ω(a, ω(b, c)) = ω(ω(a, b), c) for any a, b, c ∈ G,
• Existence of Neutral Element. There exists e ∈ G such that ω(e, a) =
ω(a, e) = a for every a ∈ G,
• Existence of Inverse. For any a ∈ G there exists b ∈ G such that ω(a, b) =
ω(b, a) = e.
23:A Uniqueness of Neutral Element. In a group a neutral element is unique.
23:B Uniqueness of Inverse. For any element of a group an element inverse to
it is unique.
23:C The First Examples of Groups. Check if in each of the following situations
we have a group. What is its neutral element? How to calculate the element inverse
to a given one?
• The set G is the set Z of integers, the group operation is addition: ω(a, b) = a+b.
• The set G is the set Q
>0
of positive rational nambers, the group operation is
multiplication: ω(a, b) = ab.
• G = R, ω(a, b) = a +b.
• G = C, ω(a, b) = a +b.
• G = R 0, ω(a, b) = ab.
• G is the set of all bijection of a set A onto itself, the group operatio is composition
of bijections: ω(a, b) = a ◦ b.
23:1 Simplest Group. Can a group be empty? Can it consist of one ele
ment?
A group consisting of one element is said to be trivial.
23:2 Solving Equations. Let G be a set with associative operation ω :
G G → G. Prove that this is a group, iﬀ for any a, b ∈ G there exists
uniques elements x, y ∈ G such that ω(a, x) = b and ω(y, a) = b.
139
§23. DIGRESSION. GENERALITIES ON GROUPS 140
'23
◦
2 Additive and Multiplicative Notations
The notation above are never used! (The only exception may happen, as
here, if the deﬁnition of group is discussed.) Instead, one uses either multiplicative or
aditive notation.
Under multiplicative notation the group operation is called multiplication and
denoted as multiplication: (a, b) → ab. The neutral element is called unity and
denoted by 1. The element inverse to a is denoted by a
−1
. These notation are
borrowed from the case, say, of group of nonzero rational numbers with the usual
multiplication.
Under additive notation the group operation is called addition and denoted as
addition: (a, b) → a + b. The neutral element is called zero and denoted by 0. The
element inverse to a is denoted by −a. These notation are borrowed from the case,
say, of group of integer numbers with the usual addition.
An operation ω : GG →G is commutative provided that ω(a, b) = ω(b, a) for all
a, b ∈ G. A group with commutative group operation is called commutative or abelian.
Traditionally the additive notation is used only in the case of commutative group,
while the multiplicative notation is used both for commutative and noncommutative
cases. Below we use mostly the multiplicative notation.
23:3. Check if in each of the following situations we have a group:
(a) a singleton ¦a¦ with multiplication aa = a,
(b) the set S
n
of bijections of the set ¦1, 2, . . . , n¦ of n ﬁrst natural numbers
with composition (symmetric group of degree n,)
(c) the sets R
n
and C
n
with coordinatewise addition,
(d) the set Homeo(X) of all homeomorphisms of a topological space X with
composition,
(e) the set of invertible real n nmatrices GL(n, R) with matrix multipli
cation,
(f) the set M
n
(R) of all real n nmatrices with addition of matrices,
(g) the set of all subsets of a set X with symmetric diﬀerence
(A, B) →(A ∪ B) (A ∩ B)
,
(h) the set Z
n
of classes of natural numbers congruent modulo n with the
addition deﬁned by the addition of natural numbers,
(i) the set of complex roots of unity of degree n with multiplication,
(j) the set R
>0
of positive real numbers with multiplication,
(k) S
1
⊂ C with multiplication,
(l) the set of translations of a plane with composition.
Associativity implies that in a group any ﬁnite sequence of elements has a well
deﬁned product, which can be calculated by a sequence of pairwise multiplications
deﬁned by any placement of backets, say abcde = (ab)(c(de)). The distribution of
brackets does not matter. In the case of a sequence of three elements this is just
associativity (ab)c = a(bc).
23:D. Prove that independence of brackets in the product of any length follows from
associativity.
For an element a of a group G powers a
n
with n ∈ Z are deﬁned by formulas:
a
n+1
= a
n
a, a
0
= 1, a
−n
= (a
−1
)
n
.
§23. DIGRESSION. GENERALITIES ON GROUPS 141
23:E. Prove that a
n
is welldeﬁned and has the properties a
p
a
q
= a
p+q
and (a
p
)
q
=
a
pq
.
'23
◦
3 Homomorphisms
Recall that a map f : G → H of a group to other one is called a homomorphism,
if f(xy) = f(x)f(y) for any x, y ∈ G.
23:4. Above, in the deﬁnition of homomorphism, multiplicative notation is
used. How does this deﬁnition look in additive notation? What if one of the
groups is multiplicative, while the other is additive?
23:5. Let a be an element of a multiplicative group G. Is a map Z → G
deﬁned by formula n →a
n
a homomorphism?
23:F. Let G and H be groups. Is the contant map G → H mapping the whole G
to the neutral element of H a homomorphism? Is any other constant map G →H a
homomorphism?
23:G. A homomorphism maps the neutral element to the neutral element, and ele
ments inverse to each other to elements inverse to each other.
23:H. The identity map of a group is a homomorphism. The composition of homo
morphisms is a homomorphism.
Recall that a homomorphism is called an epimorphism, if it is surjective, monomor
phism, if it is injective, and isomorphism, if it is bijective.
23:I. The inverse map to an isomorphism is also an isomorphism.
Groups are called isomorphic, if there exists an isomorphism of one of them to
another one.
23:J. Being isomorphic is an equivalence relation.
'23
◦
4 Subgroups
A subset A of a group G is called a subgroup of G if it is invariant under the
group operation of G (i.e. ab ∈ A for any a, b ∈ A) and A, with the operation induced
by the operation in G, is a group.
For subsets A and B of a group G put AB = ¦ab [ a ∈ A, b ∈ B¦ and A
−1
=
¦a
−1
[ a ∈ A¦.
23:K. A subset A of a multiplicative group G is a subgroup of G, iﬀ AA ⊂ G and
A
−1
⊂ A.
23:6. The one element subset of a group consisting of its neutral element is
a subgroup.
23:7. Prove that a subset A of a ﬁnite group G is a subgroup, if AA ⊂ A
(the condition A
−1
⊂ A is not needed).
23:8. List all subgroups of the additive Z.
23:9. Is GL(n, R) a subgroup of M
n
(R) (for notations see 23:3)?
23:L. The image of a group homomorphism f : G →H is a subgroup of H.
§23. DIGRESSION. GENERALITIES ON GROUPS 142
23:M. Let f : G → H be a group homomorphism and K a subgroup of H. Then
f
−1
(K) is a subgroup of G. Shortly speaking:
The preimage of a subgroup under a group homomorphism is a subgroup.
The preimage of the neutral element under a group homomorphism f : G → H
is called the kernel of f and denoted by Ker f.
23:N Corollary of 23:M. Kernel of a group homomorphism is a subgroup.
23:O. A group homomorphism is a monomorphism, iﬀ its kernel is trivial.
23:P. Intersection of any collection of subgroups of a group is a subgroup of this
group.
A subgroup of a group G is said to be generated by a subset S ⊂ G if this is the
smallest subgroup of G which contains S.
23:Q. The subgroup generated by S is the intersection of all the subgroups of G
that contain S. On the other hand, this is the set of all the elements which can be
obtained as products of elements of S and elements inverse to elements of S.
The elements of a set that generates G are called generators of G. A group
generated by an element is said to be cyclic.
23:R. A cyclic (multiplicative) group consists of powers of its generator. (I.e., if G
is a cyclic group and a is its generator then G = ¦a
n
[ n ∈ Z¦.) Any cyclic group is
commutative.
23:10. A subgroup H of group G is cyclic, iﬀ there exists a homomorphism
f : Z →G such that H = f(Z).
23:S. Any subgroup of a cyclic group is cyclic.
The number of elements of a group G is called the order of G. It is denoted by
[G[.
23:T. Let G be a ﬁnite cyclic group. Then for any positive divisor d of [G[ there
exists a unique subgroup H of G with [H[ = d.
Each element of a group generates a cyclic subgroup which consists of all powers
of this element. The order of the subgroup generated by a ∈ G is called the order of
a. It can be a natural number or inﬁnity.
'24 Topological Groups
'24
◦
1 The Notion of Topological Group
A topological group is a set G equipped with both topological and group structures
such that the maps GG →G : (x, y) →xy and G →G : x →x
−1
are continuous.
24:1. Prove that if G is a group and a topological space then maps G
G → G : (x, y) → xy and G → G : x → x
−1
are continuous, iﬀ the map
GG →G : (x, y) →x
−1
y is continuous.
24:2. Prove that for a topological group G the inversion G → G : x → x
−1
is a homeomorphism.
24:3. Let G be a topological group, X a topological space, and f, g : X → G
be maps continuous at a point x
0
∈ X. Prove that maps X → G : x →
f(x)g(x) and X → G : x →(f(x))
−1
are continuous at x
0
.
24:A. Any group equipped with the discrete topological structure is a topological
group.
24:4. Is a group equipped with the indiscrete topological structure a topo
logical group?
'24
◦
2 Examples of Topological Groups
24:B. The real line R and complex line C with the addition are topological groups.
24:C. The punctured real line R 0 with the multiplication is a topological group.
24:D. The punctured complex line C 0 with the multiplication is a topological
group.
24:E. The unit circle S
1
= ¦[z[ = 1¦ ⊂ C with the multiplication is a topological
group.
24:5. Check if in each of the following situations we have a topological group:
(a) the spaces R
n
, C
n
and H
n
with coordinatewise addition,
(b) the sets M
n
(R), M
n
(C) and M
n
(H) of all nn matrices with real, com
plex, and, respectively, quaternion elements with the product topology
(we identify M
n
(R) with R
n
2
, M
n
(C) with C
n
2
and M
n
(H) with H
n
2
)
under the entrywise addition,
(c) the set GL(n, R) of invertible real nnmatrices with the matrix multi
plication and the topology induced by the inclusion GL(n, R) ⊂ M
n
(R),
(d) the set GL(n, C) of invertible complex n nmatrices with the matrix
multiplication and the topology induced by the inclusion GL(n, C) ⊂
M
n
(C),
(e) the set GL(n, H) of invertible quaternionic nnmatrices with the ma
trix multiplication and the topology induced by the inclusion GL(n, H) ⊂
M
n
(H),
'24
◦
3 SelfHomeomorphisms Making a Topological Group Homogeneous
Recall that the maps of a group G to itself deﬁned by formula x → xa
−1
and
x → ax, respectively, are called (right and left) translations and denoted by R
a
and
L
a
.
143
§24. TOPOLOGICAL GROUPS 144
24:F. Any translation of a topological group is a homeomorphism.
Recall that the conjugation of a group G by a ∈ G is the map G →G : x →a
−1
xa.
24:G. Conjugation of a topological group by any its element is a homeomorphism.
Recall that for given subsets A, B of a group G, the set ¦ab : a ∈ A, b ∈ B¦ is
denoted by AB, and ¦a
−1
: a ∈ A¦ is denoted by A
−1
.
24:H. If U is an open set in a topological group G then for any x ∈ G the sets xU,
Ux and U
−1
are open.
24:6. Does the same hold true for closed sets?
24:7. Prove that if U and V are subsets of a topological group G and U is
open then UV and V U are open.
24:8. Does the same hold true if one replaces all the words open by closed?
24:8.1. Which of the following sugroups of the additive group R
are closed:
(a) Z,
(b)
√
2Z,
(c) Z +
√
2Z?
24:9. Prove that if U is compact and V is closed then UV and V U are closed.
24:9.1. Let F and C be disjoint subsets of a topological group G.
If F is closed and C is compact, then there exists a neighborhood
V of unity such that CV does not intersect F.
'24
◦
4 Neighborhoods
24:I. If Γ is a neighborhood basis at the unity 1 in a topological group G then Σ =
¦aU : a ∈ G, U ∈ Γ¦ is a basis for topology of G.
A subset A of a group G is said to be symmetric if A
−1
= A.
24:J. Any neighborhood of unity of a topological group contains a symmetric neigh
borhood of unity.
24:K. For any neihgborhood U of 1 of a topological group there exists a neighborhood
V of 1 such that V V ⊂ U.
24:10. For any neihgborhood U of 1 of a topological group and any natural
number n there exists a symmetric neighborhood V of 1 such that V
n
⊂ U.
24:11. Let G be a group and Σ be a collection of its subsets. Prove that there
exists a unique topology on G such that G with this topology is a topological
group and Σ is its neighborhood basis at the unity, iﬀ Σ satisﬁes the following
ﬁve conditions:
(a) each U ∈ Σ contains the unity of G,
(b) for every x ∈ U ∈ Σ there exists V ∈ Σ such that xV ⊂ U,
(c) for each U ∈ Σ there exists V ∈ Σ such that V
−1
⊂ U,
(d) for each U ∈ Σ there exists V ∈ Σ such that V V ⊂ U,
(e) for every x ∈ G and U ∈ Σ there exists V ∈ Σ such that V ⊂ x
−1
Ux.
24:L. Enigma. For what reasons 24:K is similar to the triangle inequality?
§24. TOPOLOGICAL GROUPS 145
'24
◦
5 Separaion Axioms
24:M. A topological group is Hausdorﬀ, iﬀ it satisﬁes the ﬁrst separation axiom, iﬀ
the unity is closed.
24:N. A topological group is Hausdorﬀ, iﬀ the unity is equal to the intersection of
its neighborhoods.
24:O. If the unity of a topological group G is closed, then G (as a topological space)
is regular.
24:P Corollary. For topological groups the ﬁrst three separation axioms are equiv
alent.
'24
◦
6 Countability Axioms
24:Q. If Γ is a neighborhood basis at the unity 1 in a topological group G and S ⊂ G
is dense in G, then Σ = ¦aU : a ∈ S, U ∈ Γ¦ is a basis for topology of G. Cf. 24:I
and 14.F.
24:R. A ﬁrst countable separable topological group is second countable.
24:12*. (Cf. 14:D) If a Hausdorﬀ topological group has a countable base
at the unity, then it is metrizable.
'24
◦
7 Subgroups
Recall that a subset H of a group G such that HH = H and H
−1
= H is called
a subgroup of G. It is a group with the operation deﬁned by the group operation of
G. If G is a topological group, then H inherits also a topological structure from G.
24:S. If H is a subgroup of a topological group G, then the topological and group
structures induced from G make H a topological group.
24:13. Prove that a subgroup of a topological group is open, iﬀ it contains
an interior point.
24:14. Prove that every open subgroup of a topological group is also closed.
24:15. Prove that every closed subgroup of ﬁnite index is open.
24:16. Find an example of a subgroup of a topological group, which
(a) is closed, but not open,
(b) is neither closed, nor open.
24:17. Prove that a subgroup of a topological group is discrete, iﬀ it contains
an isolated point.
24:18. Prove that a subgroup H of a topological group G is closed, iﬀ it
is locally closed, i.e., there exists an open set U ⊂ G such that U ∩ H =
U ∩ Cl H = ∅.
24:19. Prove that if H is a nonclosed subgroup of a topological group G
then Cl H H is dense in Cl H.
24:20. The closure of a subgroup of a topological group is a subgroup.
24:21. Is it true that the interior of a subgroup of a topological group is a
subgroup?
§24. TOPOLOGICAL GROUPS 146
Recall that the smallest subgroup of a group G containing a set S is said to be
generated by S.
24:T. A connected topological group is generated by any neighborhood of the unity.
Recall that for a subgroup H of a group Gright cosets are sets Ha = ¦xa : x ∈ H¦
with a ∈ G. Analogously, sets aH are left cosets of H in G.
24:22. Let H be a subgroup of a group G. Deﬁne a relation: a ∼ b if
ab
−1
∈ H. Prove that this is an equivalence relation and the right cosets of
H in G are the equivalence classes.
24:23. What is the counterpart of 24:22 for left cosets?
The set of left cosets of H in G is denoted by G/H, the set of right cosets of H
in G, by H ` G. If G is a topological group and H is its subgroup then the sets G/H
and H ` G are provided with the quotient topology. Equipped with these topologies,
they are called spaces of cosets.
24:U. For any topological group G and its subgroup H, the natural projections
G →G/H and G →H ` G are open (i.e., the image of every open set is open).
24:24. The space of left (or right) cosets of a closed subgroup in a topological
group is regular.
'24
◦
8 Normal Subgroups
Recall that a subgroup H of a group G is said to be normal if a
−1
ha ∈ H for
all h ∈ H and a ∈ G. Normal subgroups are called also normal divisors or invariant
subgroups.
24:25. Prove that the closure of a normal subgroup of a topological group
is a normal subgroup.
24:26. The connected component of the unity of a topological group is a
closed normal subgroup.
24:27. The pathconnected component of the unity of a topological group is
a normal subgroup.
Recall that for a normal subgroup left cosets coincide with right cosets and the
set of cosets is a group with the multiplication deﬁned by formula (aH)(bH) = abH.
The group of cosets of H in G is called the quotient group or factor group of G by H
and denoted by G/H.
24:V. The quotient group of a topological group is a topological group (provided
that it is considered with the quotient topology).
24:28. The natural projection of a topological group onto its quotient group
is open.
24:29. A quotient group of a ﬁrst (or second) countable group is ﬁrst (re
spectively, second) countable.
24:30. The quotient group G/H of a topological group G is regular, iﬀ H is
closed.
24:31. Prove that if a normal subgroup H of a topological group G is open
then the quotient group G/H is discrete.
§24. TOPOLOGICAL GROUPS 147
24:32. Let G be a ﬁnite topological group. Prove that there exists a normal
subgroup H of G such that a set U ⊂ G is open, iﬀ it is a union of several
cosets of H in G.
'24
◦
9 Homomorphisms
Recall that a map f of a group G to a group H is called a (group) homomorphism
if f(xy) = f(x)f(y) for all x, y ∈ G. If G and H are topological groups then by a
homomorphism G →H one means a group homomorphism which is continuous.
24:W. A group homomorphism of a topological group to a topological group is con
tinuous, iﬀ it is continuous at 1.
Besides similar modiﬁcations, which can be summarized by the following prin
ciple: everything is assumed to respect the topological structures, the terminology of
group theory passes over without changes. In particular, the kernel Ker f of a homo
morphism f : G →H is deﬁned as the preimage of the unity of H. A homomorphism
f is a monomorphism if it is injective. This is known to be equivalent to Ker f = 1.
A homomorphism f : G → H is an epimorphism if it is surjective, i.e, its image
Imf = f(G) is the whole H.
In group theory, an isomorphism is an invertible homomorphism. Its inverse is a
homomorphism (and hence an isomorphism) automatically. In theory of topological
groups this must be included in the deﬁnition of isomorphism: an isomorphism of
topological groups is an invertible homomorphism whose inverse is also a homomor
phism. In other words, an isomorphism of topological groups is a map which is both
an algebraic homomorphism and a homeomorphism. Cf. Section '10.
24:33. An epimorphism f : G → H is open, iﬀ its injective factor, f/
S(f)
:
G/ Ker f →H, is an isomorphism.
24:34. An epimorphism of a compact topological group onto a topological
group with closed unity is open.
24:35. Prove that the quotient group R/Z of the additive group of real
numbers by the subgroup of integers is isomorphic to the multiplicative group
S
1
= ¦z ∈ C : [z[ = 1¦ of complex numbers with absolute value 1.
'24
◦
10 Local Isomorphisms
Let G and H be topological groups. A local isomorphism of G to H is a homeo
morphism f of a neighborhood U of the unity of G to a neighborhood V of the unity
of H such that
• f(xy) = f(x)f(y) for every x, y ∈ U such that xy ∈ U,
• f
−1
(zt) = f
−1
(z)f
−1
(t) for every z, t ∈ V such that zt ∈ V .
Topological groups G, H are said to be locally isomorphic if there exists a local
isomorphism of G to H.
24:X. Isomorphic topological groups are locally isomorphic.
24:Y. Additive group R of real numbers and multiplicative group S
1
of complex
numbers with absolute value 1 are locally isomorphic, but not isomorphic.
24:36. Prove that the relation of being locally isomorphic is an equivalence
relation on the class of topological groups.
§24. TOPOLOGICAL GROUPS 148
24:37. Find neighborhoods of unities in R and S
1
and a homeomorphism
between them, which satisﬁes the ﬁrst condition from the deﬁnition of local
isomorphism, but does not satisfy the second one.
24:38. Prove that for any homeomorphism between neighborhods of unities
of two topological groups, which satisﬁes the ﬁrst condition from the deﬁnition
of local isomorphism, but does not satisfy the second one, there exists a
submapping, which is a locall isomorphsm between these topological groups.
'24
◦
11 Direct Products
Let G and H be topological groups. In group theory, the product GH is given
a group structure,
1
in topology it is given a topological structure (see Secion '18).
24:Z. These two structures are compatible: the group operations in G H are
continuous with respect to the product topology.
Thus, G H is a topological group. It is called the direct product of the topo
logical groups G and H. There are canonical homomorphisms related with this: the
inclusions i
G
: G → G H : x → (x, 1) and i
H
: H → G H : x → (1, x),
which are monomorphisms, and the projections p
G
: G H → G : (x, y) → x and
p
H
: GH →H : (x, y) →y, which are epimorphisms.
24:39. Prove that the topological groups GH/
i
H
and G are isomorphic.
24:40. The product operation is both commutative and associative: GH is
(canonically) isomorphic to HG and G(HK) is canonically isomorphic
to (GH) K.
A topological group G is said to decompose into the direct product of its subgroups
A and B if the map AB →G : (x, y) →xy is an isomorphism of topological groups.
If this is the case, the groups G and AB are usually identiﬁed via this isomorphism.
Recall that a similar deﬁnition exists in ordinary group theory. The only diﬀerence
is that there the isomorphism is just an algebraic isomorphism. Moreover, in that
theory, G decomposes into the direct product of its subgroups A and B, iﬀ A and B
generate G, are normal subgroups and A ∩ B = 1. Therefore, if these conditions are
satisﬁed in the case of topological groups, then (x, y) → xy : A B → G is a group
isomorphism.
24:41. Prove that in this situation the map (x, y) → xy : A B → G is
continuous. Find an example where the inverse group isomorphism is not
continuous.
24:42. Prove that a compact Hausdorﬀ group which decomposes algebraically
into the direct product of two subgroups, decomposes also into the direct
product of these subgroups in the category of topological groups.
24:43. Prove that the multiplicative group R0 of real numbers is isomorphic
(as a topological group) to the direct product of the multiplicative group
S
0
= ¦1, −1¦ and the multiplicative group R
∗
+
= ¦x ∈ R : x > 0¦.
24:44. Prove that the multiplicative group C 0 of complex numbers is
isomorphic (as a topological group) to the direct product of the multiplicative
group S
1
= ¦z ∈ C : [z[ = 1¦ and the multiplicative group R
∗
+
.
1
Recall that the multiplication in G H is deﬁned by formula (x, u)(y, v) =
(xy, uv).
§24. TOPOLOGICAL GROUPS 149
24:45. Prove that the multiplicative group H0 of quaternions is isomorphic
(as a topological group) to the direct product of the multiplicative group
S
3
= ¦z ∈ H : [z[ = 1¦ and the multiplicative group R
∗
+
.
24:46. Prove that the subgroup S
0
= ¦1, −1¦ of S
3
= ¦z ∈ H : [z[ = 1¦ is
not a direct factor.
24:47. Find a topological group homeomorphic to RP
3
(the threedimensional
real projective space).
'25 Actions of Topological Groups
'25
◦
1 Actions of Group in Set
Must be written!
'25
◦
2 Continuous Actions
Must be written!
'25
◦
3 Orbit Spaces
Must be written!
'25
◦
4 Homogeneous Spaces
Must be written!
150
Part 2
Algebraic Topology
This part of the book can be considered as an introduction to alge
braic topology. This is a part of topology, which relates topological and
algebraic problems. The relationship is used in both directions, but re
duction of topological problems to algebra is at ﬁrst stages more useful,
since algebra is usually easier.
The relation is established according to the following scheme. One
invents a construction, which assigns to each topological space X under
consideration an algebraic object A(X). The latter may be a group, or a
ring, or a quadratic form, or algebra, etc. Another construction assigns
to a continuous mapping f : X → Y a homomorphism A(f) : A(X) →
A(Y ). The constructions should satisfy natural conditions (in particu
lar, they form a functor), which make it possible to relate topological
phenomena with their algebraic images obtained via the constructions.
There are inﬁnitely many useful constructions of this kind. In this
part we deal mostly with one of them. This is the ﬁrst one, ﬁrst from both
the viewpoints of history and its role in mathematics. It was invented
by Henri Poincar´e in the end of the nineteenth century.
CHAPTER 5
Fundamental Group and Covering Spaces
'26 Homotopy
'26
◦
1 Continuous Deformations of Maps
26.A. Is it possible to deform continuously
(a) The identity map id : R
2
→R
2
to the constant map R
2
→R
2
: x →
0,
(b) The identity map id : S
1
→S
1
to the symmetry S
1
→S
1
: x →−x
(here x is considered as a complex number, since the circle S
1
is
¦x ∈ C : [x[ = 1¦),
(c) The identity map id : S
1
→S
1
to the constant map S
1
→S
1
: x →
1,
(d) The identity map id : S
1
→ S
1
to the twofold wrapping S
1
→ S
1
:
x →x
2
,
(e) The inclusion S
1
→R
2
to a constant map,
(f) The inclusion S
1
→R
2
0 to a constant map?
26.B. Enigma. When you (tried to) solve the previous problem, what
did you mean by “deform continuously”?
This section is devoted to the notion of homotopy formalizing the
naive idea of the continuous deformation of a map.
'26
◦
2 Homotopy as Map and Family of Maps
Let f, g be continuous maps of a topological space X to a topological
space Y , and H : X I → Y a continuous map such that H(x, 0) =
f(x) and H(x, 1) = g(x) for any x ∈ X. Then f and g are said to be
homotopic, and H is called a homotopy between f and g.
For x ∈ X, t ∈ I denote H(x, t) by h
t
(x). This change of notation
results in a change of the point of view of H. Indeed, for a ﬁxed t the
formula x → h
t
(x) deﬁnes a map h
t
: X → Y and H appears to be a
family of maps h
t
enumerated by t ∈ I.
26.C. Each h
t
is continuous.
26.D. Does continuity of all h
t
imply continuity of H?
The conditions H(x, 0) = f(x) and H(x, 1) = g(x) in the deﬁnition
of homotopy above can be reformulated as h
0
= f and h
1
= g. Thus a
153
§26. HOMOTOPY 154
homotopy between f and g can be considered as a family of continuous
maps, which connects f and g. Continuity of a homotopy allows one to
say that it is a continuous family of continuous maps.
'26
◦
3 Homotopy as Relation
26.E. Homotopy of maps is an equivalence relation.
26.E.1. If f : X →Y is a continuous map then H : X I →Y deﬁned by
H(x, t) = f(x) is a homotopy between f and f.
26.E.2. If H is a homotopy between f and g then H
deﬁned by H
(x, t) =
H(x, 1 −t) is a homotopy between g and f.
26.E.3. If H is a homotopy between f and f
and H
is a homotopy between
f
and f
then H
deﬁned by
H
(x, t) =
H(x, 2t) for t ≤ 1/2,
H
(x, 2t −1) for t ≥ 1/2
is a homotopy between f and f
.
Homotopy, being an equivalence relation by 26.E, divides the set
((X, Y ) of all continuous mappings of a space X to a space Y into equiva
lence classes. The latter are called homotopy classes. The set of homotopy
classes of all continuous maps X →Y is denoted by π(X, Y ).
26.1. Prove that for any X, the set π(X, I) has a single element.
26.2. Prove that the number of elements of π(I, Y ) coincides with the num
ber of path connected components of Y .
'26
◦
4 StraightLine Homotopy
26.F. Any two continuous maps of the same space to R
n
are homotopic.
26.G. Solve the preceding problem by proving that for continuous maps
f, g : X →R
n
formula H(x, t) = (1 −t)f(x) +tg(x) deﬁnes a homotopy
between f and g.
The homotopy deﬁned in 26.G is called a straightline homotopy.
26.H. Any two continuous maps of an arbitrary space to a convex sub
space of R
n
are homotopic.
'26
◦
5 Maps to Star Convex Sets
A set A ⊂ R
n
is said to be star convex, if there exists a point b ∈ A such
that for any x ∈ A the whole segment [a, x] connecting x to a is contained in
A.
26.3. Prove that any two continuous maps of a space to a star convex sub
space of R
n
are homotopic.
§26. HOMOTOPY 155
'26
◦
6 Maps of Convex Sets
26.4. Prove that any continuous map of a convex set C ⊂ R
n
to any space
is homotopic to a constant map.
26.5. Under what conditions (formulated in terms of known topological prop
erties of a space X) any two continuous maps of any convex set to X are
homotopic?
'26
◦
7 Easy Homotopies
26.6. Prove that any nonsurjective map of an arbitrary topological space
to S
n
is homotopic to a constant map.
26.7. Prove that any two maps of a onepoint space to R
n
¦0¦ with n > 1
are homotopic.
26.8. Find two nonhomotopic maps of a onepoint space to R ¦0¦.
26.9. For various m, n, k, calculate the number of homotopy classes of maps
¦1, 2, . . . , m¦ → R
n
¦x
1
, x
2
, . . . , x
k
¦, where ¦1, 2, . . . , m¦ is equipped with
discrete topology.
26.10. Let f, g be maps of a topological space X to C 0. Prove that if
[f(x) −g(x)[ < [f(x)[ for any x ∈ X then f and g are homotopic.
26.11. Prove that for any polynomials p and q over C of the same degree in
one variable there exists r > 0 such that for any R > r formulas z → p(z)
and z → q(z) deﬁne maps of circle ¦z ∈ C : [z[ = R¦ to C 0 and these
maps are homotopic.
26.12. Let f, g be maps of an arbitrary topological space X to S
n
. Prove
that if [f(a) −g(a)[ < 2 for any a ∈ X then f is homotopic to g.
26.13. Let f : S
n
→S
n
be a continuous map. Prove that if it is ﬁxed point
free, i.e., f(x) = x for any x ∈ S
n
, then f is homotopic to the symmetry
x →−x.
'26
◦
8 Two Natural Properties of Homotopies
26.I. Let f, f
: X → Y , g : Y → B, h : A → X be continuous maps
and F : X I → Y a homotopy between f and f
. Prove that then
g ◦ F ◦ (h id
I
) is a homotopy between g ◦ f ◦ h and g ◦ f
◦ h.
26.J. Enigma. Under conditions of 26.I deﬁne a natural mapping
π(X, Y ) →π(A, B).
How does it depend on g and h? Write down all the nice properties of
this construction.
26.K. Prove that maps f
0
, f
1
: X →Y Z are homotopic iﬀ pr
Y
◦f
0
is
homotopic to pr
Y
◦ f
1
and pr
Z
◦f
0
is homotopic to pr
Z
◦ f
1
.
§26. HOMOTOPY 156
'26
◦
9 Stationary Homotopy
Let A be a subset of X. A homotopy H : XI →Y is said to be ﬁxed
or stationary on A, or, brieﬂy, to be an Ahomotopy, if H(x, t) = H(x, 0)
for all x ∈ A, t ∈ I. Maps which can be connected by an Ahomotopy
are said to be Ahomotopic.
Of course, Ahomotopic maps coincide on A. If one wants to empha
size that a homotopy is not assumed to be ﬁxed, one says that it is free.
If one wants to emphasize the opposite (that it is ﬁxed), one says that
the homotopy is relative.
Warning: there is a similar, but diﬀerent kind of homotopy, which is
also called relative. See below.
26.L. Prove that, like free homotopy, Ahomotopy is an equivalence
relation.
The classes into which Ahomotopy divides the set of continuous maps
X → Y that agree on A with a map f : A → Y are called Ahomotopy
classes of continuous extensions of f to X.
26.M. For what A is a straightline homotopy ﬁxed on A?
'26
◦
10 Homotopies and Paths
Recall that by a path in a space X we mean a continuous mapping of
the interval I into X. (See Section '12.)
26.N. Enigma. In what sense is any path a homotopy?
26.O. Enigma. In what sense does any homotopy consist of paths?
26.P. Enigma. In what sense is any homotopy a path?
26.Q. Enigma. Introduce a topology in the set ((X, Y ) of all continu
ous mappings X →Y in such a way that for any homotopy h
t
: X →Y
the map I →((X, Y ) : t →h
t
would be continuous.
Recall that the compactopen topology in ((X, Y ) is the topology generated by
the sets ¦ϕ ∈ ((X, Y ) [ ϕ(A) ⊂ B¦ for compact A ⊂ X and open B ⊂ Y .
26:A. Prove that any homotopy h
t
: X → Y deﬁnes (by the formula presented in
26.Q) a path in ((X, Y ) with compactopen topology.
26:B. Prove that if X is locally compact and regular then any path in ((X, Y ) with
compactopen topology deﬁnes a homotopy.
'26
◦
11 Homotopy of Paths
26.R. Prove that any two paths in the same space X are freely homo
topic, iﬀ their images belong to the same pathwise connected component
of X.
§26. HOMOTOPY 157
This shows that the notion of free homotopy in the case of paths is
not interesting. On the other hand, there is a sort of relative homotopy
playing a very important role. This is (0 ∪1)homotopy. This causes the
following commonly accepted deviation from the terminology introduced
above: homotopy of paths always means not a free homotopy, but a
homotopy ﬁxed on the end points of I (i.e. on 0 ∪ 1).
Notation: a homotopy class of a path s is denoted by [s].
'27 Homotopy Properties of Path Multiplication
'27
◦
1 Multiplication of Homotopy Classes of Paths
Recall (see Section '12) that paths u and v in a space X can be
multiplied, provided the initial point v(0) of v coincides with the ﬁnal
point u(1) of u. The product uv is deﬁned by
uv(t) =
u(2t), if t ≤ 1/2
v(2t −1), if t ≥ 1/2.
27.A. If a path u is homotopic to u
and a path v is homotopic to v
and
there exists product uv, then u
v
exists and is homotopic to uv.
Deﬁne a product of homotopy classes of paths u and v to be the
homotopy class of uv. So, [u][v] is deﬁned as [uv], provided uv is deﬁned.
This is a deﬁnition that requires a proof.
27.B. The product of homotopy classes of paths is welldeﬁned.
1
'27
◦
2 Associativity
27.C. Is multiplication of paths associative?
Of course, this question might be formulated with more details:
27.D. Let u, v, w be paths in the same space such that products uv
and vw are deﬁned (i.e., u(1) = v(0) and v(1) = w(0)). Is it true that
(uv)w = u(vw)?
27.1. Prove that for paths in a metric space (uv)w = u(vw) implies that u,
v, w are constant maps.
27.2. Enigma. Find nonconstant paths u, v, and w in an indiscrete space
such that (uv)w = u(vw).
27.E. Multiplication of homotopy classes of paths is associative.
27.E.1. Reformulate Theorem 27.E in terms of paths and their homotopies.
27.E.2. Find a map ϕ : I →I such that if u, v and w are paths such that
u(1) = v(0) and v(1) = w(0) then ((uv)w) ◦ ϕ = u(vw).
27.E.3. Any path in I which starts at 0 and ﬁnishes at 1 is homotopic to
id I →I.
27.E.4. Recall Problem 26.I.
27.E.5. Let u, v and w be paths in a space such that products uv and vw
are deﬁned (thus, u(1) = v(0) and v(1) = w(0)). Then (uv)w is homotopic
to u(vw).
1
Of course, when the initial point of paths of the ﬁrst class coincides with the
ﬁnal point of paths of the second class.
158
§27. HOMOTOPY PROPERTIES OF PATH MULTIPLICATION 159
If you want to understand the essence of 27.E, you have to realize that
paths (uv)w and u(vw) have the same trajectories and diﬀers by time
spent in diﬀerent fragments of the path. Therefore to ﬁnd a homotopy
between them one has to ﬁnd a continuous way to change one schedule
to the other. The lemmae above suggest a formal way of such a change,
but the same eﬀect can be achieved in many other ways.
'27
◦
3 Unit
Let a be a point of a space X. Denote by e
a
the path I →X : t →a.
27.F. Is e
a
a unit for multiplication of paths?
The same question in more detailed form:
27.G. For a path u with u(0) = a is e
a
u = u? For a path v with v(1) = a
is ve
a
= v?
27.3. Prove that e
a
u = u implies u = e
a
.
27.H. The homotopy class of e
a
is a unit for multiplication of homotopy
classes of paths.
Problems 27.F, 27.G and 27.H are similar to problems 27.C, 27.D and
27.E, respectively.
27.H.1. Enigma. Exploring this analogy, formupate and prove analogues
of lemmas 27.E.1, 27.E.2 and 27.E.5.
'27
◦
4 Inverse
Recall that for a path u there is inverse path u
−1
deﬁned by u
−1
(t) =
u(1 −t) (see Section '12).
27.I. Is the inverse path inverse with respect to multiplication of paths?
In other words:
27.J. For a path u beginning in a and ﬁnishing in b is uu
−1
= e
a
and
u
−1
u = e
b
?
27.4. Prove that for a path u with u(0) = a equality uu
−1
= e
a
implies
u = e
a
.
27.K. For any path u the homotopy class of path u
−1
is inverse to the
homotopy class of u.
27.K.1. Find a map ϕ : I →I such that (uu
−1
) = u ◦ ϕ for any path u.
27.K.2. Any path in I, which starts and ﬁnishes at 0, is homotopic to the
constant path e
0
: I →I.
§27. HOMOTOPY PROPERTIES OF PATH MULTIPLICATION 160
We see that from the algebraic viewpoint multiplication of paths is
terrible, but it deﬁnes multiplication of homotopy classes of paths, which
has nice algebraic properties. The only unfortunate property is that the
multiplication of homotopy classes of paths is not deﬁned for any two
classes.
27.L. Enigma. How to select a subset of the set of homotopy classes
of paths to obtain a group?
'28 Fundamental Group
'28
◦
1 Deﬁnition of Fundamental Group
Let X be a topological space, x
0
its point. A path in X which starts
and ends at x
0
is called a loop in X at x
0
. Denote by Ω(X, x
0
) the set
of loops in X at x
0
. Denote by π
1
(X, x
0
) the set of homotopy classes of
loops in X at x
0
.
Both Ω(X, x
0
) and π
1
(X, x
0
) are equipped with a multiplication.
28.A. For any topological space X and a point x
0
∈ X the set π
1
(X, x
0
)
of homotopy classes of loops at x
0
with multiplication deﬁned above in
Section '27 is a group.
π
1
(X, x
0
) is called the fundamental group of the space X with base
point x
0
. It was introduced by Poincar´e and that is why it is called also
Poincar´e group. The letter π in its notation is also due to Poincar´e.
'28
◦
2 Why Index 1?
The index 1 in the notation π
1
(X, x
0
) appeared later than the let
ter π. It is related to one more name of the fundamental group: the
ﬁrst (or onedimensional) homotopy group. There is an inﬁnite series of
groups π
r
(X, x
0
) with r = 1, 2, 3, . . . and the fundamental group is one of
them. The higherdimensional homotopy groups were deﬁned by Witold
Hurewicz in 1935, thirty years after the fundamental group was deﬁned.
There is even a zerodimensional homotopy group π
0
(X, x
0
), but it is
not a group, as a rule. It is the set of pathwise connected components
of X. Although there is no natural multiplication in π
0
(X, x
0
) , unless
X is equipped with some special additional structures, there is a natural
unit in π
0
(X, x
0
). This is the component containing x
0
.
Roughly speaking, the general deﬁnition of π
r
(X, x
0
) is obtained from
the deﬁnition of π
1
(X, x
0
) by replacing I with the cube I
r
.
28.B. Enigma. How to generalize problems of this section in such a
way that in each of them I would be replaced by I
r
?
'28
◦
3 High Homotopy Groups
Let X be a topological space and x
0
its point. A continuous map I
r
→X
which maps the boundary ∂I
r
of I
r
to x
0
is called a spheroid of dimension r
of X at x
0
. Two rdimensional spheroids are said to be homotopic, if they
are ∂I
r
homotopic. For spheroids u, v of X at x
0
of dimension r ≥ 1 deﬁne
their product uv by formula
uv(t
1
, t
2
, . . . , t
r
) =
u(2t
1
, t
2
, . . . , t
r
), if t
1
≤ 1/2
v(2t
1
−1, t
2
, . . . , t
r
), if t
1
≥ 1/2.
161
§28. FUNDAMENTAL GROUP 162
The set of homotopy classes of rdimensional spheroids of a space X at
x
0
is the rth (or rdimensional) homotopy group π
r
(X, x
0
) of X at x
0
. Thus,
π
r
(X, x
0
) = π(I
r
, ∂I
r
; X, x
0
).
Multiplication of spheroids induces multiplication in π
r
(X, x
0
), which makes
π
r
(X, x
0
) a group.
28.1. For any X and x
0
the group π
r
(X, x
0
) with r ≥ 2 is Abelian.
28.2. Enigma. For any X, x
0
and r ≥ 2 present group π
r
(X, x
0
) as the
fundamental group of some space.
'28
◦
4 Circular loops
Let X be a topological space, x
0
its point. A continuous map l :
S
1
→ X such that
2
l(1) = x
0
is called a (circular) loop at x
0
. Assign
to each circular loop l the composition of l with the exponential map
I →S
1
: t →e
2πit
. This is a usual loop at the same point.
28.C. Prove that any loop can be obtained in this way from a circular
loop.
Circular loops l
1
, l
2
are said to be homotopic if they are 1homotopic.
Homotopy of a circular loop not ﬁxed at x
0
is called a free homotopy.
28.D. Prove that circular loops are homotopic, iﬀ the corresponding
loops are homotopic.
28.3. What kind of homotopy of loops corresponds to free homotopy of
circular loops?
28.4. Describe the operation with circular loops corresponding to the mul
tiplication of paths.
28.5. Let U and V be circular loops with common base point U(1) = V (1)
which correspond to loops u and v. Prove that circular loop deﬁned by
formula
z →
U(z
2
), if Im(z) ≥ 0
V (z
2
), if Im(z) ≤ 0
corresponds to the product of u and v.
28.6. Outline a construction of fundamental group based on circular loops.
Similarly, highdimensional homotopy groups can be constructed not out
of homotopy classes of maps (I
r
, ∂I
r
) →(X, x
0
), but as
π(S
r
, (1, 0, . . . , 0); X, x
0
).
Another, also quite a popular way, is to deﬁne π
r
(X, x
0
) as
π(D
r
, ∂D
r
; X, x
0
).
28.7. Establish natural bijections
π(I
r
, ∂I
r
; X, x
0
) → π(D
r
, ∂D
r
; X, x
0
) →π(S
r
, (1, 0, . . . , 0); X, x
0
)
2
Recall that S
1
is considered as a subset of the plane R
2
and the latter is identiﬁed
in a canonical way with C. Hence 1 ∈ S
1
¦z ∈ C : [z[ = 1¦.
§28. FUNDAMENTAL GROUP 163
'28
◦
5 The Very First Calculations
28.E. Prove that π
1
(R
n
, 0) is a trivial group (i.e., consists of one ele
ment).
28:A. What about π
r
(R
n
, 0)?
28.F. Generalize 28.E to the situations suggested by 26.H and 26.3.
28.8. Calculate the fundamental group of an indiscrete space.
28.9. Calculate the fundamental group of the quotient space of disk D
2
obtained by identiﬁcation of each x ∈ D
2
with −x.
28.G. Prove that π
1
(S
n
, (1, 0, . . . , 0)) with n ≥ 2 is a trivial group.
Whether you have solved 28.G or not, we would recommend you con
sider problems 28.G.1, 28.G.3, 28.G.4, 28.G.5 and 28.G.6 designed to give
an approach to 28.G, warn about a natural mistake and prepare an impor
tant tool for further calculations of fundamental groups.
28.G.1. Prove that any loop s : I → S
n
, which does not ﬁll the whole S
n
(i.e., s(I) = S
n
) is homotopic to the constant loop, provided n ≥ 2. (Cf.
Problem 26.6.)
Warning: for any n there exists a loop ﬁlling S
n
. See 9:M.
28.G.2. Is a loop ﬁlling S
2
homotopic to the constant loop?
28.G.3 Corollary of Lebesgue Lemma 15.V. Let s : I →X be a path,
and Γ be an open covering of a topological space X. There exists a sequence
of points a
1
, . . . , a
N
∈ I with 0 = a
1
< a
2
< < a
N−1
< a
N
= 1 such
that s([a
i
, a
i+1
]) is contained in an element of Γ for each i.
28.G.4. Prove that if n ≥ 2 then for any path s : I → S
n
there exists a
subdivision of I into a ﬁnite number of subintervals such that the restriction
of s to each of the subintervals is homotopic, via a homotopy ﬁxed on the
endpoints of the subinterval, to a map with nowhere dense image.
28.G.5. Prove that if n ≥ 2 then any loop in S
n
is homotopic to a loop
which is not surjective.
28.G.6. Deduce 28.G from 28.G.1 and 28.G.5. Find all the points of the
proof of 28.G obtained in this way, where the condition n ≥ 2 is used.
'28
◦
6 Fundamental Group of Product
28.H. The fundamental group of the product of topological spaces is
canonically isomorphic to the product of the fundamental groups of the
factors:
π
1
(X Y, (x
0
, y
0
)) = π
1
(X, x
0
) π
1
(Y, y
0
)
28.10. Prove that π
1
(R
n
0, (1, 0, . . . , 0)) is trivial if n ≥ 3
28:B. Prove the following generalization of 28.H:
π
r
(X Y, (x
0
, y
0
)) = π
r
(X, x
0
) π
r
(Y, y
0
).
§28. FUNDAMENTAL GROUP 164
'28
◦
7 SimplyConnectedness
A nonempty topological space X is said to be simply connected or
oneconnected, if it is pathconnected and any loop in it is homotopic to
a constant map.
28.I. For a pathconnected topological space X the following statements
are equivalent:
(a) X is simply connected,
(b) any continuous map f : S
1
→X is (freely) homotopic to a constant
map,
(c) any continuous map f : S
1
→ X can be extended to a continuous
map D
2
→X,
(d) any two paths s
1
, s
2
: I → X connecting the same points x
0
and x
1
are homotopic.
The following theorem implies Theorem 28.I. However, since it treats
a single loop, it can be applied to more situations. Anyway, proving 28.I,
one proves 28.J in fact.
28.J. Let X be a topological space and s : S
1
→ X be a circular loop.
Then the following statements are equivalent:
(a) s is homotopic to the constant loop,
(b) s is freely homotopic to a constant map,
(c) s can be extended to a continuous map D
2
→X,
(d) the paths s
+
, s
−
: I → X deﬁned by formula s
ε
(t) = s(e
επit
) are
homotopic.
28.J.1. Enigma. To prove that 4 statements are equivalent, one has to
prove at least 4 implications. What implications would you choose for the
easiest proof of Theorem 28.J?
28.J.2. Does homotopy of circular loops imply that these circular loops are
free homotopic?
28.J.3. A homotopy between a map of the circle and a constant map pos
sesses a quotient map whose source space is homoeomorphic to disk D
2
.
28.J.4. Represent the problem of constructing of a homotopy between
paths s
+
and , s
−
as a problem of extension of a certain continuous map
of the boundary of a square to a continuous of the whole square.
28.J.5. For the sake of extension problem obtained as a result of Problem
28.J.4, does it help to know that circular loop S
1
→ X : t → s(e
2πit
) can
be extended to a continuous map of a disk?
28.11. Which of the following spaces are simply connected:
(a) a discrete
space;
(b) an indiscrete
space;
(c) R
n
;
(d) a convex set; (e) a star convex set; (f) S
n
;
(g) R
n
0?
§28. FUNDAMENTAL GROUP 165
28.12. Prove that a topological space X, which is presented as the union of
open simply connected sets U and V with simply connected U ∩V , is simply
connected.
28.13. Show that the assumption that U and V are open is necessary in
28.12.
28.14*. Let X be a topological space, U and V its open sets. Prove that if
U ∪V and U ∩V are simply connected, then U and V are simply connected,
too.
'28
◦
8 Fundamental Group of a Topological Group
Let G be a topological group. Given loops u, v : I → G starting at the unity
1 ∈ G, let us deﬁne a loop u v : I → G by the formula u v(t) = u(t) v(t), where
denotes the group operation in G.
28:C. Prove that the set Ω(G, 1) of all the loops in G starting at 1 equipped with
the operation is a group.
28:D. Prove that the operation on Ω(G, 1) deﬁnes a group operation on π
1
(G, 1)
and that this operation coincides with the standard group operation (deﬁned by mul
tiplication of paths).
28:D.1. For loops u, v →G starting at 1, ﬁnd (ue
1
) (e
1
v).
28:E. The fundamental group of a topological group is abelian.
28:F. Formulate and prove the analogues of Problems 28:C and 28:D for high homo
topy groups and π
0
(G, 1).
'29 The Role of Base Point
'29
◦
1 Overview of the Role of Base Point
Sometimes the choice of the base point does not matter, sometimes it
is obviously crucial, sometimes this is a delicate question. In this section
we have to clarify all the subtleties related to the base point. We begin
with preliminary formulations describing the subject in its entirety, but
without some necessary details.
Roughly, the role of base point may be described as follows:
• While the base point changes within the same pathconnected com
ponent, the fundamental group remains in the same class of isomor
phic groups.
• However, if the group is not commutative, it is impossible to ﬁnd a
natural isomorphism between fundamental groups at diﬀerent base
points even in the same pathconnected component.
• Fundamental groups of a space at base points belonging to diﬀerent
pathconnected components have nothing to do to each other.
In this section these will be demonstrated. The proof involves useful
constructions, which importance extends far outside of the frameworks
of our initial question on the role of base point.
'29
◦
2 Deﬁnition of Translation Maps
Let x
0
and x
1
be points of a topological space X, and let s be a path
connecting x
0
with x
1
. Denote by σ the homotopy class [s] of s. Deﬁne
a map T
s
: π
1
(X, x
0
) →π
1
(X, x
1
) by formula T
s
(α) = σ
−1
ασ.
29.1. Prove that for any loop a : I → X representing α ∈ π
1
(X, x
0
) and a
path s : I → X with s(0) = x
0
there exists a free homotopy H : I I → X
between a and a loop representing T
s
(α) such that H(0, t) = H(1, t) = s(t)
for t ∈ I.
29.2. Let a, b : I → X be loops which are homotopic via a homotopy H :
I I → X such that H(0, t) = H(1, t) (i.e., H is a free homotopy of loops:
at each moment t ∈ I it keeps the end points of the path coinciding). Set
s(t) = H(0, t) (hence s is the path run over by the initial point of the loop
under the homotopy). Prove that the homotopy class of b is the image of the
homotopy class of a under T
s
: π
1
(X, s(0)) → π
1
(X, s(1)).
'29
◦
3 Properties of T
s
29.A. T
s
is a (group) homomorphism. (Recall that this means that
T
s
(αβ) = T
s
(α)T
s
(β).)
166
§29. THE ROLE OF BASE POINT 167
29.B. If u is a path connecting x
0
to x
1
and v is a path connecting x
1
with x
2
then T
uv
= T
v
◦ T
u
. In other words the diagram
π
1
(X, x
0
)
Tu
−−−→ π
1
(X, x
1
)
` Tuv
Tv
π
1
(X, x
2
)
is commutative.
29.C. If paths u and v are homotopic then T
u
= T
v
.
29.D. T
ea
= id : π
1
(X, a) →π
1
(X, a)
29.E. T
s
−1 = T
−1
s
.
29.F. T
s
is an isomorphism for any path s.
29.G. For any points x
0
and x
1
lying in the same pathconnected com
ponent of X groups π
1
(X, x
0
) and π
1
(X, x
1
) are isomorphic.
'29
◦
4 Role of Path
29.H. If s is a loop representing an element σ of fundamental group
π
1
(X, x
0
) then T
s
is the internal automorphism of π
1
(X, x
0
) deﬁned by
α →σ
−1
ασ.
29.I. Let x
0
and x
1
be points of a topological space X belonging to
the same pathconnected component. Isomorphisms T
s
: π
1
(X, x
0
) →
π
1
(X, x
1
) do not depend on s, iﬀ π
1
(X, x
0
) is commutative.
'29
◦
5 High Homotopy Groups
29.3. Enigma. Guess how T
s
is generalized to π
r
(X, x
0
) with any r.
Here is another form of the same question. We put it since it contains in
its statement a greater piece of an answer.
29.4. Enigma. Given a path s : I → X with s(0) = x
0
and a spheroid
f : I
r
→X at x
0
, how to cook up a spheroid at x
1
= s(1) out of these?
29.5. Prove that for any path s : I → X and a spheroid f : I
r
→ X with
f(Fr I
r
) = s(0) there exists a homotopy H : I
r
I → X of f such that
H(Fr I
r
t) = s(t) for any t ∈ I and that the spheroid obtained by such a
homotopy is unique up to homotopy and deﬁnes an element of π
r
(X, s(1))
welldeﬁned by the homotopy class of s and the element of π
r
(X, s(0)) rep
resented by f.
Of course, a solution of 29.5 gives an answer to 29.4 and 29.3. The map
π
r
(X, s(0)) → π
r
(X, s(1)) deﬁned by 29.5 is denoted by T
s
. By 29.2 this T
s
generalizes T
s
deﬁned in the beginning of the section for the case r = 1.
29.6. Prove that the properties of T
s
formulated in Problems 29.A – 29.G
hold true in all dimensions.
§29. THE ROLE OF BASE POINT 168
'29
◦
6 In Topological Group
In a topological group G there is another way to relate π
1
(G, x
0
) with π
1
(G, x
1
):
there are homeomorphisms L
g
: G → G : x → xg and R
g
: G → G : x → gx,
so that there are the induced isomorphisms (L
x
−1
0
x1
)
∗
: π
1
(G, x
0
) → π
1
(G, x
1
) and
(R
x1x
−1
0
)
∗
: π
1
(G, x
0
) →π
1
(G, x
1
).
29:A. Let G be a topological group, s I →G be a path. Prove that
T
s
= (L
s(0)
−1
s(1)
)
∗
= (R
s(1)s(0)
−1 ) : π
1
(G, s(0)) →π
1
(G, s(1)).
29:B. Deduce from 29:A that the fundamental group of a topological group is abelian
(cf. 28:E).
29:1. Prove that the fundamental groups of the following spaces are com
mutative:
(a) the space of nondegenerate real nn matrices GL(n, R) = ¦A [ det A =
0¦;
(b) the space of orthogonal real nn matrices O(n, R) = ¦A [ A (
t
A) = 1¦;
(c) the space of special unitary complex n n matrices SU(n) = ¦A [
A (
t
¯
A) = 1, det A = 1¦
(d) RP
n
;
(e) V
k,n
= Hom(R
k
, R
n
);
29:C. Generalize 29:A and 29:B to a homogeneous space G/H.
29:D. Enigma. What are the counterparts for 29:A and 29:B and 29:C for high
homotopy groups?
CHAPTER 6
Covering Spaces and Calculation of Fundamental
Groups
'30 Covering Spaces
'30
◦
1 Deﬁnition of Covering
Let X, B topological spaces, p : X → B a continuous map. Assume
that p is surjective and each point of B possesses a neighborhood U such
that the preimage p
−1
(U) of U is presented as a disjoint union of open
sets V
α
and p maps each V
α
homeomorphically onto U. Then p : X →B
is called a covering, (of the space B), the space B is called the base of
this covering, X is called the covering space for B and the total space of
the covering. Neighborhoods like U are said to be trivially covered. The
map p is called also a covering map, or a covering projection.
30.A. Let B be a topological space and F be a discrete space. Prove
that the projection pr
B
: B F →B is a covering.
The following statement shows that in a sense locally any covering is
organized as the covering of 30.A.
30.B. A continuous surjective map p : X →B is a covering, iﬀ for each
point a of B the preimage p
−1
(a) is discrete and there exist a neighbor
hood U of a and a homeomorphism h : p
−1
(U) → U p
−1
(a) such that
p[
p
−1
(U)
= pr
U
◦ h.
However, the coverings of 30.A are not interesting. They are said to
be trivial. Here is the ﬁrst really interesting example.
30.C. Prove that R →S
1
: x →e
2πix
is a covering.
To distinguish the most interesting examples, a covering with a con
nected total space is called a covering in narrow sense. Of course, the
covering of 30.C is a covering in a narrow sense.
'30
◦
2 More Examples
30.D. R
2
→S
1
R : (x, y) →(e
2πix
, y) is a covering.
30.E. C →C 0 : z →e
z
is a covering.
30.1. Enigma. In what sense the coverings of 30.D and 30.E are the same?
Deﬁne an appropriate equivalence relation for coverings.
169
§30. COVERING SPACES 170
30.F. R
2
→S
1
S
1
: (x, y) →(e
2πix
, e
2πiy
) is a covering.
30.G. For any natural n the map S
1
→S
1
: z →z
n
is a covering.
30.2. Prove that for any natural n the map C 0 → C 0 : z → z
n
is a
covering.
30.H. For any natural p and q the map S
1
S
1
→ S
1
S
1
: (z, w) →
(z
p
, w
q
) is a covering.
30.3. Prove that if p : X → B and p
: X
→ B
are coverings, then
p p
: X X
→B B
is a covering.
30.I. The natural projection S
n
→RP
n
is a covering.
30.J. Is (0, 3) →S
1
: x →e
2πix
a covering? (Cf. 30.14.)
30.K. Is the projection R
2
→ R : (x, y) → x a covering? Indeed, why
not take an open interval (a, b) ⊂ R as a trivially covered neighborhood:
its preimage (a, b) R is the union of open intervals (a, b) ¦y¦ which
are projected homeomorphically by the projection (x, y) →x onto (a, b)?
30.4. Find coverings of M¨obius strip by cylinder.
30.5. Find nontrivial coverings of M¨obius strip by itself.
30.6. Find a covering of the Klein bottle by torus. Cf. Problem 20.14.
30.7. Find coverings of the Klein bottle by plane R
2
, cylinder S
1
R, and
a nontrivial covering by itself.
30.8. Construct a covering of the Klein bottle by R
2
. Describe explicitly the
partition of R
2
into preimages of points under this covering.
30.9. Find a covering of a sphere with any number of crosscaps by a sphere
with handles.
'30
◦
3 Local homeomorphisms versus coverings
30.10. Any covering is an open map.
1
A map f : X → Y is said to be locally homeomorphic if each point of X
has a neighborhood U such that the image f(U) is open in Y and the map
U →f(U) deﬁned by f is a homeomorphism.
30.11. Any covering is a locally homeomorphic map.
30.12. Show that there exists a locally homeomorphic map which is not a
covering.
30.13. Prove that a restriction of a locally homeomorphic map to an open
set is locally homeomorphic.
30.14. For which subsets of R is the restriction of the map of Problem 30.C
a covering.
30.15. Find a nontrivial covering X → B with X homeomorphic to B and
prove that it satisﬁes the deﬁnition of covering.
1
Remind that a map is said to be open if the image of any open set is open.
§30. COVERING SPACES 171
'30
◦
4 Number of Sheets
Let p : X →B be a covering. The cardinality (i.e., number of points)
of the preimage p
−1
(a) of a point a ∈ B is called the multiplicity of the
covering at a or the number of sheets of the covering over a.
30.L. If the base of a covering is connected then the multiplicity of the
covering at a point does not depend on the point.
In the case of covering with connected base the multiplicity is called
the number of sheets of the covering. If the number of sheets is n then
the covering is said to be nsheeted and we talk about nfold covering.
Of course, unless the covering is trivial, it is impossible to distinguish
the sheets of it, but this does not prevent us from speaking about the
number of sheets.
30.16. What are the numbers of sheets for the coverings from Section '30
◦
2?
30.17. What numbers can you realize as the number of sheets of a covering
of the M¨obius strip by the cylinder S
1
I?
30.18. What numbers can you realize as the number of sheets of a covering
of the M¨obius strip by itself?
30.19. What numbers can you realize as the number of sheets of a covering
of the Klein bottle by torus?
30.20. What numbers can you realize as the number of sheets of a covering
of the Klein bottle by itself?
30.21. What numbers can you realize as the numbers of sheets for a covering
of the Klein bottle by plane R
2
?
30.22. What numbers can you realize as the numbers of sheets for a covering
of the Klein bottle by S
1
R?
30.23. Construct a dfold covering of a sphere with p handles by a sphere
with 1 +d(p −1) handles.
30.24. Let p : X → Y and q : Y → Z be coverings. Prove that if q has
ﬁnitely many sheets then q ◦ p : x →Y is a covering.
30.25*. Is the hypothesis of ﬁniteness of the number of sheets in Problem
30.24 necesary?
'30
◦
5 Universal Coverings
A covering p : X →B is said to be universal if X is simply connected.
The appearance of word universal in this context will be explained below
in Section '36.
30.M. Which coverings of the problems stated above in this section are
universal?
'31 Theorems on Path Lifting
'31
◦
1 Lifting
Let p : X →B and f : A →B be arbitrary maps. A map g : A →X
such that p ◦ g = f is said to cover f or be a lifting of f. A lot of topo
logical problems can be phrased in terms of ﬁnding a continuous lifting
of some continuous map. Problems of this sort are called lifting problems.
They may involve additional requirements. For example, the desired
lifting has to coincide with a lifting already given on some subspace.
31.A. The identity map S
1
→ S
1
does not admit a continuous lifting
with respect to the covering R → S
1
: x → e
2πix
. (In other words, there
exists no continuous map g : S
1
→R such that e
2πig(x)
= x for x ∈ S
1
.)
'31
◦
2 Path Lifting
31.B Path Lifting Theorem. Let p : X → B be a covering, x
0
∈ X,
b
0
∈ B be points such that p(x
0
) = b
0
. Then for any path s : I → B
starting at b
0
there exists a unique path ˜ s : I → X starting at x
0
and
being a lifting of s. (In other words, there exists a unique path ˜ s : I →X
with ˜ s(0) = x
0
and p ◦ ˜ s = s.)
31.B.1 Lemma 1. Let p : X → B be a trivial covering. Then for any
continuous map f of any space A to B there exists a continuous lifting
˜
f : A →X.
31.B.2 Lemma 2. Let p : X →B be a trivial covering and x
0
∈ X, b
0
∈ B
be points such that p(x
0
) = b
0
. Then for any continuous map f of a space
A to B mapping a point a
0
to b
0
, a continuous lifting
˜
f : A → X with
˜
f(a
0
) = x
0
is unique.
31.B.3 Lemma 3.
2
Let p : X →B be a covering, A a connected space. If
f, g : A →X are continuous maps coinciding in some point and p◦f = p◦g,
then f = g.
31.1. If in the Problem 31.B.2 one replaces x
0
, b
0
and a
0
by pairs of points,
then it may happen that the lifting problem has no solution
˜
f with
˜
f(a
0
) = x
0
.
Formulate a condition necessary and suﬃcient for existence of such a solution.
31.2. What goes wrong with the Path Lifting Theorem 31.B for the local
homeomorphism of Problem 30.J?
31.3. Consider the covering C → C 0 : z → e
z
. Find liftings of the paths
u(t) = 2 −t, v(t) = (1 +t)e
2πit
, and their product uv.
31.4. Prove that any covering p : X →B with simply connected B and path
connected X is a homeomorphism.
2
This is rather a generalization of the uniqueness, than a necessary step of the
proof. But a good lemma is to clarify the real content of the proof, and a generalization
is one of the best ways to do this.
172
§31. THEOREMS ON PATH LIFTING 173
'31
◦
3 Homotopy Lifting
31.C Path Homotopy Lifting Theorem. Let p : X → B be a cov
ering, x
0
∈ X, b
0
∈ B be points such that p(x
0
) = b
0
. Let u, v : I →B be
paths starting at b
0
and ˜ u, ˜ v : I →X be the lifting paths for u, v starting
at x
0
. If the paths u and v are homotopic then the covering paths ˜ u and
˜ v are homotopic.
31.D Corollary. Under the assumptions of Theorem 31.C, the covering
paths ˜ u and ˜ v have the same ﬁnal point (i.e., ˜ u(1) = ˜ v(1)).
Notice that in 31.C and 31.D paths are assumed to share the initial
point x
0
. In the statement of 31.D we emphasize that then they share
also the ﬁnal point.
31.E Corollary of 31.D. Let p : X →B be a covering and s : I →B
be a loop. If there exists a lifting ˜ s : I → X of s with ˜ s(0) = ˜ s(1) (i.e.,
there exists a covering path which is not a loop), then s is not homotopic
to a constant loop.
31.5. Prove that if a pathwise connected space B has a non trivial pathwise
connected covering space, then the fundamental group of B is not trivial.
31.6. What corollaries can you deduce from 31.5 and the examples of cov
erings presented above in Section '30?
31:A. Enigma. Is it that important in the hypothesis of Theorem 30.6 that u and
v are paths? To what class of maps can you generalize this therorem?
'31
◦
4 HighDimensional Homotopy Groups of Covering Space
31:B. Let p : X → B be a covering. Then for any continuous map s : I
n
→ B
and a lifting u : I
n−1
→ X of the restriction s[
I
n−1 there exists a unique lifting of s
extending u.
31:C. For any covering p : X → B and points x
0
∈ X, b
0
∈ B such that p(x
0
) = b
0
the homotopy groups π
r
(X, x
0
) and π
r
(B, b
0
) with r > 1 are canonically isomorphic.
31:D. Prove that homotopy groups of dimensions greater than 1 of circle, torus,
Klein bottle and M¨obius strip are trivial.
'32 Calculations of Fundamental Groups Using
Universal Coverings
'32
◦
1 Fundamental Group of Circle
For an integer n denote by s
n
the loop in S
1
deﬁned by formula
s
n
(t) = e
2πint
. The initial point of this loop is 1. Denote the homotopy
class of s
1
by α. Thus α ∈ π
1
(S
1
, 1).
32.A. The loop s
n
represents α
n
∈ π
1
(S
1
, 1).
32.B. Find the paths in R starting at 0 ∈ R and covering the loops s
n
with respect to the universal covering R →S
1
?
32.C. The homomorphism Z →π
1
(S
1
, 1) deﬁned by formula n →α
n
is
an isomorphism.
32.C.1. Formula n →α
n
deﬁnes a homomorphism Z →π
1
(S
1
, 1).
32.C.2. Rephrase the statement that the homomorphism of Theorem 32.C
is surjective in terms of loops and loop homotopies.
32.C.3. Prove that a loop s : I → S
1
starting at 1 is homotopic to s
n
if
the path ˜ s : I →R covering s and starting at 0 ∈ R ﬁnishes at n ∈ R (i.e.,
˜ s(1) = n).
32.C.4. Rephrase the statement that the homomorphism of Theorem 32.C
is injective in terms of loops and loop homotopies.
32.C.5. Prove that if loop s
n
is homotopic to constant then n = 0.
32.1. What is the image under the isomorphism of Theorem 32.C of the
homotopy class of loop t →e
2πit
2
?
Denote by deg the isomorphism inverse to the isomorphism of Theorem
32.C.
32.2. For any loop s : I → S
1
starting at 1 ∈ S
1
, the integer deg([s])
coincides with the ﬁnal point of the path starting at 0 ∈ R and covering s.
32.D Corollary of Theorem 32.C. The fundamental group of (S
1
)
n
is a free abelian group of rank n (i.e., isomorphic to Z
n
).
32.E. On torus S
1
S
1
ﬁnd two loops whose homotopy classes generate
the fundamental group of the torus.
32.F Corollary of Theorem 32.C. The fundamental group of punc
tured plane R
2
0 is an inﬁnite cyclic group.
32.3. Solve Problems 32.D – 32.F without reference to Theorems 32.C and
28.H, but using explicit constructions of the corresponding universal cover
ings.
174
§32. CALCULATIONS OF FUNDAMENTAL GROUPS 175
'32
◦
2 Fundamental Group of Projective Space
The fundamental group of the projective line is an inﬁnite cyclic
group. It is calculated in the previous subsection, since the projective
line is a circle. The zerodimensional projective space is a point, hence
its fundamental group is trivial. Calculate now the fundamental groups
of projective spaces of all the other dimensions.
Let n ≥ 2 and l : I → RP
n
be a loop covered by a path
˜
l : I → S
n
which connects two antipodal points, say the poles P
+
= (1, 0, . . . , 0) and
P
−
= (−1, 0, . . . , 0), of S
n
. Denote by λ the homotopy class of l. It is an
element of π
1
(RP
n
, (1 : 0 : : 0)).
32.G. For any n ≥ 2 group π
1
(RP
n
, (1 : 0 : : 0)) is a cyclic group of
order 2. It consists of two elements: λ and 1.
32.G.1 Lemma. Any loop in RP
n
at (1 : 0 : : 0) is homotopic either
to l or constant. This depends on whether the covering path of the loop
connects the poles P
+
and P
−
, or is a loop.
32.4. Where in the proofs of Theorem 32.G and Lemma 32.G.1 the assump
tion n ≥ 2 is used?
'32
◦
3 Fundamental Groups of Bouquet of Circles
Consider a family of topological spaces ¦X
α
¦. In each of the spaces let
a point x
α
be marked. Take the sum H
α
X
α
and identify all the marked
points. The resulting quotient space is called the bouquet of ¦X
α
¦ and
denoted by ∨
α
X
α
. Hence bouquet of q circles is a space which is a union
of q copies of circle. The copies meet in a single common point, and this
is the only common point for any two of them. The common point is
called the center of the bouquet.
Denote the bouquet of q circles by B
q
and its center by c. Let u
1
, . . . ,
u
q
be loops in B
q
starting at c and parametrizing the q copies of circle
comprising B
q
. Denote the homotopy class of u
i
by α
i
.
32.H. π
1
(B
q
, c) is a free group freely generated by α
1
, . . . , α
q
.
'32
◦
4 Algebraic Digression. Free Groups
Recall that a group G is a free group freely generated by its elements
a
1
, . . . , a
q
if:
• each its element x ∈ G can be expressed as a product of powers
(with positive or negative integer exponents) of a
1
, . . . , a
q
, i.e.,
x = a
e
1
i
1
a
e
2
i
2
. . . a
en
in
and
§32. CALCULATIONS OF FUNDAMENTAL GROUPS 176
• this expression is unique up to the following trivial ambiguity: one
may insert or delete factors a
i
a
−1
i
and a
−1
i
a
i
or replace a
m
i
by a
r
i
a
s
i
with r +s = m.
32.I. A free group is deﬁned up to isomorphism by the number of its free
generators.
The number of free generators is called the rank of the free group.
For a standard representative of the isomorphism class of free groups of
rank q one can take the group of words in alphabet of q letters a
1
, . . . , a
q
and their inverses a
−1
1
, . . . , a
−1
q
. Two words represent the same element
of the group, iﬀ they can be obtained from each other by a sequence
of insertions or deletions of fragments a
i
a
−1
i
and a
−1
i
a
i
. This group is
denoted by F(a
1
, . . . , a
q
), or just F
q
, when the notations for the generators
are not to be emphasized.
32.J. Each element of F(a
1
, . . . , a
q
) has a unique shortest representative.
This is a word without fragments that could have been deleted.
The number of letters in the shortest representative of x ∈ F(a
1
, . . . , a
q
)
is called the length of x and denoted by l(x). Of course, this number is
not well deﬁned, unless the generators are ﬁxed.
32.5. Show that an automorphism of F
q
can map x ∈ F
q
to an element with
diﬀerent length. For what value of q does such an example not exist? Is it
possible to change the length in this way arbitrarily?
32.K. A group G is a free group freely generated by its elements a
1
, . . . ,
a
q
if and only if every map of the set ¦a
1
, . . . , a
q
¦ to any group X can be
extended to a unique homomorphism G →X.
Sometimes Theorem 32.K is taken as a deﬁnition of free group. (A
deﬁnition of this sort emphasizes relations among diﬀerent groups, rather
than the internal structure of a single group. Of course, relations among
groups can tell everything about internal aﬀairs of each group.)
Now we can reformulate Theorem 32.H as follows:
32.L. The homomorphism
F(a
1
, . . . , a
q
) →π
1
(B
q
, c)
taking a
i
to α
i
for i = 1, . . . , q is an isomorphism.
First, for the sake of simplicity let us agree to restrict ourselves to the
case of q = 2. It would allow us to avoid superﬂuous complications in
notations and pictures. This is the simplest case, which really represents
the general situation. The case q = 1 is too special.
To take advantages of this, let us change notations. Put B = B
2
,
u = u
1
, v = u
2
, α = α
1
, β = α
2
.
Now Theorem 32.L looks as follows:
§32. CALCULATIONS OF FUNDAMENTAL GROUPS 177
The homomorphism F(a, b) →π(B, c) taking a to α and b to β is an
isomorphism.
This theorem can be proved like Theorems 32.C and 32.G, provided
the universal covering of B is known.
'32
◦
5 Universal Covering for Bouquet of Circles
Denote by U and V the points antipodal to c on the circles of B. Cut
B at these points, removing U and V and putting instead each of them
two new points. Whatever this operation is, its result is a cross K, which
is the union of four closed segments with a common end point c. There
appears a natural map P : K →B, which takes the center c of the cross
to the center c of B and maps homeomorphically the rays of the cross
onto halfcircles of B. Since the circles of B are parametrized by loops
u and v, the halves of each of the circles are ordered: the corresponding
loop passes ﬁrst one of the halves and then the other one. Denote by U
+
the point of P
−1
(U), which belongs to the ray mapped by P onto the
second half of the circle, and by U
−
the other point of P
−1
(U). Similarly
denote points of P
−1
(V ) by V
+
and V
−
.
The restriction of P to K ¦U
+
, U
−
, V
+
, V
−
¦ maps this set home
omorphically onto B ¦U, V ¦. Therefore P provides a covering of
B ¦U, V ¦. But it fails to be a covering at U and V : each of this
points has no trivially covered neighborhood. Moreover, the preimage
of each of these points consists of 2 points (the end points of the cross),
where P is not even a local homeomorphism. To recover, we may attach
a copy of K at each of the 4 end points of K and extend P in a natural
way to the result. But then new 12 end points, where the map is not
a local homeomorphism, appear. Well, we repeat the trick and recover
the property of being a local homeomorphism at each of the new 12 end
points. Then we have to do this at each of the new 36 points, etc. But
if we repeat this inﬁnitely many times, all the bad points are turned to
nice ones.
3
3
This sounds like a story about a battle with Medusa Gorgon, but the happy
ending demonstrates that modern mathematicians have a magic power of the sort
that the heros of myths and tales could not dream of. Indeed, we meet Medusa K
with 4 heads, cut oﬀ all the heads, but, according to the old tradition of the genre,
3 new heads appear in place of each of the original heads. We cut them oﬀ, and the
story repeats. We do not even try to prevent this multiplication of heads. We just
cut them oﬀ. But contrary to the real heros of tales, we act outside of Time and
hence have no time limitations. Thus after inﬁnite repetitions of the exercise with an
exponentially growing number of heads we succeed! No heads left!
This is a typical success story about an inﬁnite construction in mathematics.
Sometimes, as in our case, such a construction can be replaced by a ﬁnite one, but
which deals with inﬁnite objects. However, there are important constructions, in
which an inﬁnite fragment is unavoidable.
§32. CALCULATIONS OF FUNDAMENTAL GROUPS 178
32.M. Formalize the construction of a covering for B described above.
Consider F(a, b) as a discrete topological space. Take K F(a, b). It
can be thought of as a collection of copies of K enumerated by elements
of F(a, b). Topologically this is a disjoint sum of the copies, since F(a, b)
is equipped with discrete topology. In KF(a, b) identify points (U
−
, g)
with (U
+
, ga) and (V
−
, g) with (V
+
, gb) for each g ∈ F(a, b). Denote the
resulting quotient space by X.
32.N. The composition of the natural projection K F(a, b) →K and
P : K →B deﬁnes a continuous quotient map p : X →B.
32.O. p : X →B is a covering.
32.P. X is pathconnected. For any g ∈ F(a, b) there exists a path
connecting (c, 1) with (c, g) and covering loop obtained from g by sub
stituting a by u and b by v.
32.Q. X is simply connected.
'32
◦
6 Fundamental groups of some ﬁnite
topological spaces
32.6. Consider a topological space X which consists of points a, b, c and d
with topology deﬁned by base formed of sets ¦a¦, ¦c¦, ¦a, b, c¦ and ¦c, d, a¦.
Construct an inﬁnitesheeted covering (in the narrow sense) of X. Calculate
π
1
(X).
32.7. Find a nonsimplyconnected topological space with ﬁnitely many points.
32.8. What is the smallest possible number of points of a topological space
with a nontrivial fundamental group?
32.9. What groups can be realized as the fundamental group of a ﬁnite topo
logical space with the minimal number of points for which the fundamental
group can be nontrivial?
32.10. Let X be a topological space, which can be presented as a union of
open connected sets U and V . Prove that if U ∩ V is disconnected then X
has a connected inﬁnitefold covering space.
32.11. Find a ﬁnite topological space with a nonabelian fundamental group.
What is the minimal number of points in such a space?
32.12. Prove that if under conditions of 32.10 U∩V contains at least three
connected components then the fundamental group of X is nonabelian and,
moreover, it can be mapped epimorphically onto a free group of rank 2.
32.13*. Find a ﬁnite topological space with fundamental group Z
2
.
CHAPTER 7
Fundamental Group and Mappings
'33 Induced Homomorphisms
and Their First Applications
'33
◦
1 Homomorphisms Induced by a Continuous Map
Let f : X → Y be a continuous map of a topological space X to
a topological space Y . Let x
0
∈ X and y
0
∈ Y be points such that
f(x
0
) = y
0
. The latter property of f is expressed by saying that f maps
pair (X, x
0
) to pair (Y, y
0
) and writing f : (X, x
0
) →(Y, y
0
).
Denote by f
the map Ω(X, x
0
) →Ω(Y, y
0
) deﬁned by formula f
(s) =
f ◦ s. This map assigns to a loop its composition with f.
33.A. f
maps homotopic loops to homotopic loops.
Therefore f
induces a map π
1
(X, x
0
) → π
1
(Y, y
0
). The latter is
denoted by f
∗
.
33.B. f
∗
: π(X, x
0
) → π
1
(Y, y
0
) is a homomorphism for any continuous
map f : (X, x
0
) →(Y, y
0
).
f
∗
: π(X, x
0
) →π
1
(Y, y
0
) is called a homomorphism induced by f.
33.C. Let f : (X, x
0
) →(Y, y
0
) and g : (Y, y
0
) →(Z, z
0
) be (continuous)
maps. Then
(g ◦ f)
∗
= g
∗
◦ f
∗
: π
1
(X, x
0
) →π
1
(Z, z
0
).
33.D. Let f, g : (X, x
0
) → (Y, y
0
) be continuous maps homotopic via a
homotopy ﬁxed at x
0
. Then f
∗
= g
∗
.
33.E. Enigma. How to generalize Theorem 33.D to the case of freely
homotopic f and g?
33.F. Let f : X → Y be a continuous map, x
0
and x
1
points of X
connected by a path s : I → X. Denote f(x
0
) by y
0
and f(x
1
) by y
1
.
Then the diagram
π
1
(X, x
0
)
f∗
−−−→ π
1
(Y, y
0
)
Ts
T
f◦s
π
1
(X, x
1
)
f∗
−−−→ π
1
(Y, y
1
)
179
§33. INDUCED HOMOMORPHISMS AND APPLICATIONS 180
is commutative, i.e., T
f◦s
◦ f
∗
= f
∗
◦ T
s
.
33.1. Prove that the map C 0 → C 0 : z → z
3
is not homotopic to the
identity map C 0 →C 0 : z →z.
33.2. Let X be a subset of R
n
. Prove that a if a continuous map f : X →Y
is extentable to a continuous map R
n
→Y then f
∗
: π
1
(X, x
0
) →π
1
(Y, f(x
0
))
is the trivial homomorphism (i.e., maps everything to 1) for any x
0
∈ X.
33.3. Prove that a Hausdorﬀ space, which contains an open set homeomor
phic to S
1
S
1
(1, 1), has an inﬁnite noncyclic fundamental group.
33.3.1. Prove that a space X satisfying the conditions of 33.3 can be
continuously mapped to a space with inﬁnite noncyclic fundamen
tal group in such a way that the map would induce an epimorphism
of π
1
(X) onto this inﬁnite group.
33.4. Prove that the fundamental group of the space GL(n, C) of complex
n nmatrices with nonzero determinant is inﬁnite.
33.4.1. Construct continuous maps S
1
→ GL(n, C) → S
1
, whose
composition is the identity.
'33
◦
2 Fundamental Theorem of Algebra
Here our goal is to prove the following theorem, which at ﬁrst glance
has no relation to fundamental group.
33.G Fundamental Theorem of Algebra. Every polynomial of a
positive degree in one variable with complex coeﬃcients has a complex
root.
With more details:
Let p(z) = z
n
+a
1
z
n−1
+ +a
n
be a polynomial of degree n > 0 in
z with complex coeﬃcients. Then there exists a complex number w such
that p(w) = 0.
Although it is formulated in an algebraic way and called “The Fun
damental Theorem of Algebra,” it has no purely algebraic proof. Its
proofs are based either on topological arguments or use complex analy
sis. This is because the ﬁeld C of complex numbers cannot be described
in purely algebraic terms: all its descriptions involve a sort of completion
construction, cf. Section '17.
33.G.1 Reduction to Problem on a Map. Deduce Theorem 33.G from
the following statement:
For any complex polynomial p(z) of a positive degree the zero belongs
to the image of the map C → C : z → p(z). In other words, the formula
z →p(z) does not deﬁne a map C →C 0.
33.G.2 Estimate of Remainder. Let p(z) = z
n
+a
1
z
n−1
+ +a
n
be a
complex polynomial, q(z) = z
n
and r(z) = p(z) − q(z). Then there exists
a positive number R such that [r(z)[ < [q(z)[ = R
n
for any z with [z[ = R
§33. INDUCED HOMOMORPHISMS AND APPLICATIONS 181
33.G.3 Lemma on Lady with Doggy. (Cf. 26.10.) A lady q(z) and her
dog p(z) walk on punctured plane C0 periodically (i.e., say, with z ∈ S
1
).
Prove that if the lady does not let the dog to run further than by [q(z)[
from her then the doggy loop S
1
→ C 0 : z → p(z) is homotopic to the
lady loop S
1
→C 0 : z →q(z).
33.G.4 Lemma for Dummies. (Cf. 26.11.) If f : X →Y is a continuous
map and s : S
1
→ X is a loop homotopic to the trivial one then f ◦ s :
S
1
→Y is also homotopic to trivial.
'33
◦
3 Generalization of Intermediate Value Theorem
33.H. Enigma. How to generalize Intermediate Value Theorem 11.T
to the case of maps f : D
n
→R
n
?
33.I. Is Intermediate Value Theorem 11.T equivalent to the following
statement:
Let f : D
1
→ R
1
be a continuous map. If 0 ∈ f(S
0
) and submap f[ :
S
0
→R
1
¦0¦ of f induces a nonconstatn map π
0
(S
0
) →π
0
(R
1
¦0¦)
then there exists a point x ∈ D
1
such that f(x) = 0?
33.J. Enigma. Suggest a generaliation of Intermediate Value Theorem
to maps D
n
→R
n
which would generalize its reformulation 33.I.
33.K. Let f : D
n
→ R
n
be a continuous map. If f(S
n−1
) does not
contain 0 ∈ R
n
and the submap f[ : S
n−1
→ R
n
¦0¦ of f induces a
nonconstant map
π
n−1
(S
n−1
) →π
n−1
(R
n
¦0¦),
then there exists a point x ∈ D
1
such that f(x) = 0.
Usability of Theorem 33.K is impeded by a condition, which is diﬃcult
to check, if n > 0. For n = 1 this is still possible in the frameworks of the
theory developed above.
33.5. Let f : D
2
→ R
2
be a continuous map which leaves ﬁxed each point
of the bounding circle S
1
. Then f(D
2
) ⊃ D
2
.
33.6. Let f : D
2
→ R
2
be a continuous map. If f(S
1
) does not contain
a ∈ R
2
and the circular loop f[ : S
1
→R
2
a deﬁnes a nontrivial element of
π
1
(R
2
a) then there exists x ∈ D
2
such that f(x) = a.
33.7. Let f : R
2
→ R
2
be a continuous map and there exists a real number
m such that [f(x) −x[ ≤ m for any x ∈ R
2
. Prove that f is a surjection.
33.8. Let u, v : I →I I be two paths such that u(0) = (0, 0), u(1) = (1, 1)
and v(0) = (0, 1), v(1) = (1, 0). Prove that u(I) ∩ v(I) = ∅.
33.8.1. Let u, v be as in 33.8. Denote by w the map I
2
→ R
2
deﬁned by w(x, y) = u(x) − v(y). Prove that 0 ∈ R
2
is a value of
w.
33.9. Let C be a smooth simple closed curve on the plane with two inﬂection
points. Prove that there is a line intersecting C in four points a, b, c, d with
segments [a, b], [b, c] and [c, d] of the same length.
§33. INDUCED HOMOMORPHISMS AND APPLICATIONS 182
'33
◦
4 Winding Number
As we know (see 32.F), the fundamental group of the punctured plane R
2
0
is Z. There are two isomorphisms which diﬀer by multiplication by −1. We choose
the one which maps the homotopy class of the loop t → (cos 2πt, sin2πt) to 1 ∈ Z.
In terms of circular loops, the isomorphism means that to any loop f : S
1
→ R
2
0
we associate an integer. It is the number of times the loop goes arround 0 in the
counterclockwise direction.
Now we change the viewpoint in this consideration, and ﬁx the loop, but vary
the point. Let f : S
1
→R
2
be a circular loop and x ∈ R
2
f(S
1
). Then f deﬁnes an
element of π
1
(R
2
x) = Z (we choose basically the same identiﬁcation of π
1
(R
2
x)
with Z assigning 1 to the homotopy class of t →x + (cos 2πt, sin 2πt)). This number
is denoted by ind(f, x) and called winding number or index of x with respect to f.
33:A. Formula x → ind(u, x) deﬁnes a locally constant function on R
2
u(S
1
).
33:1. Let f : S
1
→ R
2
be a loop and x, y ∈ R
2
f(S
1
). Prove that if
ind(f, x) = ind(f, y) then any path connecting x and y in R
2
meets f(S
1
).
33:2. Prove that if u(S
1
) is contained in a disk, and a point x is not, then
ind(u, x) = 0.
33:3. Find the set of values of function ind : R
2
u(S
1
) →Z for the following
loops u:
a) u(z) = z; b) u(z) = ¯ z; c) u(z) = z
2
; d) u(z) = z+z
−1
+z
2
−z
−2
(here z ∈ S
1
⊂ C).
33:4. Choose several loops u : S
1
→R
2
such that u(S
1
) is a bouquet of two
circles. Calculate winding number for various points with respect to these
loops.
33:5. Find a loop f : S
1
→R
2
such that there exist points x, y ∈ R
2
f(S
1
)
with ind(f, x) = ind(f, y), but belonging to diﬀerent connected conmponents
of R
2
f(S
1
).
33:6. Prove that for any ray R radiating from x meets f(S
1
) at least in
[ ind(f, x)[ points (i. e., the number of points in f
−1
(R) is not less than
[ ind(f, x)[).
33:B. If u : S
1
→ R
2
is a restriction of a continuous map F : D
2
→ R
2
and
ind(u, x) = 0 then x ∈ F(D
2
).
33:C. If u and v are circular loops in R
2
with common base point (i. e., u(1) = v(1))
and uv is their product then ind(uv, x) = ind(u, x)+ind(v, x) for any x ∈ R
2
uv(S
1
).
33:D. Let u and v be circular loops in R
2
, and x ∈ R
2
(u(S
1
) ∪ v(S
1
)). If there
exists a (free) homotopy u
t
, t ∈ I connecting u and v such that x ∈ R
2
u
t
(S
1
) for
each t ∈ I then ind(u, x) = ind(v, x).
33:E. Let u : S
1
→C be a circular loop and a ∈ C
2
u(S
1
). Then
ind(u, a) =
1
2πi
S
1
[u(z) −a[
u(z) −a
dz.
'33
◦
5 BorsukUlam Theorem
33:F OneDimensional BorsukUlam. For each continuous map f : S
1
→ R
1
there exists x ∈ S
1
such that f(x) = f(−x).
§33. INDUCED HOMOMORPHISMS AND APPLICATIONS 183
33:G TwoDimensional BorsukUlam. For each continuous map f : S
2
→ R
2
there exists x ∈ S
2
such that f(x) = f(−x).
33:G.1 Lemma. If there exists a continuous map f : S
2
→R
2
with f(x) =
f(−x) for any x ∈ S
2
then there exists a continuous map ϕ : RP
2
→RP
1
inducing a nonzero homomorphism π
1
(RP
2
) →π
1
(RP
1
).
33:7. Prove that at each instant of time, there is a pair of antipodal points on
the earth’s surface where the pressures and also the temperatures are equal.
Theorems 33:F and 33:G are special cases of the following general theorem. We
do not assume the reader to be ready to prove Theorem 33:H in the full generality,
but is there another easy special case?
33:H BorsukUlam Theorem. For each continuous map f : S
n
→R
n
there exists
x ∈ S
n
such that f(x) = f(−x).
'34 Retractions and Fixed Points
'34
◦
1 Retractions and Retracts
A continuous map of a topological space onto a subspace is called
a retraction if the restriction of the map to the subspace is the identity
mapping. In other words, if X is a topological space, A ⊂ X then
ρ : X →A is a retraction if it is continuous and ρ[A = id
A
.
34.A. Let ρ be a continuous map of a space X onto its subspace A.
Then the following statements are equivalent:
(a) ρ is a retraction,
(b) ρ(a) = a for any a ∈ A,
(c) ρ ◦ in = id
A
,
(d) ρ : X →A is an extension of the identity mapping A →A.
A subspace A of a space X is said to be a retract of X if there exists
a retraction X →A.
34.B. Any onepoint subset is a retract.
Twopoint set may be a nonretract.
34.C. Any subset of R consisting of two points is not a retract of R.
34.1. If A is a retract of X and B is a retract of A then B is a retract of X.
34.2. If A is a retract of X and B is a retract of Y then A B is a retract
of X Y .
34.3. A closed interval [a, b] is a retract of R.
34.4. An open interval (a, b) is not a retract of R.
34.5. What topological properties of ambient space are inherited by a re
tract?
34.6. Prove that a retract of a Hausdorﬀ space is closed.
34.7. Prove that the union of Y axis and the set ¦(x, y) ∈ R
2
: x > 0, y =
sin
1
x
¦ is not a retract of R
2
and moreover is not a retract of any of its
neighborhoods.
34.D. S
0
is not a retract of D
1
.
The role of the notion of retract is clariﬁed by the following theorem.
34.E. A subset A of a topological space X is a retract of X, iﬀ any
continuous map A →Y to any space Y can be extended to a continuous
map X →Y .
'34
◦
2 Fundamental Group and Retractions
34.F. If ρ : X → A is a retraction, i : A → X is the inclusion and
x
0
∈ A, then ρ
∗
: π
1
(X, x
0
) → π
1
(A, x
0
) is an epimorphism and i
∗
:
π
1
(A, x
0
) →π
1
(X, x
0
) is a monomorphism.
184
§34. RETRACTIONS AND FIXED POINTS 185
34.G. Enigma. Which of the two statements of Theorem 34.F (about
ρ
∗
or i
∗
) is easier to use for proving that a set A ⊂ X is not a retract of
X?
34.H Borsuk Theorem in Dimension 2. S
1
is not a retract of D
2
.
34.8. Is the projective line a retract of the projective plane?
The following problem is more diﬃcult than 34.H in the sense that its
solution is not a straightforward consequence of Theorem 34.F, but rather
demands to reexamine the arguments used in proof of 34.F.
34.9. Prove that the boundary circle of M¨obius band is not a retract of
M¨obius band.
34.10. Prove that the boundary circle of a handle is not a retract of the
handle.
The Borsuk Theorem in its whole generality cannot be deduced like
Theorem 34.H from Theorem 34.F. However, it can be proven using a
generalization of 34.F to higher homotopy groups. Although we do not
assume that you can successfully prove it now relying only on the tools
provided above, we formulate it here.
34.I Borsuk Theorem. Sphere S
n−1
is not a retract of ball D
n
.
At ﬁrst glance this theorem seems to be useless. Why could it be
interesting to know that a map with a very special property of being
retraction does not exists in this situation? However in mathematics non
existence theorems may be closely related to theorems, which may seem
to be more attractive. For instance, Borsuk Theorem implies Brower
Theorem discussed below. But prior to this we have to introduce an
important notion related to Brower Theorem.
'34
◦
3 FixedPoint Property.
Let f : X →X be a continuous map. A point a ∈ X is called a ﬁxed
point of f if f(a) = a. A space X is said to have the ﬁxedpoint property
if any continuous map X → X has a ﬁxed point. Fixed point property
means solvability of a wide class of equations.
34.11. Prove that the ﬁxed point property is a topological property.
34.12. A closed interval [a, b] has the ﬁxed point property.
34.13. Prove that if a topological space has ﬁxed point property then each
its retract also has the ﬁxedpoint property.
34.14. Prove that if topological spaces X and Y have ﬁxed point property,
x
0
∈ X and y
0
∈ Y , then X H Y/
x
0
∼ y
0
also has the ﬁxed point property.
§34. RETRACTIONS AND FIXED POINTS 186
34.15. Prove that any ﬁnite tree (i.e., a connected space obtained from a
ﬁnite collection of closed intervals by some identifying of their end points such
that deleting of an internal point of each of the segments makes the space
disconnected, see '38
◦
3) has the ﬁxedpoint property. Is this statement true
for inﬁnite trees?
34.16. Prove that R
n
with n > 0 does not have the ﬁxed point property.
34.17. Prove that S
n
does not have the ﬁxed point property.
34.18. Prove that RP
n
with odd n does not have the ﬁxed point property.
34.19*. Prove that CP
n
with odd n does not have the ﬁxed point property.
Information. RP
n
and CP
n
with any even n have the ﬁxed point
property.
34.J Brower Theorem. D
n
has the ﬁxed point property.
34.J.1. Deduce from Borsuk Theorem in dimension n (i.e., from the state
ment that S
n−1
is not a retract of D
n
) Brower Theorem in dimension n
(i.e., the statement that any continuous map D
n
→D
n
has a ﬁxed point).
'35 Homotopy Equivalences
'35
◦
1 Homotopy Equivalence as Map
Let X and Y be topological spaces, f : X → Y and g : Y → X
continuous maps. Consider compositions f ◦ g : Y →Y and g ◦ f : X →
X. They would be equal to the corresponding identity maps, if f and g
were homeomorphisms inverse to each other. If f ◦ g and g ◦ f are only
homotopic to the identity maps then f and g are said to be homotopy
inverse to each other. If a continuous map possesses a homotopy inverse
map then it is called homotopy invertible or a homotopy equivalence.
35.A. Prove the following properties of homotopy equivalences:
(a) any homeomorphism is a homotopy equivalence,
(b) a map homotopy inverse to a homotopy equivalence is a homotopy
equivalence,
(c) the composition of homotopy equivalences is a homotopy equiva
lence.
35.1. Find a homotopy equivalence that is not a homeomorphism.
'35
◦
2 Homotopy Equivalence as Relation
Topological spaces X and Y are said to be homotopy equivalent if
there exists a homotopy equivalence X →Y .
35.B. Homotopy equivalence of topological spaces is an equivalence re
lation.
The classes of homotopy equivalent spaces are called homotopy types.
Thus homotopy equivalent spaces are said to be of the same homotopy
type.
35.2. Prove that homotopy equivalent spaces have the same number of path
connected components.
35.3. Prove that homotopy equivalent spaces have the same number of con
nected components.
35.4. Find inﬁnite series of topological spaces belonging to the same homo
topy type, but pairwise nonhomeomorphic.
'35
◦
3 Deformation Retraction
A retraction ρ, which is homotopy inverse to the inclusion, is called a
deformation retraction. Since ρ is a retraction, one of the two conditions
from the deﬁnition of homotopy inverse maps is satisﬁed automatically:
its composition with the inclusion ρ ◦ in is equal to the identity id
A
. The
other condition says that in ◦ ρ is homotopic to the identity id
X
.
If X admits a deformation retraction onto A, then A is called a de
formation retract of X.
187
§35. HOMOTOPY EQUIVALENCES 188
'35
◦
4 Examples
35.C. Circle S
1
is a deformation retract of R
2
0
35.5. Prove that M¨obius strip is homotopy equivalent to circle.
35.6. Prove that a handle is homotopy equivalent to a union of two circles
intersecting in a single point.
35.7. Prove that a handle is homotopy equivalent to a union of three arcs
with common end points (i.e., letter θ).
35.8. Classify letters of Latin alphabet up to homotopy equivalence.
35.D. Prove that a plane with s points deleted is homotopy equivalent
to a union of s circles intersecting in a single point.
35.E. Prove that the union of a diagonal of a square and the contour of
the same square is homotopy equivalent to a union of two circles inter
secting in a single point.
35.9. Prove that the space obtained from S
2
by identiﬁcation of a two (dis
tinct) points is homotopy equivalent to the union of a twodimensional sphere
and a circle intersecting in a single point.
35.10. Prove that the space ¦(p, q) ∈ C : z
2
+pz +q has two distinct roots¦
of quadratic complex polynomials with distinct roots is homotopy equivalent
to the circle.
35.11. Prove that the space GL(n, R) of invertible n n real matrices is
homotopy equivalent to the subspace O(n) consisting of orthogonal matrices.
35.12. Enigma. Is there any relation between a solution of the preceding
problem and the GramSchmidt orthogonalization? Can the GramSchmidt
orthogonalization algorithm be considered a deformation retraction?
'35
◦
5 Deformation Retraction Versus Homotopy Equivalence
35.F. Spaces of Problem 35.E cannot be embedded one to another. On
the other hand, they can be embedded as deformation retracts to plane
with two points removed.
Deformation retractions comprise a special type of homotopy equiv
alences. They are easier to visualize. However, as follows from 35.F,
homotopy equivalent spaces may be such that none of them can be em
bedded to the other one, and hence none of them is homeomorphic to a
deformation retract of the other one. Therefore deformation retractions
seem to be not suﬃcient for establishing homotopy equivalences.
Though it is not the case:
35.13*. Prove that any two homotopy equivalent spaces can be embedded
as deformation retracts to the same topological space.
§35. HOMOTOPY EQUIVALENCES 189
'35
◦
6 Contractible Spaces
A topological space X is said to be contractible if the identity map id :
X →X is homotopic to a constant map.
35.14. Show that R and I are contractible.
35.15. Prove that any contractible space is pathconnected.
35.16. Prove that the following three statements about a topological space
X are equivalent:
(a) X is contractible,
(b) X is homotopy equivalent to a point,
(c) there exists a deformation retraction of X onto a point,
(d) any point a of X is a deformation retract of X,
(e) any continuous map of any topological space Y to X is homotopic to a
constant map,
(f) any continuous map of X to any topological space Y is homotopic to a
constant map.
35.17. Is it right that if X is a contractible space then for any topological
space Y
(a) any two continuous maps X →Y are homotopic?
(b) any two continuous maps Y →X are homotopic?
35.18. Check if spaces of the following list are contractible:
(a) R
n
,
(b) a convex subset of R
n
,
(c) a star convex subset of R
n
,
(d) ¦(x, y) ∈ R
2
: x
2
−y
2
≤ 1¦,
(e) a ﬁnite tree (i.e., a connected space obtained from a ﬁnite collection
of closed intervals by some identifying of their end points such that
deleting of an internal point of each of the segments makes the space
disconnected, see '38
◦
3.)
35.19. Prove that X Y is contractible, iﬀ both X and Y are contractible.
'35
◦
7 Fundamental Group and Homotopy Equivalences
35.G. Let f : X → Y and g : Y → X be homotopy inverse maps,
x
0
∈ X and y
0
∈ Y be points such that f(x
0
) = y
0
and g(y
0
) = x
0
and, moreover, the homotopies relating f ◦ g to id
Y
and g ◦ f to id
X
are
ﬁxed at y
0
and x
0
, respectively. Then f
∗
and g
∗
are inverse to each other
isomorphisms between groups π
1
(X, x
0
) and π
1
(Y, y
0
).
35.H Corollary. If ρ : X → A is a strong deformation retraction,
x
0
∈ A, then ρ
∗
: π
1
(X, x
0
) → π
1
(A, x
0
) and in
∗
: π
1
(A, x
0
) → π
1
(X, x
0
)
are isomorphisms inverse to each other.
35.20. Calculate the fundamental group of the following spaces:
(a) R
3
R
1
, (b) R
N
R
n
, (c) R
3
S
1
, (d) R
N
S
n
,
(e) S
3
S
1
, (f) S
N
S
k
, (g) RP
3
RP
1
, (h) handle,
(i) M¨obius band, (j) sphere with s handles,
(k) Klein bottle with a point re
moved,
(l) M¨obius band with s holes.
§35. HOMOTOPY EQUIVALENCES 190
35.21. Prove that the boundary of the M¨obius band standardly embedded
in R
3
(see 20.18) could not be the boundary of a disk embedded in R
3
in such
a way that its interior does not intersect the band.
35.22. Calculate the fundamental group of the space of all the complex
polynomials ax
2
+ bx + c with distinct roots. Calculate the fundamental
group of the subspace of this space consisting of polynomials with a = 1.
35.23. Enigma. Can you solve 35.22 along deriving of the formular for
roots of quadratic trinomial?
35.I. What if the hypothesis of Theorem 35.G were weakened as follows:
g(y
0
) = x
0
and/or the homotopies relating f ◦ g to id
Y
and g ◦ f to id
X
are not ﬁxed at y
0
and x
0
, respectively? How would f
∗
and g
∗
be related?
Would π
1
(X, x
0
) and π
1
(Y, y
0
) be isomorphic?
'36 Covering Spaces via Fundamental Groups
'36
◦
1 Homomorphisms Induced by Covering Projections
36.A. Let p : X → B be a covering, x
0
∈ X, b
0
= p(x
0
). Then
p
∗
: π
1
(X, x
0
) →π
1
(B, b
0
) is a monomorphism. Cf. 31.C.
The image of the monomorphism p
∗
: π
1
(X, x
0
) → π
1
(B, b
0
) induced
by a covering projection p : X →B is called the group of covering p with
base point x
0
.
36.B. Enigma. Is the group of covering determined by the covering?
36.C Group of Covering vs Lifting Loops. Describe loops in the
base space of a covering, whose homotopy classes belong to the group of
the covering, in terms provided by Path Lifting Theorem 31.B.
36.D. Let p : X → B be a covering, let x
0
, x
1
∈ X belong to the same
pathcomponent of X, and b
0
= p(x
0
) = p(x
1
). Then p
∗
(π
1
(X, x
0
)) and
p
∗
(π
1
(X, x
1
)) are conjugate subgroups of π
1
(B, b
0
) (i.e., there exists an
α ∈ π
1
(B, b
0
) such that p
∗
(π
1
(X, x
1
)) = α
−1
p
∗
(π
1
(X, x
0
))α).
36.E. Let p : X → B be a covering, x
0
∈ X, b
0
= p(x
0
). Let α ∈
π
1
(B, b
0
). Then there exists x
1
∈ p
−1
(b
0
) such that p
∗
(π
1
(X, x
1
)) =
α
−1
p
∗
(π
1
(X, x
0
))α.
36.F. Let p : X →B be a covering in a narrow sense and G ⊂ π
1
(B, b
0
)
be the group of this covering with a base point x
0
. A subgroup H ⊂
π
1
(B, b
0
) is a group of the same covering, iﬀ it is conjugate to G.
'36
◦
2 Number of Sheets
36.G Number of Sheets and Index of Subgroup. Let p : X → B
be a covering in narrow sense with ﬁnite number of sheets. Then the
number of sheets is equal to the index of the group of this covering.
36.H Sheets and Right Cosets. Let p : X → B be a covering in
narrow sense, b
0
∈ B, x
0
∈ p
−1
(b
0
). Construct a natural bijection of
p
−1
(b
0
) and the set p
∗
(π
1
(X, x
0
))`π
1
(B, b
0
) of right cosets of the group
of the covering in the fundamental group of the base space.
36.1 Number of Sheets in Universal Covering. The number of sheets
of a universal covering equals the order of the fundamental group of the base
space.
36.2 NonTrivial Covering Means NonTrivial π
1
. Any topological
space, which has a nontrivial pathconnected covering space, has a nontrivial
fundamental group.
36.3. What numbers can appear as the number of sheets of a covering of the
M¨obius strip by the cylinder S
1
I?
191
§36. COVERING SPACES VIA FUNDAMENTAL GROUPS 192
36.4. What numbers can appear as the number of sheets of a covering of the
M¨obius strip by itself?
36.5. What numbers can appear as the number of sheets of a covering of the
Klein bottle by torus?
36.6. What numbers can appear as the number of sheets of a covering of the
Klein bottle by itself?
36.7. What numbers can appear as the numbers of sheets for a covering of
the Klein bottle by plane R
2
?
36.8. What numbers can appear as the numbers of sheets for a covering of
the Klein bottle by S
1
R?
36:A Action of π
1
in Fiber. Let p : X → B be a covering, b
0
∈ B. Construct a
natural right action of π
1
(B, b
0
) in p
−1
(b
0
).
36:B. When the action in 36:A is transitive?
'36
◦
3 Hierarchy of Coverings
Let p : X → B and q : Y → B be coverings, x
0
∈ X, y
0
∈ Y and
p(x
0
) = q(y
0
) = b
0
. One says that q with base point y
0
is subordinate to
p with base point x
0
if there exists a map ϕ : X →Y such that q ◦ ϕ = p
and ϕ(x
0
) = y
0
. In this case the map ϕ is called a subordination.
36.I. A subordination is a covering map.
36.J. If a subordination exists, then it is unique. Cf. 31.B.
Coverings p : X →B and q : Y →B are said to be equivalent if there
exists a homeomorphism h : X → Y such that p = q ◦ h. In this case h
and h
−1
are called equivalencies
36.K. If two coverings are mutually subordinate, then the corresponding
subordinations are equivalencies.
36.L. The equivalence of coverings is, indeed, an equivalence relation in
the set of coverings with a given base space.
36.M. Subordination deﬁnes a partial order in the set of equivalence
classes of coverings with a given base.
36.9. What equivalence class of coverings is minimal (i. e., subordinate to
all other classes)?
36.N. Let p : X → B and q : Y → B be coverings, x
0
∈ X, y
0
∈ Y
and p(x
0
) = q(y
0
) = b
0
. If q with base point y
0
is subordinate to p with
base point x
0
then the group of covering p is contained in the group of
covering q, i.e. p
∗
(π
1
(X, x
0
)) ⊂ q
∗
(π
1
(Y, y
0
)).
§36. COVERING SPACES VIA FUNDAMENTAL GROUPS 193
'36
◦
4 Existence of subordinations
A topological space X is said to be locally pathconnected if for each point a ∈ X
and each neighborhood U of a there is a neighborhood V ⊂ U which is pathconnected.
36:1. Find a path connected, but not locally path connected topological
space.
36:C. Let B be a locally pathconnected space, p : X → B and q : Y → B be
coverings in narrow sense, x
0
∈ X, y
0
∈ Y and p(x
0
) = q(y
0
) = b
0
. If p
∗
(π
1
(X, x
0
)) ⊂
q
∗
(π
1
(Y, y
0
)) then q is subordinate to p.
36:C.1. Under the conditions of 36:C, if paths u, v : I →X have the same
initial point x
0
and a common ﬁnal point, then the paths which cover p ◦u
and p ◦ v and have the same initial point y
0
also have the same ﬁnal point.
36:C.2. Under the conditions of 36:C, the map X → Y deﬁned by 36:C.1
(guess, what is this map!) is continuous.
36:2. Construct an example, which proves that the hypothesis of local path
connectedness in 36:C.2 and 36:C is necessary.
36:D. Two coverings, p : X → B and q : Y → B, with a common locally path
connected base are equivalent, iﬀ for some x
0
∈ X and y
0
∈ Y with p(x
0
) = q(y
0
) = b
0
the groups p
∗
(π
1
(X, x
0
)) and q
∗
(π
1
(Y, y
0
)) are conjugate in π
1
(B, b
0
).
36:3. Construct an example, which proves that the hypothesis of local path
connectedness of the base in 36:D is necessary.
'36
◦
5 Micro Simply Connected Spaces
A topological space X is said to be micro simply connected if each point a ∈ X
has a neighborhood U such that the inclusion homomorphism π
1
(U, a) →π
1
(X, a) is
trivial.
36:4. Any simply connected space is micro simply connected.
36:5. Find a micro simply connected, but not simply connected space.
A topological space is called locally contractible at point a, if any neigh
borhood U of a contains a neighborhood V such that the inclusion V →U is
homotopic to a constant map. A topological space is called locally contractible,
if it is locally contractible at each of its points.
36:6. Any ﬁnite topological space is locally contractible.
36:7. Any locally contractible space is micro simply connected.
36:8. Find a space which is not micro simply connected.
In literature micro simply connectedness is called also weak local sim
ply connectedness, while a strong local simply connectedness is the following
property: any neighborhood U of any point contains a neighborhood V such
that any loop contained in V is homotopic to constant in U.
36:9. Find a micro simply connected space, which is not strong locally simply
connected.
§36. COVERING SPACES VIA FUNDAMENTAL GROUPS 194
'36
◦
6 Existence of Coverings
36:E. Any space, which has a universal covering space, is micro simply connected.
36:F Existence of covering with a given group. If a topological space B is path
connected, locally path connected and micro simply connected then for any b
0
∈ B and
any subgroup π of π
1
(B, b
0
) there exists a covering p : X → B and a point x
0
∈ X
such that p(x
0
) = b
0
and p
∗
(π
1
(X, x
0
)) = π.
36:F.1. Let under conditions of Theorem 36:F there exists a covering p :
X → B satifying all the requirements of this theorem. For each x ∈ X,
describe all paths in B which can be represented as images under p of paths
connecting x
0
to x in X.
36:F.2. Does the solution of Proble 36:F.1 deﬁne an equivalence relation
in the set of all paths in B, starting at b
0
, such that there is a bijection
between the set X and the set of equivalence classes?
36:F.3. Describe a topology in the set of equivalence classes from 36:F.2
so that the natural bijection between X and this set would be a homeo
morphism.
36:F.4. Prove that the reconstruction of X and p : X → B provided by
problems 36:F.1 – 36:F.4 under the assumptions of Theorem 36:F give rise
to a covering which existence is laimed by Theorem 36:F.
36:G Classiﬁcation of Coverings Over a Good Space. There is a onetoone
correspondence between classes of equivalent coverings (in narrow sense) over a path
connected, locally path connected, and micro simply connected space B with a base
point b
0
, on one hand, and congruence classes of subgroups of π
1
(B, b
0
), on the other
hand. This correspondence identiﬁes the hierarhy of coverings (ordered by subordina
tion) with the hierarhy of subgroups (ordered by incluion).
Under the correspondence of Theorem 36:G, the trivial subgroup corresponds to
the covering with simply connected covering space. Since this is the covering which
subordinates any other covering with the same base space, it is said to be universal.
36:10. Describe all the coverings of the following spaces up to equivalence
and subordination:
(a) circle S
1
;
(b) punctured plane R
2
0;
(c) torus S
1
S
1
;
(d) M¨obius strip;
(e) Klein Bottle;
(f) four point digital circle (the space formed by 4 points, a, b, c, d with the
base of open sets formed by ¦a¦, ¦c¦, ¦a, b, c¦ and ¦c, d, a¦).
'36
◦
7 Automorphisms of Covering
A homeomorphism ϕ : X →X is called an automorphism of covering p : X →B,
if p ◦ ϕ = p.
36:H. Automorphisms of a covering form a group (with repect to compositions).
Denote the group of automorphisms of a covering p : X →B by Aut(p).
§36. COVERING SPACES VIA FUNDAMENTAL GROUPS 195
36:I. An automorphism ϕ : X → X of covering p : X → B is recovered from the
image ϕ(x
0
) of any x
0
∈ X. Cf. 36.J.
36:J. Any twofold covering has a nontrivial automorphism.
36:11. Find a threefold covering which does not have a nontrivial auto
morphism.
Let G be a group and H its subgroup. Recall that normiliser Nr(H) of H is the
subset of G consisting of g ∈ G such that g
−1
Hg = H. This is a subgroup of G,
which contains H as a normal subgroup. So, Nr(H)/
H
is a group.
36:K. Let p : X → B be a covering, x
0
∈ X and b
0
= p(x
0
). Construct a map
π
1
(B, b
0
) →p
−1
(b
0
), which would induce a bujection of the set p
∗
(π
1
(X, x
0
))`π
1
(B, b
0
)
of right cosets onto p
−1
(b
0
).
36:L. Show that the set of images of a point x
0
under all automorphisms of covering
p : X → B is mapped by the bijection p
∗
(π
1
(X, x
0
))`π
1
(B, b
0
) → p
−1
(b
0
) from 36:K
to the group Nr(p
∗
(π
1
(X, x
0
)))/
p
∗
(π
1
(X, x
0
))
.
36:M. For any covering p : X → B in the narrow sense, there is a natural in
jective map Aut(p) to the group Nr(p
∗
(π
1
(X, x
0
)))/
p
∗
(π
1
(X, x
0
))
. This map is an
antihomomorphism.
1
36:N. Under assumptions of Theorem 36:M, if the base space B is locally path con
nected then the antihomomorphism Aut(p) →Nr(p
∗
(π
1
(X, x
0
)))/
p
∗
(π
1
(X, x
0
))
is bi
jective.
'36
◦
8 Regular Coverings
36:O Regularity of Covering. Let p : X → B be a covering in a narrow sense,
b
0
∈ B, x
0
∈ p
−1
(b
0
). The following conditions are equivalent:
(a) p
∗
π
1
(X, x
0
) is a normal subgroup of π
1
(B, b
0
);
(b) p
∗
π
1
(X, x) is a normal subgroup of π
1
(B, p(x)) for each x ∈ X;
(c) all the groups p
∗
π
1
(X, x) for x ∈ p
−1
(b) are the same;
(d) for any loop s : I →B either every path in X covering s is a loop (independent
on the its initial point) or none of them is a loop;
(e) the automorphism group acts transitively on p
−1
(b
0
).
A covering satisfying to (any of) the equivalent conditions of Theorem 36:O is
said to be regular.
36:P. The coverings R → S
1
: x → e
2πix
and S
1
→ S
1
: z → z
n
for integer n > 0
are regular.
36:Q. The automorphism group of a regular covering p : X → B is naturally anti
isomorphic to the quotient group π
1
(B, b
0
)/
p
∗
π
1
(X, x
0
)
of the group π
1
(B, b
0
) by the
group of the covering for any x
0
∈ p
−1
(b
0
).
36:R. Any twofold covering is regular.
1
Recall that a map ϕ : G →H of a group G to a group H is called antihomomor
phism, if ϕ(ab) = ϕ(a)ϕ(b) for any a, b ∈ G.
§36. COVERING SPACES VIA FUNDAMENTAL GROUPS 196
36:12. Which coverings considered in Problems of Section '30 are regular?
Is out there any nonregular covering?
36:13. Find a threefold nonregular covering of a bouquet of two circles.
36:14. Let p : X →B be a regular covering, Y ⊂ X, C ⊂ B and q : Y → C
be a submap of p. Prove that if q is a covering then this covering is regular.
'36
◦
9 Lifting and Covering Maps
36:S. Enigma. Let p : X →B and f : Y → B be continuous maps. Let x
0
∈ X and
y
0
∈ Y be points such that p(x
0
) = f(y
0
). Formulate in terms of homomorphisms
p
∗
: π
1
(X, x
0
) →π
1
(B, p(x
0
)) and f
∗
: π
1
(Y, y
0
) →π
1
(B, f(y
0
)) a necessary condition
for existence of a lifting
¯
f : Y → X of f such that
¯
f(y
0
) = x
0
. Find an example
in which this condition is not suﬃcient. What additional assumptions can make it
suﬃcient?
36:T Theorem on Lifting a Map. Let p : X → B be a covering in the narrow
sense and f : Y →B be a continuous map. Let x
0
∈ X and y
0
∈ Y be points such that
p(x
0
) = f(y
0
). If Y is a locally path connected space and f
∗
π(Y, y
0
) ⊂ p
∗
π(X, x
0
),
then there exists a unique continuous map
¯
f : Y → X such that p ◦
¯
f = f and
¯
f(y
0
) = x
0
.
36:U. Let p : X →B and q : Y →C be coverings in the narrow sense and f : B → C
be a continuous map. Let x
0
∈ X and y
0
∈ Y be points such that fp(x
0
) = q(y
0
). If
there exists a continuous map F : X → Y such that fp = qF and F(x
0
) = y
0
then
f
∗
p
∗
π
1
(X, x
0
) ⊂ q
∗
π
1
(Y, y
0
).
36:V Theorem on Covering of a Map. Let p : X → B and q : Y → C be
coverings in the narrow sense and f : B → C be a continuous map. Let x
0
∈ X
and y
0
∈ Y be points such that fp(x
0
) = q(y
0
). If Y is locally path connected and
f
∗
p
∗
π
1
(X, x
0
) ⊂ q
∗
π
1
(Y, y
0
) then there exists a unique continuous map F : X → Y
such that fp = qF and F(x
0
) = y
0
.
CHAPTER 8
Cellular Techniques
'37 Cellular Spaces
'37
◦
1 Deﬁnition of Cellular Spaces
In this section we study a class of topological spaces, which play an
important role in algebraic topology. Their role in the context of this
book is more restricted: this is the class of spaces for which we learn how
to calculate the fundamental group.
This class of spaces was introduced by J.H.C.Whitehead. He called
these spaces CWcomplexes, and they are known under this name. How
ever, for many reasons it is not a good name. For very rare exceptions
(one of which is CWcomplex, other is simplicial complex), the word com
plex is used nowadays for various algebraic notions, but not for spaces.
We have decided to use the term cellular space instead of CW
complex, following D. B. Fuchs and V. A. Rokhlin, Beginner’s Course
in Topology: Geometric Chapters. Berlin; New York: SpringerVerlag,
1984. [?].
A zerodimensional cellular space is just a discrete space. Points of
a 0dimensional cellular space are also called (zerodimensional) cells or
0cells.
A onedimensional cellular space is a space, which can be obtained as
follows. Take any 0dimensional cellular space X
0
. Take a family of maps
ϕ
α
: S
0
→ X
0
. Attach to X
0
by ϕ
α
the sum of a family of copies of D
1
(indexed by the same indices α as the maps ϕ
α
):
X
0
∪
ϕα
(H
α
D
1
).
The images of the interior parts of copies of D
1
are called (open) 1
dimensional cells, or 1cells, or onecell, or edges. The subsets obtained
out of D
1
are called closed 1cells. The cells of X
0
(i.e., points of X
0
)
are also called vertices. Open 1cells and 0cells comprise a partition of a
onedimensional cellular space. This partition is included in the notion
of cellular space, i.e., a onedimensional cellular space is a topological
space equipped with a partition, which can be obtained in this way.
197
§37. CELLULAR SPACES 198
Onedimensional cellular spaces are associated also with the word
graph. However, rather often this word is used for objects of other
classes. For example, in this way one can call onedimenional cellular
spaces, in which attaching maps of diﬀerent onecells are prohibitted to
coincide, or the boundary of a onecell is prohibited to consist of a single
vertex. When onedimensional cellular spaces are to be considered any
way, despite of this terminological disregard, they are called multigraphs
or pseudographs. Moreover, sometimes one includes into the notion of
graph an additional structure. Say, a choice of orientation on each edge.
Certainly, all these variations contradict to a general tendency in math
ematical terminology to call in a simpler way decent objects of a more
general nature, passing to more complicated terms along with adding
structures and imposing restrictions. However, in this speciﬁc situation
there is no hope to implement that tendency. Any attempt to ﬁx a
meaning for the word graph apparently contributes only to the chaos,
and we just keep this word away from important formulations, using it
as a short informal synonym for more formal term of onedimeninsional
cellular space. (Other overused common words, like curve and surface,
also deserve this sort of caution.)
A twodimensional cellular space is a space, which can be obtained
as follows. Take any cellular space X
1
of dimension ≤ 1. Take a family
of continuous
1
maps ϕ
α
: S
1
→ X
1
. Attach to X
1
by ϕ
α
the sum of a
family of copies of D
2
:
X
1
∪
ϕα
(H
α
D
2
).
The images of the interior parts of copies of D
2
are called open 2
dimensional cells, or 2cells, or faces. The cells of X
1
are also considered
as cells of the 2dimensional cellular space. Open cells of both kinds
comprise a partition of a 2dimensional cellular space. This partition is
included in the notion of cellular space, i.e., a twodimensional cellular
space is a topological space equipped with a partition, which can be
obtained in the way described above. The set obtained out of a copy of
the whole D
2
is called a closed 2cell.
A cellular space of dimension n is deﬁned in a similar way: This is a
space equipped with a partition. It can be obtained from a cellular space
X
n−1
of dimension < n by attaching a family of copies of ball D
n
by a
family of continuous maps of their boundary spheres:
X
n−1
∪
ϕα
(H
α
D
n
).
The images of interior parts of the attached ndimensional balls are
called (open) ndimensional cells, or ncells. The images of the whole
1
Above, in the deﬁnition of 1dimensional cellular space, the attaching maps ϕ
α
also were continuous, although their continuity was not required, since any map of
S
0
to any space is continuous.
§37. CELLULAR SPACES 199
ndimensional balls are called closed ncells. Cells of X
n−1
are also
considered as cells of the ndimensional cellular space.
A cellular space is obtained as a union of increasing sequence of cellular
spaces X
0
⊂ X
1
⊂ ⊂ X
n
⊂ . . . obtained in this way from each other.
The sequence may be ﬁnite or inﬁnite. In the latter case topological
structure is introduced by saying that the cover of the union by X
n
’s
is fundamental, i.e., that a set U ⊂ ∪
∞
n=0
X
n
is open, iﬀ its intersection
U ∩ X
n
with each X
n
is open in X
n
.
The partition of a cellular space into its open cells is called a cellular
decomposition. The union of all cells of dimension ≤ n of a cellular space
X is called the ndimensional skeleton of X. This term may be misleading,
since ndimensional skeleton may be without cells of dimension n, hence
it may coincide with (n − 1)dimensional skeleton. Thus ndimensional
skeleton may have dimension < n. Therefore it is better to speak about
nth skeleton or nskeleton. Cells of dimension n are called also ncells. A
cellular space is said to be ﬁnite if it contains a ﬁnite number of cells. A
cellular space is said to be countable if it contains a countable number of
cells. A cellular space is said to be locally ﬁnite, if any of its points has a
neighborhood which intersects a ﬁnite number of cells.
Let X be a cellular space. A subspace A ⊂ X, which can be presented
both as a union of closed cells and a union of open cells, is called a cellular
subspace of X. Of course, it is provided with a partition into the open
cells of X contained in A. Obviously, the kskeleton of a cellular space
X is a cellular subspace of X.
37.A. Prove that a cellular subspace of a cellular space is a cellular
space.
'37
◦
2 First Examples
37.B. A cellular space consisting of two cells, one 0dimensional and one
ndimensional, is homeomorphic to S
n
.
37.C. Represent D
n
with n > 0 as a cellular space made of three cells.
37.D. A cellular space consisting of a single zerodimensional cell and q
onedimensional cells is a bouquet of q circles.
37.E. Represent torus S
1
S
1
as a cellular space with one 0cell, two
1cells, and one 2cell.
37.F. How to obtain a presentation of torus S
1
S
1
as a cellular space
with 4 cells from a presentation of S
1
as a cellular space with 2 cells?
37.1. Prove that if X and Y are ﬁnite cellular spaces then X Y can be
equipped in a natural way with a structure of ﬁnite cellular space.
37.2*. Does the statement of 37.1 remain true if one skips the ﬁniteness
condition in it? If yes, prove; if no, ﬁnd an example when the product is not
a cellular space.
§37. CELLULAR SPACES 200
37.G. Represent sphere S
n
as a cellular space such that spheres S
0
⊂
S
1
⊂ S
2
⊂ ⊂ S
n−1
are its skeletons.
37.H. Represent RP
n
as a cellular space with n + 1 cells. Describe the
attaching maps of its cells.
37.3. Represent CP
n
as a cellular space with n + 1 cells. Describe the
attaching maps of its cells.
37.4. Represent the following topological spaces as cellular ones
(a) handle; (b) M¨obius strip; (c) S
1
I,
(d) sphere with p
handles;
(e) sphere with p
crosscaps.
37.5. What is the minimal number of cells in a cellular space homeomorphic
to
(a) M¨obius strip; (b) sphere with p
handles;
(c) sphere with p
crosscaps?
37.6. Find a cellular space, in which a closure of a cell is not equal to a union
of other cells. What is the minimal number of cells in a space containing a
cell of this sort?
37.7. Consider a disjoint sum of a countable collection of copies of closed
interval I and identify the copies of 0 in all of them. Represent the result
(which is the bouquet of the countable family of intervals) as a countable
cellular space. Prove that this space is not ﬁrst countable.
37.I. Represent R
1
as a cellular space.
37.8. Prove that for any two cellular spaces homeomorphic to R
1
there exists
a homeomorphism between them mapping each cell of one of them homeo
morphically onto a cell of the other one.
37.J. Represent R
n
as a cellular space.
Denote by R
∞
the union of the sequence of Euclidean spaces R
0
⊂
R
1
⊂ ⊂ R
n
⊂ canonically included to each other: R
n
= ¦x ∈ R
n+1
:
x
n+1
= 0¦. Equip R
∞
with the topological structure, for which the spaces
R
n
comprise a fundamental cover.
37.K. Represent R
∞
as a cellular space.
'37
◦
3 Further TwoDimensional Examples
Let us consider a class of 2dimensional cellular spaces, which ad
mit a simple combinatorial description. Each space of this class can be
presented as a quotient space of a ﬁnite family of convex polygons by
identiﬁcation of sides via aﬃne homeomorphisms. The identiﬁcation of
vertices is deﬁned by the identiﬁcation of the sides. The quotient space is
naturally equipped with decomposition into 0cells, which are the images
of vertices, 1cells, which are the images of sides, and faces, the images
of the interior parts of the polygons.
§37. CELLULAR SPACES 201
To describe such a space, one needs, ﬁrst, to show, what sides are to
be identiﬁed. Usually this is indicated by writing the same letters at the
sides that are to be identiﬁed. There are only two aﬃne homeomorphisms
between two closed intervals. To specify one of them, it is enough to show
orientations of the intervals which are identiﬁed by the homeomorphism.
Usually this is done by drawing arrows on the sides. Here is a description
of this sort for the standard presentation of torus S
1
S
1
as the quotient
space of square:
It is possible to avoid a picture by a description. To do this, go around
the polygons counterclockwise writing down the letters, which stay at
the sides of polygon along the contour. The letters corresponding to the
sides, whose orientation is opposite to the counterclockwise direction,
put with exponent −1. This gives rise to a collection of words, which
contains a suﬃcient information about the family of polygons and the
partition. For instance, the presentation of torus shown above is encoded
by the word ab
−1
a
−1
b.
37.9. Prove that:
(a) word a
−1
a describes a cellular space homeomorphic to S
2
,
(b) word aa describes a cellular space homeomorphic to RP
2
,
(c) word aba
−1
b
−1
c describes a handle,
(d) word abcb
−1
describes cylinder S
1
I,
(e) each of the words aab and abac describe M¨obius strip,
(f) word abab describes a cellular space homeomorphic to RP
2
,
(g) each of the words aabb and ab
−1
ab describe Klein bottle,
(h) word
a
1
b
1
a
−1
1
b
−1
1
a
2
b
2
a
−1
2
b
−1
2
. . . a
g
b
g
a
−1
g
b
−1
g
.
describes sphere with g handles,
(i) word a
1
a
1
a
2
a
2
. . . a
g
a
g
describes sphere with g crosscaps.
'37
◦
4 Simplicial spaces
Recall that in '19
◦
10 it was introduced a class of topological spaces, simplicial
spaces. Each simplicial space is equipped with a partition to subsets, called open
simplices, which are really homeomorphic to open simplices of Euclidean space.
37:A. Any simplicial space is cellular and its partition to open simplicies coincides
with the corresponding partition to open cells.
'37
◦
5 Topological Properties of Cellular Spaces
37:B. Closed cells constitute a fundamental cover of a cellular space.
37:C. If A is cellular subspace of a cellular space X then A is closed in X.
§37. CELLULAR SPACES 202
37:D. Prove that any compact subset of a cellular space intersects a ﬁnite number
of cells.
37:E Corollary. A cellular space is compact, iﬀ it is ﬁnite.
37:F. Any cell of a cellular space is contained in a ﬁnite cellular subspace of this
space.
37:G. Any compact subset of a cellular space is contained in a ﬁnite cellular subspace.
37:H. A cellular space is separable, iﬀ it is countable.
37:I. Any pathconnected component of a cellular space is a cellular subspace.
37:J. Any pathconnected component of a cellular space is both open and closed. It
is a connected component.
37:K. A cellular space is connected, iﬀ it is path connected.
37:L. Any connected locally ﬁnite cellular space is countable.
37:M. A cellular space is connected, iﬀ its 1skeleton is connected.
37:N. Any cellular space is normal.
'37
◦
6 Embedding to Euclidean Space
37.L. Any countable 0dimensional cellular space can be embedded into
R.
37.M. Any countable locally ﬁnite 1dimensional cellular space can be
embedded into R
3
.
37.10. Find a 1dimensional cellular space, which you cannot embed into
R
2
. (We do not ask to prove that it is impossible to embed.)
37.N. Any ﬁnite dimensional countable locally ﬁnite cellular space can
be embedded into Euclidean space of suﬃciently high dimension.
37.N.1. Let X and Y be topological spaces such that X can be embedded
into R
p
and Y can be embedded into R
q
. Let A be a closed subset of
Y . Assume that A has a neighborhood U in Y such that there exists a
homeomorphism h : Cl U → A I mapping A to A 0. Let ϕ : A → X
be any continuous map. Then there exists an embedding of X ∪
ϕ
Y into
R
p+q+1
.
37.N.2. Let X be a locally ﬁnite countable kdimensional cellular space
and A be its (k −1)skeleton. Prove that if A can be embedded to R
p
then
X can be embedded into R
p+k+1
.
37.O. Any countable locally ﬁnite cellular space can be embedded into
R
∞
.
37.P. Any countable locally ﬁnite cellular space is metrizable.
§37. CELLULAR SPACES 203
'37
◦
7 Euler Characteristic
Let X be a ﬁnite cellular space. Let c
i
(X) denote the number of its
cells of dimension i. Euler characteristic of X is the alternating sum of
c
i
(X):
χ(X) = c
0
(X) −c
1
(X) +c
2
(X) − + (−1)
i
c
i
(X) +. . .
.
37:O. Prove that Euler characteristic is additive in the following sense: for any
cellular space X and its ﬁnite cellular subspaces A and B
χ(A ∪ B) = χ(A) +χ(B) −χ(A ∩ B).
37:P. Prove that Euler characteristic is multiplicative in the following sense: for any
ﬁnite cellular spaces X and Y the Euler characteristic of their product X Y is
χ(X)χ(Y ).
'37
◦
8 Collaps
Let X be a cellular space, e and f its open cells of dimensions n and
n −1, respectively. Suppose:
• the attaching map ϕ
e
: S
n−1
→ X
n−1
of the cell e deﬁnes a homeo
morphism ϕ
−1
e
(f) →f,
• f does not meet images of attaching maps of cells, distinct from e,
• the cell e is disjoint from the image of attaching map of any cell.
37.Q. X (e ∪ f) is a cellular subspace of X.
37.R. X (e ∪ f) is a deformation retract of X.
We say that X (e ∪f) is obtained from X by an elementary collaps,
and write X `X (e ∪ f).
If a cellular subspace A of a cellular space X is obtained from X by
a sequence of elementary collapses, then we say that X is collapsed onto
A and also write X `A.
37.S. Collapsing does not change the Euler characteristic: if X is a
ﬁnite cellular space and X `A, then χ(A) = χ(X).
37.11. Let A be cellular space of dimension n, let ϕ : S
n
→A and ψ : S
n
→
A be continuous maps. Prove that if ϕ and ψ are homotopic then spaces
X
ϕ
= A ∪
ϕ
D
n+1
and X
ψ
= A ∪
ψ
D
n+1
are homotopy equivalent.
37.12. Let X be a space obtained from a circle by attaching of two copies
of disk by maps S
1
→ S
1
: z → z
2
and S
1
→S
1
: z → z
3
, respectively. Find
a cellular space homotopy equivalent to X with smallest possible number of
cells.
37.13. Enigma. Generalize the preceding problem.
§37. CELLULAR SPACES 204
'37
◦
9 Generalized collaps
Let, as above, X be a cellular space, e and f be its open cells of
dimensions n and n − 1, respectively, and let the attaching map ϕ
e
:
S
n
→ X
n−1
of e deﬁne a homeomorphism ϕ
−1
e
(f) → f. (Unlike in the
preceding section, we assume neither that f is disjoint from the images
of attaching maps of cells diﬀerent from e, nor that e is disjoint from
the images of attaching maps of whatever cells.) Let χ
e
: D
n
→ X be a
natural map of D
n
onto the closure of e, which maps the interior of D
n
homeomorphicaly onto e and coincides on the boundary sphere with ϕ
e
.
(This map is called a characteristic map of e.
Furthermore, let ψ : D
n
→ S
n−1
ϕ
−1
e
(f) be a deformation re
traction. Since ϕ
−1
e
(f) is homeomorphic to an open (n − 1)ball, the
latter condition is not very restrictive. For example, if ϕ
−1
e
(f) is hemi
sphere ¦(x
1
, . . . , x
n
) ∈ S
n−1
[ x
n
> 0¦, then for ψ one can take the
projection of D
n
along the nth coordinate axis onto closed hemisphere
¦(x
1
, . . . , x
n
) ∈ S
n−1
[ x
n
≤ 0¦.
37.T. Under these conditions, quotient space X/
[χ
e
(x) ∼ ϕ
e
(ψ(x))]
of
X is a cellular space, which cells are the images under the natural pro
jections of all the cells of X, besides e and f.
Cellular space X/
[χ
e
(x) ∼ ϕ
e
(ψ(x))]
is said to be obtained by cancel
lation of cells e and f.
37.U. The natural projection X →X/
[χ
e
(x) ∼ ϕ
e
(ψ(x))]
is a homotopy
equivalence.
37.U.1. Find a cellular subspace Y of a cellular space X such that the
projection Y → Y/
[χ
e
(x) ∼ ϕ
e
(ψ(x))]
would be a homotopy equivalence
by Theorem 37.R.
37.U.2. Extend inclusion map Y (e ∪ f) → Y to a map homotopically
inverse to projection X → X/
[χ
e
(x) ∼ ϕ
e
(ψ(x))]
, and extend homotopies
which exist by 37.U.1 to homotopies which prove that this map is homo
topically inverse to the projection above.
37:Q. A cellular space X, which contains a closed 1cell e homeomorphic to I, is
homotopy equivalent to cellular space X/
e
obtained by contraction of e.
37:R. Any connected cellular space is homotopy equivalent to a cellular space with
one point 0skeleton.
37.14. Prove that the torus S
1
S
1
with two disks attached along meridian
¦1¦ S
1
and parallel S
1
¦1¦, respectively, is homotopy equivalent to S
2
.
37.15. Prove that the quotient space
CP
2
/
[(z
0
: z
1
: z
2
) ∼ (z
0
: z
1
: z
2
)]
§37. CELLULAR SPACES 205
of the complex projective plane CP
2
is homotopy equivalent to S
4
.
'38 OneDimensional Cellular Spaces
'38
◦
1 Homotopy Classiﬁcation
38.A. Any connected ﬁnite 1dimensional cellular space is homotopy
equivalent to a bouquet of circles.
38.A.1 Lemma. Let X be a 1dimensional cellular space, and e its 1cell,
which is attached by an injective map S
0
→ X
0
(i.e., it has two distinct
end points). Prove that the natural projection X → X/
e
is a homotopy
equivalence. Describe the homotopy inverse map explicitly.
38.B. A ﬁnite connected cellular space X of dimension one is homotopy
equivalent to the bouquet of 1 −χ(X) circles and its fundamental group
is a free group of rank 1 −χ(X).
38.C Corollary. The Euler charateristic of a ﬁnite connected one
dimensional cellular space is invariant under homotopy equivalence. It is
not greater than one. It equals one, iﬀ the space is homotopy equivalent
to point
38.D Corollary. The Euler characteristic of a ﬁnite onedimensional
cellular space is not greater than the number of its connected components.
It is equal to this number, iﬀ each of its connected components is homo
topy equivalent to point
38.E Homotopy Classiﬁcation of Finite 1Dimensional Cel
lular Spaces. Finite connected onedimensional cellular spaces are ho
motopy equivalent, iﬀ their fundamental groups are isomorphic, iﬀ their
Euler characteristics are equal.
38.1. The fundamental group of a 2sphere punctured at n points is a free
group of rank n −1.
38.2. Prove that the Euler characteristic of a cellular space homeomorphic
to S
2
is equal to 2.
38.3 The Euler Theorem. For any convex polyhedron in R
3
, the sum of
the number of its vertices and the number of its faces equals the number of
its edges plus two.
'38
◦
2 Dividing Cells
38:A. In a connected cellular space each connected component of the complement of
an edge meets the closure of the edge. The complement has at most two connected
component.
A complete local characterization of a vertex in a onedimensional cellular space
is its valency. This is the totla number of points in the preimages of the vertex under
attaching maps of all onedimensional cells of the space.
38:B. Each connected component of the complement of a vertex in a connected one
dimensional cellular space contains an edge with boundary containing the vertex. The
complement of a vertex of valency m has at most m connected components.
206
§38. ONEDIMENSIONAL CELLULAR SPACES 207
38:C. In any ﬁnite connected onedimensional cellular space there exists a vertex
with connected complement.
'38
◦
3 Trees and Forests
A onedimensional cellular space is called a tree, if it is connected and the com
plement of any its (open) 1cell is not connected. A onedimensional cellular space is
called a forest, if each of its connected components is a tree.
38:D. Any cellular subspace of a forest is a forest. In particular, any connected
cellular subspace of a tree is a tree.
38:E. In a tree the complement of an edge consists of two connected components.
38:F. In a tree the complement of a vertex of valency m consists of m connected
components.
38:G. In a ﬁnite tree there exists a vertex of valency one.
38:H. Any ﬁnite tree can be collapsed to a point and has Euler characteristic one.
38:I. Prove that any point of a tree is its deformation retract.
38:J. Any ﬁnite onedimensional cellular space, which can be collapsed to a point, is
a tree.
38:K. In any ﬁnite onedimensional cellular space the sum of valencies of all vertices
is equal to the number of edges multiplied by two.
38:L. A ﬁnite connected onedimensional cellular space with Euler characteristic one
has a vertex of valency one.
38:M. A ﬁnite connected onedimensional cellular space with Euler characteristic
one can be collapsed to a point.
'38
◦
4 Simple Paths
Let X be a onedimensional cellular space. A simple path of length n in X is
a ﬁnite sequence (v
1
, e
1
, v
2
, e
2
, . . . , e
n
, v
n+1
), formed of vertices v
i
and edges e
i
of X
such that each term appears in it only once and the boundary of every edge e
i
consists
of the preceding and susequent vertices v
i
and v
i+1
. Vertex v
1
is said to be initial
and v
n+1
, ﬁnal vertex. The simple path is said to connect these vertices. They are
connected by a path I →X, which is a topological embedding with image contained
in the union of all the cells involved in the simple path. The union of these cells is a
cellular subspace of X. It is called a simple broken line.
38:N. In a connected onedimensional cellular space any two vertices can be connected
by a simple path.
38:O Corollary. In a connected onedimensional cellular space X any two points
can be connected by a path I →X, which is a topological embedding.
38:1. Does there exist a pathconnected space in which there exist distinct
points which cannot be connected by a path which is a topological embedding.
38:2. Can you ﬁnd a Hausdorﬀ space with this property?
38:P. A connected onedimensional cellular space X is a tree, iﬀ there exists no
topological embedding S
1
→X.
§38. ONEDIMENSIONAL CELLULAR SPACES 208
38:Q. In a onedimensional cellular space X there exists a loop S
1
→ X not homo
topic to a constant loop, iﬀ there exists a topological embedding S
1
→X.
38:R. A onedimensional cellular space is a tree, iﬀ any two distinct vertices can be
connected in it by a unique simple path.
38:3. Prove that any ﬁnite tree has ﬁxed point property.
Cf. 34.12, 34.13 and 34.14.
38:4. Is this true for any tree; for any ﬁnite connected onedimensional cel
lular space?
'38
◦
5 Maximal Trees
A cellular subspace A of a cellular space X is called a maximal tree
of X if A is a tree and is not contained in any other cellular subspace
B ⊂ X, which is a tree.
38.F. Any ﬁnite connected onedimensional cellular space contains a
maximal tree.
38.G. Prove that a cellular subspace A of a cellular space X is a maximal
tree, iﬀ it is a tree and contains all vertices of X.
A maximal tree is called also a spanning tree. Theorem 38.G provides
the reason for this.
38.H. Prove that a cellular subspace A of a cellular space X is a maximal
tree, iﬀ it is a tree and the quotient space X/
A
is a bouquet of circles.
38.I. Let X be a onrdimensional cellular space and A its cellular sub
space. Prove that if A is a tree then the natural projection X →X/
A
is
a homotopy equivalence.
Problems 38.F, 38.I and 38.H provide yet another proof of Theorem
38.A.
'39 Fundamental Group of a Cellular Space
'39
◦
1 OneDimensional Cellular Spaces
39.A. The fundamental group of a connected ﬁnite onedimensional cel
lular space X is a free group of rank 1 −χ(X).
39.B. Let X be a ﬁnite connected 1dimensional cellular space, T a
maximal tree of X and x
0
∈ T. For each 1cell e ⊂ X T choose a loop
s
e
, which starts at x
0
, goes inside T to e, then goes once along e and
then comes back to x
0
in T. Prove that π
1
(X, x
0
) is freely generated by
homotopy classes of s
e
.
'39
◦
2 Generators
39.C. Let A be a topological space, x
0
∈ A. Let ϕ : S
k−1
→A be a con
tinuous map, X = A∪
ϕ
D
k
. If k > 1, then the inclusion homomorphism
π
1
(A, x
0
) →π
1
(X, x
0
) is surjective. Cf. 28.G.5, 28.G.4.
39.D. Let X be a cellular space, x
0
its 0cell and X
1
the 1skeleton of
X. Then the inclusion homomorphism
π
1
(X
1
, x
0
) →π
1
(X, x
0
)
is surjective.
39.E. Let X be a ﬁnite cellular space, T a maximal tree of X
1
and
x
0
∈ T. For each cell e ⊂ X
1
T choose a loop s
e
, which starts at x
0
,
goes inside T to e, then goes once along e and then comes back to x
0
in
T. Prove that π
1
(X, x
0
) is generated by homotopy classes of s
e
.
39.1. Deduce Theorem 28.G from Theorem 39.D.
39.2. Find π
1
(CP
n
).
'39
◦
3 Relators
Let X be a cellular space, x
0
its 0cell. Denote by X
n
the nskeleton
of X. Recall that X
2
is obtained from X
1
by attaching copies of disk
D
2
by continuous maps ϕ
α
: S
1
→ X
1
. The attaching maps are circular
loops in X
1
. For each α choose a path s
α
: I → X
1
connecting ϕ
α
(1)
with x
0
. Denote by N the normal subgroup of π
1
(X, x
0
) generated (as a
normal subgroup
2
.) by elements
T
sα
[ϕ
α
] ∈ π
1
(X
1
, x
0
).
39.F. N does not depend on the choice of paths s
α
.
2
Recall that a subgroup is said to be normal if it coincides with conjugate sub
groups. The normal subgroup generated by a set A is the minimal normal subgroup
containing A. As a subgroup, it is generated by elements of A and elements conju
gate to them. This means that each element of this normal subgroup is a product of
elements conjugate to elements of A
209
§39. FUNDAMENTAL GROUP OF A CELLULAR SPACE 210
39.G. N coincides with the kernel of the inclusion homomorphism
i
∗
: π
1
(X
1
, x
0
) →π
1
(X, x
0
).
39.G.1 Lemma 1. N ⊂ Ker i
∗
, cf. 28.J (c).
39.G.2 Lemma 2. Let p
1
: Y
1
→ X
1
be a covering with covering group
N. Then for any α and a point y ∈ p
−1
1
(ϕ
α
(1)) there exists a lifting
¯ ϕ
α
: S
1
→Y
1
of ϕ
α
with ¯ ϕ
α
(1) = y.
39.G.3 Lemma 3. Let Y
2
be a cellular space obtained by attaching copies
of disk to Y
1
by all liftings of attaching maps ϕ
α
. Then there exists a map
p
2
: Y
2
→X
2
extending p
1
and this is a covering.
39.G.4 Lemma 4. Any loop s : I →X
1
realizing an element of the kernel
of the inclusion homomorphism π
1
(X
1
, x
0
) → π
1
(X
2
, x
0
) (i.e., homotopic
to constant in X
2
) is covered by a loop of Y
2
. The covering loop is contained
in Y
1
.
39.G.5 Lemma 5. N coincides with the kernel of the inclusion homomor
phism π
1
(X
1
, x
0
) →π
1
(X
2
, x
0
).
39.G.6 Lemma 6. Attaching maps of ncells with n ≥ 3 are lifted to any
covering space. Cf. 31:B, 31:C.
39.G.7 Lemma 7. Covering p
2
: Y
2
→ X
2
can be extended to a covering
of the whole X.
39.G.8 Lemma 8. Any loop s : I → X
1
realizing an element of Ker i
∗
(i.e., homotopic to constant in X) is covered by a loop of Y . The covering
loop is contained in Y
1
.
'39
◦
4 Writing Down Generators and Relators
Theorems 39.E and 39.G imply the following prescription for writ
ing down presentation for the fundamental group of a ﬁnite dimensional
cellular space by generators and relators:
Let X be a ﬁnite cellular space, x
0
its 0cell. Let T a maximal tree
of 1skeleton of X. For each 1cell e ⊂ T of X choose a loop s
e
, which
starts at x
0
, goes inside T to e, then goes once along e and then comes
back to x
0
in T. Let g
1
, . . . , g
m
be the homotopy classes of these loops.
Let ϕ
1
, . . . , ϕ
n
: S
1
→X
1
be attaching maps of 2cells of X. For each ϕ
i
choose a path s
i
connecting ϕ
i
(1) with x
0
in 1skeleton of X. Express the
homotopy class of the loop s
−1
i
ϕ
i
s
i
as a product of powers of generators
g
j
. Let r
1
, . . . , r
n
are the words in letters g
1
, . . . , g
m
obtained in this
way. The fundamental group of X is generated by g
1
, . . . , g
m
, which are
subject to deﬁning relators r
1
= 1, . . . , r
n
= 1.
39.H. Check that this rule gives correct answers in the cases of RP
n
and S
1
S
1
for the cellular presentations of these spaces provided in
Problems 37.H and 37.E.
§39. FUNDAMENTAL GROUP OF A CELLULAR SPACE 211
'39
◦
5 Fundamental Groups of Basic Surfaces
39.I. The fundamental group of a sphere with g handles admits presen
tation
¦a
1
, b
1
, a
2
, b
2
, . . . a
g
, b
g
: a
1
b
1
a
−1
1
b
−1
1
a
2
b
2
a
−1
2
b
−1
2
. . . a
g
b
g
a
−1
g
b
−1
g
= 1¦.
39.J. The fundamental group of a sphere with g crosscaps admits pre
sentation
¸
a
1
, a
2
, . . . a
g
: a
2
1
a
2
2
. . . a
2
g
= 1
¸
.
39.K. Fundamental groups of spheres with diﬀerent number of handles
are not isomorphic.
When one needs to prove that two ﬁnitely presented groups are not iso
morphic, one of the ﬁrst natural moves is to abelianize the groups. Recall
that to abelianize a group G means to quotient it out by the commutator
subgroup. The commutator subgroup [G, G] is the normal subgroup gen
erated by commutators a
−1
b
−1
ab for all a, b ∈ G. Abelianization means
adding relations that ab = ba for any a, b ∈ G.
Abelian ﬁnitely generated groups are well known. Any ﬁnitely gener
ated abelian group is isomorphic to a product of a ﬁnite number of cyclic
groups. If the abelianized groups are not isomorphic then the original
groups are not isomorphic as well.
39.K.1. Abelianized fundamental group of a sphere with g handles is a free
abelian group of rank 2g (i.e., is isomorphic to Z
2g
).
39.K.2. Fundamental groups of spheres with diﬀerent number of crosscaps
are not isomorphic.
39.K.3. The abelianized fundamental group of a sphere with g crosscaps is
isomorphic to Z
g−1
Z
2
.
39.L. Spheres with diﬀerent numbers of handles are not homotopy equiv
alent.
39.M. Spheres with diﬀerent numbers of crosscaps are not homotopy
equivalent.
39.N. A sphere with handles is not homotopy equivalent to a sphere with
crosscaps.
If X is a pathconnected space then the abelianized fundamental
group of X is called the 1dimensional (or ﬁrst) homology group of X
and denoted by H
1
(X). If X is not pathconnected then H
1
(X) is the
direct sum of the ﬁrst homology groups of all pathconnected components
of X. Thus 39.K.1 can be rephrased as follows: if F
g
is a sphere with g
handles then H
1
(F
g
) = Z
2g
.
§39. FUNDAMENTAL GROUP OF A CELLULAR SPACE 212
'39
◦
6 Seifert  van Kampen Theorem
Let X be a connected cellular space, A and B its cellular subspaces which cover
X. Denote A ∩ B by C.
39:A. How fundamental groups of X, A, B and C are related?
39:B Seifert  van Kampen Theorem. Suppose A, B, and C are connected. Let
x
0
∈ C,
π
1
(A, x
0
) = ¦α
1
, . . . , α
p
: ρ
1
= 1, . . . , ρ
r
= 1¦,
π
1
(B, x
0
) = ¦β
1
, . . . , β
q
: σ
1
= 1, . . . , σ
s
= 1¦,
and π
1
(C, x
0
) be generated by γ
1
, . . . γ
t
. Let the images of γ
i
under the inclusion
homomorphisms π
1
(C, x
0
) → π
1
(A, x
0
) and π
1
(C, x
0
) → π
1
(B, x
0
) be expressed as
ξ
i
(α
1
, . . . , α
p
) and η
i
(β
1
, . . . , β
q
), respectively. Then
π
1
(X) = ¦α
1
, . . . , α
p
, β
1
, . . . , β
q
:
ρ
1
= 1, . . . , ρ
r
= 1, σ
1
= 1, . . . , σ
s
= 1,
ξ
1
= η
1
, . . . , ξ
t
= η
t
¦.
39:C. Let X, A, B and C be as above. Suppose A, B are simply connected and C
consists of two path connected components. Prove that π
1
(X) is isomorphic to Z.
To write details: van Kampen published much more gen
eral theorem!
'40 OneDimensional Homology and Cohomology
'40
◦
1 Why and What for
Sometimes the fundamental group contains too much information to deal with,
and it is handy to ignore a part of this information. A regular way to do this is to
use instead of the fundamental group some of its natural quotient groups. One of
them, the abelianized fundamental group, was introduced and used in Section '39 to
prove, in particular, that spheres with diﬀerent numbers of handles are not homotopy
equivalent, see Problems 39.K, 39.K.139.K.3 and 39.L.
In this Section we will study the onedimensional homology and its closest rela
tives. Usually they are studied in the framework of homology theory together with
their highdimensional generalizations. This general theory requires much more al
gebra and takes more time and eﬀorts. On the other hand, onedimensional case is
useful on its own, involves a lot of speciﬁc details and provides a geometric intuition,
which is useful, in particular, for studying the highdimensional homology.
'40
◦
2 OneDimensional Integer Homology
Recall that for a pathconnected space X the abelianized fundamental group of
X is called its onedimensional homology group and denoted by H
1
(X). If X is
an arbitrary topological space then H
1
(X) is the direct sum of the onedimensional
homology groups of all the connected components of X.
40:1. Find H
1
(X) for the following spaces
(a) M¨obius strip,
(b) handle,
(c) sphere with p handles and r holes,
(d) sphere with p crosscaps r holes,
(e) the complement in R
3
of the circles ¦(x, y, z) ∈ R
3
[ z = 0, x
2
+y
2
= 1¦
and ¦(x, y, z) ∈ R
3
[ x = 0, z
2
+ (y −1)
2
= 1¦,
(f) the complement in R
3
of the circles ¦(x, y, z) ∈ R
3
[ z = 0, x
2
+y
2
= 1¦
and ¦(x, y, z) ∈ R
3
[ z = 1, x
2
+y
2
= 1¦,
The name of H
1
(X) appears often with the adjective integer or expression with
coeﬃcients in Z, so it comes as onedimensional integer homology group of X, or one
dimensional homology group of X with coeﬃcients in Z. This is done to distinguish
H
1
(X) from its genegalizations, onedimensional homology groups with coeﬃcients in
any abelian group G. The case of G = Z
2
is considered below, but we will not study
these generalizations in full generality.
The group operation in H
1
(X) (as well as in other homology groups) is written
additively and called addition. Thus the product of loops represents the sum of the
homology classes represented by the loops multiplied.
Few more new words. An element of a homology group is called a homology class.
The homology classes really admit several interpretations as equivalence classes of
objects of various nature. For example, according to the deﬁnition we start with, a
homology class is a coset consisting of elements of the fundamental group. In turn,
each element of the fundamental group consists of loops. Thus, we can think of a
homology class as of a set of loops.
213
§40. ONEDIMENSIONAL HOMOLOGY AND COHOMOLOGY 214
'40
◦
3 Zero Homologous Loops and Disks with Handles
A loop which belongs to the zero homology class is said to be zerohomologous.
Loops, which belong to the same homology class, are said to be homologous to each
other.
40:A ZeroHomologous Loop. Let X be a topological space. A circular loop s :
S
1
→ X is zerohomologous, iﬀ there exist a continuous map f of a disk D with
handles (i.e., a sphere with a hole and handles) to X and a homeomorphism h of S
1
onto the boundary circle of D such that f ◦ h = s.
40:A.1. In the fundamental group of a disk with handles, a loop, whose
homotopy class generates the fundamental group of the boundary circle, is
homotopic to a product of commutators of meridian and longitude loops
of the handles.
A homotopy between a loop and a product of commutators of loops
can be thought of as an extension of the loop to a continuous map of a
sphere with handles and a hole.
'40
◦
4 Description of H
1
(X) in Terms of Free Circular Loops
Factorization by the commutator subgroup kills the diﬀerence between translation
maps deﬁned by diﬀerent paths. Therefore the abelianized fundamental groups of a
pathconnected space can be naturally identiﬁed. Hence each free loop deﬁnes a
homology class. This suggests that H
1
(X) can be deﬁned starting with free loops,
rather than loops at a base point.
40:B. On the sphere with two handles and three holes shown in Figure 1 the sum
of the homology classes of the three loops, which go counterclockwise arround the
three holes, is zero.
Figure 1. Sphere with two handles and three holes. The
boundary circles of the holes are equipped with arrows
showing the counterclockwise orientation.
40:C ZeroHomologous Collections of Loops. Let X be a pathwise connected
space and s
1
, . . . , s
n
: S
1
→ X be a collection of n free loops. Prove that the sum
of homology classes of s
1
, . . . , s
n
is equal to zero, iﬀ there exist a continuous map
f : F →X, where F is a sphere with handles and n holes, and embeddings i
1
, . . . , i
n
:
S
1
→ F parametrizing the boundary circles of the holes in the counterclockwise
direction (as in Figure 1) such that s
k
= f ◦ i
k
for k = 1, . . . , n.
§40. ONEDIMENSIONAL HOMOLOGY AND COHOMOLOGY 215
40:D Homologous Collections of Loops. In a topological space X any class
ξ ∈ H
1
(X) can be represented by a ﬁnite collection of free circular loops. Collections
¦u
1
, . . . , u
p
¦ and ¦v
1
, . . . , v
q
¦ of free circular loops in X deﬁne the same homology
class, iﬀ there exist a continuous map f : F → X, where F is a disjoint sum of
several spheres with handles and holes with the total number of holes equal p + q,
and embeddings i
1
, . . . , i
p+q
: S
1
→ F parametrizing the boundary circles of all the
holes of F in the counterclockwise direction such that u
k
= f ◦ i
k
for k = 1, . . . , p
and v
−1
k
= f ◦ i
k+p
for k = 1, . . . , q.
'40
◦
5 Homology and Continuous Maps
Let X be a path connected topological space with a base point x
0
∈ X. The
factorization map π
1
(X, x
0
) →H
1
(X) is usually called the Hurewicz homomorphism
3
and denoted by H. If X is not path connected and X
0
is its path connected component
containing x
0
, then the inclusion X
0
→ X deﬁnes an isomorphism in : π
1
(X
0
, x
0
) →
π
1
(X, x
0
). On the other hand, H
1
(X
0
) is contained in H
1
(X) as a direct summand.
This allows one to deﬁne the Hurewicz homomorphism π
1
(X, x
0
) → H
1
(X) as a
composition of the Hurewicz homomorphism H : π
1
(X
0
, x
0
) → H
1
(X
0
) (which is
already deﬁned above), isomorphism in
−1
: π
1
(X, x
0
) → π
1
(X
0
, x
0
) (inverse to the
inclusion isomorphism), and inclusion H
1
(X
0
) →H
1
(X).
40:E. Let f : (X, x
0
) → (Y, y
0
) be a continuous map. If X is path connected, then
the diagram
π
1
(X, x
0
) −−−−→
f∗
π(Y, y
0
)
H
H
H
1
(X) H
1
(Y )
is completed in a unique way to a commutative diagram
π
1
(X, x
0
) −−−−→
f∗
π(Y, y
0
)
H
H
H
1
(X) −−−−→ H
1
(Y )
The homomorphism H
1
(X) →H
1
(Y ) completing the diagram in 40:E is denoted
by the same symbol f
∗
as the homomorphism f
∗
: π
1
(X, x
0
) → π
1
(Y, y
0
). It is also
called a homomorphism induced by f.
40:F. Extend the deﬁnition of f
∗
: H
1
(X) → H
1
(Y ) given in 40:E to the case when
X is not path connected.
40:G. For any continuous map f : X → Y and any loop ϕ : S
1
→ X, the image
under f
∗
: H
1
(X) → H
1
(Y ) of the homology class represented by ϕ is the homology
class represented by f ◦ ϕ.
3
Witold Hurewicz has introduced a high dimensional generalization of this ho
momorphism, π
n
(X, x
0
) → H
n
(X), which we cannot discuss here for you are not
assumed to be familiar with H
n
(X). The homomorphism π
1
(X, x
0
) →H
1
(X) should
be rather attributed to Henry Poincar´e, although the group H
1
(X) was introduced
long after he died.
§40. ONEDIMENSIONAL HOMOLOGY AND COHOMOLOGY 216
40:2. Look through '33, '34, '35, '36 and '39 and ﬁnd all the theorems
about homomorphisms of fundamental groups which gives rise to similar the
orems about homomorphisms of onedimensional homology groups. In which
applications the fundamental groups can be replaced by onedimensional ho
mology groups?
40:3 Homology Group of a Cellular Space. Deduce from the calcula
tion of the fundamental group of a cellular space (see '39) an algorithm for
calculation of H
1
(X) for a cellular space X.
'40
◦
6 OneDimensional Cohomology
Let X be a pathconnected topological space and G a commutative group.
40:H. The homomorphisms π
1
(X, x
0
) →G comprise a commutative group in which
the group operation is the pointwise addition.
The group Hom(π
1
(X, x
0
), G) of all the homomorphisms π
1
(X, x
0
) →G is called
onedimensional cohomology group of X with coeﬃcients in Gand denoted by H
1
(X; G).
For an arbitrary topological space X, the onedimensional cohomology group of X
with coeﬃcients in G is deﬁned as the direct product of onedimensional cohomology
group with coeﬃcients in G of all the pathconnected components of X.
40:I Cohomology via Homology. H
1
(X; G) = Hom(H
1
(X), G).
The following subsection is to be rewritten when
the section on classiﬁcation of coverings will be done!
'40
◦
7 Cohomology and Classiﬁcation of Regular Coverings
Recall that a covering p : X → B is a regular Gcovering if X is a Gspace, in
which the orbits of the action of G are the ﬁbers of p and G acts eﬀectively on each
of them. Regular Gcovering may be with disconnected total space. For example,
X G →X is a regular Gcovering.
For any loop s : I →B in the base B of a regular Gcovering p : X →B there is a
map M
s
: p
−1
(s(0)) →p
−1
(s(0)) assigning to x ∈ p
−1
(s(0)) the ﬁnal point of the path
covering s
−1
and beginning at x. This map is called the monodromy transformation
of p
−1
(s(0)) deﬁned by s. It coincides with action of one of the elements of G. In this
way a homomorphismπ
1
(B) →G is deﬁned. It is called the monodromy representaion
of the fundamental group. Thus any regular Gcovering of X deﬁnes a cohomology
class belonging to H
1
(X; G).
40:J Cohomology and Regular Coverings. This map is a bijection of the set of
all the regular Gcoverings of X onto H
1
(X; G).
40:4 Addition of GCoverings. What operation on the set of regular G
coverings corresponds to addition of cohomology classes?
'40
◦
8 Integer Cohomology and Maps to S
1
Let X be a topological space and f : X → S
1
a continuous map. It induces
a homomorphism f
∗
: H
1
(X) → H
1
(S
1
) = Z. Therefore it deﬁnes an element of
H
1
(X; Z).
§40. ONEDIMENSIONAL HOMOLOGY AND COHOMOLOGY 217
40:K. This construction deﬁnes a bijection of the set of all the homotopy classes of
maps X →S
1
onto H
1
(X; Z).
40:L Addition of Maps to Circle. What operation on the set of homotopy classes
of maps to S
1
corresponds to the addition in H
1
(X; Z)?
40:M. What regular Zcovering of X corresponds to a homotopy class of mappings
X →S
1
under the compositions of the bijections described in 40:K and 40:J
'40
◦
9 OneDimensional Homology Modulo 2
Here we deﬁne yet another natural quotient group of the fundamental group. It
is even simpler than H
1
(X).
For a pathconnected X, consider the quotient group of π
1
(X) by the normal
subgroup generated by squares of all the elements of π(X). It is denoted by H
1
(X; Z
2
)
and called onedimensional homology group of X with coeﬃcients in Z
2
or the ﬁrst Z
2

homology group of X. For an arbitrary X, the group H
1
(X; Z
2
) is deﬁned as the sum
of onedimensional homology group with coeﬃcients in Z
2
of all the pathconnected
components of X.
Elements of H
1
(X; Z
2
) are called onedimensional homology classes modulo 2 or
onedimensional homology classes with coeﬃcients in Z
2
. They can be thought of as
classes of elements of the fundamental groups or classes of loops. A loop deﬁning the
zero homology class modulo 2 is said to be zerohomologous modulo 2.
40:N. In a disk with crosscaps the boundary loop is zerohomologous modulo 2.
40:O Loops ZeroHomologous Modulo 2. Prove that a circular loop s : S
1
→
X is zerohomologous modulo 2, iﬀ there exist a continuous map f of a disk with
crosscaps D to X and a homeomorphism h of S
1
onto the boundary circle of D such
that f ◦ h = s.
40:P. If a loop is zerohomologous then it is zerohomologous modulo 2.
40:Q Homology and Mod 2 Homology. H
1
(X; Z
2
) is commutative for any X,
and can be obtained as the quotient group of H
1
(X) by the subgroup of all even
homology classes, i.e. elements of H
1
(X) of the form 2ξ with ξ ∈ H
1
(X). Each
element of H
1
(X; Z
2
) is of order 2 and H
1
(X; Z
2
) is a vector space over the ﬁeld of
two elements Z
2
.
40:5. Find H
1
(X; Z
2
) for the following spaces
(a) M¨obius strip,
(b) handle,
(c) sphere with p handles,
(d) sphere with p crosscaps,
(e) sphere with p handles and r holes,
(f) sphere with p crosscaps and r holes,
(g) the complement in R
3
of the circles ¦(x, y, z) ∈ R
3
[ z = 0, x
2
+y
2
= 1¦
and ¦(x, y, z) ∈ R
3
[ x = 0, z
2
+ (y −1)
2
= 1¦,
(h) the complement in R
3
of the circles ¦(x, y, z) ∈ R
3
[ z = 0, x
2
+y
2
= 1¦
and ¦(x, y, z) ∈ R
3
[ z = 1, x
2
+y
2
= 1¦,
40:6 Z
2
Homology of Cellular Space. Deduce from the calculation of
the fundamental group of a cellular space (see Section '39) an algorithm for
calculation of the onedimensional homology group with Z
2
coeﬃcients of a
cellular space.
§40. ONEDIMENSIONAL HOMOLOGY AND COHOMOLOGY 218
40:R Collections of Loops Homologous Mod 2. Let X be a topological space.
Any class ξ ∈ H
1
(X; Z
2
) can be represented by a ﬁnite collection of free circular loops
in X. Collections ¦u
1
, . . . , u
p
¦ and ¦v
1
, . . . , v
q
¦ of free circular loops in X deﬁne the
same homology class modulo 2, iﬀ there exist a continuous map f : F →X, where F
is a disjoint sum of several spheres with crosscaps and holes with the total number of
holes equal p +q, and embeddings i
1
, . . . , i
p+q
: S
1
→F parametrizing the boundary
circles of all the holes of F such that u
k
= f ◦ i
k
for k = 1, . . . , p and v
k
= f ◦ i
k+p
for k = 1, . . . , q.
40:7. Compare 40:R with 40:D. Why in 40:R the counterclockwise direction
has not appeared? In what other aspects 40:R is simpler than 40:D and why?
40:S Duality Between Mod 2 Homology and Cohomology.
H
1
(X; Z
2
) = Hom(H
1
(X; Z
2
), Z
2
) = Hom
Z2
(H
1
(X; Z
2
), Z
2
)
for any space X. If H
1
(X; Z
2
) is ﬁnite then H
1
(X; Z
2
) and H
1
(X; Z
2
) are ﬁnite
dimensional vector spaces over Z
2
dual to each other.
40:8. A loop is zerohomologous modulo 2 in X, iﬀ it is covered by a loop
in any twofold covering space of X.
40:T. Enigma. Homology Modulo n? Generalize all the theory above about Z
2

homology to deﬁne and study Z
n
homology for any natural n.
Part 3
Manifolds
This part is devoted to study of the most important topological
spaces. These spaces provide a scene for most of geometric branches
of mathematics.
CHAPTER 9
Bare Manifolds
'41 Locally Euclidean Spaces
'41
◦
1 Deﬁnition of Locally Euclidean Space
Let n be a nonnegative integer. A topological space X is called a
locally Euclidean space of dimension n if each point of X has a neigh
borhood homeomorphic either to R
n
or R
n
+
. Recall that R
n
+
= ¦x ∈ R
n
:
x
1
≥ 0¦, it is deﬁned for n ≥ 1.
41.A. The notion of 0dimensional locally Euclidean space coincides
with the notion of discrete topological space.
41.B. Prove that the following spaces are locally Euclidean:
(a) R
n
,
(b) any open subset of R
n
,
(c) S
n
,
(d) RP
n
,
(e) CP
n
,
(f) R
n
+
,
(g) any open subset of R
n
+
,
(h) D
n
,
(i) torus S
1
S
1
,
(j) handle,
(k) sphere with handles,
(l) sphere with holes,
(m) Klein bottle,
(n) sphere with crosscaps.
41.1. Prove that an open subspace of a locally Euclidean space of dimension
n is a locally Euclidean space of dimension n.
41.2. Prove that a bouquet of two circles is not locally Euclidean.
41.C. If X is a locally Euclidean space of dimension p and Y is a locally
Euclidean space of dimension q then X Y is a locally Euclidean space
of dimension p +q.
'41
◦
2 Dimension
41.D. Can a topological space be simultaneously a locally Euclidean
space of dimension both 0 and n > 0?
221
§41. LOCALLY EUCLIDEAN SPACES 222
41.E. Can a topological space be simultaneously a locally Euclidean
space of dimension both 1 and n > 1?
41.3. Prove that any nonempty open connected subset of a locally Euclidean
space of dimension 1 can be made disconnected by removing two points.
41.4. Prove that any nonempty locally Euclidean space of dimension n >
1 contains a nonempty open set, which cannot be made disconnected by
removing any two points.
41.F. Can a topological space be simultaneously a locally Euclidean
space of dimension both 2 and n > 2?
41.G. Let U be an open subset of R
2
and a p ∈ U. Prove that π
1
(U¦p¦)
admits an epimorphism onto Z.
41.H. Deduce from 41.G that a topological space cannot be simultane
ously a locally Euclidean space of dimension both 2 and n > 2.
We see that dimension of locally Euclidean topological space is a
topological invariant at least for the cases when it is not greater than
2. In fact, this holds true without that restriction. However, one needs
some technique to prove this. One possibility is provided by dimen
sion theory, see, e.g., W. Hurewicz and H. Wallman, Dimension Theory
Princeton, NJ, 1941. Other possibility is to generalize the arguments
used in 41.H to higher dimensions. However, this demands a knowledge
of highdimensional homotopy groups.
41.5. Deduce that a topological space cannot be simultaneously a locally Eu
clidean space of dimension both n and p > n from the fact that π
n−1
(S
n−1
) =
Z. Cf. 41.H
'41
◦
3 Interior and Boundary
A point a of a locally Euclidean space X is said to be an interior
point of X if a has a neighborhood (in X) homeomorphic to R
n
. A point
a ∈ X, which is not interior, is called a boundary point of X.
41.6. Which points of R
n
+
have a neighborhood homeomorphic to R
n
+
?
41.I. Formulate a deﬁnition of boundary point independent of a deﬁni
tion for interior point.
Let X be a locally Euclidean space of dimension n. The set of all
interior points of X is called the interior of X and denoted by int X.
The set of all boundary points of X is called the boundary of X and
denoted by ∂X.
These terms (interior and boundary) are used also with diﬀerent
meaning. The notions of boundary and interior points of a set in a
topological space and the interior part and boundary of a set in a topo
logical space are introduced in general topology, see Section '7. They
§41. LOCALLY EUCLIDEAN SPACES 223
have almost nothing to do with the notions discussed here. In both senses
the terminology is classical, which is impossible to change. This does not
create usually a danger of confusion.
Notations are not as commonly accepted as words. We take an easy
opportunity to select unambiguous notations: we denote the interior part
of a set A in a topological space X by Int
X
A or Int A, while the interior
of a locally Euclidean space X is denoted by int X; the boundary of a
set in a topological space is denoted by symbol Fr, while the boundary
of locally Euclidean space is denoted by symbol ∂.
41.J. For a locally Euclidean space X the interior int X is an open dense
subset of X, the boundary ∂X is a closed nowhere dense subset of X.
41.K. The interior of a locally Euclidean space of dimension n is a lo
cally Euclidean space of dimension n without boundary (i.e., with empty
boundary; in symbols: ∂(int X) = ∅).
41.L. The boundary of a locally Euclidean space of dimension n is a
locally Euclidean space of dimension n −1 without boundary (i.e., with
empty boundary; in symbols: ∂(∂X) = ∅).
41.M. int R
n
+
⊃ ¦x ∈ R
n
: x
1
> 0¦ and
∂R
n
+
⊂ ¦x ∈ R
n
: x
1
= 0¦.
41.7. For any x, y ∈ ¦x ∈ R
n
: x
1
= 0¦, there exists a homeomorphism
f : R
n
+
→R
n
+
with f(x) = y.
41.N. Either ∂R
n
+
= ∅ (and then ∂X = ∅ for any locally Euclidean
space X of dimension n), or ∂R
n
+
= ¦x ∈ R
n
: x
1
= 0¦.
In fact, the second alternative holds true. However, this is not easy
to prove for any dimension.
41.O. Prove that ∂R
1
+
= ¦0¦.
41.P. Prove that ∂R
2
+
= ¦x ∈ R
2
: x
1
= 0¦. (Cf. 41.G.)
41.8. Deduce that a ∂R
n
+
= ¦x ∈ R
n
: x
1
= 0¦ from π
n−1
(S
n−1
) = Z. (Cf.
41.P, 41.5)
41.Q. Deduce from ∂R
n
+
= ¦x ∈ R
n
: x
1
= 0¦ for all n ≥ 1 that
int(X Y ) = int X int Y
and
∂(X Y ) = (∂(X) Y ) ∪ (X ∂Y ).
The last formula resembles Leibniz formula for derivative of a product.
41.R. Enigma. Can this be a matter of chance?
41.S. Prove that
§41. LOCALLY EUCLIDEAN SPACES 224
(a) ∂(I I) = (∂I I) ∪ (I ∂I),
(b) ∂D
n
= S
n−1
,
(c) ∂(S
1
I) = S
1
∂I = S
1
HS
1
,
(d) the boundary of M¨obius strip is homeomorphic to circle.
41.T Corollary. M¨obius strip is not homeomorphic to cylinder S
1
I.
'42 Manifolds
'42
◦
1 Deﬁnition of Manifold
A topological space is called a manifold of dimension n if it is
• locally Euclidean of dimension n,
• second countable,
• Hausdorﬀ.
42.A. Prove that the three conditions of the deﬁnition are independent
(i.e., there exist spaces not satisfying any one of the three conditions and
satisfying the other two.)
42.A.1. Prove that R∪
i
R, where i : ¦x ∈ R : x < 0¦ →R is the inclusion,
is a nonHausdorﬀ locally Euclidean space of dimension one.
42.B. Check whether the spaces listed in Problem 41.B are manifolds.
A compact manifold without boundary is said to be closed. As in the
case of interior and boundary, this term coincides with one of the basic
terms of general topology. Of course, the image of a closed manifold under
embedding into a Hausdorﬀ space is a closed subset of this Hausdorﬀ
space (as any compact subset of a Hausdorﬀ space). However absence
of boundary does not work here, and even noncompact manifolds may
be closed subsets. They are closed in themselves, as any space. Here
we meet again an ambiguity of classical terminology. In the context of
manifolds the term closed relates rather to the idea of a closed surface.
'42
◦
2 Components of Manifold
42.C. A connected component of a manifold is a manifold.
42.D. A connected component of a manifold is pathconnected.
42.E. A connected component of a manifold is open in the manifold.
42.F. A manifold is the sum of its connected components.
42.G. The set of connected components of any manifold is countable.
If the manifold is compact, then the number of the components is ﬁnite.
42.1. Prove that a manifold is connected, iﬀ its interior is connected.
42.H. The fundamental group of a manifold is countable.
'42
◦
3 Making New Manifolds out of Old Ones
42.I. Prove that an open subspace of a manifold of dimension n is a
manifold of dimension n.
42.J. The interior of a manifold of dimension n is a manifold of dimen
sion n without boundary.
225
§42. MANIFOLDS 226
42.K. The boundary of a manifold of dimension n is a manifold of di
mension n −1 without boundary.
42.2. The boundary of a compact manifold of dimension n is a closed man
ifold of dimension n −1.
42.L. If X is a manifold of dimension p and Y is a manifold of dimension
q then X Y is a manifold of dimension p +q.
42.M. Prove that a covering space (in narrow sense) of a manifold is a
manifold of the same dimension.
42.N. Prove that if the total space of a covering is a manifold then the
base is a manifold of the same dimension.
42.O. Let X and Y be manifolds of dimension n, A and B components
of ∂X and ∂Y respectively. Then for any homeomorphism h : B → A
the space X ∪
h
Y is a manifold of dimension n.
42.O.1. Prove that the result of gluing of two copy of R
n
+
by the identity
map of the boundary hyperplane is homeomorphic to R
n
.
42.P. Let X and Y be manifolds of dimension n, A and B closed subsets
of ∂X and ∂Y respectively. If A and B are manifolds of dimension n−1
then for any homeomorphism h : B →A the space X ∪
h
Y is a manifold
of dimension n.
'42
◦
4 Double
42.Q. Can a manifold be embedded into a manifold of the same dimen
sion without boundary?
Let X be a manifold. Denote by DX the space X ∪
id
∂X
X obtained
by gluing of two copies of X by the identity mapping id
∂X
: ∂X → ∂X
of the boundary.
42.R. Prove that DX is a manifold without boundary of the same di
mension as X.
DX is called the double of X.
42.S. Prove that a double of a manifold is compact, iﬀ the original
manifold is compact.
'42
◦
5 Collars and Bites
Let X be a manifold. An embedding c : ∂X I → X such that c(x, 0) = x for
each x ∈ ∂X is called a collar of X. A collar can be thought of as a neighborhood of
the boundary presented as a cylinder over boundary.
42:A. Every manifold has a collar.
§42. MANIFOLDS 227
Let U be an open set in the boundary of a manifold X. For a continuous
function ϕ : ∂X →R
+
with ϕ
−1
(0, ∞) = U set
B
ϕ
= ¦(x, t) ∈ ∂X R
+
: t ≤ ϕ(x)¦.
A bite on X at U is an embedding b : B
ϕ
→ X with some ϕ : ∂X → R
+
such that b(x, 0) = x for each x ∈ ∂X.
This is a generalization of collar. Indeed, a collar is a bite at U = ∂X
with ϕ = 1.
42:A.1. Prove that if U ⊂ ∂X is contained in an open subset of X home
omorphic to R
n
+
, then there exists a bite of X at U.
42:A.2. Prove that for any bite b : B →X of a manifold X the closure of
X b(B) is a manifold.
42:A.3. Let b
1
: B
1
→ X be a bite of X and b
2
: B
2
→ Cl(X b
1
(B
1
))
be a bite of Cl(X b
1
(B
1
)). Construct a bite b : B → X of X with
b(B) = b
1
(B
1
) ∪ b
2
(B
2
).
42:A.4. Prove that if there exists a bite of X at ∂X then there exists a
collar of X.
42:B. For any two collars c
1
, c
2
: ∂X I → X there exists a homeomorphism h :
X →X with h(x) = x for x ∈ ∂X such that h ◦ c
1
= c
2
.
This means that a collar is unique up to homeomorphism.
42:B.1. For any collar c : ∂XI →X there exists a collar c
: ∂XI →X
such that c(x, t) = c
(x, t/2).
42:B.2. For any collar c : ∂X I →X there exists a homeomorphism
h : X →X ∪
x→(x,1)
∂X I
with h(c(x, t)) = (x, t).
'43 Isotopy
'43
◦
1 Isotopy of Homeomorphisms
Let X and Y be topological spaces, h, h
: X →Y homeomorphisms.
A homotopy h
t
: X →Y , t ∈ [0, 1] connecting h and h
(i.e., with h
0
= h,
h
1
= h
) is called an isotopy between h and h
if h
t
is a homeomorphism
for each t ∈ [0, 1]. Homeomorphisms h, h
are said to be isotopic if there
exists an isotopy between h and h
.
43.A. Being isotopic is an equivalence relation on the set of homeomor
phisms X →Y .
43.B. Find a topological space X such that homotopy between homeo
morphisms X →X does not imply isotopy.
This means that isotopy classiﬁcation of homeomorphisms can be
more reﬁned than homotopy classiﬁcation of them.
43.1. Classify homeomorphisms of circle S
1
to itself up to isotopy.
43.2. Classify homeomorphisms of line R
1
to itself up to isotopy.
The set of isotopy classes of homeomorphisms X → X (i.e. the
quotient of the set of selfhomeomorphisms of X by isotopy relation) is
called the mapping class group or homeotopy group of X.
43.C. For any topological space X, the mapping class group of X is a
group under the operation induced by composition of homeomorphisms.
43.3. Find the mapping class group of the union of the coordinate lines in
the plane.
43.4. Find the mapping class group of the union of bouquet of two circles.
'43
◦
2 Isotopy of Embeddings and Sets
Homeomorphisms are topological embeddings of special kind. The
notion of isotopy of homeomorphism is extended in an obvious way to the
case of embeddings. Let X and Y be topological spaces, h, h
: X → Y
topological embeddings. A homotopy h
t
: X → Y , t ∈ [0, 1] connecting
h and h
(i.e., with h
0
= h, h
1
= h
) is called an (embedding) isotopy
between h and h
if h
t
is an embedding for each t ∈ [0, 1]. Embeddings
h, h
are said to be isotopic if there exists an isotopy between h and h
.
43.D. Being isotopic is an equivalence relation on the set of embeddings
X →Y .
A family A
t
, t ∈ I of subsets of a topological space X is called an
isotopy of the set A = A
0
, if the graph Γ = ¦(x, t) ∈ X I [ x ∈ A
t
¦ of
the family is ﬁbrewise homeomorphic to the cylinder A I, i. e. there
228
§43. ISOTOPY 229
exists a homeomorphism AI →Γ mapping A¦t¦ to Γ∩X ¦t¦ for
any t ∈ I. Such a homeomorphism gives rise to an isotopy of embeddings
Φ
t
: A → X, t ∈ I with Φ
0
= in, Φ
t
(A) = A
t
. An isotopy of a subset
is also called a subset isotopy. Subsets A and A
of the same topological
space X are said to be isotopic in X, if there exists a subset isotopy A
t
of A with A
= A
1
.
43.E. It is easy to see that this is an equivalence relation on the set of
subsets of X.
As it follows immediately from the deﬁnitions, any embedding isotopy
determines an isotopy of the image of the initial embedding and any
subset isotopy is accompanied with an embedding isotopy. However the
relation between the notions of subset isotopy and embedding isotopy is
not too close because of the following two reasons:
(a) an isotopy Φ
t
accompanying a subset isotopy A
t
starts with the
inclusion of A
0
(while arbitrary isotopy may start with any embed
ding);
(b) an isotopy accompanying a subset isotopy is determined by the sub
set isotopy only up to composition with an isotopy of the identity
homeomorphism A →A(an isotopy of a homeomorphism is a special
case of embedding isotopies, since homeomorphisms can be consid
ered as a sort of embeddings).
An isotopy of a subset A in X is said to be ambient, if it may be
accompanied with an embedding isotopy Φ
t
: A → X extendible to an
isotopy
˜
Φ
t
: X →X of the identity homeomorphism of the space X. The
isotopy
˜
Φ
t
is said to be ambient for Φ
t
. This gives rise to obvious reﬁne
ments of the equivalence relations for subsets and embeddings introduced
above.
43.F. Find isotopic, but not ambiently isotopic sets in [0, 1].
43.G. If sets A
1
, A
2
⊂ X are ambiently isotopic then the complements
X A
1
and X A
2
are homeomorphic and hence homotopy equivalent.
43.5. Find isotopic, but not ambiently isotopic sets in R.
43.6. Prove that any isotopic compact subsets of R are ambiently isotopic.
43.7. Find isotopic, but not ambiently isotopic compact sets in R
3
.
43.8. Prove that any two embeddings S
1
→ R
3
are isotopic. Find embed
dings S
1
→R
3
that are not ambiently isotopic.
'43
◦
3 Isotopies and Attaching
43:A. Any isotopy h
t
: ∂X →∂X extends to an isotopy H
t
: X →X.
43:B. Let X and Y be manifolds of dimension n, A and B components of ∂X and
∂Y respectively. Then for any isotopic homeomorphisms f, g : B → A the manifolds
X ∪
f
Y and X ∪
g
Y are homeomorphic.
§43. ISOTOPY 230
43:C. Let X and Y be manifolds of dimension n, let B be a compact subset of ∂Y . If
B is a manifold of dimension n−1 then for any embeddings f, g : B →∂X ambiently
isotopic in ∂X the manifolds X ∪
f
Y and X ∪
g
Y are homeomorphic.
'43
◦
4 Connected Sums
43.H. Let X and Y be manifolds of dimension n, and ϕ : R
n
→ X,
ψ : R
n
→Y be embeddings. Then
X ϕ(Int D
n
) ∪
ψ(S
n
)→Xϕ(Int D
n
):ψ(a)→ϕ(a)
Y ψ(Int D
n
)
is a manifold of dimension n.
This manifold is called a connected sum of X and Y .
43.I. Show that the topological type of the connected sum of X and Y
depends not only on the topological types of X and Y .
43.J. Let X and Y be manifolds of dimension n, and ϕ : R
n
→ X,
ψ : R
n
→ Y be embeddings. Let h : X → X be a homeomorphism.
Then the connected sums of X and Y deﬁned via ψ and ϕ, on one hand,
and via ψ and h ◦ ϕ, on the other hand, are homeomorphic.
43.9. Find pairs of manifolds connected sums of which are homeomorphic
to
(a) S
1
,
(b) Klein bottle,
(c) sphere with three crosscaps.
43.10. Find a disconnected connected sum of connected manifolds. Describe,
under what circumstances this can happen.
'44 OneDimensional Manifolds
'44
◦
1 ZeroDimensional Manifolds
This section is devoted to topological classiﬁcation of manifolds of
dimension one. We skip the case of 0dimensional manifolds due to trivi
ality of the problem. Indeed, any 0dimensional manifold is just a count
able discrete topological space, and the only topological invariant needed
for topological classiﬁcation of 0manifolds is the number of points: two
0dimensional manifolds are homeomorphic, iﬀ they have the same num
ber of points.
The case of 1dimensional manifolds is also simple, but it requires
more detailed consideration.
'44
◦
2 Reduction to Connected Manifolds
Since each manifold is the sum of its connected components, two man
ifolds are homeomorphic if and only if there exists a onetoone correspon
dence between their components such that the corresponding components
are homeomorphic. Therefore for topological classiﬁcation of nmanifolds
it suﬃces to classify only connected nmanifolds.
'44
◦
3 Examples
44.A. What connected 1manifolds do you know?
(a) Do you know any closed connected 1manifold?
(b) Do you know a connected compact 1manifold, which is not closed?
(c) What noncompact connected 1manifolds do you know?
(d) Is there a noncompact connected 1manifolds with boundary?
44.B. Fill the following table with pluses and minuses.
Manifold X Is X compact? Is ∂X empty?
S
1
R
1
I
R
1
+
'44
◦
4 Statements of Main Theorems
44.C. Any connected manifold of dimension 1 is homeomorphic to one
of the following for manifolds:
• circle S
1
,
• line R
1
,
• interval I,
231
§44. ONEDIMENSIONAL MANIFOLDS 232
• halfline R
1
+
.
This theorem may be splitted into the following four theorems:
44.D. Any closed connected manifold of dimension 1 is homeomorphic
to circle S
1
.
44.E. Any noncompact connected manifold of dimension 1 without bound
ary is homeomorphic to line R
1
.
44.F. Any compact connected manifold of dimension 1 with nonempty
boundary is homeomorphic to interval I.
44.G. Any noncompact connected manifold of dimension one with non
empty boundary is homeomorphic to halfline R
1
+
.
'44
◦
5 Lemma on 1Manifold Covered with Two Lines
44.H Lemma. Any connected manifold of dimension 1 covered with two
open sets homeomorphic to R
1
is homeomorphic either to R
1
, or S
1
.
Let X be a connected manifold of dimension 1 and U, V ⊂ X be its
open subsets homeomorphic to R. Denote by W the intersection U ∩ V .
Let ϕ : U →R and ψ : V →R be homeomorphisms.
44.H.1. Prove that each connected component of ϕ(W) is either an open
interval, or an open ray, or the whole R.
44.H.2. Prove that a homeomorphism between two open connected subsets
of R is a (strictly) monotone continuous function.
44.H.3. Prove that if a sequence x
n
of points of W converges to a point
a ∈ U W then it does not converge in V .
44.H.4. Prove that if there exists a bounded connected component C of
ϕ(W) then C = ϕ(W), V = W, X = U and hence X is homeomorphic to
R.
44.H.5. In the case of connected W and U = V , construct a homeomor
phism X →R which takes:
• W to (0, 1),
• U to (0, +∞), and
• V to (−∞, 1).
44.H.6. In the case of W consisting of two connected components, construct
a homeomorphism X →S
1
, which takes:
• W to ¦z ∈ S
1
: −1/
√
2 < Im(z) < 1/
√
2¦,
• U to ¦z ∈ S
1
: −1/
√
2 < Im(z)¦, and
• V to ¦z ∈ S
1
: Im(z) < 1/
√
2¦.
§44. ONEDIMENSIONAL MANIFOLDS 233
'44
◦
6 Without Boundary
44.D.1. Deduce Theorem 44.D from Lemma 44.G.
44.E.1. Deduce from Lemma 44.G that for any connected noncompact
onedimensional manifold X without a boundary there exists an embedding
X →R with open image.
44.E.2. Deduce Theorem 44.E from 44.E.1.
'44
◦
7 With Boundary
44.F.1. Prove that any compact connected manifold of dimension 1 can be
embedded into S
1
.
44.F.2. List all connected subsets of S
1
.
44.F.3. Deduce Theorem 44.F from 44.F.2, and 44.F.1.
44.G.1. Prove that any noncompact connected manifold of dimension 1
can be embedded into R
1
.
44.G.2. Deduce Theorem 44.G from 44.G.1.
'44
◦
8 Consequences of Classiﬁcation
44.I. Prove that connected sum of closed 1manifolds is deﬁned up home
omorphism by topological types of summands.
44.J. Which 0manifolds bound a compact 1manifold?
'44
◦
9 Mapping Class Groups
44.K. Find the mapping class groups of
(a) S
1
,
(b) R
1
,
(c) R
1
+
,
(d) [0, 1].
44.1. Find the mapping class group of an arbitrary 1manifold with ﬁnite
number of components.
'45 TwoDimensional Manifolds: General Picture
'45
◦
1 Examples
45.A. What connected 2manifolds do you know?
(a) List closed connected 2manifold that you know.
(b) Do you know a connected compact 2manifold, which is not closed?
(c) What noncompact connected 2manifolds do you know?
(d) Is there a noncompact connected 2manifolds with nonempty bound
ary?
45.1. Construct nonhomeomorphic noncompact connected manifolds of di
mension two without boundary and with isomorphic inﬁnitely generated fun
damental group.
For notions relevant to this problem see what follows.
'45
◦
2 Ends and Odds
Let X be a noncompact Hausdorﬀ topological space, which is a union of an
increasing sequence of its compact subspaces
C
1
⊂ C
2
⊂ ⊂ C
n
⊂ ⊂ X.
Each connected component U of XC
n
is contained in some connected component of
XC
n−1
. A decreasing sequence U
1
⊃ U
2
⊃ ⊃ U
n
⊃ . . . of connected components
of
(X C
1
) ⊃ (X C
2
) ⊃ ⊃ (X C
n
) ⊃ . . .
respectively is called an end of X with respect to C
1
⊂ ⊂ C
n
⊂ . . . .
45:A. Let X and C
n
be as above, D be a compact set in X and V a connected
component of X D. Prove that there exists n such that D ⊂ C
n
.
45:B. Let X and C
n
be as above, D
n
be an increasing sequence of compact sets of
X with X = ∪
∞
n=1
D
n
. Prove that for any end U
1
⊃ ⊃ U
n
⊃ . . . of X with respect
to C
n
there exists a unique end V
1
⊃ ⊃ V
n
⊃ . . . of X with respect to D
n
such
that for any p there exists q such that V
q
⊂ U
p
.
45:C. Let X, C
n
and D
n
be as above. Then the map of the set of ends of X with
respect to C
n
to the set of ends of X with respect to D
n
deﬁned by the statement of
45:B is a bijection.
Theorem 45:C allows one to speak about ends of X without specifying a system
of compact sets
C
1
⊂ C
2
⊂ ⊂ C
n
⊂ ⊂ X
with X = ∪
∞
n=1
C
n
. Indeed, 45:B and 45:C establish a canonical onetoone corre
spondence between ends of X with respect to any two systems of this kind.
45:D. Prove that R
1
has two ends, R
n
with n > 1 has one end.
45:E. Find the number of ends for the universal covering space of the bouquet of two
circles.
45:F. Does there exist a 2manifold with a ﬁnite number of ends which cannot be
embedded into a compact 2manifold?
234
§45. TWODIMENSIONAL MANIFOLDS: GENERAL PICTURE 235
45:G. Prove that for any compact set K ⊂ S
2
with connected complement S
2
K
there is a natural map of the set of ends of S
2
K to the set of connected components
of K.
Let W be an open set of X. The set of ends U
1
⊃ ⊃ U
n
⊃ . . . of X such that
U
n
⊂ W for suﬃciently large n is said to be open.
45:H. Prove that this deﬁnes a topological structure in the set of ends of X.
The set of ends of X equipped with this topological structure is called the space
of ends of X. Denote this space by c(X).
45.1.1. Construct noncompact connected manifolds of dimension two with
out boundary and with isomorphic inﬁnitely generated fundamental group,
but with nonhomeomorphic spaces of ends.
45.1.2. Construct noncompact connected manifolds of dimension two with
out boundary and with isomorphic inﬁnitely generated fundamental group,
but with diﬀerent number of ends.
45.1.3. Construct noncompact connected manifolds of dimension two with
out boundary with isomorphic inﬁnitely generated fundamental group and
the same number of ends, but with diﬀerent topology in the space of ends.
45.1.4. Let K be a completely disconnected closed set in S
2
. Prove that
the map c(S
2
K) →K deﬁned in 45:G is continuous.
45.1.5. Construct a completely disconnected closed set K ⊂ S
2
such that
this map is a homeomorphism.
45:I. Prove that there exists an uncountable family of pairwise nonhomeomorphic
connected 2manifolds without boundary.
The examples of noncompact manifolds dimension 2 presented above
show that there are too many noncompact connected 2manifolds. This
makes impossible any really useful topological classiﬁcation of noncompact
2manifolds. Theorems reducing the homeomorphism problem for 2
manifolds of this type to the homeomorphism problem for their spaces
of ends do not seem to be useful: spaces of ends look not much simpler
than the surfaces themselves.
However, there is a special class of noncompact 2manifolds, which
admits a simple and useful classiﬁcation theorem. This is the class of
simply connected noncompact 2manifolds without boundary. We post
pone its consideration to section ??. Now we turn to the case, which is
the simplest and most useful for applications.
'45
◦
3 Closed Surfaces
45.B. Any connected closed manifold of dimension two is homeomorphic
either to sphere S
2
, or sphere with handles, or sphere with crosscaps.
§45. TWODIMENSIONAL MANIFOLDS: GENERAL PICTURE 236
Recall that according to Theorem 39.L the basic surfaces represent
pairwise distinct topological (and even homotopy) types. Therefore, 39.L
and 45.B together give topological and homotopy classiﬁcations of closed
2dimensional manifolds.
We do not recommend to prove Theorem 45.B immediately and, es
pecially, in the form given above. All known proofs of 45.B can be de
composed into two main stages: ﬁrstly, a manifold under consideration
is equipped with some additional structure (like triangulation or smooth
structure); then using this structure a required homeomorphism is con
structed. Although the ﬁrst stage appears in the proof necessarily and
is rather diﬃcult, it is not useful outside the proof. Indeed, any closed
2manifold, which we meet in a concrete mathematical context, is either
equipped, or can be easily equipped with the additional structure. The
methods of imposing the additional structure are much easier, than a
general proof of existence for this structure in arbitrary 2manifold.
Therefore, we suggest for the ﬁrst case to restrict ourselves to the
second stage of the proof of Theorem 45.B, prefacing it with general
notions related to the most classical additional structure, which can be
used for this purpose.
'46 Triangulations
'46
◦
1 Triangulations of Surfaces
By an Euclidean triangle we mean the convex hall of three noncollinear
points of Euclidean space. Of course, it is homeomorphic to disk D
2
, but
not only the topological structure is relevant for us now. The boundary
of a triangle contains three distinguished points, its vertices, which di
vide the boundary into three pieces, its sides. A topological triangle in a
topological space X is an embedding of an Euclidean triangle into X. A
vertex (respectively, side) of a topological triangle T → X is the image
of a vertex ( respectively, side) of T in X.
A set of topological triangles in a 2manifold X is a triangulation of
X provided the images of these triangles form a fundamental cover of X
and any two of the images either are disjoint or intersect in a common
side or in a common vertex.
46.A. Prove that in the case of compact X the former condition (about
fundamental cover) means that the number of triangles is ﬁnite.
46.B. Prove that the condition about fundamental cover means that the
cover is locally ﬁnite.
'46
◦
2 Triangulation as cellular decomposition
46.C. A triangulation of a 2manifold turns it into a cellular space, 0
cells of which are the vertices of all triangles of the triangulation, 1cells
are the sides of the triangles, and 2cells are the interiors of the triangles.
This result allows us to apply all the terms introduced above for cellu
lar spaces. In particular, we can speak about skeletons, cellular subspaces
and cells. However,in the latter two cases we rather use terms triangu
lated subspace and simplex. Triangulations and terminology related to
them appeared long before cellular spaces appeared. Therefore in this
context the adjective cellular is replaced usually by adjectives triangulated
or simplicial.
'46
◦
3 Two Properties of Triangulations of Surfaces
46.D Unramiﬁed. Let E be a side of a triangle involved into a trian
gulation of a 2manifold X. Prove that there exist at most two triangles
of this triangulation for which E is a side. Cf. 41.G, 41.H and 41.P.
46.E Local strong connectedness. Let V be a vertex of a triangle
involved into a triangulation of a 2manifold X and T, T
be two triangles
of the triangulation adjacent to V . Prove that there exisits a sequence
T = T
1
, T
2
, . . . , T
n
= T
of triangles of the triangulation such that V is a
237
§46. TRIANGULATIONS 238
vertex of each of them and triangles T
i
, T
i+1
have common side for each
i = 1, . . . , n −1.
'46
◦
4 Scheme of Triangulation
Triangulations
present a surface
combinatorially.
Let X be a 2manifold and T a triangulation of X. Denote the set of vertices of
T by V . Denote by Σ
2
the set of triples of vertices, which are vertices of a triangle
of T . Denote by Σ
1
the set of pairs of vertices, which are vertices of a side of T . Put
Σ
0
= S. This is the set of vertices of T . Put Σ = Σ
2
∪ Σ
1
∪ Σ
0
. The pair (V, Σ) is
called the (combinatorial) scheme of T .
46:A. Prove that the combinatorial scheme (V, Σ) of a triangulation of a 2manifold
has the following properties:
(a) Σ is a set consisting of subsets of V ,
(b) each element of Σ consists of at most 3 elements of V ,
(c) threeelement elements of Σ cover V ,
(d) any subset of an element of Σ belongs to Σ,
(e) intersection of any collection of elements of Σ belongs to Σ,
(f) for any twoelement element of Σ there exist exactly two threeelement elements
of Σ containing it.
Recall that objects of this kind appeared above, in Section '19
◦
10. Let V be
a set and Σ is a set of ﬁnite subsets of V . The pair (V, Σ) is called a triangulation
scheme if
• any subset of an element of Σ belongs to Σ,
• intersection of any collection of elements of Σ belongs to Σ,
• any one element subset of V belongs to Σ.
For any simplicial scheme (V, Σ) in '19
◦
10 a topological space S(V, Σ) was con
structed. This is, in fact, a cellular space, see 37:A.
46:B. Prove that if (V, Σ) is the combinatorial scheme of a triangulation of a 2
manifold X then S(V, Σ) is homeomorphic to X.
46:C. Let (V, Σ) be a triangulation scheme such that
(a) V is countable,
(b) each element of Σ consists of at most 3 elements of V ,
(c) threeelement elements of Σ cover V ,
(d) for any twoelement element of Σ there exist exactly two threeelement elements
of Σ containing it
Prove that (V, Σ) is a combinatorial scheme of a triangulation of a 2manifold.
'46
◦
5 Examples
46.1. Consider the cover of torus obtained in the obvious way from the cover
of the square by its halves separated by a diagonal of the square.
Is it a triangulation of torus? Why not?
46.2. Prove that the simplest triangulation of S
2
consists of 4 triangles.
§46. TRIANGULATIONS 239
46.3*. Prove that a triangulation of torus S
1
S
1
contains at least 14
triangles, and a triangulation of the projective plane contains at least 10
triangles.
'46
◦
6 Subdivision of a Triangulation
A triangulation o of a 2manifold X is said to be a subdivision of a
triangulation T , if each triangle of o is contained in some triangle
1
of T .
Then o is also called a reﬁnement of T .
There are several standard ways to subdivide a triangulation. Here
is one of the simplest of them. Choose a point inside a triangle τ, call
it a new vertex, connect it by disjoint arcs with vertices of τ and call
these arcs new edges. These arcs divide τ to three new triangles. In
the original triangulation replace τ by these three new triangles. This
operation is called a star subdivision centered at τ. See Figure 6.
τ
Figure 1. Star subdivision centered at triangle τ
.
46.F. Give a formal description of a star subdivision centered at a tri
angle τ. I.e., present it as a change of a triangulation thought of as a
collection of topological triangles. What three embeddings of Euclidean
triangles are to replace τ? Show that the replacement gives rise to a tri
angulation. Describe the corresponding operation on the combinatorial
scheme.
Here is another subdivision deﬁned locally. One adds a new vertex
taken on an edge ε of a given triangulation. One connects the new vertex
by two new edges to the vertices of the two tringles adjacent to ε. The
new edges divide these triangles, each to two new triangles. The rest of
triangles of the original triangulation are not aﬀected. This operation is
called a star subdivision centered at ε. See Figure 7.
1
Although triangles which form a triangulation of X have been deﬁned as topo
logical embeddings, we hope that a reader guess that when one of such triangles is
said to be contained in another one this means that the image of the embedding which
is the former triangle is contained in the image of the other embedding which is the
latter.
§46. TRIANGULATIONS 240
ε
Figure 2. Star subdivision centered at edge ε.
46.G. Give a formal description of a star subdivision centered at edge
ε. What four embeddings of Euclidean triangles are to replace the topo
logical triangles with edge ε? Show that the replacement gives rise to a
triangulation. Describe the corresponding operation on the combinato
rial scheme.
46.4. Find a triangulation and its subdivision, which cannot be presented
as a composition of star subdivisions at edges or triangles.
46.5*. Prove that any subdivision of a triangulation of a compact surface
can be presented as a composition of a ﬁnite sequences ao star subdivisions
centered at edges or triangles and operations inverse to such subdivisions.
By a baricentric subdivision of a triangle we call a composition of a
star subdivision centered at this tringle followed by star subdivisions at
each of its edges. See Figure 3.
Figure 3. Baricentric subdivision of a triangle.
Baricentric subdivision of a triangulation of 2manifold is a subdivision
which is a simultaneous baricentric subdivision of all triangles of this
triangulation. See Figure 4.
ε
Figure 4. Baricentric subdivision of a triangulation.
§46. TRIANGULATIONS 241
46.H. Establish a natural onetoone correspondence between vertices
of a baricentric subdivision a simplices (i.e., vertices, edges and triangles)
of the original tringulation.
46.I. Establish a natural onetoone correspondence between triangles of
a baricentric subdivision and triples each of which is formed of a triangle
of the original triangulation, an edge of this triangle and a vertex of this
edge.
The expression baricentric subdivision has appeared in a diiferent context, see
Section '19. Let us relate the two notions sharing this name .
46:D Baricentric subdivision of a triangulation and its scheme. Prove that
the combinatorial scheme of the baricentric subdivision of a triangulation of a 2
manifold coincides with the baricentric subdivision of the scheme of the original tri
angulation (see '19
◦
11).
'46
◦
7 Homotopy Type of Compact Surface with NonEmpty
Boundary
46.J. Any compact connected triangulated 2manifold with nonempty
boundary collapses to a onedimensional simplicial subspace.
46.K. Any compact connected triangulated 2manifold with nonempty
boundary is homotopy equivalent to a bouquet of circles.
46.L. The Euler characteristic of a triangulated compact connected 2
manifold with nonempty boundary does not depend on triangulation. It
is equal to 1 − r, where r is the rank of the onedimensional homology
group of the 2manifold.
46.M. The Euler characteristic of a triangulated compact connected 2
manifold with nonempty boundary is not greater than 1.
46.N. The Euler characteristic of a triangulated closed connected 2
manifold with nonempty boundary is not greater than 2.
'46
◦
8 Triangulations in dimension one
By an Euclidean segment we mean the convex hall of two diﬀerent
points of a Euclidean space. It is homeomorphic to I. A topological
segment or topological edge in a topological space X is a topological
embedding of an Euclidean segment into X. A set of topological segments
in a 1manifold X is a triangulation of X if the images of these topological
segments constitute a fundamental cover of X and any two of the images
either are disjoint or intersect in one common end point.
Traingulations of 1manifolds are similar to triangulations of 2manifolds
considered above.
§46. TRIANGULATIONS 242
46.O. Find counterparts for theorems above. Which of them have no
counterpart? What is a counterpart for the property 46.D? What are
counterparts for star and baricentric subdivisions?
46.P. Find homotopy classiﬁcation of triangulated compact 1manifolds
using arguments similar to the ones from Section '46
◦
7. Compare with
the topological classiﬁcation of 1manifolds obtained in Section '44.
46.Q. What values take the Euler characteristic on compact 1manifolds?
46.R. What is relation of the Euler characteristic of a compact triangu
lated 1manifold X and the number of ∂X?
46.S. Triangulation of a 2manifold X gives rise to a triangulation of
its boundary ∂X. Namely, the edges of the triangualtion of ∂X are the
sides of triangles of the original triangulation which lie in ∂X.
'46
◦
9 Triangualtions in higher dimensions
46.T. Generalize everything presented above in this section to the case
of manifolds of higher dimensions.
'47 Handle Decomposition
'47
◦
1 Handles and Their Anatomy
Together with triangulations, it is useful to consider representations
of a manifold as a union of balls of the same dimension, but adjacent to
each other as if they were thickening of cells of a cellular space
A space D
p
D
n−p
is called a (standard) handle of dimension n and
index p. Its subset D
p
¦0¦ ⊂ D
p
D
n−p
is called the core of handle
D
p
D
n−p
, and a subset ¦0¦ D
n−p
⊂ D
p
D
n−p
is called its cocore.
The boundary ∂(D
p
D
n−p
) = of the handle D
p
D
n−p
can be presented
as union of its base D
p
S
n−p−1
and cobase S
p−1
D
n−p
.
47.A. Draw all standard handles of dimensions ≤ 3.
A topological embedding h of the standard handle D
p
D
n−p
of
dimension n and index p into a manifold of the same dimansion n is
called a handle of dimension n and index p. The image under h of Int D
p
Int D
n−p
is called the interior of h, the image of the core h(D
p
¦0¦) of
the standard handle is called the core of h, the image h(¦0¦ D
n−p
) of
cocore, the cocore, etc.
'47
◦
2 Handle Decomposition of Manifold
Let X be a manifold of dimension n. A collecton of ndimensional
handles in X is called a handle decomposition of X, if
(a) the images of these handles constitute a locally ﬁnite cover of X,
(b) the interiors of these handles are pairwise disjoint,
(c) the base of each of the handles is contained in the union of cobases
of the handles of smaller indices.
Let X be a manifold of dimension n with boundary. A collection of
ndimensional handles in X is called a handle decomposion of X modulo
boundary, if
(a) the images of these handles constitute a locally ﬁnite cover of X,
(b) the interiors of these handles are pairwise disjoint,
(c) the base of each of the handles is contained in the union of ∂X and
cobases of the handles of smaller indices.
A composition of a handle h : D
p
D
n−p
→X with the homeomor
phism of transposition of the factors D
p
D
n−p
→D
n−p
D
p
turns the
handle h of index p into a handle of the same dimension n, but of the
complementary index n−p. The core of the handle turns into the cocore,
while the base, to cobase.
47.B. Composing each handle with the homeomorphism transposing the
factors turns a handle decomposition of manifold into a handle decom
position modulo boundary of the same manifold. Vice versa, a handle
243
§47. HANDLE DECOMPOSITION 244
decomposition modulo boundary turns into a handle decomposition of
the same manifold.
Handle decompositions obtained from each other in this way are said
to be dual to each other.
47.C. Enigma. For ndimensional manifold with boundary split into
two (n − 1)dimensional manifolds with disjoint closures, deﬁne handle
decomposition modulo one of these manifolds so that the dual handle de
composition would be modulo the complementary part of the boundary.
47.1. Find handle decompositions with a minimal number of handles for the
following manifolds:
(a) circle S
1
; (b) sphere S
n
; (c) ball D
n
(d) torus S
1
S
1
; (e) handle; (f) cylinder S
1
I;
(g) M¨obius band; (h) projective plane
RP
2
;
(i) projective space
RP
n
;
(j) sphere with p
handles;
(k) sphere with p
crosscaps;
(l) sphere with n
holes.
'47
◦
3 Handle Decomposition and Triangulation
Let X be a 2manifold, τ its triangulation, τ
its baricentric subdivi
sion, and τ
the baricentric subdivision of τ
. For each simplex S of τ
denote by H
S
the union of all simplices of τ
which contain the unique
vertex of τ
that lies in
S. Thus, if S is a vertex then H
S
is the union
of all triangles of τ
containing this vertex, if S is an edge then H
S
is the
union all of the triangles of τ
which intersect with S but do not contain
any of its vertices, and, ﬁnally, if S is a triangle of τ then H
S
is the union
of all triangles of τ
which lie in S but do not intersect its boundary.
47.D Handle Decomposition out of Triangulation. Sets H
S
con
stitute a handle decomposition of X. The index of H
S
equals the dimen
sion of S.
Figure 5. Construction of a handle decomposition from
a triangulation.
47.E. Can every handle decomposition of a 2manifold be constructed
from a triangulation as indicated in 47.D?
§47. HANDLE DECOMPOSITION 245
47.F. How to triangulate a 2manifold which is equipped with a handle
decomposition?
'47
◦
4 Regular Neighborhoods
Let X be a 2manifold, τ its triangulation, and A be a simplicial
subspace of X. The union of all those simplices of the double baricen
tric subdivision τ
of τ which intersect A is called the regular or second
baricentric neighborhood of A (with respect to τ).
Of course, usually regular neighborhood is not open in X, since it is
the union of simplices, wich are closed. So, it is a neighborhood of A
only in wide sence (its interior contains A).
47.G. A regular neighborhood of A in X is a 2manifold. It coincides
with the union of handles corresponding to the simplices contained in A.
These handles constitute a handle decomposition of the regular neigh
borhood.
47.H Collaps Induces Homemorphism. Let X be a triangulated
2manifold and A ⊂ X be its triangulated subspace. If X ` A then X
is homeomorphic to a regular neighborhood of A.
47.I. Any triangulated compact connected 2manifold with nonempty
boundary is homeomorphic to a regular neighborhood of some of its 1
dimensional triangulated subspaces.
47.J. In a triangulated 2manifold, any triangulated subspace which is
a tree has regular neighborhood homeomorphic to disk.
47.K. In a triangulated 2manifold, any triangulated subspace homeo
morphic to circle has regular neighborhood homeomorphic either to the
M¨obius band or cylinde S
1
I.
In the former case the circle is said to be onesided, in the latter,
twosided.
'47
◦
5 Cutting 2Manifold Along a Curve
47.L Cut Along Curve. Let F be a triangulated surface and C ⊂ F
be a compact onedimensional manifold contained in the 1skeleton of
F and satisfying condition ∂C = ∂F ∩ C. Prove that there exists a
2manifold T and surjective map p : T →F such that:
(a) p[ : T p
−1
(C) →F C is a homeomorphism,
(b) p[ : p
−1
(C) →C is a twofold covering.
47.M Uniqueness of Cut. The 2manifold T and map p which exist
according to Theorem 50:I, are unique up to homeomorphism: if
˜
T and ˜ p
are other 2manifold and map satisfying the same hypothesis then there
exists a homeomorphism h :
˜
T →T such that p ◦ h = ˜ p.
§47. HANDLE DECOMPOSITION 246
The 2manifold T described in 50:I is called the result of cutting of
F along C. It is denoted by F C. This is not at all the complement
F C, although a copy of F C is contained in F C as a dense subset
homotopy equivalent to the whole F C.
47.N Triangulation of Cut Result. F C possesses a unique trian
gulation such that the natural map F C →F maps homeomorphically
edges and triangles of this triangulation onto edges and, respectivly, tri
angles of the original triangulation of F.
47.O. Let X be a triangulated 2manifold, C be its triangulated sub
space homeomorphic to circle, and let F be a regular neighborhood of C
in X. Prove
(a) F C consists of two connected components, if C is twosided on
X, it is connected if C is onesided;
(b) the inverse image of C under the natural map X C →X consists
of two connected components if C is twosided on X, it is connected
if C is onesided on X.
This proposition discloses the meaning of words onesided and two
sided circle on a 2manifold. Indeed, both connected components of the
result of cutting of a regular neighborhood, and connected components
of the inverse image of the circle can claim its right to be called a side of
the circle or a side of the cut.
47.2. Describe the topological type of F C for the following F and C:
(a) F is sphere S
2
, and C is its equator;
(b) F is a M¨obius strip, and C is its middle circle (deformation retract);
(c) F = S
1
S
1
, C = S
1
1;
(d) F is torus S
1
S
1
standardly embedded into R
3
, and C is the trefoil
knot lying on F, that is ¦(z, w) ∈ S
1
S
1
[ z
2
= w
3
¦;
(e) F is a M¨obius strip, C is a segment: ﬁndtwo topologically diﬀerent
position of C on F and describe F C for each of them;
(f) F = RP
2
, C = RP
1
.
(g) F = RP
2
, C is homeomorphic to circle: ﬁnd two topologically diﬀerent
position C on F and describe F C for each of them.
47.P Euler Characteristic and Cut. Let F be a triangulated com
pact 2manifold and C ⊂
F be a closed onedimensional contained in
the 1skeleton of the triangulation of F. Then χ(F C) = χF.
47.Q. Find the Euler characteristic of F C, if ∂C = ∅.
47.R Generalized Cut (Incise). Let F be a triangulated 2manifold
and C ⊂ F be a compact 1dimensional manifold contained in 1skeleton
of F and satisfying condition ∂F ∩ C ⊂ ∂C. Let D = C (∂C ∂F).
Prove that there exist a 2manifold T and sujective continuous map p :
T →F such that:
(a) p[ : T p
−1
(D) →F D is a homeomorphism,
§47. HANDLE DECOMPOSITION 247
(b) p[ : p
−1
(D) →D is a twofold covering.
47.S Uniqueness of Cut. The 2manifold T and map p, which exist
according to Theorem 47.R, ae unique up to homeomorphism: if
˜
T and ˜ p
are other 2manifold and map satisfying the same hypothesis then there
exists a homeomorphism h :
˜
T →T such that p ◦ h = ˜ p.
The 2Manifold T described in 47.R is also called the result of cutting
of F along C and denoted by F C.
47.3. Show that if C is a segment contained in the interior of a 2manifold
F then F C is homeomorphic to F Int B, where B is the subset of
F
homeomorphic to disk.
47.4. Show that if C is a segment such that one of its end points is in
F
and the other one is on ∂F then F C is homeomorphic to F.
'48 Orientations
'48
◦
1 Orientations of Edges and Triangles
By an orientation of a segment one means an order of its end points
(which of them is its initial point and which is the ﬁnal one). By an
orientation of a triangle one means a collection of orientations of its sides
such that each vertex of the triangle is the ﬁnal point for one of the sides
adjacent to it and initial point for the other side. Thus, an orientation of
a triangle deﬁnes an orientation on each of its sides. A segment admits
two orientations. A triangle also admits two orientations: one is obained
from another one by change of the orientation on each side of the triangle.
Therefore an orientation of any side of a triangle deﬁnes an orientation
of the triangle.
Vertices of an oriented triangle are cyclicly ordered: a vertex A follows
immediately the vertex B which is the initial vertex of the side which
ﬁnishes at A. Similarly the sides of an oriented triangle are cyclicly
ordered: a side a follows immediately the side b which ﬁnal end point is
the initial point of a.
Vice versa, each of these cyclic orders deﬁnes an orientation of the
triangle.
'48
◦
2 Orientation of Triangulation
An orientation of a triangulation of a 2manifold is a collection of ori
entations of all tringles constituting the triangulation such that for each
edge the orientations deﬁned on it by the orientations of the two adja
cent triangles are opposite to each other. A triangulation is said to be
orientable, if it admits an orientation.
48.A Number of Triangulations. A triangulation of a connected
2manifold is either nonorientable or admits exactly two orientations.
These two orientations are opposite to each other. Each of them can be
recovered from the orientation of any triangle involved in the triangula
tion.
48.B Lifting of Triangulation. Let B be a triangulated surface and
p : X →B be a covering. Can you equip X with a triangulation?
48.C. Let B be an oriented triangulated surface and p : X → B be a
covering. Equip X with a triangulation such that p maps each simplex
of it homeomorphically onto a simplex of the original triangulation of B.
Is this triangulation orientable?
48.D. Let X be a triangulated surface, C ⊂ X be a 1dimensional
manifold contained in 1skeleton of X.
248
'49 Classical Approach to Topological Classiﬁcation
of Compact surfaces
'49
◦
1 Families of Polygons
The problems considered above show that triangulations provide a
combinatorial description of 2dimensional manifolds, but this descrip
tion is usually too bulky. Here we will study other, more practical way
to present 2dimensional manifolds combinatorially. The main idea is to
use larger building blocks.
Let T be a collection of convex polygons P
1
, P
2
, . . . . Let the sides
of these polygons be oriented and paired oﬀ. Then we say that this is
a family of polygons. There is a natural quotient space of the sum of
polygons involved in a family: one identiﬁes each side with its pairmate
by a homeomorphism, which respects the orientations of the sides. This
quotient space is called just the quotient of the family.
49.A. Prove that the quotient of the family of polygons is a 2manifold
without boundary.
49.B. Prove that the topological type of the quotient of a family does not
change when the homeomorphism between the sides of a distinguished
pair is replaced by other homeomorphism which respects the orientations.
49.C. Prove that any triangulation of a 2manifold gives rise to a family
of polygon whose quotient is homeomorphic to the 2manifold.
A family of polygons can be described combinatorially: Assign a letter
to each distinguished pair of sides. Go around the polygons writing down
the letters assigned to the sides and equipping a letter with exponent −1
if the side is oriented against the direction in which we go around the
polygon. At each polygon we write a word. The word depends on the
side from which we started and on the direction of going around the
polygon. Therefore it is deﬁned up to cyclic permutation and inversion.
The collection of words assigned to all the polygons of the family is called
a phrase associated with the family of polygons. It describes the family to
the extend suﬃcient to recovering the topological type of the quotient.
49.1. Prove that the quotient of the family of polygons associated with
phrase aba
−1
b
−1
is homeomorphic to S
1
S
1
.
49.2. Identify the topological type of the quotient of the family of polygons
associated with phrases
(a) aa
−1
;
(b) ab, ab;
(c) aa;
(d) abab
−1
;
(e) abab;
(f) abcabc;
249
§49. CLASSICAL APPROACH TO TOPOLOGICAL CLASSIFICATIONOF COMPACT SURFACES 250
(g) aabb;
(h) a
1
b
1
a
−1
1
b
−1
1
a
2
b
2
a
−1
2
b
−1
2
. . . a
g
b
g
a
−1
g
b
−1
g
;
(i) a
1
a
1
a
2
a
2
. . . a
g
a
g
.
49.D. A collection of words is a phrase associated with a family of poly
gons, iﬀ each letter appears twice in the words.
A family of polygons is called irreducible if the quotient is connected.
49.E. A family of polygons is irreducible, iﬀ a phrase associated with it
does not admit a division into two collections of words such that there is
no letter involved in both collections.
'49
◦
2 Operations on Family of Polygons
Although any family of polygons deﬁnes a 2manifold, there are many
families deﬁning the same 2manifold. There are simple operations which
change a family, but do not change the topological type of the quotient of
the family. Here are the most obvious and elementary of these operations.
(a) Simultaneous reversing orientations of sides belonging to one of the
pairs.
(b) Select a pair of sides and subdivide each side in the pair into two
sides. The orientations of the original sides deﬁne the orderings of
the halves. Unite the ﬁrst halves into one new pair of sides, and
the second halves into the other new pair. The orientations of the
original sides deﬁne in an obvious way orientations of their halves.
This operation is called 1subdivision. In the quotient it eﬀects in
subdivision of a 1cell (which is the image of the selected pair of
sides) into two 1cells. This 1cells is replaced by two 1cells and
one 0cell.
(c) The inverse operation to 1subdivision. It is called 1consolidation.
(d) Cut one of the polygons along its diagonal into two polygons. The
sides of the cut constitute a new pair. They are equipped with an
orientation such that gluing the polygons by a homeomorphism re
specting these orientations recovers the original polygon. This oper
ation is called 2subdivision. In the quotient it eﬀects in subdivision
of a 2cell into two new 2cells along an arc whose endpoints are
0cells (may be coinciding). The original 2cell is replaced by two
2cells and one 1cell.
(e) The inverse operation to 2subdivision. It is called 2consolidation.
'49
◦
3 Topological and Homotopy Classiﬁcation of Closed Sur
faces
49.F Reduction Theorem. Any ﬁnite irreducible family of polygons
can be reduced by the ﬁve elementary operations to one of the following
standard families:
§49. CLASSICAL APPROACH TO TOPOLOGICAL CLASSIFICATIONOF COMPACT SURFACES 251
(a) aa
−1
(b) a
1
b
1
a
−1
1
b
−1
1
a
2
b
2
a
−1
2
b
−1
2
. . . a
g
b
g
a
−1
g
b
−1
g
(c) a
1
a
1
a
2
a
2
. . . a
g
a
g
for some natural g.
49.G Corollary. Any triangulated closed connected manifold of dimen
sion 2 is homeomorphic to either sphere, or sphere with handles, or sphere
with crosscaps.
Theorems 49.G and 39.L provide classiﬁcations of triangulated closed
connected 2manifolds up to homeomorphisms and homotopy equiva
lence.
49.F.1 Reduction to Single Polygon. Any ﬁnite irreducible family of
polygons can be reduced by elementary operations to a family consisting
of a single polygon.
49.F.2 Cancellation. A family of polygons corresponding to a phrase
containing a fragment aa
−1
or a
−1
a, where a is any letter, can be trans
formed by elementary operations to a family corresponding to the phrase
obtained from the original one by erasing this fragment, unless the latter
is the whole original phrase.
49.F.3 Reduction to Single Vertex. An irreducible family of polygons
can be turned by elementary transformations to a polygon such that all its
vertices are projected to a single point of the quotient.
49.F.4 Separation of Crosscap. A family corresponding to a phrase
consisting of a word XaY a, where X and Y are words and a is a letter,
can be transformed to the family corresponding to the phrase bbY
−1
X.
49.F.5. If a family, whose quotient has a single vertex in the natural cell de
composition, corresponds to a phrase consisting of a word XaY a
−1
, where
X and Y are nonempty words and a is a letter, then X = UbU
and
Y = V b
−1
V
.
49.F.6 Separation of Handle. A family corresponding to a phrase con
sisting of a word UbU
aV b
−1
V
a
−1
, where U, U
, V , and V
are words
and a, b are letters, can be transformed to the family presented by phrase
dcd
−1
c
−1
UV
V U
.
49.F.7 Handle plus Crosscap Equals 3 Crosscaps. A family corre
sponding to phrase aba
−1
b
−1
ccX can be transformed by elementary trans
formations to the family corresponding to phrase abdbadX.
'49
◦
4 Recognizing Closed Surfaces
49.H. What is the topological type of the 2manifold, which can be
obtained as follows: Take two disjoint copies of disk. Attach three parallel
strips connecting the disks and twisted by π. The resulting surface S has
a connected boundary. Attach a copy of disk along its boundary by
a homeomorphism onto the boundary of the S. This is the space to
recognize.
§49. CLASSICAL APPROACH TO TOPOLOGICAL CLASSIFICATIONOF COMPACT SURFACES 252
49.I. Euler characteristic of the cellular space obtained as quotient of a
family of polygons is invariant under homotopy equivalences.
49.3. How can 49.I help to solve 49.H?
49.4. Let X be a closed connected surface. What values of χ(X) allow to
recover the topological type of X? What ambiguity is left for other values of
χ(X)?
'49
◦
5 Orientations
By an orientation of a segment one means an ordering of its end
points (which one of them is initial and which one is ﬁnal). By an orien
tation of a polygon one means orientation of all its sides such that each
vertex is the ﬁnal end point for one of the adjacent sides and initial for
the other one. Thus an orientation of a polygon includes orientation of
all its sides. Each segment can be oriented in two ways, and each polygon
can be oriented in two ways.
An orientation of a family of polygons is a collection of orientations
of all the polygons comprising the family such that for each pair of sides
one of the pairmates has the orientation inherited from the orientation
of the polygon containing it while the other pairmate has the orientation
opposite to the inherited orientation. A family of polygons is said to be
orientable if it admits an orientation.
49.5. Which of the families of polygons from Problem 49.2 are orientable?
49.6. Prove that a family of polygons associated with a word is orientable iﬀ
each letter appear in the word once with exponent −1 and once with exponent
1.
49.J. Orientability of a family of polygons is preserved by the elementary
operations.
A surface is said to be orientable if it can be presented as the quotient
of an orientable family of polygons.
49.K. A surface S is orientable, iﬀ any family of polygons whose quotient
is homeomorphic to S is orientable.
49.L. Spheres with handles are orientable. Spheres with crosscaps are
not.
'49
◦
6 More About Recognizing Closed Surfaces
49.7. How can the notion of orientability and 49.J help to solve 49.H?
49.M. Two closed connected manifolds of dimension two are homeo
morphic iﬀ they have the same Euler characteristic and either are both
orientable or both nonorientable.
§49. CLASSICAL APPROACH TO TOPOLOGICAL CLASSIFICATIONOF COMPACT SURFACES 253
'49
◦
7 Compact Surfaces with Boundary
As in the case of onedimensional manifolds, classiﬁcation of compact
twodimensional manifolds with boundary can be easily reduced to the
classiﬁcation of closed manifolds. In the case of onedimensional mani
folds it was very useful to double a manifold. In twodimensional case
there is a construction providing a closed manifold related to a compact
manifold with boundary even closer than the double.
49.N. Contracting to a point each connected component of the boundary
of a twodimensional compact manifold with boundary gives rise to a
closed twodimensional manifold.
49.8. A space homeomorphic to the quotient space of 49.N can be con
structed by attaching copies of D
2
one to each connected component of the
boundary.
49.O. Any connected compact manifold of dimension 2 with nonempty
boundary is homeomorphic either to sphere with holes, or sphere with
handles and holes, or sphere with crosscaps and holes.
49.P. Enigma. Generalize orientabilty to the case of nonclosed man
ifolds of dimension two. (Give as many generalization as you can and
prove that they are equivalent. The main criterium of success is that the
generalized orientability should help to recognize the topological type.)
49.Q. Two compact connected manifolds of dimension two are homeo
morphic iﬀ they have the same Euler characteristic, are both orientable
or both nonorientable and their boundaries have the same number of con
nected components.
'49
◦
8 Simply Connected Surfaces
49:A Theorem
∗
. Any simply connected noncompact manifold of dimension two
without boundary is homeomorphic to R
2
.
'50 OneDimensional mod2Homology of Surfaces
'50
◦
1 Polygonal Paths on Surface
Let F be a triagulated surface. A path s : I →F is said to be polygonal if s(I) is
contained in the onedimensional skeleton of the triangulation of F, the preimage of
any vertex of the triangulation is ﬁnite, and the restriction of s to a segment between
any two consequitive points which are mapped to vertices is an aﬃne homeomorphism
onto an edge of the triangulation. In terms of kinematics, a polygonal path represents
a moving point, which goes only along edges, does not stay anywhere, and, whenever
it appears on an edge, it goes along the edge with a constant speed to the opposite
endpoint. A circular loop l : S
1
→ F is said to be polygonal if the corresponding
path I
t→exp(2πit)
−→ S
1
l
−→F is polygonal.
50:A. Let F be a triagulated surface. Any path s : I → F connecting vertices of
the triangulation is homotopic to a polygonal path. Any circular loop l : S
1
→ F is
freely homotopic to a polygonal one.
A polygonal path is a combinatorial object:
50:B. To describe a polygonal path up to homotopy, it is enough to specify the order
in which it passes through vertices.
On the other hand, pushing a path to the onedimensional skeleton can create
new double points. Some edges may appear several time in the same edge.
50:1. Let F be a triangulated surface and α be an element of π
1
(F) diﬀerent
from 1. Prove that there exists a natural N such that for any n ≥ N each
polygonal loop representing α
n
passes through some edge of the triangulation
more than once.
'50
◦
2 Subdivisions of Triangulation
To avoid a congestion of paths on edges, one can add new edges, i.e., subdivide
the triangulation. Although an elementary operation on families of polygons applied
to a triangulation, gives rise to a family, which is not a triangulation, making several
elementary operations, one can get a new triangulation with more edges.
One triangulation of a surface is called a reﬁnement of another one if each triangle
of the former is contained in a triagle of the latter. There are several standard ways
to construct a reﬁnement of a triangulation.
For example, add a new vertex, which is located inside of a triangle τ of a given
triangulation, connect it with the vertices of this triangle with segments, which are
three new edges. The triangle is subdivided into three new triangles. The other tri
angles of the original triangulations are kept intact. This is called the star subdivision
centered at τ. See Figure 6.
Another kind of local subdivision: add a new vertex located on an edge ε of a
given triangulation, connect by new edges this vertex to the vertices opposite to ε of
the triangles adjacent to ε. Each of the adjacent triangles is subdived into two new
triangles. Leave the other triangles intact. This is a star subdivision centered at ε.
See Figure 7.
254
§50. ONEDIMENSIONAL mod2HOMOLOGY OF SURFACES 255
τ
Figure 6. Star subdivision centered at triangle τ
ε
Figure 7. Star subdivision centered at edge ε
50:2. Construct a triangulation and its subdivision which cannot be obtained
as a composition of star subdivisions centered at edges and triangles.
50:3. Prove that a subdivision of a triangulation of a compact surface can
be presented as a result of a ﬁnite sequence of star sudivisions centered at
triangles and edges and operations inverse to operations of these types.
'50
◦
3 Bringing Loops to General Position
50:C. Let F be a triangulated and u, v polygonal circular loops on F. Then there
exist a subdivision of the triangulation of F and polygonal loops u
, v
homotopic to
u and v, respectively, such that u
(I) ∩ v
(I) is ﬁnite.
50:D. Let F be a triangulated and u a polygonal circular loop on F. Then there
exist a subdivision of the triangulation of F and a polygonal loop v homotopic to u
such that v maps the preimage v
−1
(ε) of any edge ε ⊂ v(I) homeomorphically onto
ε. (In other words, v passes along each edge at most once).
Let u, v be polygonal circular loops on a triangulated surface F and a be an
isolated point of u(I) ∩ v(I). Suppose u
−1
(a) and v
−1
(a) are one point sets. One
says that u intersects v translversally at a if there exist a neighborhood U of a in F
and a homeomorphism U → R
2
which maps u(I) ∩ U onto the xaxes and v(I) ∩ U
to yaxes.
Polygonal circular loops u, v on a triangulated surface are said to be in general
position to with respect each other, if u(I)∩v(i) is ﬁnite, for each point a ∈ u(i)∩v(I)
each of the sets u
−1
(a) and v
−1
(a) contains a single point and u, v are transversal at
a.
§50. ONEDIMENSIONAL mod2HOMOLOGY OF SURFACES 256
50:E. Any two circular loops on a triangulated surface are homotopic to circular
loops, which are polygonal with respect to some subdivision of the triangulation and
in general position with respect to each other.
For a map f : X →Y denote by S
k
(f) the set
¦a ∈ X [ f
−1
f(a) consists of k elements¦
and put
S(f) = ¦a ∈ X [ f
−1
f(a) consists of more than 1 element¦.
A polygonal circular loop l on a triangulated surface F is said to be generic if
(a) S(l) is ﬁnite,
(b) S(l) = S
2
(l),
(c) at each a ∈ l(S
2
(l)) the two branches of s(I) intersecting at a are transversal,
that is a has a neighborhood U in F such that there exists a homeomorphism
U → R
2
mapping the images under s of the connected components of s
−1
(U)
to the coordinate axis.
50:F. Any circular loop on a triangulated surface is homotopic to a circular loop,
which is polygonal with respect to some subdivision of the triangulation and generic.
Generic circular loops are especially suitable for graphic representation, because
the image of a circular loop deﬁnes it to a great extend:
50:G. Let l be a generic polygonal loop on a triangulated surface. Then any generic
polygonal loop k with k(S
1
) = l(S
1
) is homotopic in l(S
1
) to either l or l
−1
.
Thus, to describe a generic circular loop up to a reparametrization homotopic to
identity, it is suﬃcient to draw the image of the loop on the surface and specify the
direction in which the loop runs along the image.
The image of a generic polygonal loop is called a generic (polygonal) closed
connected curve. A union of a ﬁnite collection of generic closed connected polygo
nal curves is called a generic (polygonal) closed curve. A generic closed connected
curve without double points (i.e., an embedded oriented circle contained in the one
dimensional skeleton of a triangulated surface) is called a simple polygonal closed
curve.
The adjective closed in the deﬁnitions above appears because there is a version
of the deﬁnitions with (nonclosed) paths instead of loops.
50:H. Enigma. What modiﬁcations in Problems 50:C – 50:G and corresponding
deﬁnitions should be done to replace loops by paths everywhere?
By a generic polygonal curve we will mean a union of a ﬁnite collection of pairwise
disjoint images of generic polygonal loops and paths.
'50
◦
4 Cutting Surface Along Curve
50:I Cutting Surface Along Curve. Let F be a triangulated twodimensional
manifold and C ⊂ F a onedimensional manifold contained in the 1skeleton of the
triangulation of F. Assume that ∂C = ∂F ∩ C. Prove that there exists a two
dimensional manifold T and a surjective continuous map p : T →F such that:
(a) p[ : T p
−1
(C) →F C is a homeomorphism,
(b) p[ : p
−1
(C) →C is a twofold covering.
§50. ONEDIMENSIONAL mod2HOMOLOGY OF SURFACES 257
Such T and p are unique up to a homeomorphism: if
˜
T and ˜ p are other manifold
and mapping satisfying the same conditions then there exists a homeomorphism h :
˜
T →T such that p ◦ h = ˜ p.
The surface T described in 50:I is called the result of cutting F along C. It is
denoted by F C. This is not the complement F C, though a copy of F C is
contained in F C as a dense subset, which is homotopy equivalent to the whole
F C.
50:J Triangulation of F C. There exists a unique triangulation of F C such
that the natural map F C →F maps edges onto edges and triangles onto triangles
homeomorphically.
50:4. Describe the topological type of F C for the following F and C:
(a) F is M¨obius band, C its core circle (deformation retract);
(b) F = S
1
S
1
, C = S
1
1;
(c) F is S
1
S
1
standardly embedded into R
3
, C the trefoil knot on F,
that is ¦(z, w) ∈ S
1
S
1
[ z
2
= w
3
¦;
(d) F is M¨obius band, C is a segment: show that there are two possible
placements of C in F and describe F C for both of them;
(e) F = RP
2
, C = RP
1
.
(f) F = RP
2
, C is homeomorphic to circle: show that there are two possible
placements of C in F and describe F C for both of them.
50:5 Euler Characteristic and Cutting. Find the Euler characteristic of
F C when ∂C = ∅. What if ∂C = ∅?
'50
◦
5 Curves on Surfaces and TwoFold Coverings
Let F be a twodimensional triangulated surface and C ⊂ F a manifold of di
mension one contained in the 1skeleton of the triangulation of F. Let ∂C = ∂F ∩C.
Since the preimage
˜
C of C under the natural projection F C → F is a twofold
covering space of C, there is an involution τ :
˜
C →
˜
C which is the only nontrivial
automorphism of this covering. Take two copies of F C and identify each x ∈
˜
C in
one of them with τ(x) in the other copy. The resulting space is denoted by F
≈C
.
50:K. The natural projection F C →F deﬁnes a continuous map F
≈C
→F. This
is a twofold covering. Its restriction over F C is trivial.
'50
◦
6 OneDimensional Z
2
Cohomology of Surface
By 40:J, a twofold covering of F can be thought of as an element of H
1
(F; Z
2
).
Thus any onedimensional manifold C contained in the 1skeleton of F and such that
∂C = ∂F ∩ C deﬁnes a cohomology class of F with coeﬃcients in Z
2
. This class is
said to be realized by C.
50:L. The cohomology class with coeﬃcients in Z
2
realized by C in a compact surface
F is zero, iﬀ C divides F, that is, F = G ∪ H, where G and H are compact two
dimensional manifolds with G∩ H = C.
Recall that the cohomology group of a pathconnected space X with coeﬃcients
in Z
2
is deﬁned above in Section '40 as Hom(π
1
(X), Z
2
).
§50. ONEDIMENSIONAL mod2HOMOLOGY OF SURFACES 258
50:M. Let F be a triangulated connected surface, let C ⊂ F be a manifold of
dimension one with ∂C = ∂F ∩ C contained in the 1skeleton of F. Let l be a
polygonal loop on F which is in general position with respect to C. Then the value
which the cohomology class with coeﬃcients in Z
2
deﬁned by C takes on the element
of π
1
(F) realized by l equals the number of points of l ∩ C reduced modulo 2.
'50
◦
7 OneDimensional Z
2
Homology of Surface
50:N Z
2
Classes via Simple Closed Curves. Let F be a triangulated connected
twodimensional manifold. Every homology class ξ ∈ H
1
(F; Z
2
) can be represented
by a polygonal simple closed curve.
50:O. A Z
2
homology class of a triangulated twodimensional manifold F represented
by a polygonal simple closed curve A ⊂ F is zero, iﬀ there exists a compact two
dimensional manifold G ⊂ F such that A = ∂G.
Of course, the “if” part of 50:O follows straightforwardly from 40:O.
The “only if” part requires trickier arguments.
50:O.1. If A is a polygonal simple closed curve on F, which does not
bound in F a compact 2manifold, then there exists a connected compact
1manifold C ⊂ F with ∂C = ∂F ∩ C, which intersects A in a single point
transversally.
50:O.2. Let F be a twodimensional triangulated surface and C ⊂ F a
manifold of dimension one contained in the 1skeleton of the triangulation
of F. Let ∂C = ∂F ∩ C. Any polygonal loop f : S
1
→F, which intersects
C in an odd number of points and transversally at each of them, is covered
in F
≈C
by a path with distinct endpoints.
50:O.3. See 40:8.
'50
◦
8 Poincar´e Duality
To be written!
'50
◦
9 OneSided and TwoSided Simple Closed Curves on Surfaces
To be written!
'50
◦
10 Orientation Covering and First StiefelWhitney Class
To be written!
'50
◦
11 Relative Homology
To be written!
'51 Surfaces Beyond Classiﬁcation
To be written!
'51
◦
1 Genus of Surface
To be written!
'51
◦
2 Systems of disjoint curves on a surface
To be written!
'51
◦
3 Polygonal Jordan and Sch¨onﬂies Theorems
To be written!
'51
◦
4 Polygonal Annulus Theorem
To be written!
'51
◦
5 Dehn Twists
To be written!
'51
◦
6 Coverings of Surfaces
To be written!
'51
◦
7 Branched Coverings
To be written!
'51
◦
8 Mapping Class Group of Torus
To be written! Lifting homeomorphisms to the universal covering space.
Nielsen and Baer Theorems for torus. GL(2, Z). Dehn twists along meridian and
longitude and relation between them. Center of the mapping class group.
'51
◦
9 Braid Groups
To be written!
259
'52 ThreeDimensional Manifolds
To be written!
'52
◦
1 Poincar´e Conjecture
'52
◦
2 Lens Spaces
'52
◦
3 Seifert Manifolds
'52
◦
4 Fibrations over Circle
'52
◦
5 Heegaard Splitting and Diagrams
260
§52. THREEDIMENSIONAL MANIFOLDS 261
Proofs and Comments
27.A Let u
t
: I →X be a homotopy between u and u
and v
t
: I →
X be a homotopy between v and v
. Then paths u
t
v
t
with t ∈ [0, 1] form
a homotopy between uv and u
v
.
27.B This is a straightforward reformulation of 27.A.
34.C The line is connected. Therefore its retract (being its contin
uous image) must be connected, too. However a pair of points in the line
is not connected.
This book includes basic material on general topology, introduces algebraic topology via the fundamental group and covering spaces, and provides a background on topological and smooth manifolds. It is written mainly for students with a limited experience in mathematics, but determined to study the subject actively. The material is presented in a concise form, proofs are omitted. Theorems, however, are formulated in detail, and the reader is expected to treat them as problems.
Foreword
Genre, Contents and Style of the Book The core of the book is the material usually included in the Topology part of the two year Geometry lecture course at the Mathematical Department of St. Petersburg University. It was composed by Vladimir Abramovich Rokhlin in the sixties and has almost not changed since then. We believe this is the minimum topology that must be mastered by any student who has decided to become a mathematician. Students with research interests in topology and related ﬁelds will surely need to go beyond this book, but it may serve as a starting point. The book includes basic material on general topology, introduces algebraic topology via its most classical and elementary part, the theory of the fundamental group and covering spaces, and provides a background on topological and smooth manifolds. It is written mainly for students with a limited experience in mathematics, but who are determined to study the subject actively. The core material is presented in a concise form; proofs are omitted. Theorems, however, are formulated in detail. We present them as problems and expect the reader to treat them as problems. Most of the theorems are easy to ﬁnd elsewhere with complete proofs. We believe that a serious attempt to prove a theorem must be the ﬁrst reaction to its formulation. It should precede looking for a book where the theorem is proved. On the other hand, we want to emphasize the role of formulations. In the early stages of studying mathematics it is especially important to take each formulation seriously. We intentionally force a reader to think about each simple statement. We hope that this will make the book inconvenient for mere skimming. The core material is enhanced by many problems of various sorts and additional pieces of theory. Although they are closely related to the main material, they can be (and usually are) kept outside of the standard lecture course. These enhancements can be recognized by wider margins, as the next paragraph.
iii
FOREWORD
iv
The problems, which do not comprise separate topics and are intended exclusively to be exercises, are typeset with small face. Some of them are very easy and included just to provide additional examples. Few problems are diﬃcult. They are to indicate relations with other parts of mathematics, show possible directions of development of the subject, or just satisfy an ambitious reader. Problems, whose solutions seem to be the most diﬃcult (from the authors’ viewpoint), are marked with a star, as in many other books. Further, we want to deliver additional pieces of theory (with respect to the core material) to more motivated and advanced students. Maybe, a mathematician, who does not work in the ﬁelds geometric in ﬂavor, can aﬀord the luxury not to know some of these things. Maybe, students studying topology can postpone this material to their graduate study. We would like to include this in graduate lecture courses. However, quite often it does not happen, because most of the topics of this sort are rather isolated from the contents of traditional graduate courses. They are important, but more related to the material of the very ﬁrst topology course. In the book these topics are intertwined with the core material and exercises, but are distinguishable: they are typeset, like these lines, with large face, theorems and problems in them are numerated in a special manner described below. Exercises and illustrative problems to the additional topics are typeset with wider margins and marked in a diﬀerent way.
Thus, the whole book contains four layers: the core material, exercises and illustrative problems to the core material, additional topics, exercises and illustrative problems to additional topics. The text of the core material is typeset with large face and smallest margins. • • • •
The text of problems elaborating on the core material is typeset with small face and larger margins. The text of additional topics is typeset is typeset with large face as the problems elaborating on the core material. The text of problems illustrating additional topics is typeset with small face and larger margins.
Therefore the book looks like a Russian folklore doll, matreshka composed of several dolls sitting inside each other. We apologize for being nonconventional in this and hope that it may help some readers and does not irritate the others too much. The whole text of the book is divided into sections. Each section is divided into subsections. Each of them is devoted to a single topic and consists of deﬁnitions, commentaries, theorems, exercises, problems, and riddles.
FOREWORD
v
By a riddle we mean a problem of a special sort: its solution is not contained in the formulation. One has to guess a solution, rather than deduce it. 0.A. Theorems, exercises, problems and riddles belonging to the core material are marked with pairs consisting of the number of section and a letter separated with a dot. The letter identiﬁes the item inside the section.
0.1. Exercises, problems, and riddles, which are not included in the core, but are closely related to it (and typeset with small face) are marked with pairs consisting of the number of the section and the number of the item inside the section. The numbers in the pair are separated also by a dot.
Theorems, exercises, problems and riddles related to additional topics are enumerated independently inside each section and denoted similarly.
0:A. The only diﬀerence is that the components of pairs marking the items are separated by a colon (rather than dot).
We assume that the reader is familiar with naive set theory, but anticipate that this familiarity may be superﬁcial. Therefore at points where set theory is especially crucial we make settheoretic digressions maintained in the same style as the rest of the book. Advices to the Reader Since the book contains a summary of elementary topology, you may use the book while preparing for an examination (especially, if the exam reduces to solving a collection of problems). However, if you attend lectures on the subject, it would be much wiser to read the book prior to the lectures and prove theorems before the lecturer gives the proofs. We think that a reader who is able to prove statements of the core of the book, does not need to solve all the other problems. It would be reasonable instead to look through formulations and concentrate on the most diﬃcult problems. The more diﬃcult the theorems of the main text seem to you, the more carefully you should consider illustrative problems, and the less time you should waste with problems marked with stars. Keep in mind that sometimes a problem which seems to be diﬃcult is followed by easier problems, which may suggest hints or serve as technical lemmas. A chain of problems of this sort is often concluded with a problem which suggests a return to the theorem, once you are armed with the lemmas. Most of our illustrative problems are easy to invent, and, moreover, if you study the subject seriously, it is always worthwhile to invent problems of this sort. To develop this style of studying mathematics while solving
FOREWORD vi our problems one should attempt to invent one’s own problems and solve them (it does not matter if they are similar to ours or not). some problems presented in this book are not easy to invent. . Of course.
and Answers vii . De Morgan Formulas § 2◦ 7 Properties of Closed Sets § 2◦ 8 Being Open or Closed § 2◦ 9 Cantor Set § 2 ◦ 10 Characterization of Topology in Terms of Closed Sets § 2 ◦ 11 Topology and Arithmetic Progressions § 2 ◦ 12 Neighborhoods Proofs and Comments Hints. Solutions. and Answers § 2. General Topology iii iii v 1 3 3 3 4 4 5 5 6 6 6 6 7 7 9 10 11 12 12 12 13 13 13 14 14 14 15 15 15 16 16 17 Chapter 1. Open and Closed Sets § 2◦ 6 SetTheoretic Digression. Generalities § 1.Contents Foreword Genre. Solutions. Comments. Comments. Topology in a Set § 2◦ 1 Deﬁnition of Topological Space § 2◦ 2 The Simplest Examples § 2◦ 3 The Most Important Example: Real Line § 2◦ 4 Additional Examples § 2◦ 5 Using New Words: Points. Contents and Style of the Book Advices to the Reader Part 1. Advises. Prove Inclusions § 1◦ 9 Inclusion Versus Belonging § 1 ◦ 10 Deﬁning a Set by a Condition § 1 ◦ 11 Intersection and Union § 1 ◦ 12 Diﬀerent Diﬀerences Proofs and Comments Hints. Digression on Sets Sets and Elements § 1◦ 1 § 1◦ 2 Equality of Sets The Empty Set § 1◦ 3 § 1◦ 4 Basic Sets of Numbers Describing a Set by Listing of Its Elements § 1◦ 5 § 1◦ 6 Subsets § 1◦ 7 Properties of Inclusion § 1◦ 8 To Prove Equality of Sets. Advises.
and Answers § 5. Metric Spaces § 4◦ 1 Deﬁnition and First Examples ◦ §4 2 Further Examples Balls and Spheres § 4◦ 3 § 4◦ 4 Subspaces of a Metric Space § 4◦ 5 Surprising Balls Segments (What Is Between) § 4◦ 6 Bounded Sets and Balls § 4◦ 7 § 4◦ 8 Norms and Normed Spaces § 4◦ 9 Metric Topology § 4 ◦ 10 Openness and Closedness of Balls and Spheres § 4 ◦ 11 Metrizable Topological Spaces § 4 ◦ 12 Equivalent Metrics § 4 ◦ 13 Ultrametric § 4 ◦ 14 Operations with Metrics § 4 ◦ 15 Distance Between Point and Set § 4 ◦ 16 Distance Between Sets § 4 ◦ 17 Asymmetrics Proofs and Comments Hints. Bases § 3◦ 1 Deﬁnition of Base § 3◦ 2 When a Collection of Sets is a Base § 3◦ 3 Bases for Plane § 3◦ 4 Subbases § 3◦ 5 Inﬁniteness of the Set of Prime Numbers § 3◦ 6 Hierarchy of Topologies Proofs and Comments Hints. Advises. Comments.CONTENTS viii § 3. Advises. Comments. Ordered Sets § 5◦ 1 Strict Orders § 5◦ 2 NonStrict Orders § 5◦ 3 Relation between Strict and NonStrict Orders § 5◦ 4 Cones § 5◦ 5 Position of an Element with Respect to a Set Total Orders § 5◦ 6 § 5◦ 7 Topologies Deﬁned by a Total Order Poset Topology § 5◦ 8 § 5◦ 9 How to Draw a Poset § 5 ◦ 10 Cyclic Orders in Finite Set § 5 ◦ 11 Cyclic Orders in Inﬁnite Sets § 5 ◦ 12 Topology of Cyclic Order Proofs and Comments Hints. and Answers § 4. Advises. Comments. and Answers 19 19 19 19 20 20 20 21 21 23 23 23 24 24 25 25 25 25 26 26 27 27 27 28 28 28 29 30 31 34 34 34 35 35 36 37 37 38 39 40 41 42 42 43 . Solutions. Solutions. Solutions.
SetTheoretic Digression. Solutions. Advises. and Answers § 8. Cl. Solutions. Subspaces § 6◦ 1 Topology for a subset of a space § 6◦ 2 Relativity of Openness § 6◦ 3 Agreement on Notations of Topological Spaces Proofs and Comments Hints.CONTENTS ix § 6. Advises. Comments. Fr § 7 ◦ 11 Characterization of Topology by Closure or Interior Operations ◦ § 7 12 Dense Sets § 7 ◦ 13 Nowhere Dense Sets § 7 ◦ 14 Limit Points and Isolated Points § 7 ◦ 15 Locally Closed Sets Proofs and Comments Hints. Continuous Maps § 9◦ 1 Deﬁnition and Main Properties of Continuous Maps § 9◦ 2 Reformulations of Deﬁnition More Examples § 9◦ 3 § 9◦ 4 Behavior of Dense Sets Local Continuity § 9◦ 5 § 9◦ 6 Properties of Continuous Functions § 9◦ 7 Continuity of Distances Isometry § 9◦ 8 § 9◦ 9 GromovHausdorﬀ distance § 9 ◦ 10 Contractive maps § 9 ◦ 11 Monotone maps 45 45 45 46 46 47 48 48 48 48 49 49 49 50 50 51 51 51 51 52 52 53 53 53 55 55 55 56 56 57 57 58 58 59 59 59 60 60 61 61 61 62 62 . and Answers § 7. Exterior and Boundary Points § 7◦ 2 Interior and Exterior § 7◦ 3 Closure § 7◦ 4 Closure in a Metric Space § 7◦ 5 Frontier Closure and Interior with Respect to a Finer Topology § 7◦ 6 § 7◦ 7 Properties of Interior and Closure § 7◦ 8 Compositions of Closure and Interior Sets with Common Frontier § 7◦ 9 § 7 ◦ 10 Convexity and Int. Position of a Point with Respect to a Set § 7◦ 1 Interior. Maps § 8◦ 1 Maps and the Main Classes of Maps § 8◦ 2 Image and Preimage § 8◦ 3 Identity and Inclusion § 8◦ 4 Composition § 8◦ 5 Inverse and Invertible § 8◦ 6 Submappings § 9. Comments.
Homeomorphisms § 10 ◦ 1 Deﬁnition and Main Properties of Homeomorphisms § 10 ◦ 2 Homeomorphic Spaces § 10 ◦ 3 Role of Homeomorphisms § 10 ◦ 4 More Examples of Homeomorphisms § 10 ◦ 5 Examples of Homeomorphic Spaces § 10 ◦ 6 Examples of Nonhomeomorphic Spaces § 10 ◦ 7 Homeomorphism Problem and Topological Properties § 10 ◦ 8 Information: Nonhomeomorphic Spaces § 10 ◦ 9 Embeddings § 10 ◦ 10 Equivalence of Embeddings § 10 ◦ 11 Information Chapter 2. and Answers § 10.CONTENTS x § 9 ◦ 12 Functions on Cantor Set and SquareFilling Curves § 9 ◦ 13 Sets Deﬁned by Systems of Equations and Inequalities § 9 ◦ 14 SetTheoretic Digression. Comments. Topological Properties § 11. Covers § 9 ◦ 15 Fundamental Covers Hints. PathConnectedness § 12 ◦ 1 Paths § 12 ◦ 2 PathConnected Spaces § 12 ◦ 3 PathConnected Sets § 12 ◦ 4 Properties of PathConnected Sets § 12 ◦ 5 PathConnected Components § 12 ◦ 6 PathConnectedness Versus Connectedness § 12 ◦ 7 PolygonConnectedness § 12 ◦ 8 Connectedness of Some Sets of Matrices § 13. Advises. Separation Axioms § 13 ◦ 1 The Hausdorﬀ Axiom § 13 ◦ 2 Limits of Sequence 62 63 64 64 65 67 67 67 67 68 69 72 72 72 73 73 74 75 75 75 75 76 76 77 77 77 78 79 79 79 80 81 81 81 82 82 82 83 83 84 85 85 85 . Solutions. Connectedness § 11 ◦ 1 Deﬁnitions of Connectedness and First Examples § 11 ◦ 2 Connected Sets § 11 ◦ 3 Properties of Connected Sets § 11 ◦ 4 Connected Components § 11 ◦ 5 Totally Disconnected Spaces § 11 ◦ 6 Frontier and Connectedness § 11 ◦ 7 Connectedness and Continuous Maps § 11 ◦ 8 Connectedness on Line § 11 ◦ 9 Intermediate Value Theorem and Its Generalizations § 11 ◦ 10 Dividing Pancakes § 11 ◦ 11 Induction on Connectedness § 11 ◦ 12 Applications to Homeomorphism Problem § 12.
Countability Axioms § 14 ◦ 1 SetTheoretic Digression. Sequential Compactness § 17 ◦ 1 Sequential Compactness Versus Compactness § 17 ◦ 2 In Metric Space § 17 ◦ 3 Completeness and Compactness § 17 ◦ 4 NonCompact Balls in Inﬁnite Dimension § 17 ◦ 5 pAdic Numbers § 17 ◦ 6 Induction on Compactness § 17 ◦ 7 Spaces of Convex Figures 86 86 86 87 88 88 89 89 90 90 90 91 91 92 92 93 94 94 94 95 95 95 96 96 97 97 98 99 99 99 100 100 101 101 101 101 103 103 103 104 104 104 105 105 . Local Compactness and Paracompactness § 16 ◦ 1 Local Compactness § 16 ◦ 2 OnePoint Compactiﬁcation § 16 ◦ 3 Proper Maps § 16 ◦ 4 Locally Finite Collections of Subsets § 16 ◦ 5 Paracompact Spaces § 16 ◦ 6 Paracompactness and Separation Axioms § 16 ◦ 7 Partitions of Unity § 16 ◦ 8 Application: Making Embeddings from Pieces § 17. Compactness § 15 ◦ 1 Deﬁnition of Compactness § 15 ◦ 2 Terminology Remarks § 15 ◦ 3 Compactness in Terms of Closed Sets § 15 ◦ 4 Compact Sets § 15 ◦ 5 Compact Sets Versus Closed Sets § 15 ◦ 6 Compactness and Separation Axioms § 15 ◦ 7 Compactness in Euclidean Space § 15 ◦ 8 Compactness and Continuous Maps § 15 ◦ 9 Closed Maps § 15 ◦ 10 Norms in Rn § 16. Countability § 14 ◦ 2 Second Countability and Separability § 14 ◦ 3 Embedding and Metrization Theorems § 14 ◦ 4 Bases at a Point § 14 ◦ 5 First Countability § 14 ◦ 6 Sequential Approach to Topology § 14 ◦ 7 Sequential Continuity § 15.CONTENTS xi § 13 ◦ 3 Coincidence Set and Fixed Point Set § 13 ◦ 4 Hereditary Properties § 13 ◦ 5 The First Separation Axiom § 13 ◦ 6 The Kolmogorov Axiom § 13 ◦ 7 The Third Separation Axiom § 13 ◦ 8 The Fourth Separation Axiom § 13 ◦ 9 Niemytski’s Space § 13 ◦ 10 Urysohn Lemma and Tietze Theorem § 14.
Quotient Spaces § 19 ◦ 1 SetTheoretic Digression. Topological Constructions § 18. Sums of Sets § 20 ◦ 12 Sums of Spaces § 20 ◦ 13 Attaching Space § 20 ◦ 14 Basic Surfaces § 21. Partitions and Equivalence Relations ◦ § 19 2 Quotient Topology § 19 ◦ 3 Topological Properties of Quotient Spaces § 19 ◦ 4 SetTheoretic Digression. Projective Spaces 107 109 109 109 110 110 111 111 112 113 114 114 115 115 116 116 117 117 117 118 119 119 121 121 121 122 123 124 124 125 125 126 126 126 126 127 128 130 . Quotients and Maps § 19 ◦ 5 Continuity of Quotient Maps § 19 ◦ 6 Closed Partitions § 19 ◦ 7 Open Partitions § 19 ◦ 8 SetTheoretic Digression: Splitting a transitive relation into equivalence and partial order ◦ § 19 9 Finite Topological Spaces § 19 ◦ 10 Simplicial schemes § 19 ◦ 11 Baricentric Subdivision of a Poset § 20.CONTENTS xii Problems for Tests Chapter 3. Multiplication § 18 ◦ 1 SetTheoretic Digression. Zoo of Quotient Spaces § 20 ◦ 1 Tool for Identifying a Quotient Space with a Known Space ◦ § 20 2 Tools for Describing Partitions § 20 ◦ 3 Entrance to the Zoo § 20 ◦ 4 Transitivity of Factorization § 20 ◦ 5 M¨bius Strip o ◦ § 20 6 Contracting Subsets § 20 ◦ 7 Further Examples § 20 ◦ 8 Klein Bottle § 20 ◦ 9 Projective Plane § 20 ◦ 10 You May Have Been Provoked to Perform an Illegal Operation ◦ § 20 11 SetTheoretic Digression. Product of Sets § 18 ◦ 2 Product of Topologies § 18 ◦ 3 Topological Properties of Projections and Fibers § 18 ◦ 4 Cartesian Products of Maps § 18 ◦ 5 Properties of Diagonal and Other Graphs § 18 ◦ 6 Topological Properties of Products § 18 ◦ 7 Representation of Special Spaces as Products § 19.
Algebraic Topology 151 153 153 153 153 154 154 154 Chapter 5. Fundamental Group and Covering Spaces § 26. Topological Groups 143 § 24 ◦ 1 The Notion of Topological Group 143 ◦ § 24 2 Examples of Topological Groups 143 § 24 ◦ 3 SelfHomeomorphisms Making a Topological Group Homogeneous 143 § 24 ◦ 4 Neighborhoods 144 § 24 ◦ 5 Separaion Axioms 145 § 24 ◦ 6 Countability Axioms 145 ◦ § 24 7 Subgroups 145 § 24 ◦ 8 Normal Subgroups 146 ◦ § 24 9 Homomorphisms 147 § 24 ◦ 10 Local Isomorphisms 147 ◦ § 24 11 Direct Products 148 § 25. Generalities on Groups 139 § 23 ◦ 1 The Notion of Group 139 ◦ § 23 2 Additive and Multiplicative Notations 140 141 § 23 ◦ 3 Homomorphisms § 23 ◦ 4 Subgroups 141 § 24. Homotopy § 26 ◦ 1 Continuous Deformations of Maps § 26 ◦ 2 Homotopy as Map and Family of Maps § 26 ◦ 3 Homotopy as Relation § 26 ◦ 4 StraightLine Homotopy § 26 ◦ 5 Maps to Star Convex Sets . Actions of Topological Groups 150 § 25 ◦ 1 Actions of Group in Set 150 ◦ § 25 2 Continuous Actions 150 § 25 ◦ 3 Orbit Spaces 150 § 25 ◦ 4 Homogeneous Spaces 150 Part 2. Digression. Spaces of Continuous Maps § 22 ◦ 1 Sets of Continuous Mappings § 22 ◦ 2 Topologies on Set of Continuous Mappings § 22 ◦ 3 Topological Properties of Mapping Spaces § 22 ◦ 4 Metric Case § 22 ◦ 5 Interactions With Other Constructions § 22 ◦ 6 Mappings X × Y → Z and X → C(Y. A Touch of Topological Algebra 138 § 23. Z) 130 131 131 134 134 134 135 135 136 136 Chapter 4.CONTENTS xiii § 21 ◦ 1 Real Projective Space of Dimension n § 21 ◦ 2 Complex Projective Space of Dimension n § 21 ◦ 3 Quaternionic Projective Spaces § 22.
Calculations of Fundamental Groups Using Universal Coverings 174 174 § 32 ◦ 1 Fundamental Group of Circle . Theorems on Path Lifting 172 ◦ § 31 1 Lifting 172 § 31 ◦ 2 Path Lifting 172 § 31 ◦ 3 Homotopy Lifting 173 § 31 ◦ 4 HighDimensional Homotopy Groups of Covering Space173 § 32. Covering Spaces 169 ◦ § 30 1 Deﬁnition of Covering 169 § 30 ◦ 2 More Examples 169 § 30 ◦ 3 Local homeomorphisms versus coverings 170 § 30 ◦ 4 Number of Sheets 171 171 § 30 ◦ 5 Universal Coverings § 31. Homotopy Properties of Path Multiplication § 27 ◦ 1 Multiplication of Homotopy Classes of Paths § 27 ◦ 2 Associativity § 27 ◦ 3 Unit § 27 ◦ 4 Inverse § 28. 155 155 155 156 156 156 158 158 158 159 159 161 161 161 161 162 163 163 164 165 166 166 166 166 167 167 168 Covering Spaces and Calculation of Fundamental Groups 169 § 30. The Role of Base Point § 29 ◦ 1 Overview of the Role of Base Point § 29 ◦ 2 Deﬁnition of Translation Maps § 29 ◦ 3 Properties of Ts § 29 ◦ 4 Role of Path § 29 ◦ 5 High Homotopy Groups § 29 ◦ 6 In Topological Group Chapter 6. Fundamental Group § 28 ◦ 1 Deﬁnition of Fundamental Group § 28 ◦ 2 Why Index 1? § 28 ◦ 3 High Homotopy Groups § 28 ◦ 4 Circular loops § 28 ◦ 5 The Very First Calculations § 28 ◦ 6 Fundamental Group of Product § 28 ◦ 7 SimplyConnectedness § 28 ◦ 8 Fundamental Group of a Topological Group § 29.CONTENTS xiv § 26 ◦ 6 Maps of Convex Sets § 26 ◦ 7 Easy Homotopies § 26 ◦ 8 Two Natural Properties of Homotopies § 26 ◦ 9 Stationary Homotopy § 26 ◦ 10 Homotopies and Paths § 26 ◦ 11 Homotopy of Paths § 27.
Cellular Spaces § 37 ◦ 1 Deﬁnition of Cellular Spaces § 37 ◦ 2 First Examples § 37 ◦ 3 Further TwoDimensional Examples § 37 ◦ 4 Simplicial spaces § 37 ◦ 5 Topological Properties of Cellular Spaces 197 197 197 199 200 201 201 . Induced Homomorphisms and Their First Applications 179 ◦ § 33 1 Homomorphisms Induced by a Continuous Map 179 § 33 ◦ 2 Fundamental Theorem of Algebra 180 § 33 ◦ 3 Generalization of Intermediate Value Theorem 181 182 § 33 ◦ 4 Winding Number § 33 ◦ 5 BorsukUlam Theorem 182 § 34. Homotopy Equivalences 187 § 35 ◦ 1 Homotopy Equivalence as Map 187 ◦ § 35 2 Homotopy Equivalence as Relation 187 § 35 ◦ 3 Deformation Retraction 187 § 35 ◦ 4 Examples 188 § 35 ◦ 5 Deformation Retraction Versus Homotopy Equivalence188 § 35 ◦ 6 Contractible Spaces 189 § 35 ◦ 7 Fundamental Group and Homotopy Equivalences 189 § 36. Fundamental Group and Mappings 179 § 33. Cellular Techniques § 37.CONTENTS xv § 32 ◦ 2 § 32 ◦ 3 § 32 ◦ 4 § 32 ◦ 5 § 32 ◦ 6 Fundamental Group of Projective Space Fundamental Groups of Bouquet of Circles Algebraic Digression. Free Groups Universal Covering for Bouquet of Circles Fundamental groups of some ﬁnite topological spaces 175 175 175 177 178 Chapter 7. 185 § 35. Retractions and Fixed Points 184 ◦ 184 § 34 1 Retractions and Retracts § 34 ◦ 2 Fundamental Group and Retractions 184 ◦ § 34 3 FixedPoint Property. Covering Spaces via Fundamental Groups 191 ◦ § 36 1 Homomorphisms Induced by Covering Projections 191 § 36 ◦ 2 Number of Sheets 191 § 36 ◦ 3 Hierarchy of Coverings 192 § 36 ◦ 4 Existence of subordinations 193 § 36 ◦ 5 Micro Simply Connected Spaces 193 § 36 ◦ 6 Existence of Coverings 194 § 36 ◦ 7 Automorphisms of Covering 194 195 § 36 ◦ 8 Regular Coverings § 36 ◦ 9 Lifting and Covering Maps 196 Chapter 8.
Locally Euclidean Spaces § 41 ◦ 1 Deﬁnition of Locally Euclidean Space § 41 ◦ 2 Dimension § 41 ◦ 3 Interior and Boundary § 42. Fundamental Group of a Cellular Space § 39 ◦ 1 OneDimensional Cellular Spaces § 39 ◦ 2 Generators § 39 ◦ 3 Relators § 39 ◦ 4 Writing Down Generators and Relators § 39 ◦ 5 Fundamental Groups of Basic Surfaces § 39 ◦ 6 Seifert .van Kampen Theorem § 40. OneDimensional Homology and Cohomology § 40 ◦ 1 Why and What for § 40 ◦ 2 OneDimensional Integer Homology § 40 ◦ 3 Zero Homologous Loops and Disks with Handles § 40 ◦ 4 Description of H1 (X) in Terms of Free Circular Loops § 40 ◦ 5 Homology and Continuous Maps § 40 ◦ 6 OneDimensional Cohomology § 40 ◦ 7 Cohomology and Classiﬁcation of Regular Coverings § 40 ◦ 8 Integer Cohomology and Maps to S 1 § 40 ◦ 9 OneDimensional Homology Modulo 2 Part 3.CONTENTS xvi § 37 ◦ 6 Embedding to Euclidean Space § 37 ◦ 7 Euler Characteristic § 37 ◦ 8 Collaps § 37 ◦ 9 Generalized collaps § 38. Manifolds 202 203 203 204 206 206 206 207 207 208 209 209 209 209 210 211 212 213 213 213 214 214 215 216 216 216 217 219 221 221 221 221 222 225 225 225 225 226 226 228 228 228 Chapter 9. Bare Manifolds § 41. Isotopy § 43 ◦ 1 Isotopy of Homeomorphisms § 43 ◦ 2 Isotopy of Embeddings and Sets . Manifolds § 42 ◦ 1 Deﬁnition of Manifold § 42 ◦ 2 Components of Manifold § 42 ◦ 3 Making New Manifolds out of Old Ones § 42 ◦ 4 Double § 42 ◦ 5 Collars and Bites § 43. OneDimensional Cellular Spaces § 38 ◦ 1 Homotopy Classiﬁcation § 38 ◦ 2 Dividing Cells § 38 ◦ 3 Trees and Forests § 38 ◦ 4 Simple Paths § 38 ◦ 5 Maximal Trees § 39.
Orientations 248 ◦ § 48 1 Orientations of Edges and Triangles 248 § 48 ◦ 2 Orientation of Triangulation 248 § 49. OneDimensional Manifolds 231 ◦ § 44 1 ZeroDimensional Manifolds 231 § 44 ◦ 2 Reduction to Connected Manifolds 231 ◦ § 44 3 Examples 231 § 44 ◦ 4 Statements of Main Theorems 231 ◦ § 44 5 Lemma on 1Manifold Covered with Two Lines 232 § 44 ◦ 6 Without Boundary 233 § 44 ◦ 7 With Boundary 233 § 44 ◦ 8 Consequences of Classiﬁcation 233 ◦ § 44 9 Mapping Class Groups 233 § 45. Triangulations 237 § 46 ◦ 1 Triangulations of Surfaces 237 ◦ § 46 2 Triangulation as cellular decomposition 237 § 46 ◦ 3 Two Properties of Triangulations of Surfaces 237 § 46 ◦ 4 Scheme of Triangulation 238 § 46 ◦ 5 Examples 238 § 46 ◦ 6 Subdivision of a Triangulation 239 ◦ § 46 7 Homotopy Type of Compact Surface with NonEmpty Boundary 241 ◦ § 46 8 Triangulations in dimension one 241 § 46 ◦ 9 Triangualtions in higher dimensions 242 § 47. Classical Approach to Topological Classiﬁcation of Compact surfaces 249 ◦ § 49 1 Families of Polygons 249 250 § 49 ◦ 2 Operations on Family of Polygons ◦ § 49 3 Topological and Homotopy Classiﬁcation of Closed Surfaces 250 251 § 49 ◦ 4 Recognizing Closed Surfaces ◦ § 49 5 Orientations 252 . TwoDimensional Manifolds: General Picture 234 § 45 ◦ 1 Examples 234 § 45 ◦ 2 Ends and Odds 234 235 § 45 ◦ 3 Closed Surfaces § 46.CONTENTS xvii § 43 ◦ 3 Isotopies and Attaching 229 § 43 ◦ 4 Connected Sums 230 § 44. Handle Decomposition 243 § 47 ◦ 1 Handles and Their Anatomy 243 ◦ § 47 2 Handle Decomposition of Manifold 243 § 47 ◦ 3 Handle Decomposition and Triangulation 244 ◦ § 47 4 Regular Neighborhoods 245 § 47 ◦ 5 Cutting 2Manifold Along a Curve 245 § 48.
ThreeDimensional Manifolds § 52 ◦ 1 Poincar´ Conjecture e § 52 ◦ 2 Lens Spaces § 52 ◦ 3 Seifert Manifolds § 52 ◦ 4 Fibrations over Circle § 52 ◦ 5 Heegaard Splitting and Diagrams Proofs and Comments 252 253 253 254 254 254 255 256 257 257 258 258 258 258 258 259 259 259 259 259 259 259 259 259 259 260 260 260 260 260 260 261 .CONTENTS xviii § 49 ◦ 6 More About Recognizing Closed Surfaces § 49 ◦ 7 Compact Surfaces with Boundary § 49 ◦ 8 Simply Connected Surfaces § 50. OneDimensional mod2Homology of Surfaces § 50 ◦ 1 Polygonal Paths on Surface § 50 ◦ 2 Subdivisions of Triangulation § 50 ◦ 3 Bringing Loops to General Position § 50 ◦ 4 Cutting Surface Along Curve § 50 ◦ 5 Curves on Surfaces and TwoFold Coverings § 50 ◦ 6 OneDimensional Z2 Cohomology of Surface § 50 ◦ 7 OneDimensional Z2 Homology of Surface § 50 ◦ 8 Poincar´ Duality e § 50 ◦ 9 OneSided and TwoSided Simple Closed Curves on Surfaces § 50 ◦ 10 Orientation Covering and First StiefelWhitney Class § 50 ◦ 11 Relative Homology § 51. Surfaces Beyond Classiﬁcation § 51 ◦ 1 Genus of Surface § 51 ◦ 2 Systems of disjoint curves on a surface § 51 ◦ 3 Polygonal Jordan and Sch¨nﬂies Theorems o ◦ § 51 4 Polygonal Annulus Theorem § 51 ◦ 5 Dehn Twists § 51 ◦ 6 Coverings of Surfaces § 51 ◦ 7 Branched Coverings § 51 ◦ 8 Mapping Class Group of Torus § 51 ◦ 9 Braid Groups § 52.
Part 1 General Topology .
while proofs of all theorems are extremely simple. Nowadays studying general topology really resembles studying a language rather than mathematics: one needs to learn a lot of new words. but it is not very active. which treats the basic notions related to continuity. More speciﬁcally. We have to warn students. one of its most important components: the language of settheoretic topology. Do not hurry to fall in love with it too seriously. This ﬁeld may seam to be charming. It hardly provides as much room for exciting new research as most of other ﬁelds. do not let an imprinting happen. it was completed a long time ago.The goal of this part of the book is to teach the language of mathematics. for they play the role of rules regulating usage of words. . A permanent usage in the capacity of a common mathematical language has polished its system of deﬁnitions and theorems. for whom this is one of the ﬁrst mathematical subjects. As a research ﬁeld. the theorems are numerous. The term general topology means: this is the topology that is needed and used by most mathematicians. On the other hand.
It is formed by them. which we would like to consider unnecessary. This is a part of the common mathematical language. The gathering is performed in minds and is not accompanied with any action in the physical world. As soon as the group has been created and assigned with a name. The sign ∈ is a version of Greek letter epsilon. A set consists of its elements. are used in the same sense. may be included into other groups. such as class. and hence should be used instead of the word set cautiously. The ﬁrst words in this language are set and element. We would not be able to say anything about topology without this part (see the next section to make sure that this is not an exaggeration). § 1 ◦ 1 Sets and Elements In any intellectual activity. we write x ∈ A and say x belongs to A and A contains x. one may expect that naive set theory becomes familiar to a student when she or he studies Calculus or Algebra. An object included into the collection is called an element of the set. too. Sometimes other words. since each of these words is associated in the modern mathematics with a more special meaning. Its subject is the ﬁrst basic notions of the naive set theory. If x is an element of a set A. Naturally. it may be subject of thoughts and arguments and. a slightly misleading name. please. in particular. To diversify wording. but even more profound than general topology. because this is rather a language. To make formulas more 3 . which is the ﬁrst letter of the Latin word element. glance through this section and move to the next one.CHAPTER 1 Generalities § 1 Digression on Sets We begin with a digression. In Mathematics there is an elaborated system of notions which organizes and regulate creating of those groups and manipulating them. the subjects which usually precede topology. one of the most profound action is gathering objects into groups. family and group. but it is not quite safe. If this is what really happened to you. the word set is replaced by the word collection. than a theory. y a set we understand an arbitrary collection of various objects. This system is called the naive set theory.
It is a little more than just a name: it is a declaration of our wish to think about this collection of things as of entity and not to go into details about the nature of its memberselements. the formula x ∈ A is allowed to be written also as A x. § 1 ◦ 3 The Empty Set Thus. when we say that a line is a set of points. maybe unintentionally. or a rare oldfashioned exception from the common mathematical terminology (sometimes the expression under the sign of integral is called an inﬁnitesimal element. We may think of sets as boxes. For example. The cost of this lightness is that such a box is not more than the collection of elements placed inside. On the other hand. an element may not be without a set. This manifests most sharply in the following principle: two sets are considered equal. To state that x is not an element of A.g. When you call something an element. in turn. a lack of interest to whatever organization of the elements of this set. we write x ∈ A or A x. whose element is this one. in old texts lines. DIGRESSION ON SETS 4 ﬂexible. it is not good to use the word element as a replacement for other. For example. just to distinguish them from the rest of the world. the distance between points. However a set may be without elements. but as long as we consider them elements. They are even overused. It is nothing but a collection of its elements. which can be built eﬀortlessly around elements. the order of points on the line) separately from the notion of line. we indicate that two lines coincide if and only if they consist of the same points. In modern Mathematics the words set and element are very common and appear in most of texts. unless we deal with nonmathematical term (like chemical element). but a more meaningful similarity to the inequality symbols < and > is emphasized. When one calls something a set.§ 1. we commit ourselves to consider all the relations between points on a line (e. Euclid’s famous book on Geometry is called Elements. In this sense the word set has slightly disparaging meaning. more meaningful words. So the origin of notation is ignored. because a set is deﬁned completely by its elements. the set. may also be sets. Elements. they play the role of atoms with their own original nature ignored. this shows. It is called . The word element makes sense only in a combination with the word set. planes and other geometric images are called elements). § 1 ◦ 2 Equality of Sets A set is deﬁned by its elements. This set is unique. There are instances when it is not appropriate to use them. should be clear. There is a set which has no element. if and only if they have the same elements.
In other words. 3} and {3. which are so important that have their own unique names and notation. this would not be wise. x of its elements is denoted by symbol {a. 1. 2}}. b.. like Λ. How many elements do the following sets contain? 1) {1. . 2. whose elements are listed. 2) {∅} ∈ {{∅}}. . This is any set which can be presented as {a} for some a. . 7) {{∅}. both positive (that is natural numbers) and negative and the zero. For example. quite often one cannot say a priori if there are repetitions or not. ∅}. 5. {1.2. 1}. 3x − 1} for x ∈ R. 1. The set of all natural numbers. . 3) {{2}}. . A} means the set which consists of three elements. x. 2) {1. is denoted by Z. 1}. also were in use. the list of objects enclosed in a curly brackets denotes the set. 3. For example. they don’t. The set of all the rational numbers (add to the 2 integers those numbers which can be presented by fractions.4. The set of all the real numbers (obtained by adjoining √ to rational numbers the numbers like 2 and π = 3. etc. − 7 ) 5 is denoted by Q. 2 and 123.. 5) {1. The set of all integer numbers.14 . {1. whatever these three letters denote.§ 1. . while for other values. . {∅}}. the elements of the list may depend on parameter. . . Is {{∅}} a singleton? Notice that sets {1. 2. 3) ∅ ∈ {{∅}}? A set consisting of a single element is called a singleton. 8) {x. x}. 1. and under certain values of the parameters some entries of the list coincide. . In fact. Which of the following formulas are correct: 1. like 3 . The set of complex numbers is denoted by C. ∅}. . a. 6) {{∅}. . since they consists of the same elements. but ∅ has become common. 2. as it often happens to temptations to prohibit something. DIGRESSION ON SETS 5 the empty set and denoted by ∅. . there are few other sets.1. 2. 2. § 1 ◦ 4 Basic Sets of Numbers Besides ∅. 2. 4. 4) {{1}. . 1. § 1 ◦ 5 Describing a Set by Listing of Its Elements The set presented by the list a. 1. Notation {a. Other notations. What is {∅}? How many elements does it contain? 1) ∅ ∈ {∅. However. . is denoted by N. b. At ﬁrst glance. 123} denotes the set which consists of numbers 1. i.e. ) is denoted by R. x and A.3. a list with repetition of elements is never needed. {∅}}. There arises even a temptation to prohibit usage of lists with repetitions in such a notation. 2} are equal.
∅ ⊂ A for any set A. Any set includes itself: A ⊂ A holds true for any A. if and only if A ⊂ B and B ⊂ A. This word is used when one does not want to consider the obvious subsets (which are called improper). . the inclusion signs are not completely true counterparts of the inequality signs < and >.§ 1. say A and B. 1. 1. the empty set is present in each set as a subset. we need from time to time to prove that two sets. Thus. Prove that if A ⊂ B then a ≤ b.A. They are closer to ≤ and ≥. DIGRESSION ON SETS 6 § 1 ◦ 6 Subsets If A and B are sets and every element of A belongs also to B. A subset of A diﬀerent from ∅ and A is called a proper subset of A. we say that A is a subset of B. while A ⊂ B means that A is made of some of the elements of B. Notice that there is no number a satisfying inequality a < a. § 1 ◦ 9 Inclusion Versus Belonging Despite this obvious relation between the notions of belonging ∈ and inclusion ⊂ and similarity of the symbols ∈ and ⊂. x ∈ A.D Transitivity of Inclusion. are equal. which may have emerged in quite diﬀerent ways. and write A ⊂ B or B ⊃ A. The inclusion signs ⊂ and ⊃ recall the inequality signs < and > for a good reason: in the world of sets the inclusion signs are obvious counterparts for the signs of inequalities. Prove Inclusions Working with sets. 1. § 1 ◦ 7 Properties of Inclusion 1. The most common way to do this is provided by the following theorem.E Criterium of Equality for Sets.B Reﬂexivity of Inclusion. Indeed. the concepts are very diﬀerent.F. If A. Let a set A consists of a elements. 1. or B includes A. then A ⊂ C. A ∈ B means that A is one of the elements of B (that is one of indivisible pieces comprising B). if and only if {x} ⊂ A. In other words. A ⊂ B and B ⊂ C. B and C are sets. and a set B of b elements.C The Empty Set Is Everywhere. Thus. A = B. each set A has two obvious subsets: the empty set ∅ and A itself. 1. § 1 ◦ 8 To Prove Equality of Sets.
H NonTransitivity of Belonging. .§ 1. Thus. The union of sets A and B is the set consisting of those elements each of which belongs to at least one of these sets. Sets A and B are said to be disjoint. B and C such that A ∈ B and B ∈ C. while A ∈ A for any reasonable A.B. if their intersection is empty. This is a reasonable description of the set. and eventually may be lucky to solve the equation and get the list of its solutions. . Cf. It can be described by formula A ∪ B = {x  x ∈ A or x ∈ B}. that is elements belonging both to A and B. For example. it is easy to say: “the set of all the solutions of the following equation” and write down the equation. while B ∈ B. (b) {x ∈ N  x < 0}. § 1 ◦ 10 Deﬁning a Set by a Condition As we know (see Section § 1 ◦ 5). . At least.G NonReﬂexivity of Belonging. i. 1. 1. . Cf.e. However the latter may be diﬃcult and should not prevent us from discussing the set. Thus we see another way for description of a set: to formulate the properties which distinguish the elements of the set among elements of some wider and already known set.e. Here is the corresponding notation: the subset of a set A consisting of elements x which satisfy condition P (x) is denoted by {x ∈ A  P (x)}. Present the following sets by lists of their elements (i. A ∪ B = ∅.5. }) (a) {x ∈ N  x < 5}. Construct sets A. It is denoted by A ∩ B and can be described by formula A ∩ B = {x  x ∈ A x ∈ B}. b. 1. Having accepted it. a set can be described by presenting a list of its elements. be not the easiest one. 1. Construct sets A and B such that A ∈ A.. but A ∈ C. belonging is not reﬂexive.D. DIGRESSION ON SETS 7 In particular. while inclusion is. This simplest way may be not available or. The union of A and B is denoted by A ∪ B. § 1 ◦ 11 Intersection and Union The intersection of sets A and B is the set consisting of their common elements. studying its properties. 1. (c) {x ∈ Z  x < 0}. we may start speaking on the set. A ⊂ A. at least. it is unambiguous. in the form {a.. Yet another diﬀerence: belonging is not transitive.
This set is denoted by ∩A∈Γ A or A∈Γ A. How are related to each other the notions of system of equations and intersection of sets? 1. This set is denoted by ∪A∈Γ A or A∈Γ A . For any sets A. The intersection of the sets belonging to Γ is the set formed by elements which belong to every set. B} C = A ∩ B and C = A ∪ B. Prove that for any set A A ∩ A = A. A ∪ A = A. B and C (1) (2) (A ∩ B) ∪ C = (A ∪ C) ∩ (B ∪ C). For any sets A. maybe.K.I Commutativity of ∩ and ∪. A B A B A B A∩B A∪B Figure 1. For any sets A and B A∩B = B∩A A ∪ B = B ∪ A. intersection and union of arbitrarily large (in particular. 1. Enigma. Prove that for any sets A and B A ∩ B = A. C∈Γ C∈Γ 1. iﬀ A ∪ B = B. Indeed. One deﬁnes A ∩ B ∩ C = (A ∩ B) ∩ C = A ∩ (B ∩ C) and A ∪ B ∪ C = (A ∪ B) ∪ C = A ∪ (B ∪ C).§ 1. without reference to intersection or union of two sets.7. iﬀ A ∪ ∅ = A and A ∩ ∅ = ∅. 1. The notions of intersection and union of arbitrary collection of sets generalize the notions of intersection and union of two sets: for Γ = {A. 1. their intersection A ∩ B and union A ∪ B. belonging to Γ.6. inﬁnite) collection of sets can be deﬁned directly. Associativity allows us do not care about brackets and sometimes even omit them. to both of them. However. B and C (A ∩ B) ∩ C = A ∩ (B ∩ C) and (A ∪ B) ∪ C = A ∪ (B ∪ C). let Γ be a collection of sets.J Associativity of ∩ and ∪.8. DIGRESSION ON SETS 8 Here the conjunction or should be understood in the inclusive way: the statement “x ∈ A or x ∈ B” means that x belongs to at least one of the sets A and B. but. 1. (A ∪ B) ∩ C = (A ∩ C) ∪ (B ∩ C) . Disks A and B. 1. A ⊂ B. Similarly. the union of the sets belonging to Γ is the set formed by elements which belong to at least one of the sets belonging to Γ.L Two Distributivities.
10. Such comics are called Venn diagrams.N Yet Another Pair of Distributivities. 1. If A ⊃ B.M. C) = (A ∩ B) 1. Here it is not assumed that A ⊃ B. Draw Venn diagrams illustrating all formulas below in this section. the set A B is called also the complement of B in A. Prove that A ∩ (B The set (A B) ∪ (B A) is called the symmetric diﬀerence of sets A and B.L is illustrated by a sort of comics. B and C. Prove that A (A B) = A ∩ B for any sets A and B. In Figure 2 the ﬁrst of two equalities of Theorem 1. Then A∩ B= B∈Γ B∈Γ (A ∩ B) and A∪ B= B∈Γ B∈Γ (A ∪ B). Let A be a set and Γ be a set consisting of sets.14. if and only if A 1. 1.12. The lefthand side (A ∩ B) ∪ C of the equality (1) and the sets A ∪ C B ∪ C. 1. Generalize Theorem 1. Prove that for any sets A and B A B = (A ∪ B) (A ∩ B) . (A ∩ C) for any sets A. Enigma. B = ∅. DIGRESSION ON SETS 9 A C B = A C ∩ B C (A ∩ B) ∪ C = (A ∪ C) ∩ (B ∪ C) Figure 2. Prove that A ⊂ B.11. It is denoted by A B. § 1 ◦ 12 Diﬀerent Diﬀerences A diﬀerence A B of sets A and B is the set of those elements of A which do not belong to B. 1.9.§ 1. Draw a Venn diagram illustrating (2). Prove (1) and (2) tracing all the details of the proofs in Venn diagrams. whose intersection is the righthand side of the equation (1). Prove that for any sets A and B their union A ∪ B can be represented as the union of the following three sets: A B. B A and A ∩ B. 1. 1. A ∪ B. and that these sets are pairwise disjoint.13. They are very useful and we strongly recommend to draw them for each formula about sets.L to the case of arbitrary collection of sets.
which would include a theory of counting.) Therefore we can start counting of elements of B with counting the elements of A. because ∅ has no elements at all.B Recall that. B C (B ∪ C)? Proofs and Comments 1. if there are some elements of B which are not in A. Does the following equality hold true for any sets A.§ 1. counting will continue. ﬁrst.C Recall that. The counting of elements of A will be done. What does it mean that A consists of a elements? It means. But there is no such an element in ∅. because there is no elements in ∅. say. DIGRESSION ON SETS 10 A B A B A B B A A B A B Figure 3.18. This is correct. Thus we need to prove that any element of ∅ belongs to A.15 Associativity of Symmetric Diﬀerence. B and C.A The question is so elementary that it is diﬃcult to ﬁnd more elementary facts. 1. and the last element will get number a. A ⊂ B means that each element of A is an element of B. 1. How can it happen that ∅ is not a subset of A? It could happen. and then. let us use it without proof. whether this can be wrong. . and in which this is a theorem.16. by the deﬁnition of inclusion. Thus the number of elements in A is less than or equal to the number of elements in B. It is known that the result does not depend on the order in which we count. B and C (A B) C = A (B C). only if there was an element of ∅ which would not be an element of A. Therefore the statement that we have to prove can be rephrased as follows: each element of A is an element of A. that we can count elements of A one by one assigning to them numbers 1. 1. Prove that for any sets A. Prove that (A sets A. But since we have no doubts in this fact. (A B) ∪ C = (A ∪ C) B) ∩ C = (A ∩ C) (B ∩ C) for any 1. let us resort to a question. 3. by the deﬁnition of inclusion.17 Distributivity. one can develop the set theory. Enigma. Diﬀerences of disks A and B. 2. This is a tautology. Find a symmetric deﬁnition of symmetric diﬀerence (A B) C of three sets and generalize it to any ﬁnite collection of sets. (In fact. 1. A ⊂ B means that each element of A is an element of B. If you are not satisﬁed with this argument (since it sounds too crazy). which a proof can be based on. 1.
5 (a) {1. If we consider the set of all sets. . −3. 1. or A = N. {{∅}} is a singleton. Since b ⊂ C. (b) {}. it follows that x ∈ B. −2. 1. B = {{1}} and C = {{{1}}}. −6. This is what we had to prove. 2 1. 2. Of course. B ∈ C. 2. Hints. (c) {−1. −4. 1. Advises. It would not appear in real problems. 3.8 The set of solutions for a system of equations is equal to the intersection of the sets of solutions of individual equations belonging to the system. Since A ⊂ B. unless you think really globally.4 2. 2 for x = 1 2 and 1 if x = 1 . {{1}}}. On the other hand. 1. A ⊂ B means that each element of A belongs to B and B ⊂ A means that each element of B belongs to A.§ 1. Hence if A = B then A ⊂ B and B ⊂ A. Comments. why not to consider the set Y of all the sets X such that X ∈ X? Does Y belongs to itself? If Y ∈ Y then Y ∈ Y . If Y ∈ Y then Y ∈ Y since Y is the set of ALL the sets X such that X ∈ X. Hence A and B have the same elements..H Take A = {1}. B = {{1}}.G It is easy to construct a set A with A ∈ A.e. since each element X of Y has the property that X ∈ X. It is more diﬃcult to construct sets A. 4}. 3. Take for B the set of all sets. 2. Solutions. There are good reasons for this. . which means that they are equal. 1. Take A = {1}.2 1) and 2) are correct. and A ∈ C. or A = {1}. . DIGRESSION ON SETS 11 1.E We have already seen that A ⊂ A. A set B such that B ∈ B is a strange creature.D We have to prove that each element of A is an element of C. 2.} 1. An easy way to avoid this paradox is to prohibit consideration of sets with the property X ∈ X. Take A = ∅. this element itself is the empty set and contains no element. −5. This contradiction shows that our deﬁnition of Y does not make sense.1 The set {∅} consists of one element. 1.. 3) is not. The the set of all sets is not a legitimate set. . 1. but the set {∅} consists of a single element ∅. and Answers 1. B and C such that A ∈ B. . 1. Mathematicians avoid such sets. . Let x ∈ A. which is the empty set ∅. x ∈ B) implies x ∈ C.3 Yes. C = {{1}. the latter (i.
X. {a}.3. b. Of course. b}. {b.e. is it not? Here are slightly less trivial examples. +∞) with a ≥ 0. X.1. satisfy the axioms of topological structure: (a) ∅. X. X. {b}. c}.3 (a) by . Prove that Ω is a topological structure. c. {a. 2. this does not mean that Ω coincides with the subject of topology. i. {a. Check that this is a topological space. Which of the following collections of its subsets are topological structures in X. c.1 is called an arrow. This is a topological structure. {b. but everything in this subject is related to Ω.§ 2 Topology in a Set § 2 ◦ 1 Deﬁnition of Topological Space Let X be a set. 12 1 the arrow: ∞ . X. (c) ∅. • an element of X is called a point of this topological space. We denote the space of 2. Let Ω be a collection of its subsets such that: (a) the union of a collection of sets. The conditions in the deﬁnition above are called the axioms of topological structure. are not important. 2. {a. Let X be a plane. which are elements of Ω. (c) the empty set ∅ and the whole X belong to Ω. (b) ∅. (b) the intersection of a ﬁnite collection of sets. {a. i. Ω). • an element of Ω is called an open set of the topological space (X. belongs to Ω. and Ω consists of ∅.. c}. § 2 ◦ 2 The Simplest Examples A discrete topological space is a set with the topological structure which consists of all the subsets.2. Both of these spaces. It is a sort of toy space made of 4 points. but provide good simple examples. d}? The space of 2. An indiscrete topological space is the opposite example. {b}. 2. d}. d}. Ω) is called a topological space. c. Is this a topological structure? 2.B. Let X be the ray [0. It consists only of X and ∅. in which the topological structure is the most meager. {a.A. 2. and all open disks with center at the origin. Thus Ω is important: it is called by the same word as the whole branch of mathematics. Let Σ consist of ∅. which are elements of Ω. b. d}. +∞). all axioms of topological structure hold true. and all the rays (a..2. belongs to Ω. Let X consist of four elements: X = {a. {a}. as well as the space of 2. Then • Ω is called a topological structure or just a topology 1 in X. b}. • the pair (X.e.
Let X be R. 1}} a topological structure in {0.6.5 is denoted by RT1 and called the line with T1 topology . Is Ω a topological structure? 2. Is {{a} ∪ U : U ∈ Ω} ∪ {∅} a topological structure in Y ? 2. Oﬀen in German. 2. This space is called usually the real line and the structure is referred to as the canonical or standard topology in R. Let (X. Ω be the set of unions of all intervals (a.4. {0}..e. Is the set {∅. X A set F ⊂ X is said to be closed in the space (X. Is Ω a topological structure? The space of 2.5. elements of X are called points. § 2 ◦ 5 Using New Words: Points. Ω) be a topological space and Y be the set obtained from X by adding a single element a. 2 The letter Ω stands for the letter O which is the initial of the words with the same meaning: Open in English. say. This is the topological structure which is always meant when R is considered as a topological space (unless other topological structure is explicitly speciﬁed).7.8. Otkrytyj in Russian. .D. Ω). List all the topological structures in a twoelement set. 1}. b ∈ R.6.7 is a particular point topology. {0. Ouvert in French. and Ω consists of empty set and complements of all ﬁnite subsets of R. 2. the topology in Y is called a particular point topology or topology of everywhere dense point.C. for a topological space (X. Let X be R. in {0. TOPOLOGY IN A SET 13 § 2 ◦ 3 The Most Important Example: Real Line Let X be the set R of all real numbers. 2.§ 2.2 2. b) with a. X F ∈ Ω). § 2 ◦ 4 Additional Examples 2. 1}? In Problem 2. if topology Ω discrete. and elements of Ω are called open sets. Reformulate the axioms of topological structure using the words open set wherever possible. Open and Closed Sets Recall that. and Ω consists of empty set and all the inﬁnite subsets of R. it is called also the topology of connected pair of points or Sierpinski topology. The topology in Problem 2. Check if Ω satisﬁes the axioms of topological structure. Ω) if its complement F is open (i.
. nor closed.. Then (3) X A∈Γ A= (X A∈Γ A) (4) X A∈Γ A= (X A∈Γ A). Enigma. § 2 ◦ 7 Properties of Closed Sets 2. TOPOLOGY IN A SET 14 § 2 ◦ 6 SetTheoretic Digression. (b) union of any ﬁnite number of closed sets is closed. Let Γ be an arbitrary collection of subsets of a set X. Give an explicit description of closed sets in (a) a discrete space. unions and intersections. while the union of any collection of open sets is open. 2. Prove that: (a) the intersection of any collection of closed sets is closed. and closed simultaneously. Along the same lines. (They are not exact antonyms in everyday usage. (b) an indiscrete space.9.e.H. Is a closed segment [a. (c) empty set and the whole space (i. 2. while the intersection of any collection of closed sets is closed.§ 2. The main diﬀerence is that the intersection of an inﬁnite collection of open sets does not have to be necessarily open. too). Concepts of closed and open sets are similar in a number of ways. 2.F.E. . Formula (4) is deduced from (3) in one step. Find examples of sets. the union of an inﬁnite collection of closed sets is not necessarily closed. (b) are neither open. b] closed in R. Find such a formulation. which contains in a symmetric way sets and their complements. the underlying set of the topological structure) are closed. is it not? These formulas are nonsymmetric cases of a single formulation. § 2 ◦ 8 Being Open or Closed Notice that the property of being closed is not a negation of the property of being open. (d) (e) RT1 .10. which (a) are both open. (c) the arrow. 2. De Morgan Formulas 2.G.
K is the set of real numbers which 3 in the positional system with base 3 are presented as 0. 3 (c) K does not intersect 3s+1 . 1 (b) K does not intersect 3 .11. 2 . Solving this problem. Prove that (a) K is contained in [0. Prove that the halfopen interval [0. Consider the following property of a subset F of the set N of natural numbers: there exists N ∈ N such that F does not contain an arithmetic progression of length greater than N . in which these collections are complete sets of closed sets. Try to draw K.15.1. ak . . then F is the set of all closed sets of a topological space (which one?). It has a lot of remarkable properties and is involved in numerous problems below. (c) ∅ and X belong to F . § 2 ◦ 11 Topology and Arithmetic Progressions 2.§ 2. 3s+2 for any integers k and s. you probably are not able to avoid the following combinatorial theorem. . 2:A. 1] with an inﬁnite family of open intervals removed.a1 a2 . § 2 ◦ 10 Characterization of Topology in Terms of Closed Sets 2. . Prove that if a collection F of subsets of X satisﬁes the following conditions: (a) the intersection of any family of sets from F belongs to F . (b) the union of any ﬁnite number sets from F belongs to F . In other words.2. TOPOLOGY IN A SET 15 2. n=1 2:A. 2. 2. Find a geometric description of K. without digit 1. 3k 3k 2:A.14. Prove. 1) is neither open nor closed in R. Prove that the set A = {0} ∪ § 2 ◦ 9 Cantor Set Let K be the set of real numbers which can be presented as sums of series of the ak form ∞ k=1 k with ak = 0 or 2. 2. that subsets with this property together with the whole N form a collection of closed subsets in some topology in N.12.16*. 2:B.13. List all collections of subsets of a threeelement set such that there exist topologies. Prove that K is a closed set in the real line. 1 n ∞ is closed in R. but can be presented as either the union of closed sets or intersection of open sets. The set K is called the Cantor set.3. . Present K as [0. 1]. 2:A. Prove that every open set of the real line is a union of disjoint open intervals. .
we n=1 numbered all the rational numbers). Since it belongs to the union. Proofs and Comments 2.17 Van der Waerden’s Theorem*. either A or {1. . try to answer the following question (which has nothing to do with the problem under consideration. The intersection of any ﬁnite collection of open sets is open. it is. it belongs to X. Exactly in the same way one checks the second axiom. though). 2. this is right. TOPOLOGY IN A SET 16 2. § 2 ◦ 12 Neighborhoods By a neighborhood of a point one means any open set containing this point. ∅ ⊂ X and X ⊂ X. (e) connected pair of points. 2. Give an explicit description of all neighborhoods of a point in (a) a discrete space. . 2. For every n ∈ N there exists N ∈ N such that for any A ⊂ {1. If A ⊂ X for each A ∈ Γ then ∪A∈Γ A ⊂ X. (f) particular point topology. (b) an indiscrete space. which contains all (!) the rational numbers. Then the union is also empty.18. (d) . Here we can list all the collections of sets that we need to consider. e.§ 2. Then the intersection equals X. N }. r + 2−n ) = R. Finally. Indeed. N } A contains an arithmetic progression of length n. The empty set and the whole space are open.A What should we check? The ﬁrst axiom reads here that the union of any collection of subsets of X is a subset of X? Well. .B Yes. If it is not there. Prove that (r − 2−n . Analysts and French mathematicians (following N. Let {rn }∞ = Q (i. of course. If one of the united sets is X then the union is X. take arbitrary point b ∈ ∪A∈Γ A. Therefore if A and B are intervals then the righthand side is a union of intervals. then what is? Only the whole X. (c) the arrow. and since A ⊂ X. it belongs to at least one of A ∈ Γ. Bourbaki) prefer a wider notion of neighborhood: they use this word for any set containing a neighborhood in the sense above. What if it is not there? Then what is there? Empty set. show that A∈Γ A∩ B= B∈Σ A∈Γ. 2. . although this is a union of some intervals. . With intersections the situation is simialr.B∈Σ (A ∩ B). . 2. . at most.C First. If one of the sets to intersect is ∅ the the intersection is ∅. . 2.D The union of any collection of open sets is open. . If you think that a set which is a union of intervals is too simple. please..
{b.2 Yes. c. In a discrete space any set is both open and closed. +∞) is not a topological structure. +∞). 2. d}.10 The following sets are closed (a) in a disctrete space: all sets. 2. 2. bα ) then 1 ∈ (aα0 . {c. d} is not. while {a. bα ) = (0. Finally. 1] = ∪(aα . c. Find two elements of the third collection such that there intersection does not belong to it. +∞) is open. Semiopen interval is neither open nor closed on the line. 1]. (b) in an indiscrete: only those which are also open. (c) in the arrow: ∅. d} = {a} ∪ {b.G In any topological space the empty set and the whole space are both open and closed. it is. Hints.§ 2. the whole space and segments of the form [0. {a. (e) in RT1 : all ﬁnite sets and the whole R.3 The main point here is to realize that the axioms of topological structure are conditions on the collection of subsets and if these conditions are satisﬁed then the collection is called a topological structure. {b. 1] = (−∞. and Answers 2. +∞) = (inf aα . Why is not (0. since it may happen that ∪[aα . Advises. Solutions.11 Here it is important to overcome the feeling that the question is completely obvious. it is not diﬃcult to see that all the unions and intersections of elements of the ﬁrst collection still belong to the ﬁrst collection. ∅. d}. {b. that is the empty set and the whole space. By this you would prove that this is not a topology. d}. +∞) = (a0 .H Yes. b. it is. because the sets {a}. d}. and it follows that ∪(aα . hence bα0 > 1.1 The solution is based on the equality ∪(aα . Similarly R (0. +∞) . The second collection is not a topological structure. because R [a. 0] ∪ (1. Comments. Prove it. Cf. bα0 ) for some α0 . 2. also the next problem. 1] open? If (0. 2. (d) in : sets X. +∞) (ﬁnd an example). A proof coincides almost literally with the solution of the preceding problem. {d}. a) ∪ (b. By the way the collection of closed rays [a.E (a) x ∈ ∩A∈Γ (X A) ⇐⇒ ∀ A x ∈ X A ⇐⇒ ∀ λ x ∈ Aλ ⇐⇒ x ∈ ∪A∈Γ A ⇐⇒ / / x∈X ∪A∈Γ A (b) Replace both sides of the formula by their complements in X and put B = X A. b] = (−∞. 2. a]. TOPOLOGY IN A SET 17 2. d} are contained in it.
2.14 are obviously satisﬁed. Let sets A and B do not contain arithmetic progression of length ≥ n. and must solve it on your own. 2.16 The conditions (a) and (c) from 2. To prove (b). ∞ (0. 2. If the set A ∪ B contained a suﬃciently long progression.15 Control indication: there number of such collections is 14. 2. On the other hand.17.18 By this point you have to learn already everything needed for solving this problem. Please. let us use 2. don’t be lazy. . 1] = n=1 n+1 1 0.14 Check that Ω = {U  X U ∈ F } is a topological structure. in one of the original sets there would be a progression of length n.§ 2.1 = n n n=1 ∞ . TOPOLOGY IN A SET 18 is not open. .
3. disks without its boundary circles). § 3 ◦ 3 Bases for Plane Consider the following three collections of subsets of R2 : • Σ2 which consists of all possible open disks (i. some set of Σ containing this point (cf.6. all intervals form a base for the real line.A.e. Prove that every element of Σ2 is a union of elements of Σ∞. A collection Σ of open sets is a base for the topology. Prove that any base of the canonical topology in R can be diminished. Are there diﬀerent topological structures with the same base? 3. Find some bases of topology of . iﬀ X is a union of sets of Σ and intersection of any two sets of Σ is a union of sets in Σ. (d) the arrow. Σ1 is a base for some topological structure in R2 . • Σ1 which consists of all possible open squares with sides parallel to the bisectors of the coordinate angles. For instance. respectively.3.B. Prove that intersection of any two elements of Σ1 is a union of elements of Σ1. and that the structures deﬁned by these collections coincide.. (a) a discrete space.1. Show that the second condition in 3. A collection Σ of open sets is called a base for a topology. Describe all topological structures having exactly one base. squares without their sides and vertices) with sides parallel to the coordinate axis. 3. if each nonempty open set is a union of sets belonging to Σ.§ 3 Bases § 3 ◦ 1 Deﬁnition of Base Usually the topological structure is presented by describing its part.C. 3. (b) (c) an indiscrete space. • Σ∞ which consists of all possible open squares (i.A). Σ∞.e. Prove that each of the collections Σ2.2. 3. together with any of its points.) 3. (Squares of Σ∞ and Σ1 are deﬁned by inequalities max{x−a. 3.. 3. 3.5. iﬀ for any open set U and any point x ∈ U there is a set V ∈ Σ such that x ∈ V ⊂ U. § 3 ◦ 2 When a Collection of Sets is a Base 3. 19 .7. Try to choose the bases as small as possible. 3.B (on intersection) is equivalent to the following: the intersection of any two sets of Σ contains. A collection Σ of subsets of a set X is a base for some topology in X.4. which is suﬃcient to recover the whole structure. y−b} < ρ and x − a + y − b < ρ.
D. Let Σ1 and Σ2 be bases for topological structures Ω1 and Ω2 in a set X. k ∈ N} i=1 of all ﬁnite intersections of sets belonging to ∆ is a base for Ω. 3. § 3 ◦ 5 Inﬁniteness of the Set of Prime Numbers 3.) . Ω) be a topological space. Enigma.7).§ 3.8. § 3 ◦ 6 Hierarchy of Topologies If Ω1 and Ω2 are topological structures in a set X such that Ω1 ⊂ Ω2 then Ω2 is said to be ﬁner than Ω1 . Formulate a necessary and suﬃcient condition for two bases to be equivalent without explicit mentioning of topological structures deﬁned by the bases. 3. 3. provided the collection Σ = {V  V = ∩k Wi . Wi ∈ ∆. § 3 ◦ 4 Subbases Let (X. BASES 20 Figure 1. Enigma. 3. Prove that for any set X a collection ∆ of its subsets is a subbase of a topology in X. Bases deﬁning the same topological structure are said to be equivalent. Find necessary and suﬃcient condition for Ω1 ⊂ Ω2 in terms of the bases Σ1 and Σ2 without explicit referring to Ω1 and Ω2 (cf. A collection ∆ of its open subsets is called a subbase for Ω. iﬀ ∆ = ∅ and X = ∪W ∈∆ W . 3. 3.10. among all topological structures in the same set the indiscrete topology is the coarsest topology. Prove that all inﬁnite arithmetic progressions consisting of natural numbers form a base for some topology in N. and Σ1 must satisfy the condition you are looking for.9. For instance. Show that T1 topology (see Section § 2) is coarser than the canonical topology in the real line. Elements of Σ∞ (left) and Σ1 (right). Using this topology prove that the set of all prime numbers is inﬁnite. (Cf. and Ω1 coarser than Ω2 . Σ∞.7: bases Σ2.11.12. is it not? 3. and the discrete topology is the ﬁnest one.
Notice that ∪x∈U Vx ⊂ U. On the other hand. too. Vice versa. Advises. Let us prove the second axiom (the intersection of two open sets is open).β (Aα ∩ Bβ ). U = ∪x∈U Vx . 3. Aα . Solutions. c}. {a. Obviously. assume that Σ is a collection of subsets of X such that X is a union of sets belonging to Σ and the intersection of any two sets belonging to Σ is a union of sets belonging to Σ. Aα ∩ Bβ can be presented as union of elements of Σ. X is a union of sets belonging to Σ. d}. since the union of some unions is a union. not! A topological structure is recovered from its base as the set of unions of all collections of sets which belong to the base. 3. Comments. we need to check only the part concerning the whole X. Hints. b.B Assume that Σ is a base of a topology. By the assumption. each point x ∈ U is contained in its Vx and hence in ∪x∈U Vx .§ 3. and show that Σ is a base of Ω. Therefore U ⊂ ∪x∈U Vx . (why?) (b) For a base in one can take. Then X. Ω1 ⊂ Ω2 . Each point x ∈ U is contained in some of these sets. In the third axiom. The intersection of any two sets belonging to Σ is open. Now recall that Ω1 = Ω2 ⇐⇒ Ω1 ⊂ Ω2 and Ω2 ⊂ Ω1 . BASES 21 Proofs and Comments 3. being an open set in any topology.1 Of course.D Let Σ1 and Σ2 be bases of topological structures Ω1 and Ω2 in a set X. and Answers 3. The ﬁrst axiom is obviously satisﬁed. Let us prove that the set of unions of all the collections of elements of Σ satisﬁes the axioms of topological structure. It is contained in U. 3. say. since it participates in a union which is equal to U. Vice versa. can be presented as a union of some sets belonging to GS. Bβ ∈ Σ. {b}. the intersection U ∩ V can be presented in this form. Let U = ∪α Aα V = ∪β Bβ . Thus. iﬀ ∀ U ∈ Σ1 ∀ x ∈ U ∃ V ∈ Σ2 : x ∈ V ⊂ U. {a}. and since. Such a set can be chosen as V . since Vx ⊂ U for each x ∈ U. Present U as a union of elements of Σ.A Let Σ be a base of Ω and U ∈ Ω. . For this we need to prove that any U ∈ Ω can be represented as a union of elements of Σ. by the assumpiton. Then U ∩ V = (∪α Aα ) ∩ (∪β Bβ ) = ∪α. c. assume that for any U ∈ Ω and any point x ∈ U there exists a set V ∈ Σ such that x ∈ V ⊂ U. {a. For each point x ∈ U choose according to the assumption a set Vx ∈ Σ such that x ∈ Vx ⊂ U and consider ∪x∈U Vx .2 (a) A discrete space admits a base consisting of all onepoint subsets of the space and this base is minimal. therefore it also can be presented as a union of base sets.
+∞)}r∈Q+ is a base. In such a space any open set cannot be represented as a union of two open sets distinct from it. i + 2d. So. . (r. U is not involved into these unions. . 3p. So. BASES 22 (c) In indiscrete space the minimal base is formed by a single set. . .10 Since the sets {i. Each of the intervals. 3. the space should contain only ﬁnite number of open sets.} cover N {1}.5. . 3. where Bα ∈ B. . the set {1} would be open. In particular. i = 1.}. To prove that the structures deﬁned.B and 3. you should prove that R {xi }n is open in the canonical topology of the line. iﬀ ∀ x ∈ A ∃ Bx ∈ B : x ∈ Bx ⊂ A. . . are open.§ 3. . 3p. for each prime number p the set {p. which are shorter than distance between some two points of U . But it cannot presented as union of arithmetic progressions. .6 for Σ2 and Σ∞ . pairwise disjoint and cover the whole N.6. the whole space.12 Ω1 ⊂ Ω2 . . i=1 3. Is it enough to prove that a disk is a union of squares? What is the simplest way to do this (cf.4 We will show that removing of any element from any base of the standard topology of the line gives a base of the same topology! Let U be an arbitrary element of a base.} is closed. 2p. d. it follows that each of them is closed. belonging to Ω1 ) belongs also to Ω2 . 3. say.6)? 3. 3. Therefore. our advice concerning 3. +∞). can be presented as a union of sets of the base under consideration. Hence if the set of prime numbers was ﬁnite. . iﬀ ∀ U ∈ Σ1 ∀ x ∈ U ∃ V ∈ Σ2 : x ∈ V ⊂ U . All together the sets of the form {p. . . Moreover.7 The statement: “B is a base of a topological structure” is equivalent to the following: the set of unions of all collections of sets belonging to B is a topological structure. 3. 2p.3 The whole topological structure is its own base.e.. you need to prove that a union of disks can be presented as a union of squares and vice versa. in turn.9 Observe that intersection of arithmetic progressions is an arithmetic progression. 3. It can be presented as a union of open intervals. Hence open sets are linearly ordered by inclusion. by bases Σ1 and Σ2 . the question is when this is the only base.5 and 3. since it is not contained in so short intervals. Hence U is a union of elements of the base distinct from U and it can be replaced by this union in a presentation of an open set as a union of elements of the base. 3.11 Inclusion Ω1 ⊂ Ω2 means that a set open in the ﬁrst topology (i. (d) In the arrow {[0. i + d. you need to prove analogues of 3. Σ1 is a base of some topology by 3. since otherwisean open set could be obtained as a union of inﬁnite increasing sequence of open sets.6 In solution of each of these problems the following easy lemma may help: A = Bα . We would need at least two such intervals.
iﬀ x = y. where ρ is a metric in X. y) ≤ ρ(x.. x) for every x. These metrics are called Euclidean. y) → 0. y) → i=1 xi − yi  p 1 p . Metric of 4.C–4. 4. 23 . yi ≥ 0. n i=1 (xi (x. y) → maxi=1. Prove that o n i=1 n 1/p n q yi i=1 1 q 1/q xi y i ≤ xp i i=1 1 and p if xi .B. Prove that Rn × Rn → R + : (x. if x = y Metrics 4. an .. a2 . 4. y) → 4. for any positive a1 .C. (c) ρ(x. y) → n i=1 − yi)2 is a metric.B is a special case of metric 4. z ∈ X. . Metric 4. Prove that ρ(p) is a metric for any p ≥ 1.3.1 H¨lder Inequality.3. y. Prove that Rn × Rn → R + : (x. ρ). The pair (X. z) + ρ(z. . (b) ρ(x.. if (a) ρ(x.n xi − yi  is a metric.. q > 0 + = 1. is called a metric space. Prove that R × R → R + : (x.2 is ρ(1). y ∈ X.2 are included in inﬁnite series of the metrics n ρ (p) : (x. p.§ 4 Metric Spaces § 4 ◦ 1 Deﬁnition and First Examples A function ρ : X × X → R + = { x ∈ R  x ≥ 0 } is called a metric (or distance) in X. 4. xi − yi  is a metric.A. .C. Metrics in Rn introduced in 4. y) = ρ(y. and metric of 4. p ≥ 1. y) → x − y is a metric.1. y) = 0. if x = y. 4. 4. § 4 ◦ 2 Further Examples 4. 1.C is ρ(2). .2. y) for every x.C are always meant when R and Rn are considered as metric spaces unless another metric is speciﬁed explicitly. Prove that for any set X ρ : X × X → R+ : is a metric. The condition (c) is triangle inequality. Prove that Rn × Rn → R + : (x.B and 4.1 can be denoted by ρ(∞) and adjoined to the series since n p→+∞ lim ap i i=1 1 p = max ai . metric of 4.
§ 4 ◦ 4 Subspaces of a Metric Space If (X. The last two statements clarify the origin of terms sphere and ball (in the context of metric spaces). What if r < ρ(x. Check that D 1 is the segment [−1. D 2 is a disk. They are considered as metric spaces (with the metric restricted from Rn ). ρ) be a metric space. Enigma. and let r be a positive real number.4. The sets (5) (6) (7) Dr (a) = { x ∈ X  ρ(a. Prove that for points x and a of any metric space and any r > ρ(a. y ∈ l(p) the series converges and that ∞ ∞ i=1 xi − yi p (x. S 1 is a circle. D 3 is a ball. respectively. are called. then the restriction of metric ρ to A × A is a metric in A. ρ) is a metric space and A ⊂ X.C) are denoted by symbols D n and S n−1 and called ndimensional ball and (n − 1)dimensional sphere. let a be its point.D. 4.. 4. y) → is a metric in l(p) . Σ∞. ρ A×A) is a metric space.5.§ 4. It is called a subspace of (X. S 2 is a sphere.E in this case? ...2. ∞ such that the series i=1 xp converges. closed ball. x) ≤ r }. i=1 xi − yi  p 1 p . Prove that for any two elements x. ρ). METRIC SPACES 24 4. Some properties of balls and spheres in arbitrary metric space resemble familiar properties of planar disks and circles and spatial balls and spheres. a)? What is an analogue for the statement of Problem 4. 1]. x) = r } Br (a) = { x ∈ X  ρ(a. p≥1 § 4 ◦ 3 Balls and Spheres Let (X. S 0 is the pair of points {−1. x) < r }. x) Dr−ρ(a. open ball.x) (x) ⊂ Dr (a).6. Sr (a) = { x ∈ X  ρ(a. ρ) with center at a and radius r. 4.E. Enigma. 1}. 4. and Σ1 from Section § 3? For a real number p ≥ 1 denote by l(p) the set of sequences x = {xi }i=1. How is this related to Σ2. The ball D1 (0) and sphere S1 (0) in Rn (with Euclidean metric. and sphere of the space (X. and (A. see 4.
8. 2 4.12 look like with ρ of 4.10. y) → x − y The vector space equipped with a norm is called a normed space.) § 4 ◦ 7 Bounded Sets and Balls A subset A of a metric space (X. x) + ρ(x. METRIC SPACES 25 § 4 ◦ 5 Surprising Balls However in other metric spaces balls and spheres may have rather surprising properties. ρ : X × X → R + : (x.11. and S 2 (a) in the space of 4. Find a metric space and two balls in it such that the ball with the smaller radius contains the ball with the bigger one and does not coincide with it. 4.15. 4. 4. What is the relation between the minimal radius of such a ball and diam(A)? § 4 ◦ 8 Norms and Normed Spaces Let X be a vector space (over R).9. iﬀ it is contained in a ball. Prove that in 4. The greatest lower bound of such d is called the diameter of A and denoted by diam(A). b) }. 4. What is the minimal number of points in the space which is required to be constructed in 4. Find D1 (a).§ 4. Prove that if x → x is a norm then is a metric. { x ∈ Rn  ρ(a. (b) λx = λx for any λ ∈ R and x ∈ X. y ∈ X.2 (cf. 4.13. ρ) is said to be bounded.14.1 and 4. 4. iﬀ x = 0. D 1 (a). . Prove that a set A is bounded.F. (c) x + y ≤ x + y for any x. if there is a number d > 0 such that ρ(x. § 4 ◦ 6 Segments (What Is Between) 4. b) = ρ(a. y) < d for any x. b ∈ Rn can be described as where ρ is the Euclidean metric. The metric deﬁned by the norm as in 4.9 the big radius does not exceed double the smaller radius.9. What are balls and spheres in R2 with metrics of 4. Prove that the segment with end points a.A. y ∈ A. 4.1 and 4.12.15 turns the normed space into the metric one in a canonical way.2? (Consider the case n = 2 if it appears to be easier. 4. Function X → R + : x → x is called a norm if (a) x = 0. How do the sets deﬁned as in 4.4)? 1 4.7.
3. in this space). so it makes sense only for subsets of spaces. when one says about open and closed sets. iﬀ it contains together with any its point a ball with center at this point.H.G. Recall that a set A is said to be convex if for any x. which is open (with respect to the metric topology). The collection of all open balls in the metric space is a base for some topology (cf. In any metric space. Prove that a closed ball is closed (with respect to the metric topology). which is closed (with respect to the metric topology). in fact. Of course.x) (x) for any point a. Find a closed ball. 4.G.I.§ 4. 3. this deﬁnition is based on the notion of segment. y ∈ A the segment connecting x.A. 4. METRIC SPACES 26 4. 4. 4. A set is open in a metric space.20.17. Find a sphere. 4. It is said to be induced by the metric. This is the case in vector and aﬃne spaces over R 3 . real number r > 0 and point x ∈ Br (a). Prove that every ball in the normed space is a convex3 set symmetric with respect to the center of the ball. Prove that the standard topological structure in R introduced in Section § 2 is induced by metric (x. Br (a) ⊃ Br−ρ(a. and that this norm is uniquely deﬁned by this ball.1 Lemma.B and 4.16.23.21. neighborhoods.19.18*. which is symmetric with respect to its center and is not contained in any aﬃne space except Rn itself. Look through the problems of this section and ﬁgure out which of the metric spaces involved are. Find an open ball. 4. y) → x − y. normed vector spaces. Prove that a sphere is closed.E).24. 4. § 4 ◦ 10 Openness and Closedness of Balls and Spheres 4. is the unit ball with respect to some norm. § 4 ◦ 9 Metric Topology 4. which is open. What topological structure is induced by the metric of 4. Prove that every convex closed bounded set in Rn . y is contained in A. This topology is called metric topology. where the notion of segment connecting two point is deﬁned. 4. This topological structure is always meant whenever the metric space is considered as a topological one (for instance. etc.22. 4.A? 4.
4. c) are equal). y) for any x. z). Generally speaking the inverse is not true.K.27 can be weakened. y ∈ X. ρC (f. y) ≤ Cρ1 (x. Prove that the following two metrics ρ1 . C > 0 such that cρ1 (x. ρ(b. b).25. 4:A.26.1. x∈[0. 1] → R are not equivalent: 1 ρ1 (f.e. Which topological spaces described in Section § 2 are metrizable? § 4 ◦ 12 Equivalent Metrics Two metrics in the same set are said to be equivalent if they induce the same topology.2 is ultrametric. METRIC SPACES 27 § 4 ◦ 11 Metrizable Topological Spaces A topological space is said to be metrizable if its topological structure is induced by some metric. ρ(a. 4:C. ρ2 in X are equivalent if there are numbers c. c two of the three distances ρ(a.3 are equivalent.27. 4. 4. y)} for any x. Check that only one metric in 4. A metric space (X.J.A–4. A ﬁnite space is metrizable iﬀ it is discrete. and 4. . ρC in the set of all continuous functions [0. 4. g) = 0 f (x) − g(x) dx. ρ) with ultrametric ρ is called an ultrametric space. How? 4. An indiscrete space is not metrizable unless it consists of a single point (it has too few open sets). Prove that in an ultrametric space all triangles are isosceles (i. 4.30. Enigma. 4.1] Is it true that topological structure deﬁned by one of them is ﬁner than another? § 4 ◦ 13 Ultrametric A metric ρ is called an ultrametric if it satisﬁes to ultrametric triangle inequality : ρ(x. Which one? 4:B. y) ≤ max{ρ(x. z. 4.31*.23) but also open. Hence the condition of the equivalence of metrics formulated in 4.. 4. Metrics ρ(p) in Rn deﬁned right above Problem 4. Prove that metrics ρ1 .29. c). g) = max f (x) − g(x) . y) ≤ ρ2 (x. 4.2 equivalent? 4.C. Prove that in a ultrametric space spheres are not only closed (cf. ρ(z.28. Are the metrics of 4. for any three points a. b.§ 4. y.
§ 4. METRIC SPACES
28
The most important example of ultrametric is padic metric in the set Q of all rational numbers. Let p be a prime number. For x, y ∈ Q, present the diﬀerence x − y as r pα , where r, s, and α are integers, and r, s are relatively prime with p. Put s ρ(x, y) = p−α . 4:D. Prove that this is an ultrametric.
§ 4 ◦ 14 Operations with Metrics
4.32. Prove that if ρ1 , ρ2 are metrics in X then ρ1 + ρ2 and max{ρ1 , ρ2 } are ρ1 also metrics. Are the functions min{ρ1 , ρ2 }, , and ρ1 ρ2 metrics? ρ2 4.33. Prove that if ρ : X × X → R + is a metric then (a) function ρ(x, y) (x, y) → 1 + ρ(x, y) is a metric; (b) function (x, y) → f ρ(x, y) is a metric, if f satisﬁes the following conditions: (1) f (0) = 0, (2) f is a monotone increasing function, and (3) f (x + y) ≤ f (x) + f (y) for any x, y ∈ R. 4.34. Prove that metrics ρ and ρ are equivalent. 1+ρ
§ 4 ◦ 15 Distance Between Point and Set Let (X, ρ) be a metric space, A ⊂ X, b ∈ X. The inf{ ρ(b, a)  a ∈ A } is called a distance from the point b to the set A and denoted by ρ(b, A). 4.L. Let A be a closed set. Prove that ρ(b, A) = 0, iﬀ b ∈ A.
4.35. Prove that ρ(x, A) − ρ(y, A) ≤ ρ(x, y) for any set A and points x, y of the same metric space. § 4 ◦ 16 Distance Between Sets Let A and B be bounded subsets in the metric space (X, ρ). Put dρ (A, B) = max sup ρ(a, B), sup ρ(b, A) .
a∈A b∈B
This number is called the Hausdorﬀ distance between A and B. 4:E. Prove that the Hausdorﬀ distance in the set of all bounded subsets of a metric space satisﬁes the conditions (b) and (c) of the deﬁnition of metric. 4:F. Prove that for every metric space the Hausdorﬀ distance is a metric in the set of its closed bounded subsets.
§ 4. METRIC SPACES
29
Let A and B be bounded polygons in the plane4 . Put d∆ (A, B) = S(A) + S(B) − 2S(A ∩ B), where S(C) is the area of polygon C. 4:G. Prove that d∆ is a metric in the set of all plane bounded polygons. We will call d∆ the area metric. 4:H. Prove that in the set of all bounded plane polygons the area metric is not equivalent to the Hausdorﬀ metric. 4:I. Prove that in the set of convex bounded plane polygons the area metric is equivalent to the Hausdorﬀ metric. § 4 ◦ 17 Asymmetrics A function ρ : X × X → R + is called an asymmetric in set X, if (a) ρ(x, y) = 0 and ρ(y, x) = 0, iﬀ x = y; (b) ρ(x, y) ≤ ρ(x, z) + ρ(z, y) for any x, y, z ∈ X. Thus, an asymmetric satisﬁes the conditions a and c of the deﬁnition of metric, but does not satisfy condition b. An example of asymmetric taken from “the real life”: the shortest length of path from one point to another by a car in a city in which there exist one way streets. 4:J. Prove that if ρ : X × X → R + is an asymmetric then the function (x, y) → ρ(x, y) + ρ(y, x) is a metric in X. Let A and B be bounded subsets of a metric space (X, ρ). The number aρ (A, B) = supb∈B ρ(b, A) is called the asymmetric distance from A to B. 4:K. aρ in the set of nounded subsets of a metric space satisﬁes the triangle inequality from the deﬁnition of asymmetric. 4:L. In a metric space (X, ρ), a set B is contained in all the closed sets containing A, iﬀ aρ (A, B) = 0. 4:M. Prove that aρ is an asymmetric in the set of all bounded closed subsets of a metric space (X, ρ) . Let A and B be polygons on the plane. Put a∆ (A, B) = S(B) − S(A ∩ B) = S(B where S(C) is the area of polygon C. 4:1. Prove that a∆ is an asymmetric in the set of all planar polygons. Although we assume that the notion of bounded polygon is wellknown from elementary geometry, recall the deﬁnition. A bounded plane polygon is a set of the points of a simple closed polygonal line and the points surrounded by this line. By a simple closed polygonal line we mean a cyclic sequence of segments such that each of them starts at the point where the previous one ﬁnishes and these are the only pairwise intersections of the segments.
4
A),
§ 4. METRIC SPACES
30
A pair (X, ρ), where ρ is an asymmetric in X, is called an asymmetric space. Of course, any metric space is an asymmetric space, too. In an asymmetric space, balls (open and closed) and spheres are deﬁned like in a metric space, see § 4 ◦ 3. 4:N. The set of all open balls of an asymmetric space is a base of some topology. This topology is said to be generated by the asymmetric. 4:2. Prove that formula a(x, y) = min(x − y, 0) deﬁnes an asymmetric in [0, ∞), and that the topology generated by this asymmetric coincides with the arrow topology, see § 2 ◦ 2.
Proofs and Comments 4.A Indeed, it makes sense to check that all the conditions of the deﬁnition of metric is satisﬁed for each combination of points x, y z. 4.B Triangle inequality in this case looks as follows x − y ≤ x − z + z − y. Put a = x − z, b = z − y. This turns the triangle ineguality to a wellknown inequality a + b ≤ a + b. 4.C As in the solution of Problem 4.B, the triangle inequality can be n n n 2 2 2 rewritten as follows: i=1 ai + i=1 bi . By two i=1 (ai + bi ) ≤ squaring followed by an obvious simpliﬁcation, this inequality is reduced to the wellknown Cauchy inequality ( ai bi )2 ≤ a2 b2 . i i 4.F Show that if d = diam A and a ∈ A then A ⊂ Dd (a). Vice versa: diam Dd (a) ≤ 2d (cf. 4.11).
4.G.1 We have to prove that any point y ∈ Br−ρ(a,x) (x) belongs to Br (a). In terms of distances, this means that ρ(y, a) < r, if ρ(y, x) < r − ρ(a, x) and ρ(a, x) < r. By the triangle inequality, ρ(y, a) ≤ ρ(y, x) + ρ(x, a). Replacing in the righthand side of the latter inequality the ﬁrst summand by a greater number r − ρ(a, x), we get the desired inequality. 4.G It is claimed that Ω = {∪Br (x)  r > 0, x ∈ X} is a topological structure. This follows from Lemma 4.G.1 and Theorems 3.B and 3.C. 4.H For this metric, the balls are open intervals. Each open interval in R appears as a ball. The standard topology in R is deﬁned by the base consisting of all open intervals. 4.I If a set contains together with any of its points a ball with center at this point, this set is the union of those balls. Thus, it is open in the metric topology. If a ∈ U, where U is open, then a ∈ Br (x) and Br−ρ(a,x) (a) ∈ Br (x) ⊂ U, see Lemma 4.G.1. 4.J An indiscrete space does not have enough open sets. For x, y ∈ X and r = ρ(x, y) > 0, the ball Dr (x) is not empty and does not coincide with the whole space.
§ 4. METRIC SPACES
31
4:A Clearly, the metric in 4.A is an ultrametric. The other metrics are not: for each of them you can ﬁnd points x, y, z such that ρ(x, y) = ρ(x, z) + ρ(z, y). 4:B The deﬁnition of ultrametric implies that no one of pairwise distances between points a, b, c can be greater than each of the other two. 4:C By 4:B, if y ∈ Sr (x) and r > s > 0 then Bs (y) ⊂ Sr (x). 4:D + y = max{ρ(x, z), ρ(z, y)}.
r pα1 s1 1
Let x − z =
r2 α2 −α1 p s2
r1 α1 p , s1
=
r2 α2 p and α1 s2 α1 r1 s2 +r2 s1 pα2 −α1 p , hence s1 s2
z−y =
≤ α2 . Then: x −
p(x, y) ≤ p−α1 =
4.L Condition ρ(b, A) = 0 means that each ball centered at b meets A. In turn, this means that b does not belong to the complement of A (since A is closed).
Hints, Comments, Advises, Solutions, and Answers
4.2 Cf. 4.B. 4.4 Look for an answer in 4.7. 4.7 Squares with sides parallel to the coordinate axes and bisectors of the coordinate angles, respectively. 4.8 D1 (a) = X, D1/2 (a) = {a}, S1/2 (a) = ∅. 4.9 For example, X = D1 (0) ⊂ R1 , and D3/2 (5/6) ⊂ D1 (0). 4.10 Three points suﬃce. 4.11 Let R > r and DR (b) ⊂ Dr (a). Take c ∈ DR (b) and use the triangle inequality ρ(b, c) ≤ ρ(b, a) + ρ(a, c). 4.12 Put u = b − x and t = x − a. The Cauchy inequality becomes equality, iﬀ the vectors u and t has the same direction, i.e., x lies on the segment connecting a and b. 4.13 For metric ρ(p) with p > 1 this set coincides with the segment connecting a and b, and for metric ρ(1) it is a rectangular parallelipiped whose opposite vertices are those two points. 4.14 See the proof of 4.F. 4.19 The discrete one. 4.20 Let us just remind you that you need to prove that X {x  ρ(x, a) > r} is open. 4.23 Use the obvious equality X the result of 4.20. Sr (a) = Br (a) ∪ (X Dr (a) =
Dr (a)) and
4.K For x ∈ X put r = min{p(x, y)y ∈ X Dr (x)?
x}. Which points are in
§ 4. METRIC SPACES
32
4.25 Only line and discrete spaces. 4.26 According to 3.7, for n = 2 metrics ρ(2) , ρ(1) ρ(∞) are equivalent; similar arguments work for n > 2, too. However in this case it is more convenient to use the result of the next problem: to show that for any pair of metrics ρ(p) (1 ≤ p ≤ ∞) there exist appropriate constants c and C, required in 4.27. 4.27 First, let us prove that Ω2 ⊂ Ω1 , provided ρ2 (x, y) ≤ Cρ1 (x, y). In(ρ ) (ρ2 ) deed, inequality ρ2 ≤ Cρ1 implies Br 1 (a) ⊂ BCr . Now let us use Theorem 4.I. Inequality cρ1 (x, y) ≤ ρ2 (x, y) can be represented as ρ1 (x, y) ≤ 1 ρ2 (x, y). c Hence Ω1 ⊂ Ω2 . 4.28 Metrics ρ1 (x, y) = x − y ρ2 (x, y) = arctan x − y on the line are equivalent, but obviously there is no number C such that ρ1 ≤ Cρ2 . 4.29 Metrics ρ1 , ρ2 are equivalent, if there exist c, C, d > 0, such that ρ1 (x, y) ≤ d implies cρ1 (x, y) ≤ ρ2 (x, y) ≤ Cρ1 (x, y). 4.31 Ω1 ⊂ Ω2 , because ρ1 (f, g) ≤ ρC (f, g). On the other hand, there is (ρ ) no ball centered at the origin for metric ρ1 , which would not ﬁt to B1 C (0), since ∀ ε > 0 ∃ f f  < ε, max f  ≥ 1.
[0,1]
4.32 Clearly in all the cases the only thing which is to be proved and is not completely obvious is the triangle inequality. Forρ1 + ρ2 it is obvious either. Furthermore ρ1 (x, y) ≤ ρ1 (x, z) + ρ1 (z, y) ≤
max{ρ1 (x, z), ρ2 (x, z)} + max{ρ1 (y, z), ρ2 (y, z)}.
A similar inequality holds true for ρ2 (x, y), therefore max{ρ1 , ρ2 } is a metρ1 ric. Construct examples which would prove that neither min{ρ1 , ρ2 }, nor ρ2 , nor ρ1 ρ2 is a metric (for this it would be enough to ﬁnd three points with appropriate pairwise distances). 4.33 The latter statement is quite obvious. The ﬁrst and the second t ones follow from the last one for f (t) = 1+t and f (t) = min 1, t, respectively. Thus it suﬃes to check that these function satisfy the conditions of the last statement.
ρ 4.34 Since 1+ρ ≤ ρ, and for ρ(x, y) ≤ 1 inequality holds true, the statement follows from 4.29. 1 2 ρ(x, y)
≤
ρ(x,y) 1+ρ(x,y)
4:E Condition (b) is obviously satisﬁed. Put r(A, B) = sup ρ(a, B), so
a∈A
that dρ (A, B) = max{r(A, B), r(B, A)}. To prove that (c) is also satisﬁed, it suﬁcies to prove that r(A, C) ≤ r(A, B) + r(B, C) for any sets A, B, C ⊂ X. One can easily see that ρ(a, C) ≤ ρ(a, b) + ρ(b, C) for all a ∈ A, b ∈ B. Hence ρ(a, C) ≤ ρ(a, b) + r(B, C), and therefore ρ(a, C) ≤ inf b∈B ρ(a, b) + r(B, C) = ρ(a, B) + r(B, C) ≤ r(A, B) + r(B, C), which implies the desired inequality. 4:F By 4:E, dρ satisﬁes (b) (c) from the deﬁnition of metric. From 4.L it follows that if the Hausdorﬀ metric between closed sets A and B equals zero then A ⊂ B and B ⊂ A, i. e. A = B. Thus dρ satisﬁes (a). 4:G d∆ (A, B) is the area of the symmetric diﬀerence of A and B, i. e. the area of A∆B = (A B) ∪ (B A). The ﬁrst two axioms of metric
METRIC SPACES 33 are obviously satisﬁed.§ 4. B⊂ . Prove the triangle inequality using inclusion A (C B) ∪ (A C).
A binary relation in a set X is called a strict partial order.A Antisymmetry. if such a notation is used. Let be a strict partial order in a set X. or just nonstrict order.§ 5 Ordered Sets This section is devoted to orders. The advantage of the latter two ways of reading is that then the relation is not associated too closely to the inequality between real numbers. Many relations are denoted by special symbols. Is the relation ab a nonstrict partial order in the set Z of integers? 34 . virtually all ﬁnite topological spaces appear in this way. § 5 ◦ 1 Strict Orders A binary relation in a set X is a set of ordered pairs of elements of X. This is a source of interesting and important examples of topological spaces. y ∈ X such that both x y and y x hold true. <. They are structures in sets and occupy in Mathematics a position almost as profound as topological structures. or x ∼ y. or ≡.1. • Antisymmetry : If a b and b a then a = b for any a. 5.B. • Reﬂexivity : a a for any a ∈ X. ⊂. or ∼. or just strict order. 5.D. c ∈ X. b. partially ordered sets give rise to natural topological structures. Like metric spaces. like . 5. b. In the case. etc. c ∈ X. we write x y. that is a subset R ⊂ X × X. if it satisﬁes the following two conditions: • Irreﬂexivity : There is no a ∈ X such that a a holds. 5. b ∈ X. In the set N of natural numbers the relation ab (a divides b) is a nonstrict partial order. we will focus on relations between structures of these two types. if it satisﬁes the following three conditions: • Transitivity : If a b and b c then a c for any a. ≤. § 5 ◦ 2 NonStrict Orders Binary relation in a set X is called a nonstrict partial order. >. or x y. There exist no x. • Transitivity : a b and b c imply a c for any a. Formula a b is read sometimes as “a is less than b” or “b is greater than a”. there is a tradition to write xRy instead of writing (x. y) ∈ R. 5. Relation ≤ in R is a nonstrict order. Relation < in the set R of real numbers is a strict order. So. This generalizes the notation for classical binary relations =. or .C. After a short general introduction. As we will see later (in Section § 19). but often it is read as “a is followed by a” or “a precedes b”.
For each strict order . which is almost never understood as notation solely for strict inclusion.§ 5. if either a b or a = b. A set equipped with a partial order (either strict or nonstrict) is called a partially ordered set or poset. One says that the strict order nonstrict order . 5. instead of a strict partial order one can use the corresponding nonstrict one . We have already met with a similar phenomenon in topology: open and closed sets in a topological space deﬁne each other and provide diﬀerent ways of presenting a topological structure. nobody introduces notation for a strict divisibility. Although it would be handy to keep both of them available. strict and nonstrict orders are deﬁned by each other.E. This relation is a nonstrict order. strict or nonstrict. prevails in each concrete case is a matter of convenience. a nonstrict one by symbol .G is associated to the original 5. nonstrict orders conquer situation by situation. § 5 ◦ 4 Cones Let (X. This relation is a strict of 5. of 5. taste and tradition. They are just diﬀerent incarnations of the same structure of order. The set {x ∈ X  a x} is called the upper cone of a. if a order. which used to denote nonstrict inclusion. More formally speaking. Thus. ) be a poset and a ∈ X. Certainly. The construction of the two preceding problems are inverse to each other: applied one after another in any order.H. Another example: symbol ⊆. § 5 ◦ 3 Relation between Strict and NonStrict Orders 5.F is associated to the . there is a relation deﬁned in the b and a = b. In the set of subsets of a set X inclusion is a nonstrict partial order. is replaced by symbol ⊂. Which of the orders. a partially ordered set is a pair (X. ORDERED SETS 35 5.F. they give the initial relation. there is a relation deﬁned in the same set as follows: a b. In abstract considerations we will use both kinds of orders. a strict partial order is denoted by symbol . One says that the nonstrict order original strict order . For each nonstrict order same set as follows: a b. For instance. ) formed by a set X and a strict partial order in X.G. and the set {x ∈ X  x a} the — lower cone of .
An element b of a set A is called its maximal element. ) be a poset. (c) Ca = Cb implies a = b. which is deﬁned as follows: a C b. is the time) there is a relation event (x1 .3. x3 . Let C ⊂ R3 be a set. Let (X. if b ∈ A and c b for every c ∈ A. ˜ c(t − t) ≥ (x1 − x1 )2 + (x2 − x2 )2 + (x3 − x3 )2 . 0) such that P ∩ C = (0.J Cones Determine an Order. Prove that any convex cone C in R3 with vertex (0. x2 . x1 . b is said to be the smallest element of A. + + + CX (b) ⊂ CX (a). are the spatial coordinates and the fourth one. 5. Similarly. 0. In the spacetime R4 of special relativity theory (where points represent moment point events. What properties of C would imply that C would be a partial order in R3 ? In the poset (R3 . x2 .L. (b) a ∈ Ca for each a ∈ X. if b ∈ Ca . By adding it to them. iﬀ A ⊂ CX (b). 0. x3 . t. Each set has at most one greatest and at most one smallest element. iﬀ A ⊂ CX (b). Answer the versions of questions of the preceding problem concerning twodimensional and threedimensional versions of this space. Let us write a b. ) be a poset. 0) for some plane P satisﬁes the conditions found in solution of the preceding problem. Then the relation is a nonstrict order + in X and for this order CX (a) = Ca . x2 . b is the smallest element of A. 5. + 5. The element a does not belong to its cones. We say that b is the greatest element of A.K. An element b of a set A is called its minimal element. If (a) b ∈ Ca implies Cb ⊂ Ca . Consider relation C in R3 . 5.I (a) (b) (c) Properties of Cones. if b ∈ A and b c for every c ∈ A. . the ﬁrst three coordinates. + a ∈ CX (a) for each a ∈ X.§ 5. in which the number of spatial coordinates is 1 and 2 respectively. if A does not contain an element c such that c b. what are the upper and lower cones? 5.2. This relation is deﬁned by inequality Is this a partual ordr? If yes.4. if A does not contain an element c such that b c. ORDERED SETS 36 a. 5. t). 5. § 5 ◦ 5 Position of an Element with Respect to a Set Let (X. + + CX (a) = CX (b) implies a = b. what are the upper and lower cones of an event? 5. − b is the greatest element of A. C ). Let X be an arbitrary set. if b − a ∈ C.5. t) precedes (and can inﬂuence) event (x1 . Suppose for any a ∈ X one ﬁxes a subset Ca ⊂ X. we + get completed cones: the upper completed cone or star CX (a) = {x ∈ X  − a x} and the lower completed cone CX (a) = {x ∈ X  x a}. provided b ∈ CX (a). A be its subset. x3 .
§ 5. ORDERED SETS
− 5.M. An element b of A is maximal, iﬀ A ∩ CX (b) = b. + An element b of A is minimal, iﬀ A ∩ CX (b) = b.
37
5.6. Enigma. How are the notions of maximal and greatest elements related? What can you say about a poset, in which for any subset these notions coincide?
§ 5 ◦ 6 Total Orders Please, notice: the deﬁnition of a strict order does not require that for any a, b ∈ X either a b, or b a, or a = b. This condition is called a trichotomy. In terms of the corresponding nonstrict order, it can be reformulated as follows: any two elements a, b ∈ X are comparable: either a b, or a b. A strict order which satisﬁes trichotomy is said to be total or linear. The corresponding poset is said to be linearly or totally ordered. It is called also just an ordered set.5 Some orders do satisfy trichotomy. 5.N. The order < in the set R of real numbers is linear. This is the most important example of totally ordered set. The words and images rooted in it are often extended to all totally ordered set. For example, cones are called rays, the upper cones turn to right rays, while lower cones to left rays.
+ − 5.7. A poset (X, ) is linearly ordered, iﬀ X = CX (a) ∪ CX (a) for each a ∈ X.
5.8. In the set N of natural numbers the order ab is not linear. 5.9. For which X the relation of inclusion in the set of all subsets of X is a linear order?
§ 5 ◦ 7 Topologies Deﬁned by a Total Order 5.O. Let (X, ) be a totally ordered set. The set of all its subsets consisting of all the right rays, i.e., sets of the form {x ∈ X  a x}, where a runs over X, and the set X itself is a base of a topological structure in X.
Quite a bit of confusion was brought into the terminology by Bourbaki. Then total orders were called orders, nontotal orders were called partial orders, and in occasions when it was not known if the order under consideration was total, the fact that this was unknown was explicitly stated. Bourbaki suggested to withdraw the word partial. The motivation for this was that a partial order, as a phenomenon more general than a total order, deserves a shorter and simpler name. In French literature this suggestion was commonly accepted, but in English it would imply abolishing of a nice short word poset that seems to be absolutely impossible.
5
§ 5. ORDERED SETS
38
The topological structure deﬁned by this base is called the right ray topology of linearly ordered set (X, ). Left ray topology is deﬁned similarly: it is generated by the base consisting of X and sets of the form {x ∈ X  x a} with a ∈ X.
5.10. The topology of the arrow (see. § 2) coincides with the right ray topology of halﬂine [0, ∞) equipped with the order <. 5.11. Enigma. To what extent is the assumption that the order is linear necessary in Theorem 5.O? Find a weaker condition which would imply the conclusion of Theorem 5.O and allow one to speak about the topological structure described in Problem 2.2 as the right ray topology of an appropriate partial order on the plane.
5.P. Let (X, ) be a totally ordered set. The set of its subsets consisting of X and all sets of the form {x ∈ X  a x b}, where a and b run over the whole X is a base of a topological structure in X. The topological structure deﬁned by this base is called the interval topology of a linearly ordered set.
5.12. Prove that the interval topology is the smallest topological structure containing the right ray and left ray topological structures.
5.Q. The canonical topology of the line coincides with the interval topology of (R, <). § 5 ◦ 8 Poset Topology 5.R. Let (X, ) be a poset. The set of subsets of X consisting of all the sets of form {x ∈ X  a x}, where a runs over the whole X, is a base of a topological structure in X. The topological strucuter generated by this base is called a poset topology. 5.S. In a poset topology each point a ∈ X has the smallest (with respect to inclusion) neighborhood. This is {x ∈ X  a x}. 5.T. The following properties of a topological space are equivalent: • each point has a smallest neighborhood, • the intersection of any collection of open sets is open, • the union of any collection of closed sets is closed.
A space satisfying the conditions of Theorem 5.T is called a smallest neighborhood space.6 In a smallest neighborhood space open and closed sets satisfy the same conditions. In particular, the set of all closed sets of
This class of topological spaces was introduced and studied by P. S. Aleksandrov in 1935. Aleksandrov called them discrete. Nowadays the term discrete space is used for a much narrower class of topological spaces (see Section § 2). The term smallest neighborhood space was introduced by Christer Kiselman.
6
§ 5. ORDERED SETS
39
a smallest neighborhood space is a topological structure. This structure is said to be dual to the original one. It corresponds to the opposite partial order.
5.13. How to characterize points open in a poset topology in terms of the partial order? The same question about closed points. 5.14. Describe directly open sets in the poset topology of R with order <. 5.15. In set {a, b, c, d} consider a partial order in which the strict inequalities are: c a, d c, d a, d b. Check that this is a partial order and that the poset topology coincides with the topology of , described in Problem 2.3 (a).
§ 5 ◦ 9 How to Draw a Poset Now we can explain a pictogram , which we use to denote the space introduced in Problem 2.3 (a). It describes the partial order in {a, b, c, d}, which deﬁnes the topology of this space by 5.15. Indeed, if we place the elements of the set under consideration at vertices of the graph of the pictogram, as shown in the picture, the vera tices corresponding to comparable elements occur to be b c connected by a segment or ascending broken line and the d greater element correspond to the higher vertex. In this way one can draw a scheme representing any ﬁnite poset. Elements of the poset are represented by points and a b, iﬀ the point repsenting b is above the point representing a and those points are connected either by a segment or broken line consisting of segments which connect points representing intermediate elements of a chain a c1 ··· cn b. One could connect by a segment any two points c2 corresponding to comparable elements, but this would make the diagram excessively cumbersome. This is why segments which could be recovered from the others by transitivity are not drawn. A diagram of this kind representing a poset is called its Hasse diagram. 5.U. Prove that any ﬁnite poset can be described by a Hasse diagram. 5.V. Discribe the topological structure in set Z of integers which is the poset topology deﬁned by the following Hasse diagram −5 −4 −3 −2 −1 0 1 3 5
2
4
6
The space of Problem 5.V is called the digital line or Khalimsky line. In this space each even number is closed and each odd open.
§ 5. ORDERED SETS
40
5.16. Associate to each even integer 2k the interval (2k − 1, 2k + 1) of length 2 centered at this point, and to each odd integer 2k−1, the singleton {2k−1}. Prove that a set of integers is open in the Khalimsky topology, iﬀ the union of sets associated to its elements is open in R with the standard topology. 5.17. Among the topological spaces described in Section § 2, ﬁnd all that can be obtained as posets with the poset topology. In the cases of ﬁnite sets, draw Hasse diagrams describing the corresponding partial orders.
§ 5 ◦ 10 Cyclic Orders in Finite Set Recall that a cyclic order in a ﬁnite set X is a linear order considered up to cyclic permutatuion. A linear order allows one to enumerate elements of the set X by natural numbers, so that X = {x1 , x2 , . . . , xn }. A cyclic permutation transpose the ﬁrst k elements with the last n − k elements without change of the ordering inside of each of these two parts of the set: (x1 , x2 , . . . , xk , xk+1 , xk+2 , . . . , xn ) → (xx+1 , xk+2 , . . . , xn , x1 , x2 , . . . , xk ). When we consider a cyclic order, it does not make sense to say that one of its elements is greater than other one, since an appropriate cyclic permutation put these two elements in the opposite order. However, it makes sense to say that an element is followed immediately by other one. Certainly, the very last element is followed immediately by the very ﬁrst: indeed, any nonidentity cyclic permutation puts the ﬁrst element immediately after the last one. In a cyclicly ordered ﬁnite set for each element a there is a unique element b next to a, that is which follows a immediatly. This deﬁnes a map of the set onto itself, the simplest cyclic permutation xi → xi+1 , x1 , if i < n if i = n.
This permutation acts transitively (i.e., any element is mapped to any other one by an appropriate iteration of it). 5.W. Any map T : X → X, which acts transitively in X, deﬁnes a cyclic order in X such that each a ∈ X is followed by T (a). 5.X. In a set consisting of n elements possesses exactly (n − 1)! pairwise distinct cyclic orders. In particular, in a two element set there is only one cyclic order (which so uninteresting that sometimes it is said to make no sense), and a set consisting of three elements possesses two cyclic orders.
§ 5. ORDERED SETS
41
§ 5 ◦ 11 Cyclic Orders in Inﬁnite Sets One can consider cyclic orders in an inﬁnite set. However most of what was said above cannot apply to cyclic orders in inﬁnite sets without adjustment. In particular, most of them cannot be described by showing pairs of elements which are next to each other. For example, points of a circle can be cyclically ordered clockwise (or counterclockwise), but with respect to this cyclic order no point follows is immediately by other point. Such “continuous” cyclic orders can be deﬁned almost in the same way as cyclic orders in ﬁnite sets were deﬁned above. The diﬀerence is that sometimes one cannot deﬁne cyclic permutations of the set in necessary quantity, and they have to be replaced by cyclic tranformations of the linear orders. Namely, a cyclic order is deﬁned as a linear order considered up to cyclic transformations, where by cyclic transformation of a linear order in a set X we mean a passage from to a linear order such that X splits into subsets A and B such that the restrictions of to each of them coincide, while a b and b a for any a ∈ A and b ∈ B. 5:A. Existense of a cyclic transformation turning linear orders to each other is an equivalence relation on the set of all linear orders in a set. A cyclic order in a set is an equivalence class of linear orders under the relation of existence of a cyclic transformation. 5:B. Prove that for a ﬁnite set this deﬁnition is equivalent to the deﬁnition of the preceding Section. 5:C. Prove that the cyclic “counterclockwise” order on a circle can be deﬁned along the deﬁniton of this Section, but cannot be deﬁned as a linear order modulo cyclic transformations of the set for whatever deﬁnition of cyclic transformations of circle. Describe the linear orders on circle which deﬁne this cyclic order up to cyclic transfomrations of orders. 5:D. Let A be a subset of a set X. If linear orders and on X can be obtained from each other by a cyclic transformation, than their restrictions to A can also be obtained from each other by a cyclic transformation. 5:E Corollary. A cyclic order in a set induces a welldeﬁned cyclic order in every subset of this set. 5:F. A cyclic order in a set can be recovered from cyclic orders induced by it on all its subsets consisting of three elements.
5:F.1. A cyclic order in a set X can be recovered from cyclic orders induced by it all its subsets consisting of three elements and containing a ﬁxed element a ∈ X.
Theorem 5:F allows one to describe acyclic order as a ternary relation. Namely, for a set {a, b, c} denote by [a b c] a cyclic order deﬁned by the linear order in which the inequalities presented in the brackets hold true (i.e., b follows by a and c follows by b). 5:G. Cyclic orders induced in three element subsets of X induced by the same cyclic order on X, satisfy the following requirements: (a) [a a b] is not true for any a, b ∈ X; (b) for any pairwise distinct a, b, c ∈ X either [a b c], or [b a c] is true, but both cannot be true simultaneously;
Proofs and Comments 5. then [a b d]. The second assertion follows from the deﬁnition of cone and reﬂexivity + of order.I. Indeed. Indeed. (d) if [a b c] and [a c d]. y) = x − y on S 1 ⊂ C. or a = b and b c. and. and this together with antisymmetricity implies that a = b. in the third case a b = c.F We need to check that the relation “a b.I. A cyclic order topology deﬁned by cyclic counterclockwise order of S 1 coincides with the topology deﬁned by the metric ρ(x. 5:I. or a b and b = c. Other conditions are checked similarly. ﬁnally.I The ﬁrst assertion follows from transitivity of the order. By the deﬁnition of cone. hence a c and a c. + Thereby we proved that each element of CX (b) a + the condition b ∈ CX (a) means that a + + + belongs to CX (a). Doing this. or a = b and b = c. Proving Theorem 5. a ternary relation having these four properties in a set X deﬁnes a cyclic order in X. Let a b and b c. b. In the ﬁrst case a c by transitivity of . hence a c and a c. we have shown that these properties follows from . hence a c. by the deﬁnition. Vice versa. iﬀ [b c a]. Hence CX (b) ⊂ CX (a). in the second case a = b c. ORDERED SETS 42 (c) [a b c]. Let X be a set with a cyclic order. while b. hence a = c and a c. + consider arbitrary c ∈ CX (b). CX (a) consists of b such that a b. by reﬂexivity of order.§ 5. we can use only the fact that satisﬁes the conditions from the deﬁnition of a strict order. a a. at the fourth case a = b = c.J By the preceding theorem 5. Let us check transitivity. 5. that is c ∈ + CX (a). which is what we had to prove. The third assertion follows similarly from antisymmetricity: the as+ + sumption CX (a) = CX (b) together with the second assertion implies a b and b a. § 5 ◦ 12 Topology of Cyclic Order 5:H. The topology deﬁned in 5:H is called a cyclic order topology. This meens that either a b and b c. The collection of sets which belong to the interval topology of every linear order deﬁning this cyclic order is a topological structure in X. iﬀ [c a b] a. c ∈ X. or a = b” satisﬁes the three conditions from the deﬁnition of a nonstrict order. By transitivity this implies c. 5. b c. in a poset cones have the properties which form the hypothesis of the theorem under consideration.
Consider the intersection of {x ∈ X  a x} and {x ∈ X  b x}. every set of this collection is a neighborhood.C. therefore either a b. Let c ∈ CX (a)∩ + + CX (b) and d ∈ CX (c). this is the smallest neighborhood. for it is not antisymmetric. For each of its points. + CX (c) ⊂ ∩ + CX (b). 5. Then a c d and b c d.2 The hypothesis of Theorem 5. It contains the point. Then {x ∈ X  a x} ∩ {x ∈ X  b x} = {x ∈ X  b x}.J turn into the following restrictions on C: it should be closed with respect to addition. Where can one get such a neighborhood from? How to construct it? Take all the neighborhoods of a point and consider their intersection. CX (a) and CX (b). The intersection is the union of these neighborhoods. being the intersection of all neighborhoods. . Consider the intersection of an arbitrary collection of open sets.T Equivalence of the second and third properties is proved using de Morgan formulas. they are equivalent to these conditions. The order is total. ORDERED SETS 43 the corresponding conditions from the deﬁnition of partial nonstrict order. Permute words in the proof of Theorem 5.O By Theorem 3. the smallest neighborhood of the point is contained in the intersection. 5. Hints. as 2. and Answers 5. or b a.V The minimal base of this topology consists of singletons of the form {2k−1} with k ∈ Z and threepoint sets of the form {2k−1. Let us prove that the ﬁrst property implies the second one. Thus. Now let us prove that if the intersection of any collection of open sets is open then any point has a smallest neighborhood. By the assumption. is contained in each of the neighborhoods. 5. to get a proof of Theorem 5. Comments. Hence. −11 and 1−1. it suﬃces to prove that any element of the + + intersection of two conus.§ 5. Solutions.R By Theorem 3. Indeed. where again k ∈ Z. Therefore d ∈ + CX (a) ∩ + CX (b). This neighborhood. it suﬃces to prove that the intersection of any two right rays is a union of right rays.I.F. Advises. Hence. is contained in the + intersection together with a whole cone of the same kind. this intersection is open. say. 5. Therefore its smallest neighborhood is contained in every of the sets which are to be intersected. 2k+ 1}.B. Thus. 2k. Let a b.1 No. Therefore this is a neighborhood of the point. Therefore it is open. In fact.J. each point of the intersection is contained in the intersection together with its neighborhood. 5. contain the zero and no nonidentity translation can map it bijectively onto itself. hence a + CX (a) d and b d. but −1 = 1.
Similarly. each of them would be maximal.9 The relation of inclusion in the set of all subsets of X is a linear order. ∞). 5. For any subset of a poset these notions coincide.13 A point open in a poset topology is a point maximal in the whole poset. consider a twoelement subset. one of them is greater than the other). But the greatest element is unique. Vice versa. If the two elements were incomparable. the smallest. but the opposite statements are not true. a point closed in a poset topology is a point miniimal in the whole poset. Show also that this condition holds true. Show that this condition holds true in any totally ordered set and it implies that right rays form a base of a topology. . say. 5. b and c such that a c and b c. the greatest element is maximal and the smallest one is minimal. iﬀ any two elements of the poset are comparable (i. 5. if right rays form a base of a topology. iﬀ X is either empty. the following condition: for arbitrary a. ORDERED SETS 44 5.§ 5. the empty set and the whole line.14 Rays of the forms (a. 5. b d and d c. and any minimal.11 Consider. and hence the greates. there exists an element d such that a d. This contradiction proves that the elements are comparable. Indeed. e. or singletone. ∞) and [a.6 Obviously. comparability of any two elements implies obiously that in any subset any maximal element is the greatest one.
which are closed in the subspace. 45 .§ 6 Subspaces § 6 ◦ 1 Topology for a subset of a space Let (X. Ω).F.. (a) (b) (c) (d) Describe the topological structures induced on the set N of natural numbers by the topology of the real line. Closed sets of the closed subspace are closed in the ambient space. iﬀ it is open in a neighborhood of each of its points. A set is closed in a subspace.2. Denote by ΩA the collection of sets A ∩ V . Sets. The same relation holds true for closed sets. which are open in the subspace. and A ⊂ X. if A ∈ Ω then ΩA ⊂ Ω.C. i.4 as follows: a set is open. However: 6. The canonical topology in R1 and the topology induced on R1 as a subspace of R2 coincide. It allows one to say that the property of being open is local.B. 2} by the topology of RT1 . and its elements are called open sets in A. e. § 6 ◦ 2 Relativity of Openness Sets. on N by the topology of the arrow. An open set of an open subspace is open in the ambient space. Ω) be a topological space. How to construct a base for the topology induced on A using the base for the topology in X? 6. 6. The unique set open in R1 . 2] considered as a subspace of the real line? 6. 6.E. are not necessarily closed in the ambient space. The collection ΩA is called the subspace topology or the relative topology or the topology induced on A by Ω.1. 1) open in the segment [0. ΩA is a topological structure in A. However: 6. are not necessarily open in the ambient space.4. 6. 6. iﬀ it is the intersection of the subspace and a closed subset of the ambient space. where V ∈ Ω. Is the halfopen interval [0. we can reformulate 6. on the twopoint set {1. Enigma. Prove that a set U is open in X. 6. The pair (A. ΩA ) is called a subspace of the space (X.A. which is also open in R2 . Indeed. iﬀ every its point has a neighborhood V in X such that U ∩ V is open in V .D.3. 6. is ∅. on the same set by the topology of the arrow.
Similarly one can check the second axiom.D No disk of R2 ﬁts into R. For each U ∈ Γ ﬁnd UX ∈ Ω such that U = A ∩ UX .7. i. Enigma. it is open in the line. Then the topology in A generated by metric ρ A×A coincides with the topology induced on A by the topology in X generated by metric ρ. 6. The third axiom: A = A ∩ X. being a topology in X. Proofs and Comments 6. Therefore the intersection of any open set of the plane with a line is a union of open intervals. (Here we use the fact that Ω. and X ⊃ A ⊃ B.C If a set F is closed in A then its complement A F is open in A.) Therefore A ∩ (∪U ∈Γ UX ) belongs to ΩA . Ω) one writes just X. Let (X. This is possible due to the deﬁnition of ΩA . Similarly one can prove that the intersection with A of the set closed in X is closed in A. That is why a topological space is usually denoted by the same symbol as the set of its points. A∩(X U) = A (A∩U) = A (A F ) = F . is open in X) as the union of sets open in X. ρ) be a metric space. where U is open in X. Any open set in the plane is a union of open disks. i. and ∅ = A ∩ ∅. Which of them is less obvious? § 6 ◦ 3 Agreement on Notations of Topological Spaces Diﬀerent topological structures in the same set are not considered simultaneously very often. e.§ 6.e.e. We have to prove that ∪U ∈Γ U ∈ ΩA . A F = A ∩ U. 6.5. ρ) one writes just X. 6. Ω) be a topological space. 6. The union ∪U ∈Γ UX belongs to Ω (i.6. instead of (X..A We need to check that ΩA satisﬁes the axioms of topological structure. i. 6. Show that the property of being closed is not local.. Indeed. Let Γ ⊂ ΩA be a collection of sets belonging to ΩA . Thus. . Then (ΩA )B = ΩB . Transform the union under consideration: ∪U ∈Γ U = ∪U ∈Γ (A ∩ UX ) = A ∩ (∪U ∈Γ UX ). The same is applied for metric spaces: instead of (X. Let (X. the topology induced on B by the topology induced on A coincides with the topology induced on B directly. Consider the ﬁrst axiom. What closed set cuts F on A? It is cut by X U. e. The statement 6.B The intersection of an open disk with a line is either an open interval or empty. satisﬁes the ﬁrst axiom of topological structure. and A ⊂ X. SUBSPACES 46 6.G Transitivity of Induced Topology.6 is equivalent to a couple of inclusions. 6.
if U is open. Here Br (x) is an open ball in X. being an element of ΩA . 0) on plane R2 .E If A ∈ Ω and B ∈ ΩA then B = A ∩ U. and is applied to U ∈ Ω.. 6.G The core of the proof is equality (U ∩ A) ∩ B = U ∩ B. 6. 6. 6.4 For V one can take U itself. is deﬁned by base Σ1 = A A {Br (a)  a ∈ A}. A and U. 1) ⊂ R and open disk with radius 1 and center at (0. Therefore Σ1 ⊂ Σ2 . Advises. Comments. for (n − 1. Therefore B ∈ Ω.2 Discrete. As U runs over Ω.§ 6. the right hand side of the equality (U ∩ A) ∩ B = U ∩ B runs over ΩB . k + 1. Ω = {∅. {2}.5 Consider interval (−1. where U ∈ Ω. However. k + 2. To prove the opposite implication.7 The topology induced by the metric in A. Hints. SUBSPACES 47 6.E.6 It is left to prove that elements of Σ2 are open in the topology deﬁned by Σ1 . ﬁnd V ∈ Σ1 such that x ∈ V ⊂ U. Indeed. (a) (b) (c) (d) 6. discrete. it may happen that Σ1 = Σ2 . but V . n + 1) ∩ N = {n}.. The other topology is deﬁned by base Σ2 = {A ∩ Br (x)  x ∈ X}. . 1) ∩ [0. since [0. 1) = (−1. but use 6. in turn. use problem 6. {1. For a point x of an element U of Σ2 .C instead of deﬁnition of induced topology. which belong to Ω.E. where Br (a) = {x ∈ A  ρ(x. 2}}. 6. elements of ΩB are intersections U ∩ B with U ∈ Ω.F Act as in the solution of the preceding problem 6. while the left hand side runs over (ΩA )B . is intersection U ∩ A with U ∈ Ω. and Answers 6.1 In the same way as the induced topology: if Σ is a base in X then ΣA = {A ∩ V  V ∈ Σ} is a base of the induced topology in A. Obviously A Br (a) = A ∩ Br (a) for a ∈ A. 1) is open on the line. 6.3 Yes. a) < r} is an open ball in A with center a and radius r. for this is the intersection of two sets.)}k∈N . for B ⊂ A. ΩN = {(k. Solutions. and (−1. 6. it is open. 2]. and elements of (ΩA )B are intersections V ∩ B with V ∈ ΩA . It takes place.
48 .D. Find the interior of {a. an open set.. 1) on the line with the Zariski topology. 7. It is denoted Cl A or. (b) Int Q = ∅ and (c) Int(R Q) = ∅.1. It is denoted Int A or. 7. 7. Find the interior of the interval (0. 1). going into details. iﬀ it coincides with its interior.B. 1) = (0.e. • a boundary point of the set A. A set is open. b. A ⊂ X. Every subset of topological space has closure. ClX A. Exterior and Boundary Points Let X be a topological space.A.E. It is the intersection of all closed sets containing this set. and b ∈ X. § 7 ◦ 1 Interior. IntX A. if it has a neighborhood contained in A.2.§ 7 Position of a Point with Respect to a Set This section is devoted to a further expansion of the vocabulary needed when one speaks of phenomena in a topological space. which contains any other open subset of A. 7. The point b is called • an interior point of the set A. Prove that in R: (a) Int[0. The exterior of a set is the maximal open set disjoint from A. 7. 7. if any its neighborhood intersects both A and the complement of A. if it has a neighborhood disjoint with A. It is the union of all open sets contained in this set. going into details. i. § 7 ◦ 2 Interior and Exterior The interior of a set A in a topological space X is the greatest (with respect to inclusion) open in X set contained in A. Every subset of a topological space has interior. It is obvious that the exterior of A is Int(X A). 7. The interior of a set is the union of its interior points. § 7 ◦ 3 Closure The closure of a set A is the smallest closed set containing A. • an exterior point of the set A.C. d} in space .
F. The closure of a set is the complement of its exterior. It is denoted by Fr A 7. 7. 1) = [0. Recall (see Section § 4) that the distance ρ(b. A set A is closed. A) from the point b to the set A is the inf{ ρ(b. The frontier of a set is the set of its boundary points. FrX A. A) = 0. 7. iﬀ A = Cl A. 7. then ClA B = (ClX B) ∩ A.J. 7.K.§ 7. iﬀ ρ(b. Is it true that IntA B = (IntX B) ∩ A? A point b is called an adherent point for a set A if all of its neighborhood intersect A.6. Let Ω1 . and Ω1 ⊂ Ω2 .H.9. 1]. 7. § 7 ◦ 6 Closure and Interior with Respect to a Finer Topology 7. In ﬁnd the frontier of {a}. The closure of a set is the set of its adherent points. § 7 ◦ 5 Frontier The frontier of a set A is the set Cl A or. Formulate and prove an analogous statement about interior. 7.10. Let Cli denote the closure with respect to Ωi . In formulas: Cl A = X Int(X A). Ω2 be topological structure in X. Describe the closure of a point in a poset topology. a)  a ∈ A }. Int A. 7. POSITION OF A POINT WITH RESPECT TO A SET 49 7.5.G. where X is the space and A ⊂ X. (c) Cl(R Q) = R. ρ). Prove that if A is a subspace of X. The frontier of a set A equals the intersection of the closure of A and the closure of the complement of A: Fr A = Cl A ∩ Cl(X A).L.8. Prove that a set A is closed. Find the closure of {a} in . Prove that Fr A = Fr(X A). (b) Cl Q = R. . 7. Prove that Cl1 A ⊃ Cl2 A for any A ⊂ X.4.7. more precisely. 7.I. § 7 ◦ 4 Closure in a Metric Space Let A be a subset. Find a formula for Fr A. and B ⊂ A. which is symmetric with respect to A and X A. Prove that in R: (a) Cl[0. iﬀ Fr A ⊂ A.3. 7. and b be a point of a metric space (X. 7. 7. Prove that b ∈ Cl A.
7.3. and Fr[0. keep on going in the same way. Int(A ∪ B) = Int A ∪ Int B? 7.22.17.19. 7.21. 7. Does a sphere contain the frontier of the closed ball with the same center and radius? 7.11–7.12. Is there a set A ⊂ R such that (a) A. Prove that if A ⊂ B then Int A ⊂ Int B. Does this inclusion hold true for any A and B? 7.23 The Kuratowski Problem.20. 7. § 7 ◦ 8 Compositions of Closure and Interior 7. Cl A. Does a sphere contain the frontier of the open ball with the same center and radius? 7.23.15. Int Cl A are pairwise distinct. (c) A.1. Cl A. In the example that you have found solving the previous problem an inclusion of one hand side into another one holds true. Is it true that for any sets A and B the following equalities hold true: (8) (9) Int(A ∩ B) = Int A ∩ Int B. Cl A.2.18. Cl A. 1]. Prove that Int Int A = Int A. How to ﬁnd the closure and interior of a set in this space? 7. Find Cl N.13. Int A.16. and Int A would be pairwise distinct. Find a set A ⊂ R such that the sets A. try to formulate a theorem explaining the failure.§ 7. 7. Give an example in which one of that equalities does not hold true. and when fail.23. Int Cl A are pairwise distinct? If you ﬁnd such sets.15. Study the operator Cl in a way suggested by the investigation of Int undertaken in 7.23. Cl { n  n ∈ N } . 1] in RT1 . 7. Int(0. Prove that Cl Int Cl Int A = Cl Int A. Cl Int A. 1).14. Find Int (0. 7.23. 7. 1 7. +∞) in the arrow. Find Cl{1}. Cl Int A are pairwise distinct. (b) A.11. POSITION OF A POINT WITH RESPECT TO A SET 50 § 7 ◦ 7 Properties of Interior and Closure 7. and Fr(2. How many pairwise distinct sets can one obtain out of a single set using operators Cl and Int? The following problems will help you to solve problem 7. . Int A. Find an example in which a sphere is disjoint from the closure of the open ball with the same center and radius. Int A. Int[0. 1] ∪ {2} . and Fr Q in R.
27. Is it possible to increase the number of such sets? § 7 ◦ 10 Convexity and Int.28*. for any x. y] belongs to A). Formulate a necessary and suﬃcient condition on the topology of a space which has an everywhere dense point.e. 7. A set is everywhere dense. The set Q is everywhere dense in R.26. 7. iﬀ it intersects any nonempty open set.N. Is it true that the union of everywhere dense sets is everywhere dense. Cl. Fr Recall that a set A ⊂ Rn is said to be convex if together with any two points it contains the whole interval connecting them (i. unless A is not contained in an (n − 1)dimensional aﬃne subspace of Rn . 7. Prove that a topological space is a discrete space.M. (c) Cl∗ (A ∪ B) = Cl∗ A ∪ Cl∗ B. POSITION OF A POINT WITH RESPECT TO A SET 51 § 7 ◦ 9 Sets with Common Frontier 7. 7. which have the same frontier. (d) Cl∗ Cl∗ A = Cl∗ A. A is said to be dense in B if Cl A ⊃ B. Find spaces satisfying the condition in Section § 2.29. Find an analogous system of axioms for Int. Which singletons are dense in a poset topology? 7.. 7. 7. When is Fr A convex? § 7 ◦ 11 Characterization of Topology by Closure or Interior Operations 7. iﬀ it has a unique everywhere dense set (by the way. and that the intersection of everywhere dense sets is everywhere dense? . which one). Let in the set of all subset of a set X exist an operator Cl∗ which has the following properties: (a) Cl∗ ∅ = ∅. in the arrow and in RT1 .24*. Prove that A contains a ball. 7. 7. Ω). Prove that Ω = { U ⊂ X  Cl∗ (X U ) = X U } is a topological structure. Find three sets in the real line.31. § 7 ◦ 12 Dense Sets Let A and B be sets in a topological space X. Give a characterization of everywhere dense sets in an indiscrete space. y ∈ A any point z belonging to the segment [x.34. Let A be a convex set in Rn . Prove that Cl A and Int A are convex.32.§ 7.33. (b) Cl∗ A ⊃ A.30. and Cl∗ A is the closure of a set A in the space (X. and everywhere dense if Cl A = X.25. 7. 7.
38*. A point b is called an isolated point of a set A if b ∈ A and there exists a neighborhood of b disjoint with A {b}.47.40.49. 7.41. Prove that for every set A there exists a maximal open set B in which A is dense. ..37*.O. Prove that the intersection of two open everywhere dense sets is everywhere dense. 7. Prove that in Rn (n ≥ 1) every proper algebraic set (i. 7. 7. Is this true for the boundary of an open set.46. Prove that a ﬁnite union of nowhere dense sets is nowhere dense. The extreme cases B = X and B = ∅ mean that A is either everywhere dense or nowhere dense respectively. Prove that in R a countable intersection of open everywhere dense sets is everywhere dense.43. a set deﬁned by algebraic equations) is nowhere dense.e. Is it possible to replace R here by an arbitrary topological space? 7. Can a set be everywhere dense and nowhere dense simultaneously? 7. 7.35. boundary of an arbitrary set? 7.45. 7. § 7 ◦ 14 Limit Points and Isolated Points A point b is called a limit point of a set A if any neighborhood of b intersects A {b}. Which condition in the previous problem is redundant? 7.36. iﬀ it contains all its limit points. iﬀ any neighborhood of any point x ∈ X contains a point y such that the complement of A contains y together with one of its neighborhoods. Prove that Q cannot be presented as a countable intersection of open sets dense in R. Every limit point of a set is its adherent point. 1 7. Find limit and isolated points of the set N in RT1 .48. Is R nowhere dense in R2 ? 7. 1] ∪ {2}. Prove that the frontier of a closed set is nowhere dense. 7. POSITION OF A POINT WITH RESPECT TO A SET 52 7. { n  n ∈ N } in Q and in R.42. § 7 ◦ 13 Nowhere Dense Sets A set is called nowhere dense if its exterior is everywhere dense. A set A is closed. Find limit and isolated points of the sets (0.39.Q.44. Enigma.P.§ 7. Prove that if A is nowhere dense then Int Cl A = ∅. A set A is nowhere dense in X. 7. 7. Give an example proving that an adherent point may be not a limit one. 7. What can you say about the interior of a nowhere dense set? 7.
Then Ux ⊂ Int A (for Int A is the greatest among all open sets contained in A). Therefore. (c) A is the intersection of open and closed subsets of X.1 Int{a. is open. .. 7. Comments. e. and hence coincides with its interior. while (0. POSITION OF A POINT WITH RESPECT TO A SET 53 § 7 ◦ 15 Locally Closed Sets A subset A of a topological space X is called locally closed if each of its points has a neighborhood U such that A ∩ U is closed in U (cf. b. a set. b}. Int Q = ∅. Solutions. and. In general. properties of closure can be obtained from properties of interior by replacing unions with intersections and vice versa. is open (as a union of open sets).B Let x be an interior point of A (i. e. Vice versa. is closed (as an intersection of closed sets). the set Int A itself is a neighborhood of x contained in A. Prove that the following conditions are equivalent: (a) A is locally closed in X. Proofs and Comments 7.§ 7. 1). 1) = (0.4–6. Therefore. secondly. (b) Since any interval contains irrational points.. if it is open. it is the greatest one among those sets). 7. 7. Advises. and hence x ∈ Int A. since this is really the greatest among all sets open in and contained in {a.5).C A set is the greatest one among all of its subsets. (b) A is an open subset of its closure ClX A. and Answers 7.D (a) [0. it is the greatest among all of its open subsets. the proof of Theorem 7.. therefore.A. which coincides with its interior. e. it is the smallest one among those sets). Cf. d}. 7. Vice versa. therefore Int[0. contains every open set which is contained in A (i. secondly. 1) is not open on the line. there exists an open set Ux such that x ∈ Ux ⊂ A). Q does not contain a nonempty set open in the classical topology of the line. R Q does not contain a nonempty set open in the classical topology of the line. Int(R Q) = ∅. is contained in every closed set which contains A (i. b.A The union of all open sets contained in A. 7.50.E The intersection of all closed sets containing A. ﬁrstly. 1) is. d} = {a. if x ∈ Int A. and. since the interior is open. 6. (c) Since any interval contains rational points. Hints. ﬁrstly.
. because no nonempty open set of this space is contained in (0. X H∈Ω (H ∩ A) = A ∩ H⊃B. let X = R2 .§ 7. 1). B = A. ClA B = F ⊃B.3 Indeed. The second equality may be obviously wrong. 1) on the line with the Zariski topology is empty. A F ∈ΩA F = H⊃B. Really. POSITION OF A POINT WITH RESPECT TO A SET 54 7. Then IntA B = R1 = ∅ = (IntX B) ∩ A. X H∈Ω H = A ∩ ClX B. 7.2 The interior of the interval (0. A = R1 .
the image of the preimage of a set B can diﬀer from B. 55 7 f (A) = {f (x) : x ∈ A}. 8. iﬀ B is contained in the image of f .. There are other words with the same meaning: map. Maps § 8 ◦ 1 Maps and the Main Classes of Maps A mapping f of a set X to a set Y is a triple consisting of X. It may happen that the preimage is not the only set with this property.B. Be careful with these terms: their etymology can be misleading. § 8 ◦ 2 Image and Preimage The image of a set A ⊂ X under a map f : X → Y is the set of images of all points of A. Then the following statements are equivalent: Of course. y ∈ Y such that the rule assigns y to x. f f −1 (B) ⊂ B for any map f : X → Y and B ⊂ Y . The preimage of a set B ⊂ Y under a map f : X → Y is the set of elements of X whose images belong to B. 8. One writes b = f (a). or invertible if it is surjective and injective. function. the preimage cannot be deﬁned as a set whose image is a given set.§ 8 SetTheoretic Digression. Hence.C.A. one considers a set of ordered pairs (x. Namely. or just a surjection if every element of Y is an image of at least one element of X. injection. A mapping is called a bijective map.7 which assigns to every element of X exactly one element of Y . f (X)) is called the image of f . or a → b. or f : a → b. And even if it does not diﬀer. It is a subset of the set X × Y of all ordered pairs (x. y) with x ∈ X. or X → Y . y). For example. 8. f f −1 (B) = B. bijection. Thus f −1 (B) = {a ∈ X : f (a) ∈ B}. and a rule. Y .e. or onetoone map if every element of Y is an image of not more than one element of X. The element b of Y assigned by f to an element a of X is denoted by f (a) and called the image of a under f . A mapping f : X → Y is called an injective map. Thus The image of the entire set X (i. the rule (as everything in the set theory) may be thought of as a set. This set is called the graph of f . It is denoted by f −1 (B). Let f : X → Y and B ⊂ Y such that f f −1 (B) = B. It is denoted by f (A). f f A mapping f : X → Y is called a surjective map. If f is a mapping of X to Y then one writes f : X → Y . .
or just id. f (A ∩ B) = f (A) ∩ f (B). A map f : X → Y is an injection. Do the following equalities hold true for any A. or just inclusion.§ 8. Give examples in which two of the equalities above are false. B ⊂ X and any f : X → Y: (13) (14) (15) f (A ∪ B) = f (A) ∪ f (B). iﬀ f (A) ∩ f (X A) = ∅.2. B ⊂ X ? 8. § 8 ◦ 4 Composition The composition of mappings f : X → Y and g : Y → Z is the mapping g ◦ f : X → Z deﬁned by formula x → g f (x) . 8. Replace the false equalities of 8. of A into X and denoted by in : A → X. f −1 (Y A) = X f −1 (A)? 8. MAPS 56 (a) f −1 (B) is the unique subset of X whose image equals B. B ⊂ Y : B ∩ f (A) = f f −1 (B) ∩ A .F. 8. (b) for any a1 . and subsets A ⊂ X.3. 8. f −1 f (A) = A.G. 8. It is denoted by idX .2 by correct inclusions. iﬀ for any B ⊂ Y such that f f −1 (B) = B the preimage f −1 (B) is the unique subset of X whose image equals B. f (X A) = Y f (A)? 8. Do the following equalities hold true for any A.6. or just in.E. 8. Prove that for any map f : X → Y .1. f −1 (A ∩ B) = f −1 (A) ∩ f −1 (B). 8. when there is no ambiguity. f −1 f (A) ⊃ A for any map f : X → Y and A ⊂ X. B ⊂ Y and any f : X → Y: (10) (11) (12) f −1 (A ∪ B) = f −1 (A) ∪ f −1 (B). What simple condition on f : X → Y should be imposed in order to make correct all the equalities of 8. a2 ∈ f −1 (B) the equality f (a1 ) = f (a2 ) implies a1 = a2 .4. SETTHEORETIC DIGRESSION.2 for any A. The preimage of a set B under an inclusion in : A → X is B ∩ A. § 8 ◦ 3 Identity and Inclusion The identity map of a set X is the map X → X deﬁned by formula x → x. when A and X are clear.5. 8. .D. If A is a subset of X then the map A → X deﬁned by formula x → x is called an inclusion map.
If an inverse map exists then it is unique. 8. . Let a composition g ◦ f be bijective. 8.§ 8. 8. Any abbreviation (including any restriction) of injections is injective. or submap. 8.I. B.Q. f ◦ (idX ) = f = (idX ) ◦ f for any f : X → Y .L. 8. 8. Is then f or g necessarily bijective? § 8 ◦ 5 Inverse and Invertible A map g : Y → X is said to be inverse to a map f : X → Y if g ◦ f = idX and f ◦ g = idY .7. or submapping. for which an inverse map exists. A map. 8. If a restriction of a mapping is surjective then the original mapping is surjective.N.B or even simply f . MAPS 57 8.H. If B = Y then ab f : A → B is denoted by f A.K. If the composition g ◦ f is surjective then g is surjective. SETTHEORETIC DIGRESSION.R.S. The composition of bijections is a bijection.O. If B = Y then ab f : A → Y is denoted by f A and called the restriction of f to A. 8. The composition of injections is injective. g : Y → Z. 8. h ◦ (g ◦ f ) = (h ◦ g) ◦ f for any maps f : X → Y .J. The composition of surjections is surjective.M. 8.P. If the composition g ◦ f is injective then f is injective. The restriction of a map f : X → Y to A ⊂ X is the composition of inclusion in A :→ X and f . f A = f ◦ in. iﬀ it is a bijection. 8. is said to be invertible. § 8 ◦ 6 Submappings If A ⊂ X and B ⊂ Y then for every f : X → Y such that f (A) ⊂ B there is mapping ab(f ) : A → B deﬁned by formula x → f (x) and called an abbreviation of the mapping f to A. A mapping is invertible. and h : Z → U. In other words. 8.
. a map with image consisting of a single point) is continuous. A map f : X → Y is said to be continuous if the preimage of any open subset of Y is an open subset of X.E. 9. 58 .D.H. Ω1 ) → (X. Y be topological spaces. Which maps X → Y and Y → X are continuous? 9. 9. a ﬁner topology in Y and the same topology in X. Let X be an indiscrete space and Y an arbitrary space.§ 9 Continuous Maps § 9 ◦ 1 Deﬁnition and Main Properties of Continuous Maps Let X.C. The inclusion in : A → X is continuous. The topology ΩA induced on A ⊂ X by the topology of X is the coarsest topology in A such that the inclusion mapping in : A → X is continuous with respect to it. 9.e. Ω2 ) is continuous. Which maps X → Y and Y → X are continuous? 9. Is it continuous with respect to a ﬁner topology in X and the same topology in Y .D admits a natural generalization with the inclusion map replaced by an arbitrary map f : A → X of an arbitrary set A. Find this generalization. Ω2 be topological structures in X. iﬀ the preimage of any closed set is closed. 9. 9. (a) (b) (c) (d) Let f : X → Y be a continuous map. A composition of continuous maps is continuous. Any constant map (i. Let Ω1 . Let X be a discrete space and Y an arbitrary space.3.F. A map f : X → Y is continuous. a coarser topology in Y and the same topology in X? 9.G. a coarser topology in X and the same topology in Y .2. 9. A map is continuous. The identity map of any topological space is continuous. 9. 9. Let A be a subspace of X. The statement 9. A submap of a continuous map is continuous.B. Prove that the identity mapping of X id : (X.1. 9. iﬀ Ω2 ⊂ Ω1 .A. iﬀ ab f : X → f (X) is continuous.5.4. Enigma. 9.
14.16. if x ∈ [1. 1). iﬀ f −1 (U ) is open for any U ∈ Σ.6.10. 1] such that f (A) = [0. Give an explicit characterization of continuous mappings of the arrow to itself. 2] 0. 3 − x. Which maps RT1 → RT1 are continuous? 9. 1] → [0. Induce topology on Z + and f (Z + ) from R. Are f and the map g −1 .17*. CONTINUOUS MAPS 59 § 9 ◦ 2 Reformulations of Deﬁnition 9. if x ∈ [0. Prove that a map f : X → Y is continuous.§ 9. Let Σ be a base for topology in Y . 9. Is it true that the image of nowhere dense set under a continuous map is nowhere dense. 9. 1]. x. continuous? § 9 ◦ 4 Behavior of Dense Sets 9. Let g : Z + → f (Z + ) be its submap. 2] deﬁned by formula f (x) = x. 1]? . 2] → [0. Is the map f of segment [0. iﬀ for any A ⊂ Y . x + 1.13. Formulate and prove similar criteria of continuity in terms of Int f −1 (A) and f −1 (Int A). Prove that the image of an everywhere dense set under a surjective continuous map is everywhere dense. Prove that a mapping f : X → Y is continuous. 2] (with the topology induced by the topology of the real line) into the arrow (see Section § 2) deﬁned by formula f (x) = continuous? 9.15. inverse to g. Give an explicit characterization of continuous mappings of RT1 (see Section § 2) to R.11. Does there exist a nowhere dense set A of [0. Cl f −1 (A) ⊂ f −1 (Cl A) 9.7. § 9 ◦ 3 More Examples 9.9. 2] continuous (with respect to the topology induced from the real line)? 9.12. 9. Is the mapping f : [0. Let f be a mapping of the set Z + of nonnegative numbers onto R deﬁned by formula f (x) = 1 x. 9. if x ∈ (1. if x = 0. 9. Do the same for Cl f (A) and f (Cl A).8. if x ∈ [0. 1] (with the topology induced out of the real line) and a continuous map f : [0. if x = 0.
.I. . f (a) < ε. CONTINUOUS MAPS 60 § 9 ◦ 5 Local Continuity A map f of a topological space X to a topological space Y is said to be continuous at a point a ∈ X if for every neighborhood U of f (a) there exists a neighborhood V of a such that f (V ) ⊂ U. Y be metric spaces. iﬀ for every ε > 0 there exists δ > 0 such that for every point x ∈ X inequality ρ(x. 9. .§ 9.K means that continuity introduced above coincides with the one that is usually studied in Calculus. Y be metric spaces. Find a sequence of continuous functions fi : R → R. Let X be any topological space. 9.21. and a ∈ X. x →f (x) − g(x). g(x)} deﬁnes a function R → R which is not continuous. 9. A map f : X → Y is continuous at a. A mapping f : X → Y is continuous at the point a.J. Let f. iﬀ for every ball with center at f (a) there exists a ball with center at a whose image is contained in the ﬁrst ball. . 9. x → min{f (x). 9. iﬀ it is continuous at each point of X. and a ∈ X. g : X → R be continuous. Prove that if 0 ∈ g(X) then a mapping X → R deﬁned by formula / x→ is continuous. Prove that the mappings X → R deﬁned by formulas (16) (17) (18) (19) (20) (21) are continuous. Prove that a function f : X → Rn : x → (f1 (x). x → max{f (x). § 9 ◦ 6 Properties of Continuous Functions 9. Let X.18. Theorem 9. A map f : X → Y is continuous. iﬀ all the functions fi : X → R with i = 1. x → f (x) . a) < δ implies ρ f (x). . . fn (x)) is continuous. 9. Let X. g(x)}. x →f (x)g(x).K. . (i ∈ N) such that the formula x → sup{ fi (x)  i ∈ N } f (x) g(x) x →f (x) + g(x).20. . n are continuous.19. .
9. Prove that the GromovHausdorﬀ distance is symmetric and satisﬁes the triangle inequality. For any metric spaces X and Y there exists a metric space Z such that X and Y can be isometrically embedded into Z. Every isometric embedding is injective. A bijection which is an isometric embedding is called an isometry.N. f (b) = ρ(a. R) is the subspace of M at(n × n. CONTINUOUS MAPS 61 Real p×qmatrices comprise a space M at(p×q. we can consider the Hausdorﬀ distance between their images (see. R) be a continuous map. 9. which diﬀers from Rpq only in the way of numeration of its natural coordinates (they are numerated by pairs of indices). R) : x → g(x)f (x) is a continuous map.L. Let f : X → GL(n. 9. A) (see Section § 4) is continuous. Let f : X → M at(p × q. Having embedded isometrically two metric space in a single one. R) be continuous maps. Every isometric embedding is continuous. The inﬁnum of such Hausdorﬀ distances over all pairs of isometric embeddings of metric spaces X and Y to metric spaces is called the GromovHausdorﬀ distance between X and Y . Enigma. b ∈ X. § 4 ◦ 16). For every subset A of a metric space X the function deﬁned by formula x → ρ(x. R) : x → (f (x))−1 is continuous. Does there exist metric spaces with inﬁnite GromovHausdorﬀ distance? 9:C. § 9 ◦ 9 GromovHausdorﬀ distance 9:A. R) and g : X → M at(q × r. 9:D. R).22.24.23. 9. § 9 ◦ 7 Continuity of Distances 9. A) is continuous. b) for every a. R) consisting of all the invertible matrices. In what sense the GromovHausdorﬀ distance can satisfy the ﬁrst axiom of metric? . with respect to which for every A ⊂ X the function X → R deﬁned by formula x → ρ(x. Prove that then X → M at(p × r. § 9 ◦ 8 Isometry A mapping f of a metric space X into a metric space Y is called an isometric embedding if ρ f (a).M. Prove that X → GL(n. Recall that GL(n. Prove that a topology of a metric space is the coarsest topology. 9. 9:B.§ 9.
b ∈ X with a b . . is f continuous with respect to right ray or left ray topologies? 9. • strictly monotonically decreasing or strictly antimonotone. § 9 ◦ 11 Monotone maps Let (X. § 9 ◦ 12 Functions on Cantor Set and SquareFilling Curves Recall that Cantor set K is the set of real numbers which can be presented as k sums of series of the form ∞ ak with ak = 0 or 2. f (b) ≤ Cρ(a. Let X.O. b ∈ X with a b . if f (b) f (a) for any a.28. • (nonstrictly) monotonically decreasing or antimotone. Prove that every contractive mapping is continuous.O is needed. b ∈ X.25. if f (a) f (b) for any a. 9. Show that the surjectivity condition in 9. A mapping f : X → Y is said to be H¨lder o if there exist C > 0 and α > 0 such that ρ f (a). 2k+1 Prove that γ1 : K → I is a continuous surjection. b) for every a. A map f : X → Y is said to be • (nonstrictly) monotonically increasing or just monotone. CONTINUOUS MAPS 62 § 9 ◦ 10 Contractive maps A mapping f : X → X of a metric space X is called contractive if there exists α ∈ (0. Prove that every H¨lder mapping is continuous. Y be metric spaces. A mapping of a poset to a poset is monotone. f (b) ≤ αρ(a. if f (a) f (b) for any a. o 9. b)α for every a. ) be posets. Let X and Y be linearly ordered sets.26. 9.27.O. Under conditions of Theorem 9. Draw the graph of ϕ. b ∈ X with a b . • strictly monotonically increasing or just strictly monotone. 9. ) and (Y. 9. b ∈ X. if f (b) f (a) for any a. Let γ1 be a map K → I deﬁned by ∞ k=1 ak → 3k ∞ k=1 ak . 1) such that ρ f (a). k=1 3 9:E. iﬀ it is continuous with respect to the poset topologies. With respect to the interval topology in X and Y any surjective monotone or antimonotone mapping X → Y is continuous. b ∈ X with a b.§ 9.P.
. Prove that any sequence of paths fk : I → I 2 satisfying the conditions of 9:J converges to a map f : I → I 2 (i. 8 Prove that any continuous map I → I 2 with dense image is surjective. § 9 ◦ 13 Sets Deﬁned by Systems of Equations and Inequalities 9. y) ∈ R2 : x ∈ K. CONTINUOUS MAPS 63 9:F. they are slightly diﬀerent. γ1 (y)) is a continuous surjection.T.e. √ (b) dist(fk (x). 15. this is a curve ﬁlling the square. Since then a lot of other similar examples have been found. 9:H. Peano in 1890. 9:G. Denote by K 2 the set {(x. 9:M. we have to mention that it would be more appropriate after Section § 15. Prove that the function K → K deﬁned by ∞ k=1 ak → 3k ∞ k=1 a2k 3k is continuous.) The latter map is a continuous surjection I → I 2 . Prove that the map γ3 : K → I 2 is a restriction of a continuous map. Here is a sketch of Hilbert’s construction. Prove that the map γ2 : K → K 2 deﬁned by ∞ ∞ ∞ k=1 ak → 3k k=1 a2k−1 . Problems 15. 9:J. . y ∈ K}. A curve with this property was ﬁrst constructed by G. Prove that there exists a sequence of polygonal maps fk : I → I 2 such that (a) fk connects all centers of the squares forming the obvious subdivision of I 2 into 4k equal squares with side 1/2k . 15. . You may ﬁnd a nice survey of them in a book by Hans Sagan. y) → (γ1 (x). . Though the construction sketched above is based on the same ideas as the original Peano’s construction. 2:A. fn (x) = 0 is closed. SpaceFilling Curves. 8 . Cf.K. .Q. n) be continuous mappings X → R. . 9:I. Thus.O. 9:L. . fk−1 (x)) ≤ 2/2k+1 for any x ∈ I (here dist means the metric induced on I 2 from the standard Euclidean metric of R2 ).§ 9. 9:K. Prove that the map γ3 : K → I 2 deﬁned as the composition of γ2 : K → K 2 and K 2 → I 2 : (x. Although this problem can be solved using theorems wellknown from Calculus.2. 3k k=1 a2k 3k is a continuous surjection. (Cf. Let fi (i = 1. for any x ∈ I there exists a limit f (x) = limk→∞ fk (x)) and this map is continuous and its image is dense in I 2 . Then the subset of X consisting of solutions of the system of equations f1 (x) = 0. Generalize 9:G – 9:I 9:J – 9:L to obtain a continuous surjection of I onto I n . SpringerVerlag 1994. .
X ⊂ A∈Γ A.R a ﬁnite system can be replaced by an inﬁnite one. .R. while the set consisting of solutions of the system of inequalities f1 (x) > 0. . Γ = {Q. § 9 ◦ 14 SetTheoretic Digression.30–9. fn (x) ≥ 0 is closed. Where in 9. 9. n) be continuous mappings X → R. 9. . Γ = {[0. To answer these questions.. Γ = {[0.T. . Γ is a set of all onepoint subsets of R? A cover Γ of a space X is said to be fundamental if a set U ⊂ X is open. i. . sets belonging to Γ are also said to cover X. ..X. Let fi (i = 1. 1].e.31 and 9. 2]. (d) X = R. iﬀ for every A ∈ Γ the set U ∩ A is open in A. . . 1]. There is also a more general meaning of these words. .e. 2]}.S–9. iﬀ a set U ⊂ X is open. . provided U ∩ A is open in A for every A ∈ Γ. solve the problems 9. fn (x) > 0 is open.§ 9. A covering Γ of a space X is fundamental.30. Each element of Γ inherits from X a topological structure. R Q}. 9. . When are these structures suﬃcient for recovering the topology of X? In particular. § 9 ◦ 15 Fundamental Covers Consider a cover Γ of a topological space X. i. iﬀ a set F ⊂ X is closed. provided F ∩ A is closed A for every A ∈ Γ. 2]}. (c) X = R.S. 2]. under what conditions on Γ does continuity of a map f : X → Y follow from continuity of its restrictions to elements of Γ. In this case elements of Γ are said to cover X. In this case. Then the subset of X consisting of solutions of the system of inequalities f1 (x) ≥ 0. A covering Γ of a space X is fundamental. CONTINUOUS MAPS 64 9. A collection Γ of subsets of a set Y is called a cover or a covering of a set X ⊂ Y if X is contained in the union of the sets belonging to Γ.29. Covers A collection Γ of subsets of a set X is called a cover or a covering of X if X is a union of sets of belonging to Γ. X = A∈Γ A. (b) X = [0. Is this true for the following coverings: (a) X = [0. [1. .Q and 9. (1. 9. .
while symmetricity and triangle inequality were formulated above only for functions with ﬁnite values. if two or three of the quantities involved are inﬁnite. Solutions. To deﬁne the distance bewteen points of diﬀerent copies. that this deﬁnes a metric. and A is a subspace of a space X. Prove that if a cover Γ is a reﬁnement of a cover Γ. and Answers 9:A Construct Z as the union of disjoint copies of X and Y . and Γ be a cover of X such that ΓA = { U ∩ A  U ∈ Γ } is a fundamental cover for subspace A ⊂ X for every A ∈ ∆. Advises. Every locally ﬁnite closed cover is fundamental. 9:B Yes. Prove that Γ is a fundamental cover.§ 9. 9.31. put the distance between points.32. A cover Γ is said to be a reﬁnement of a cover Γ if every element of Γ is contained in some element of Γ. then ΓA = { U ∩ A  U ∈ Γ } is a fundamental cover of A. Prove that if Γ is a fundamental cover. which belong to a copy of one of the spaces. Hints. b) + 1 for a ∈ X and b ∈ Y . to be equal to the distance between the corresponding points of the original space. Let Γ be a fundamental cover of a topological space X.. these two properties make sense if inﬁnite values are admitted.) The following construction helps to prove the triangle inequality.W. 9. b) = ρX (a. 9. and put ρ(a. 9:C Although.U. the GromovHausdorﬀ distance can be inﬁnite. (The triangle inequality should be considered satisﬁed. and closed if it consists of closed sets. choose points x0 ∈ X and y0 ∈ Y . Every ﬁnite closed cover is fundamental. x0 ) + ρY (y0 . A cover of a topological space is said to be open if it consists of open sets. really). if every point of a space X has a neighborhood V such that ΓV = { U ∩ V  U ∈ Γ } is fundamental. 9. Check (this is easy.33. 9. Let metric spaces X and Y are isometrically embedded into a metric space A. 9. Let ∆ be a fundamental cover of a topological space X. 9. as we have seen solving the previous problem.V. A cover of a topological space is said to be locally ﬁnite if every point of the space has a neighborhood intersecting only a ﬁnite number of elements of the cover.34. iﬀ the space is discrete. if only one of them is inﬁnite. CONTINUOUS MAPS 65 9. i. Every open cover is fundamental. then Γ is fundamental. In the union.X.35. and Γ is fundamental then Γ is also fundamental. A cover of a topological space consisting of singletons is fundamental.e. For example a singleton and any nonbounded space. and not satisﬁed. If the restriction of a mapping f : X → Y to each element of Γ is continuous then f is continuous. and metric spaces Y and . Prove that the property of being fundamental is local. Comments. 9.
§ 9. the most optimistic idea is that then there should exist an isometric bijection between the spaces. But this is not true. Prove that this gives a metric space and use the triangle inequality for the Hausdorﬀ distance between X. . Put distances between these points to be equal to the distances between them in B. Y and Z in this space. as one can see looking at spaces Q and R with standard distances in them. CONTINUOUS MAPS 66 Z are isometrically embedded into a metric space B. z)  y ∈ A ∩ B}. in what sense the spaces should be equal then? First. Certainly. Construct a new metric space in which A and B would be isometrically embedded meeting in Y . the answer is obvious. Compare this construction to the construction from the solution of Problem 9:A. 9:D Partially. For this. add to A all points of B A. We promise ahead of time that for compact metric spaces this is true. the GromovHausdorﬀ distance is nonnegative! But what if it is zero. Put the distance between x ∈ A B and z ∈ B A equal to inf{ρA (x. y) + ρB (y.
The inverse of a homeomorphism is a homeomorphism. iﬀ f (A) is closed in Y . Being homeomorphic is an equivalence relation.E? § 10 ◦ 3 Role of Homeomorphisms 10. Let f : X → Y be a homeomorphism. iﬀ f is a bijection and deﬁnes a bijection between the topological structures of X and Y . How is Theorem 10. A composition of homeomorphisms is a homeomorphism. if both this mapping and its inverse are continuous. This phenomenon was used as a basis for a deﬁnition of the subject of topology in the ﬁrst stages of its development. (d) f (Fr A) = Fr f (A). 1) → S 1 . 10. Therefore from the topological point of view homeomorphic spaces are completely identical: a homeomorphism X → Y establishes onetoone correspondence between all phenomena in X and Y which can be expressed in terms of topological structures. § 10 ◦ 2 Homeomorphic Spaces A topological space X is said to be homeomorphic to space Y .F. 10. iﬀ f (A) is a neighborhood of the point f (x). 10. Find a continuous bijection [0.§ 10 Homeomorphisms § 10 ◦ 1 Deﬁnition and Main Properties of Homeomorphisms An invertible mapping is called a homeomorphism. f : X → Y is a homeomorphism. which is not a homeomorphism.G.F related to 10.D.C. 10. (e) A is a neighborhood of a point x ∈ X.C–10. which is not a homeomorphism. 10. if there exists a homeomorphism X → Y . The identity map of a topological space is a homeomorphism. (f) etc.1. 10. Let f : X → Y be a homeomorphism. Then for every A ⊂ X (a) A is closed in X.I. (b) f (Cl A) = Cl f (A). Then U ⊂ X is open (in X). Enigma. Find an example of a continuous bijection.A. 10.E. 10.B. (c) f (Int A) = Int f (A).H. 10. iﬀ f (U) is open (in Y ). when the notion of 67 .
M. deﬁned topology as a part of geometry which deals with the properties preserved by homeomorphisms. Felix Klein in his famous Erlangen Program.9 where he classiﬁed various geometries that had emerged up to that time. HOMEOMORPHISMS 68 topological space had not been developed yet.L. Any surjective strictly monotone map X → Y is a homeomorphism with respect to the interval topological structures in X and Y .K. although it became a kind of program. and projective geometries.7. b] → [c. 10. Then mathematicians studied only subspaces of Euclidean spaces. Let f : X → Y be a homeomorphism.§ 10. c. 10.3. 10. Find two homeomorphic spaces X and Y and a continuous bijection X → Y . Prove that mapping f : H → H deﬁned by f (z) = homeomorphism if a c b > 0. It was a sort of dissertation presented by Klein for getting the position as a professor at Erlangen University.J. 10. Prove that a bijection R → R is a homeomorphism. 9 . 10. where a.5. d] is a homeomorphism. Find all homeomorphisms of the space (see Section § 2) to itself. 10. d az + b .N Corollary. Prove that the same holds true for a discrete space and RT1 . § 10 ◦ 4 More Examples of Homeomorphisms 10. In fact it was not assumed to be a program in the sense of being planned. is a cz + d Rx : Rn x2 {0} → Rn {0} 10. Prove that every nondegenerate aﬃne transformation of Rn is a homeomorphism. 10. which is not a homeomorphism. Prove that every bijection of an indiscrete space onto itself is a homeomorphism. aﬃne. 10. like Euclidean. Prove that inversion x→ is a homeomorphism. Lobachevsky. 10. their continuous mappings and homeomorphisms.8. Prove that every continuous bijection of the arrow onto itself is a homeomorphism. 10.2. b. Prove that every isometry (see Section § 9) is a homeomorphism. Any surjective stricly monotone function f : [a.6. Prove that for every A ⊂ X the reduction ab(f ) : A → f (A) is also a homeomorphism. Let X and Y be totally ordered sets.4. d ∈ R. iﬀ it is a monotone function. Let H = { z ∈ C  Imz > 0 } be the upper halfplane.
K 2 = {(x. (c) circle S 1 is homeomorphic to the boundary of square ∂I 2 = I 2 Int I 2 . = (e) [0. y) ∈ R2  y > 0 }.. (c) open strip { (x. i. b] for any a < b. . plane cut along the ray { y = 0. 1) }. (b) every bounded open convex nonempty set in the plane is homeomorphic to the plane. ≈. e. = (d) (−1.§ 10. (d) halfplane { (x. y > 0 }. c < y < d }.S. (f) open disk { (x. = = 10. g. § 10 ◦ 5 Examples of Homeomorphic Spaces 10.10.P. (i) { (x. y ∈ [0. = Below the homeomorphism relation is denoted by ∼ It is not a =.Q. S 1 {(0.O. (g) open rectangle { (x. 1) }. Prove that (a) closed disk D 2 is homeomorphic to square I 2 = { (x. y) ∈ R2  x > 0. Prove that the following plane ﬁgures are homeomorphic: (a) the whole plane R2 . S n {point} ∼ Rn . x ≥ 0 }. y) ∈ R2  x.e. 1)} ∼ R1 . y) ∈ R2  x ∈ (0. 1) ∼ R. y) ∈ R2  x2 + y 2 < 1 }. y) ∈ R2 : x ∈ K. HOMEOMORPHISMS 69 10. +∞) and (0. y ∈ (0. ∼. y) ∈ R2  x2 + y 2 < 1 } is homeomorphic to open square Int I 2 = { (x. 1] }. 1) }. (b) open square { (x. b] for any a < b. In other textbooks any sign close to. b) for any a < b. is used. b) ∼ (0. 1) ∼ (0. 1) }. = = = (c) (0. 1] ∼ [a. y ∈ (0. 10. 3k ∞ k=1 a2k 3k 10. (b) open disc Int D 2 = { (x. y) ∈ R2  y 2 + x > x }. commonly accepted notation. +∞). y) ∈ R2  a < x < b.R. y) ∈ R2  x. 1) ∼ [a. = 10. (h) open quadrant { (x. y) ∈ R2  x. . (c) boundary of every bounded convex set in the plane with nonempty interior is homeomorphic to S 1 . Prove that (a) [0. 1) ∼ [0. 10. Prove that (a) every bounded closed convex set in the plane with nonempty interior is homeomorphic to D 2 . (e) open halfstrip { (x. 1] ∼ (a. 1) ∼ (a. = (b) [0. Is γ2 : K → K 2 considered in Problem 9:G a homeomorphism? Recall that K is the Cantor set. but distinct from =.9. y ∈ K} and γ2 is deﬁned by ∞ k=1 ak → 3k ∞ k=1 a2k−1 . y ∈ (0. y) ∈ R2  x.
y)  x2 + y 2 > 1 }.. y)  0 ≤ x ≤ 1. 1] }. Prove that the following plane ﬁgures are homeomorphic to each other: (a) punctured plane R2 {(0. y) ∈ R2  f (x) ≤ y ≤ g(x) } is homeomorphic to a closed strip { (x. y < 1 }. 10. Prove that every closed simple (i.§ 10. i. postpone a proof of nonexistence of homeomorphisms till Section § 11. y)  0 < x < 1. (b) punctured disc { (x. (f) { (x. for a line or boundary of a domain the property of having angles is not preserved by homeomorphism.16.e. 10. Prove that the following plane ﬁgures are homeomorphic to each other: (a) { (x. . (d) { (x. Generalize the previous three problems to the case of sets in Rn with arbitrary n.. (f) plane without segment R2 [0. Prove that its complement R2 X is homeomorphic to the punctured plane. i=1 10. g : R → R such that f < g. (Make a list without repeats. y)  x.17. In which of the situations considered in 10. (g) { (x. 10. 1].e. .19. dislocation of discs D1 . y)  0 ≤ x. 0 ≤ y < 1 }.) 10. Let X ⊂ R2 simple nonclosed ﬁnite polyline. y ≥ 0 }. ±1). Prove that any two planes with n holes are homeomorphic. y)  y ∈ [0. 10.18. . y > 1 } ∼ I 2 {(±1. x = 1 }. Dn does not aﬀect on the topological type of R2 ∪n Int Di . 10. y)  0 ≤ x. 1]. Let D1 . i. . the space between their graphs { (x. 10. . (e) { (x. And now two more problems on this. The latter four problems show that angles are not essential in topology.22. 0)}. . Dn ⊂ R2 be pairwise disjoint closed discs. 0 ≤ y < 1 }. y ≤ 1 }. 10. y)  x ≥ y ≥ 0 }.S can the assumption that the set is bounded be omitted? 10. Classify up to homeomorphism all closed convex sets in the plane.21. (b) { (x. . = 10. (±1. Prove that every nonclosed simple ﬁnite unit polyline in R2 (and in Rn with n > 2) is homeomorphic to the segment [0. Prove that for continuous functions f. without selfintersections) polygon in R2 (and in Rn with n > 2) is homeomorphic to the circle S 1 . y)  x2 + y 2 ≤ 1. ±1)}. Let X ⊂ R2 be an union of several segments with a common end point. (d) plane without disc { (x. y)  a < x2 + y 2 < b } where 0 < a < b. (e) plane without square { (x.15.20. HOMEOMORPHISMS 70 10. (c) annulus { (x.14.11. Prove that the complement R2 X is homeomorphic to the punctured plane. .. prove that every such set is homeomorphic to one in the list. Prove that R2 { x. . y)  0 < x2 + y 2 < 1 }.13*.e. The complement of the union of its interior is said to be plane with n holes. (c) { (x.12. y)  x ≥ 0 }.
Arrange the following items to homeomorphism classes: a cup. 10. 10. In a spherical shell (the space between two concentric spheres) one drilled out a cylindrical hole connecting the boundary spheres.25. a wedding ring. a fork. a ﬂower pot (with hole in the bottom). 1. Figure 2. a drill. 10. a spoon. More details about this is given in Section § 20. 1)} . Prove that the rest is homeomorphic to D3 . Prove that surfaces shown in the Figure 3 are homeomorphic. a key. Prove that a mug (with handle) is homeomorphic to a doughnut. Prove that subset of the sphere S n deﬁned in standard coordinates 2 in Rn+1 by inequality x2 + x2 + · · · + x2 < x2 + · · · + xn is homeomorphic 1 2 k k+1 to Rn Rn−k .28.) 10. a bolt. a screw. a nut. Prove that surfaces shown in Figure 2 are homeomorphic (they are called handles). . a coin. HOMEOMORPHISMS 71 10.23.29*. In a spherical shell one made a hole connecting the boundary spheres and having the shape of a knotted tube (see Figure 1.). 10. a glass.30.27. (They are homeomorphic to Klein bottle with two holes. Prove that R3 S 1 ∼ R3 = R1 ∪ {(1. a saucer.26. 10. a knife. 10.24. a nail.§ 10. a plate. Prove that the rest of the shell is homeomorphic to D3 . Figure 1.
Problems 10.32.e. To prove that spaces are homeomorphic. Euclidean spaces are homeomorphic neither to balls. § 10 ◦ 8 Information: Nonhomeomorphic Spaces Euclidean spaces of diﬀerent dimensions are not homeomorphic.30 and 10. Spaces consisting of diﬀerent number of points are not homeomorphic. The spheres S p .T). 10. In particular. HOMEOMORPHISMS 72 Figure 3. In each special case the character of solution depends mainly on the answer. D q with p = q are not homeomorphic.U. the number of elements) of the set of points and the set of open sets (cf. while the other does not. 10. RT1 and the arrow are pairwise nonhomeomorphic. To prove that spaces are not homeomorphic. R.§ 10. Obvious examples of them are the cardinality (i. 10.T.. Prove that the spaces Z. Essentially this is what one usually does in this case. The balls D p . it does not suﬃce to consider any special mapping. A discrete space and an indiscrete space (which have more than one point) are not homeomorphic. Letters A and B are not homeomorphic (if the lines are absolutely thin!). Less obvious examples are the main object of the next chapter. and usually it is impossible to review all the mappings. Find two nonhomeomorphic spaces X and Y for which there exist continuous bijections X → Y and Y → X. Therefore for proving nonexistence of a homeomorphism one uses indirect arguments. Properties and characteristics which are shared by homeomorphic spaces are called topological properties and invariants. Q (with topology induced from R). nor to spheres (of any dimension). § 10 ◦ 7 Homeomorphism Problem and Topological Properties One of the classic problems of topology is the homeomorphism problem: to ﬁnd out whether two given topological spaces are homeomorphic. § 10 ◦ 6 Examples of Nonhomeomorphic Spaces 10. one ﬁnds a property or a characteristic shared by homeomorphic spaces and such that one of the spaces has it. Punctured plane .31. S q with p = q are not homeomorphic. it is enough to present a homeomorphism between them.
Prove that spaces R. Deduce from the Inverse Function Theorem (see. f2 : X → Y are said to be equivalent. Prove that Q cannot be embedded into Z. e. 1984. e. Give an example of continuous injective map. B. 10. Berlin. 10. which is not a topological embedding. any course of advanced calculus) the following statement: ∂f For any diﬀerentiable function f : Rn → Rn whose Jacobian det( ∂xi ) j does not vanish at the origin 0 ∈ Rn there exists a neighborhood U of the origin such that f U : U → Rn is an embedding and f (U ) is open. 10. A. However some of them can not be proven by techniques of this course. D. 10. { x2 + y 2 < These statements are of diﬀerent degrees of diﬃculty.§ 10.33.36. Find topological spaces X and Y such that X can be embedded into Y . The inclusion of a subspace into a space is an embedding. if there exist homeomorphisms hX : X → X and hY : Y → Y such that f2 ◦ hX = hY ◦ f1 (the latter equality may be stated as follows: the diagram −→ X −− Y h X −− Y −→ f2 f1 hX Y is commutative).X.. RT1 . and the arrow cannot be embedded into each other.g.) § 10 ◦ 9 Embeddings Continuous mapping f : X → Y is called a (topological) embedding if the submapping ab(f ) : X → f (X) is a homeomorphism.34. Beginner’s course in topology: Geometric chapters.37 Corollary of Inverse Function Theorem. 10.35. Some of them will be considered in the next section. (See. but X ∼ Y . = 10. New York: SpringerVerlag. Composition of embeddings is an embedding. (Find such an example above and create a new one.V. Rokhlin. Fuchs.W..g. Y can be embedded into X. . HOMEOMORPHISMS 73 R2 {point} is not homeomorphic to the plane with hole R2 1 }. V. § 10 ◦ 10 Equivalence of Embeddings Embeddings f1 .) 10. Can a discrete space be embedded into an indiscrete space? How about vice versa? 10.
Prove that knots f1 .§ 10. Prove that knots are equivalent. HOMEOMORPHISMS 74 An embedding of the circle S 1 into R3 is called a knot. 10.38. For instance. 10. f2 : S 1 → R3 with f1 (S 1 ) = f2 (S 1 ) are equivalent. § 10 ◦ 11 Information There are nonequivalent knots. and . .39.
8. it means that this set lies in some topological space (which should be clear from the context). 11. A partition of a set is a cover of this set with pairwise disjoint sets. 75 . A topological space is connected. in RT1 ? 11. Show that the set [0. 11.3.9. iﬀ it cannot be partitioned into two nonempty open sets. 11. 1} connected in R. 1] ∪ (2. and C ∩ ClX B = ∅.7. B ∩ ClX C = ∅. Is an indiscrete space connected? The same for the arrow and RT1 . To partition a set means to construct such a cover. is (X. and.CHAPTER 2 Topological Properties § 11 Connectedness § 11 ◦ 1 Deﬁnitions of Connectedness and First Examples A topological space X is said to be connected if it has only two subsets which are both open and closed: ∅ and the entire X. iﬀ it cannot be partitioned into two nonempty closed sets. Ω2 ) connected? If (X. Is the set {0.. Prove that A is disconnected. 11. is a connected topological space. Let A be a subset of a topological space X. and Ω2 be ﬁner than Ω1 (i.10. with the induced topology. Describe explicitly all connected discrete spaces. 11. iﬀ there exist nonempty sets B and C such that A = B ∪ C.2. 11. Ω1 ) is connected. Ω1 ⊂ Ω2 ).6. 11. Let Ω1 .5. 11.A. Describe explicitly all connected subsets of the arrow. is (X. of RT1 . Ω2 ) is connected. 11. Prove that every nonconvex subset of the real line is disconnected. 3] is disconnected in R. Is the set Q of rational numbers (with the topology induced from R) connected? The same about the set of irrational numbers.e. Ω2 be topological structures in a set X. If (X. Give a deﬁnition of disconnected subset without relying on the induced topology.4.1. 11. in the arrow. Ω1 ) connected? § 11 ◦ 2 Connected Sets When one says that a set is connected.
10–11. B be connected sets. Let A be a connected subset of a connected space X.B.16. Let {Ak }k∈Z be a family of connected sets such that Ak ∩Ak+1 = ∅ for any k ∈ Z.15. It is the union of all connected sets containing this point. The closure of a connected set is connected. Is ∞ Ak a connected set? k=1 § 11 ◦ 4 Connected Components A connected component of a space X is its maximal connected subset.D. 11. (In other words: the union of pairwise intersecting connected sets is connected. 11. Prove that if A and B are either both closed or both open sets.§ 11. Prove that if a set A is connected and A ⊂ B ⊂ Cl A. 11. then B is connected. Then λ∈Λ Aλ is connected.C. and A ∩ Cl B = ∅.) 11. Prove that A ∪ B is connected. 11. Prove that k∈Z Ak is connected. that is a connected subset. which is not contained in any other (strictly) larger connected subset of X. 11. and V ∩ A = ∅. Two connected components either are disjoint or coincide. Let {Aλ }λ∈Λ be a family of connected subsets of a space X. Let A1 ⊃ A2 ⊃ · · · be an inﬁnite descending sequence of connected spaces.5. 11.11. Let A. 11.12.G. and B ⊂ X A be an open and closed set in the topology of the subspace X A of the space X. and their union and intersection are connected then A and B are connected. Connected components are closed. Every point belongs to some connected component. Prove that A ∪ B is connected.13. . U ∩ A = ∅. either U ⊃ A.F. Assume that any two sets of this family intersect. which form a cover of X.17. or V ⊃ A. Does connectedness of A ∪ B and A ∩ B imply connectedness of A and B? 11.E. 11.18. 11. this component is unique. Compare 11. Find a topological space X and disconnected subset A ⊂ X such that for any disjoint open sets U and V . Moreover. Prove that for every disconnected set A in Rn there exist disjoint open sets U and V such that A ⊂ U ∪ V . too.14. 11. CONNECTEDNESS 76 11.12 with 11. § 11 ◦ 3 Properties of Connected Sets 11.
11.22. Let x and y belong to the same component.19. 11.25.23.§ 11. 11. § 11 ◦ 6 Frontier and Connectedness 11. § 11 ◦ 5 Totally Disconnected Spaces A topological space is called totally disconnected if each of its components consists of a single point. and X is connected then f (X) is also connected. Prove that two points are in the same component. 11.L Corollary. F ∩ A.G mean that connected components comprise a partition of the whole space.20. 11. Prove that Cantor set (see 2:A) is totally disconnected. Prove that any set.29). § 11 ◦ 7 Connectedness and Continuous Maps A continuous image of a space is its image under a continuous mapping. . Prove that if A ⊂ X.J. 11. and the multiplication by an element of the group is a continuous map. which is closed and open.H. iﬀ they belong to the same connected set.21.24. 11. A continuous image of a connected space is connected. The next theorem describes the corresponding equivalence relation.K. Let F be a connected subset of X. Let a space X has a group structure. Prove that if Fr A is a connected set then Cl A is also connected.E and 11. and F ∩ (X A) = ∅ then F ∩ Fr A = ∅.I Obvious Example. Theorems 11. 11. Prove that if A is a proper nonempty subset of a connected topological space then Fr A = ∅. CONNECTEDNESS 77 A connected component of a space X is called just a component of X. Prove that the component of unity is a normal subgroup.) 11. Let A be a subset of connected topological space. 11. Note that Q is not discrete. 11. Give an example of an uncountable closed totally disconnected subset of the line. Any discrete space is totally disconnected. (In other words if f : X → Y is a continuous map. The space Q (with the topology induced from R) is totally disconnected. 11. Connectedness is a topological property. either contains both x and y or does not contain either of them (cf.
an ∈ U and an ascending sequence bn with b1 = b. 11. 1] is connected. 11. Prove that the set of connected components of an open subset of R is countable. § 11 ◦ 8 Connectedness on Line 11. see 11. V be subsets of I with V = U V . Prove that every open subset of the real line is a union of disjoint open intervals (do not use 11. Is this set connected? Would the answer change.O. or neither of them (cf.O from this.N. 11. also 11.26 below. Cf. iﬀ there is a continuous surjection X → S 0.1 U and V are open. Deduce 11. ϕ are the polar coordinates) and circle S 1 .N with a traditional proof of Intermidiate Value Theorem is sketched in the following two problems.Q.O. Prove that R1 is connected. Describe explicitly all connected subsets of the line. CONNECTEDNESS 78 11. b ∈ V and a > b. 11.M Corollary.28.29.§ 11.13) 11.O.19).27. 11. 11. but refers to a famous Intermediate Value Theorem from calculus. Prove that every convex set in Rn is connected. Let a ∈ U . Is it connected? 11.2. then in which of them can be c? 11.R.26.O.T.O. Let U .S. Basically the same proof as a combination of 11. One is suggested by 11. with ϕ ≥ 0 (r. Consider the union of spiral r = exp 1 1 + ϕ2 . bn ∈ V such that both an and bn have the same limit c. The segment I = [0. There are several ways to prove 11. If under assumptions of 11.P.O).N. Consider the subset of the plane R2 consisting of points with both coordinates rational or both coordinates irrational. Find a space and two points belonging to its diﬀerent components such that each simultaneously open and closed set contains either both of the points. 11. 11. 11. Prove that there exists a descending sequence an with a1 = a. . 11.1. The number of connected components is a topological invariant. if the entire circle was replaced by some its subset? (Cf. A space X is not connected.
37. What if the knife is not makes cuts of a shape diﬀerent from straight line? For which shapes of the blade you can formulate and solve problems similar to 11.31. Let X be a connected space and f : X → R a continuous function. CONNECTEDNESS 79 § 11 ◦ 9 Intermediate Value Theorem and Its Generalizations The following theorem is usually included in Calculus.U Generalization. What about pancakes in Rn ? § 11 ◦ 11 Induction on Connectedness A function is said to be locally constant if each point of its source space has a neighborhood such that the restriction of the function to this neighborhood is constant. under the conditions of 11.31 – 11.31 and 11. if A is a bounded connected open set in the plane.32. Can you increase the number of regions in the counterpart of 11.§ 11.34 Dividing Pancake. 11. 11. 11. Formulate and solve counterparts of Problems 11. in a sense it is equivalent to connectedness of interval. Then f (X) is a convex subset of R.34 for regions in the threedimensional space.33? 11.T Intermediate Value Theorem. if A is a bounded open set in the plane and l is a line in the plane.36.34? 11. If. § 11 ◦ 10 Dividing Pancakes 11. Suppose two irregularly shaped pancakes lie on the same platter. b] → R Many problems which can be solved using Intermediate Value Theorem can be found in Calculus textbooks. In other words. A is connected then L is unique. Prove that any polynomial of odd degree in one variable with real coeﬃcients has at least one real root. A locally constant function on a connected set is constant. Prove that a plane pancake of any shape can be divided to four pieces of equal area by two straight cuts orthogonal to each other. In other words: if A and B are two bounded regions in the plane. A continuous function takes every value between f (a) and f (b). f : [a. Enigma.33.35. In fact. then there are two perpendicular lines which divide A into four parts having equal areas. . 11. Enigma.31. then there exists a line in the plane which divides each region in half by area. Enigma.V. 11. Any irregularly shaped pancake can be cut in half by one stroke of the knife made in any prescribed direction. show that it is possible to cut both exactly in half by one stroke of the knife. You can easily deduce it from the matterial of this section.31 – 11. In other words. 11. Here are few of them. then there exists a line L parallel to l which divides A in half by area. 11. 11.30.
Prove that square and segment are not homeomorphic.14. [0. CONNECTEDNESS 80 11.V and solve 11.42. from the LebesgueBrower Theorem on invariance of dimension (see.40. Dimension Theory Princeton. W.g.24 and 11. 11. 11.38.39.43.. Wallman. then the space has property E. The statement “Rp is not homeomorphic to Rq unless p = q” implies that S p is not homeomorphic to S q unless p = q. R1 . For more applications of induction on connectedness see 12. 11. C. which assign homeomorphic spaces to homeomorphic ones (e. . Rp and Rq are not homeomorphic unless p = q. Hurewicz and H. R1 and Rn are not homeomorphic if n > 1. for instance. Enigma. Prove that if the topology induced in a normal subgroup H of G is discrete. Let G be a group equipped with a topology such that for any g ∈ G the map G → G deﬁned by x → xgx−1 is continuous. Let E be a property of subsets of a topological space such that the union of sets with nonempty pairwise intersections inherits this property from the sets involved. 2] and [0.X.V related? 11. 11. e. 12. allow one to use connectedness for proving that some connected spaces are not homeomorphic. 11. 1] ∪ [2. then H is contained in the center of G (i.W.39 using 11. NJ.. 1941).R.45. 11. D. § 11 ◦ 12 Applications to Homeomorphism Problem Connectedness is a topological property. 11.§ 11.16 and 12. I. see Section § 9.Y. S 1 and [0.44. Give a topological classiﬁcation of the letters: A. Prove that if the space is connected and each its point has a neighborhood with property E. B. considered as subsets of the plane (the arcs comprising the letters are assumed to have zero thickness). . .18. . 12. It follows.. Prove that a circle is not homeomorphic to any subspace of R1 .41. Prove 11. and let G with this topology be connected. ∞) are pairwise nonhomeomorphic. 3] are not homeomorphic.g. How are 11. Information. and the number of connected components is a topological invariant (see Section § 10). deleting one or several points). hg = gh for any h ∈ H and g ∈ G).e. Recall that there exist continuous surjections of the segment onto square and these maps are called Peano curves. 11. Simple constructions.40 Induction on Connectedness. 11.
this notation is already used (for the inverse mapping). Prove that sphere of dimension n > 0 is pathconnected. This terminology is inspired by an image of moving point: at the moment t ∈ [0.D. 12. Prove that the zerodimensional sphere S 0 is not pathconnected. Which of the following topological spaces are pathconnected: (a) a discrete space. The point s(0) is called the initial point of a path s : I → X. Cf.B. A constant map s : I → X is called a stationary path and denoted by ea where a = s(I). v : I → X be paths such that u(1) = v(0). (b) (c) the arrow. Prove that I is pathwise connected. it is a path). § 12 ◦ 2 PathConnected Spaces A topological space is said to be pathconnected or pathwise connected . Set (22) uv(t) = u(2t). Prove that the Euclidean space of any dimension is pathwise connected. 12. if any two points can be connected in it by a path. strictly speaking.X. while s(1) is called its ﬁnal point.e. 1]. Prove that the map uv : I → X deﬁned by (§ 12 ◦ 1) is continuous (i. 12. 1] to X. RT1 .A. 12.1. 9. It is denoted by s−1 . 1] it is in s(t). inverse mappings. v(2t − 1). Path uv is called the product of paths u and v.. One says that path s connects s(0) with s(1). as a rule.C. the ambiguity of notations does not lead to confusion: in the context involving paths. 81 . Although. To tell the truth. 12. For a path s the inverse path is the path deﬁned by t → s(1 − t). do not appear. 2 ] 1 if t ∈ [ 2 .V and 9. 1 if t ∈ [0. Let u : I → X. since besides an information on trajectory of the point it contains a complete account on the movement: the schedule saying when the point goes through each point. 12. (d) ? (e) an indiscrete space.E. this is more than what is usually called a path. Recall that it is deﬁned only if the ﬁnal point u(1) of u coincides with the initial point v(0) of v.§ 12 PathConnectedness § 12 ◦ 1 Paths A path in a topological space X is a continuous mapping of the interval I = [0.
Every point belongs to a pathconnected component. .P. iﬀ any two points in it can be connected by a path s : I → X with s(I) ⊂ A. in some important situations it is even equivalent to connectedness.) 12.7.O. Let A be a subset of Euclidean space. proofs are usually easier in the case of pathconnectedness. 12.5.§ 12. 12. 12. Two pathconnected components either coincide or are disjoint. (See 12. Prove that the same holds true for a subset of an arbitrary pathconnected space.F. 12. Prove that two points belong to the same pathconnected component. § 12 ◦ 4 Properties of PathConnected Sets Pathconnectedness is very similar to connectedness.3. Unlike to the case of connectedness. Prove that if sets A and B are both closed or both open and their union and intersection are pathconnected. 12. pathconnected components may be nonclosed. then A and B are also pathconnected. 12.P). § 12 ◦ 5 PathConnected Components A pathconnected component or pathwise connected component of a topological space X is a pathconnected subset of X such that no other pathconnected subset of X contains it. However.2. cf.N. which carry over. Prove that a convex subset of Euclidean space is pathconnected. Prove that if Fr A is connected then Cl A is also connected. 12. A continuous image of a pathwise connected space is pathwise connected. 12. 12. Prove that interior and frontier of a pathconnected set may not be pathconnected and that connectedness shares this property. Prove that the set of plane convex polygons with topology deﬁned by the Hausdorﬀ metric is pathconnected.I.4.J. 12. Further. The union of a family of pairwise intersecting pathconnected sets is pathconnected. iﬀ they can be connected by a path. PATHCONNECTEDNESS 82 § 12 ◦ 3 PathConnected Sets By a pathconnected set or pathwise connected set one calls a subset of a topological space (which should be clear from the context) pathconnected as a space with the topology induced from the ambient space. 12.G. 12. 12. some properties of connectedness do not carry over pathconnectedness (see 12. 12.H. For properties.6.8.O. Prove that a subset A of a topological space X is pathconnected.
then the space is pathconnected. Prove that for any ε > 0 an εneighborhood of a connected subset of Euclidean space is pathconnected. For subsets of the real line pathconnectedness and connectedness are equivalent. Prove that deleting any point from A makes A and X disconnected (and hence. 0)}. . 12. Prove that A is pathconnected and X is connected.13. Let s : I → X be a path connecting a point of a set A with a point of X A. y) ∈ R2 : x > 0. Prove that s(I) ∩ Fr(A) = ∅. Find an example of a pathconnected set.Q. If each point of a space has a pathconnected neighborhood. 12. 12. 12. If each point of a space has a pathconnected neighborhood.§ 12. 12. 12.P. 12. iﬀ it is connected. (x. not pathconnected).10. 12. § 12 ◦ 6 PathConnectedness Versus Connectedness 12.S. 12. 12.12. Find an example of a pathconnected component that is not closed.O. y = sin 1 x 12.9.14. Draw A. PATHCONNECTEDNESS 83 12. Prove that any neighborhood of a connected subset of Euclidean space contains a pathconnected neighborhood of the same set. whose closure is not pathconnected.11. For an open subset of Euclidean space connectedness is equivalent to pathconnectedness.L. X is not pathconnected. 12. Put A= and X = A ∪ {(0. 12.N. § 12 ◦ 7 PolygonConnectedness A subset A of Euclidean space is said to be polygonconnected if any two points of A can be connected by a ﬁnite polygonal line contained in A.M. Prove that for open subsets of Euclidean space connectedness is equivalent to polygonconnectedness. then each pathconnected component is open.R.K. Any pathconnected space is connected.
12. C). 2 2 2 .. Find connected and pathconnected components of the following subspaces of the space of complex n × nmatrices: • GL(n. • Symm(n. Let X ⊂ Cn be a union of a countable collection of algebraic subsets (i.16. R) ∩ GL(n. which diﬀers from Rn only in the way of enumeration of its natural coordinates (they are numerated by pairs of indices). 12. PATHCONNECTEDNESS 84 12. X is polygon 12. C) ∩ GL(n. • Herm(n.20.19. Prove that then Rn X is polygonconnected. C) = {A : det A = 0}. that real n × nmatrices comprise a space. Let X ⊂ R2 be a countable set. ¯ • U (n. 12. Find connected and pathconnected components of the following subspaces of the space of real n × nmatrices: • GL(n. C) = {A : A · (t A) = }. § 12 ◦ 8 Connectedness of Some Sets of Matrices Recall. ¯ • Herm(n.e. • {A : A2 = }. R) = {A : det A = 0}. The same relation holds between the set of complex n×nmatrix and Cn (homeomorphic to R2n ). R) = {A : t A = A}. Let X ⊂ Rn be a union of a countable collection of aﬃne subspaces with dimensions not greater than n − 2. • Symm(n. Prove that then R2 connected.17. R) = {A : A · (t A) = }. C) = {A : t A = A}.15.18. R). subsets deﬁned by systems of algebraic equations in the standard coordinates of Cn ) Prove that then Cn X is polygonconnected. Construct a pathconnected subset A of Euclidean space such that A consists of more than one point and no two distinct points can be connected with a polygon in A. • O(n. 12.§ 12.
(c) the arrow..2. 13. and T4 . none. T3 . T2 . They are numerated and denoted by T0 . we advise you to think over all deﬁnitions and solve all simple problems.. 1] with the topology induced from R a Hausdorﬀ space? Do the points 0 and 1 possess disjoint neighborhoods? Which if any? § 13 ◦ 2 Limits of Sequence Let {an } be a sequence of points of a topological space X. 13. Letter T in these notations originates from a German word Trennungsaxiom. Any metric space is Hausdorﬀ. if for any neighborhood U of b there exists a number N such that an ∈ U for any n > N. (e) ? If the next problem holds you up even for a minute. which means separation axiom. Besides the notation T2 it has a name. Which of the following spaces are Hausdorﬀ: (a) a discrete space. (d) RT1 . Is the segment [0. respectively. the words no. A topological space satisfying it is called a Hausdorﬀ space.B.1. Prove that in the space RT1 each point is a limit of the sequence {an = n}. We restrict ourselves to the most important ﬁve of them. not. the Hausdorﬀ axiom. 13.A.§ 13 Separation Axioms The aim of this section is to consider natural restrictions on topological structure making the structure closer to being metrizable.1 § 13 ◦ 1 The Hausdorﬀ Axiom Let us start with the most important second axiom.) 13.e. Explain the meaning of the statement “ b is not a limit of sequence an ” avoiding as much as you can negations (i. 13.C. In a Hausdorﬀ space any sequence has at most one limit. T1 . A point b ∈ X is called its limit. The sequence is said to converge or tend to b as n tends to inﬁnity.D. 85 1 . This axiom is stated as follows: any two distinct points possess disjoint neighborhoods. (b) an indiscrete space. 13. A lot of separation axioms are known. etc.
How are problems 13. 13. The set of all ﬁxed points of a map f is called the ﬁxed point set of f . Prove that the ﬁxed point set of a continuous map of a Hausdorﬀ space to itself is closed. 13.E. 13. i.F. • inﬁniteness of the set of points.5.7.11.7 related? § 13 ◦ 4 Hereditary Properties A topological property is called hereditary if it is carried over from a space to its subspaces. 13.3. 13. • iﬀ all onepoint sets in X are closed. § 13 ◦ 5 The First Separation Axiom A topological space is said to satisfy the ﬁrst separation axiom T1 if each of any two points of the space has a neighborhood which does not contain the other point. Prove that if f. Then the set {x ∈ X : f (x) = g(x)} is called the coincidence set of f and g. 13. 13. 13.4. 13.9. iﬀ any point of X coincides with the intersection of all its neighborhoods.H. • ﬁniteness of the topological structure. The property of being Hausdorﬀ space is hereditary. . A topological space X satisﬁes the ﬁrst separation axiom.G. A is everywhere dense in X. In a Hausdorﬀ space any ﬁnite set is closed. • pathconnectedness? 13. 13.8. g : X → Y be maps.10. SEPARATION AXIOMS 86 § 13 ◦ 3 Coincidence Set and Fixed Point Set Let f.5. Prove that the coincidence set for two continuous maps of an arbitrary topological space to a Hausdorﬀ space is closed. Prove that a space X satisﬁes the ﬁrst separation axiom. Construct an example proving that the Hausdorﬀ condition in 13. Which of the following topological properties are hereditary: • ﬁniteness of the set of points. g : X → Y are continuous maps.6. • connectedness. Enigma.3 is essential. A metric space satisﬁes the ﬁrst separation axiom.5 is essential.3. and 13. 13. Y is Hausdorﬀ.§ 13. 13. A point x ∈ X is called a ﬁxed point of a map f : X → X if f (x) = x. Construct an example proving that the Hausdorﬀ condition in 13. • iﬀ all ﬁnite sets in X are closed.e. Any Hausdorﬀ space satisﬁes the ﬁrst separation axiom. if a space X possesses this property then any subspace of X possesses it. and f A = gA then f = g. 13.
13. among which we see the most important spaces. 13.14. Ω) is a smallest neighborhood topological space satisfying the Kolmogorov axiom then the relation “a belongs to any closed set which contain b” is a nonstrict partial order in X and the poset topology deﬁned by this partial order coincides with Ω. The following properties of a topological space X are equivalent: • X satisﬁes the Kolmogorov axiom. The ﬁrst separation axiom is hereditary. How can the ﬁrst separation axiom be weakened? A topological space is said to satisfy the Kolmogorov axiom or the zeroth separation axiom T0 . • X contains no indiscrete subspace consisting of more than one point. which is quite far away from the class of metric spaces. 13.2.§ 13. 13. Prove that if (X.K. Show that RT1 meets the ﬁrst separation axiom. . 13. Prove that in every set there exists a coarsest topological structure satisfying the ﬁrst separation axiom. 13.L. iﬀ this is a smallest neighborhood topology satisfying the Kolmogorov axiom. 13. A topology is a poset topology.M. 13. if at least one of any two distinct points of this space has a neighborhood which does not contain the other of these points. Prove that a continuous mapping of an indiscrete space to a space satisfying axiom T1 is constant.N. Express axioms of nonstrict order in terms of poset topology. but is not a Hausdorﬀ space (cf.12). Find an example showing that the ﬁrst separation axiom does not imply the Hausdorﬀ axiom.N. on one hand. 13.N. on the other hand. Enigma.I. An indiscrete space containing at least two points does not satisfy T0 . • X contains no indiscrete subspace consisting of two points.1.12. § 13 ◦ 6 The Kolmogorov Axiom The ﬁrst separation axiom emerges as a weakened Hausdorﬀ axiom. Thus. Describe this structure. Prove that if for any two distinct points a and b of a topological space X there exists a continuous map f of X to a space with the ﬁrst separation axiom such that f (a) = f (b) then X possesses the ﬁrst separation axiom. • any two diﬀerent points of X has diﬀerent closures. 13.15. 13. 13. all posets are obtained from topological spaces of a special kind. SEPARATION AXIOMS 87 13.J. posets give rise to numerous examples of topological spaces. 13. like the line with the standard topology.
b)  a ∈ A. 13. for any closed set F ⊂ X and point b ∈ X \ F there exist open sets U. Find a space satisfying the third. 13. 13. Prove that a space satisﬁes the fourth separation axiom. 13. i. A space is normal. iﬀ any neighborhood of any point contains the closure of some neighborhood of the same point. 13. 13. and B ⊂ V .T. 13. Any metric space is normal. 13. 13. Let f : X → Y be a continuous surjection such that the image of any closed set is closed.24. Prove that a space satisﬁes the third separation axiom. A topological space is called normal if it satisﬁes the ﬁrst and fourth separation axioms. iﬀ it satisﬁes the second and fourth separation axioms.e. 13.19. § 13 ◦ 8 The Fourth Separation Axiom A topological space X is said to satisfy the fourth separation axiom if any two disjoint closed sets have disjoint neighborhoods.S. 13.18. but not second separation axiom.e. Prove that if X is normal then Y is normal. B ⊂ X such that A∩B = ∅ there exist open sets U.22. but not the second separation axiom. 13. V ⊂ X such that U ∩ V = ∅. SEPARATION AXIOMS 88 § 13 ◦ 7 The Third Separation Axiom A topological space X is said to satisfy the third separation axiom if any closed set and a point of its complement have disjoint neighborhoods.20. b ∈ B} = 0.§ 13. V ⊂ X such that U ∩ V = ∅.Q. Prove that any closed subspace of a normal space is normal. 13. Find closed disjoint subsets A and B of some metric space such that inf{ρ(a. A topological space is called regular if it satisﬁes the ﬁrst and third separation axioms. A ⊂ U.O.21. for any closed sets A. iﬀ it satisﬁes the second and third separation axioms.23. and b ∈ V . F ⊂ U. Find a space which satisﬁes the fourth. Find a Hausdorﬀ space which is not regular. Prove that the third separation axiom is hereditary.16. 13. .. A normal space is regular (and hence Hausdorﬀ). i. iﬀ in any neighborhood of any closed set contains the closure of some neighborhood of the same set. Any metric space is regular.P. A regular space is Hausdorﬀ space.17. 13.R. A space is regular..
Let T be a subbase2 of the topology of Y .31 Corollary. Prove that there exists a continuous function g : X → − 3 . Let Y be a topological space satisfying the ﬁrst separation axiom. y) ∈ R2 : y = 0}. 2 .27. Let Σ be an open cover of a space X. r) and (r. What topological structure is induced on L from X? 13. 13:C Urysohn Lemma. 13. Rn . y) ∈ R2 : y > 0} equipped with the topology induced by the Euclidean metric. Let f : A → [−1. 13. see Section § 3. which is not normal. 13. SEPARATION AXIOMS 89 § 13 ◦ 9 Niemytski’s Space Denote by H the open upper halfplane {(x. 1].28. which is obtained by adjoining to the Euclidean topology the sets of the form x ∪ D. i. Would the statement of Tietze Theorem remain true if in the hypothesis the segment [−1. Prove that under the conditions of 13:D there exists a continuous function F : X → [−1.26. 13.29 Corollary.§ 13. 13. Theorem 13. 3 2 3 such that f (x) − g(x) ≤ for any x ∈ A. 1] such that F A = f . 1] such that f (x) − φ(x) ≤ ε for any x ∈ A. 1]...25. n. 13:G.30. the property of being normal is not hereditary? § 13 ◦ 10 Urysohn Lemma and Tietze Theorem 13:A*. Let A be a closed subset of a normal space X. 1] be a 1 1 continuous function. This is the Niemytski space. Let A and B be disjoint closed subsets of a normal space X. 1] where r = 2q . Prove that under the conditions of 13:D for any ε > 0 there exists a continuous function φ : X → [−1. Denote by X the union of H and its boundary line L = {(x. It can be used to clarify properties of the fourth separation axiom.22 does not extend to nonclosed subspaces. i. Prove that if there exists a bijection φ : Σ → T which preserves inclusions then there exists a continuous map f : X → Y such that f −1 (V ) = φ−1 (V ) for any V ∈ T . Prove that intervals [0. where x ∈ R1 and D is an open disc in H which is tangent to L at the point x. Prove that the Niemytski space is regular. Prove that the Niemytski space is not normal. but equip it with the topology.e. 13:F Tietze Extension Theorem. q ∈ N form a subbase for [0. 1]. a collection of open sets in [0. whose ﬁnite intersections form a base of the standard topology in [0. Then there exists a continuous function f : X → I such that f (A) = 0 and f (B) = 1. or S 2 ? Recall that a subbase of the topology of Y is a collection T of open sets of Y such that all ﬁnite intersections of sets from T form a base of topology of Y . S 1 . 1] was replaced by R. 13:E. 13:D. Embed the Niemytski space into a normal space in such a way that the complement of the image would be a single point. There exists a regular space. n 13:B. Prove that the Niemytski space is Hausdorﬀ. 13.e.
separability is equivalent to the second axiom of countability. As in the previous section. The second axiom of countability implies separability. Sometimes this term is used only for inﬁnite countable sets.1. The image of a countable set under any mapping is countable. A set of the same cardinality as a subset of the set N of natural numbers is said to be countable.2. § 14 ◦ 2 Second Countability and Separability In this section we study three restrictions on topological structure. for set of the cardinality of the whole set N of natural numbers.§ 14 Countability Axioms In this section we continue to study topological properties which are imposed additionally on a topological structure to make the abstract situation under consideration closer to special situations and hence richer in contents. Any subset of a countable set is countable. This would lead to other more serious inconveniences as well. i. Are the arrow and RT1 separable? 14. Construct an example proving that separability is not hereditary.B. while a set countable in the sense above is called at most countable.A. then this section would have to be called “At Most Countability Axioms”. A metric separable space is second countable.E.) 14. the third one has no number.F.G Corollary.3.e. For metric spaces. 14. The union of a countable family of countable sets is countable. 14. In particular. 90 . A topological space is said to satisfy the second axiom of countability or to be second countable if it has a countable base. § 14 ◦ 1 SetTheoretic Digression. Are the arrow and RT1 second countable? 14. we start from the restriction having number two. Restrictions studied in this section bound a topological structure from above: they require something to be countable. A space is called separable if it contains a countable dense set. (This is the countability axiom without a number mentioned above. The second axiom of countability is hereditary. Two of them have numbers (one and two). 14. Countability Recall that two sets are said to be of equal cardinality if there exists a bijection of one of them onto the other. if we adopted this terminology. Our terminology has the following advantageous properties. 14.D. 14. 14. This is less convenient.C. 14.
1 Then the family contains a minimal set. A neighborhood base at a or just base of X at a is a collection of neighborhoods of a such that any neighborhood of a contains a neighborhood from this collection. COUNTABILITY AXIOMS 91 14. § 14 ◦ 4 Bases at a Point Let X be a topological space. . 14. Prove that the space l2 is separable and second countable. A continuous image of a separable space is separable. Prove that a normal second countable space can be embedded into l2 .) Prove that for metric spaces separability is hereditary. Construct an example proving that a continuous image of a second countable space may be not second countable. .6. Prove that a regular second countable space is normal. 14. iﬀ it is regular.9. 14. 14:B. 14.4. Prove that in a separable space any collection of pairwise disjoint open sets is countable.) 14:D.H. a set such that no proper its subset belongs to the family. • the set of all open balls of center a and rational radii. In a metric space the following collections of balls are neighborhood bases at a point a: • the set of all open balls of center a. and a its point. 14..11. Let {Kλ } be a family of closed sets of a second countable space and let for any descending sequence K1 ⊃ K2 ⊃ .5.8. 14. § 14 ◦ 3 Embedding and Metrization Theorems 14:A. Construct a metric space which is not second countable. of sets belonging to this family the intersection ∩∞ Kn also belongs to the family.§ 14. 14. i.L. 14. 14. If Σ is a base of a space X then {U ∈ Σ : a ∈ U} is a base of X at a. 14. Prove that any set of disjoint ﬁgure eight curves in the plane is countable. .7.e. 14. 14:C. Prove that the set of components of an open set A ⊂ Rn is countable. which also covers the space. Prove that a second countable space is metrizable.K Lindel¨f Theorem.I.J. Any open cover of a second countable space o contains a countable part. Prove that Euclidean spaces and all their subspaces are separable and second countable.10 Brower Theorem*. (Use Urysohn Lemma 13:C. 14. (Cf. Prove that any base of a second countable space contains a countable part which is also a base.3. 14.
12.P. Paying a tribute to this tradition we explain here how and in what situations topological notions can be described in terms of sequences.Q. In fact. The set of limits of all sequences an with an ∈ A is called a sequential closure of A and denoted by SCl A. 14.M. What are the minimal bases at a point in the discrete and indiscrete spaces? § 14 ◦ 5 First Countability A topological space X is says to satisfy the ﬁrst axiom of countability or to be a ﬁrst countable space if it has a countable neighborhood base at each point. Prove that SCl A ⊂ Cl A. 14. except summing of series. Let A be a subset of a topological space X. where {rn } is any sequence of positive numbers converging to zero.O.14.13. The second axiom of countability implies the ﬁrst one. except for a long history descending from the XIX century studies on the foundations of analysis.4. Find a ﬁrst countable space which is not second countable. where sequences are involved in the underlying deﬁnitions. Moreover they like to talk about all topological notions relying on the notions of sequence and its limit. The exceptions which one may ﬁnd in the standard curriculum of a mathematical department can be counted on two hands. 14. If a space X is ﬁrst countable then the for any A ⊂ X the opposite inclusion Cl A ⊂ SCl A holds also true. COUNTABILITY AXIOMS 92 • the set of all open balls of center a and radii rn . Which of the following spaces are ﬁrst countable: (a) the arrow. (d) RT1 . (b) (c) a discrete space. This tradition has almost no mathematical justiﬁcation. an indiscrete space? 14. 14.N. and hence SCl A = Cl A. 3 . (Cf.) 14. 14. Find a ﬁrst countable separable space which is not second countable. 14. 14. provided you deal with topological notions. almost always3 it is more convenient to avoid sequences.§ 14. Any metric space is ﬁrst countable. § 14 ◦ 6 Sequential Approach to Topology Specialists in Mathematical Analysis love sequences and their limits.
14.15. any metric spaces) one can recover (hence. deﬁne) the closure of a set provided it is known which sequences are convergent and what the limits are.T. A map f : X → Y is said to be sequentially continuous if for any b ∈ X and a sequence an ∈ X. sequential closure and closure in X. the sequence f (an ) converges to f (b). As a consequence. Prove that in X there exists a set A with SCl A = Cl A. Describe convergent sequences.§ 14. 14. but discontinuous map. The preimage of a sequentially closed set under a sequentially continuous map is sequentially closed. Thus for mappings of a ﬁrst countable space continuity and sequential continuity are equivalent. 14. (Cf.16. If X is a ﬁrst countable space then any sequentially continuous map f : X → Y is continuous. COUNTABILITY AXIOMS 93 Therefore. knowledge of closures allows one to recover which sets are closed.R. 14. In turn. Let X be the set of real numbers equipped with the topology consisting of ∅ and complements of all countable subsets. in a second countable space (in particular.S. § 14 ◦ 7 Sequential Continuity Consider now continuity of maps along the same lines. which converges to b. 14. Any continuous map is sequentially continuous.15) . knowledge of closed sets allows one to recover open sets and all other topological notions. Construct a sequentially continuous. 14.
E. Another deviation from the terminology used here comes from Bourbaki: we do not include the Hausdorﬀ property into the deﬁnition of compactness. Ω2 ) imply compactness of (X.) Topological space is said to be compact if any of its open covers contains a ﬁnite part which covers the space. 15. 15. which Bourbaki includes.B.A. Urysohn (1898–1924). Which discrete topological spaces are compact? 15. 15. They suggested for it the term bicompactness. compactness.D. It is sort of topological counterpart for the property of being ﬁnite in the context of set theory. Is the arrow compact? Is RT1 compact? § 15 ◦ 2 Terminology Remarks Originally the word compactness was used for the following weaker property: any countable open cover contains a ﬁnite subcover. Ω1 )? And vice versa? 15. 15.C. RT1 is compact.e. iﬀ there exists an open covering which contains no ﬁnite subcovering.2. The line R is not compact. S. A topological space X is not compact. Let Ω1 ⊂ Ω2 be topological structures in X. 15.§ 15 Compactness § 15 ◦ 1 Deﬁnition of Compactness This section is devoted to a topological property. For a second countable space the original deﬁnition of compactness is equivalent to the modern one. Any ﬁnite topological space and indiscrete space are compact. according to Bourbaki it is not. which plays a very special role in topology and its applications. a topological space is compact if every open cover has a ﬁnite subcover. Thus. The term bicompactness is sometimes used (mainly by topologists of Alexandroﬀ school).1. According to our deﬁnition. i. Does compactness of (X. The modern notion of compactness was introduced by P. 94 . This notion appeared to be so successful that it has displaced the original one and even took its name. S. If Γ is a cover of X and Σ ⊂ Γ is a cover of X then GS is called a subcover (or subcovering ) of Γ. Alexandroﬀ (1896–1982) and P. (It seems though that this analogy has never been formalized.
Let A be a compact subset of a Hausdorﬀ space X and b a point of X which does not belong to A. but not only . 15.I. 15. Does it follow that A ∪ B is compact? Does it follow that A ∩ B is compact? 1 15.H.K.4.G. A ⊂ U and U ∩ V = ∅. iﬀ any centered collection of its closed sets has nonempty intersection.J. 2) compact in the arrow? 15. A collection Σ of subsets of a set X is centered. A topological space is compact. V ⊂ X such that b ∈ V . 15. 15. 15. 2) ⊂ R compact? 15. Any closed subset of a compact space is compact. Is [1. Prove that any subset of RT1 is compact.§ 15. § 15 ◦ 4 Compact Sets By a compact set one means a subset of a topological space (the latter must be clear from the context) provided it is compact as a space with the topology induced from the ambient space.3.9. 15. Is the same set [1.K. Is compactness hereditary? 15. Then there exists open sets U. COMPACTNESS 95 § 15 ◦ 3 Compactness in Terms of Closed Sets A collection of subsets of a set is said to be centered if the intersection of any ﬁnite subcollection is not empty.8. iﬀ there exists no ﬁnite Σ1 ⊂ Σ such that the complements of the sets belonging to Σ1 cover X. Prove that the set A = {0} ∪ { n }∞ in R is compact.5. . Let A and B be compact subsets of a topological space X. 15. 15.L Lemma to 15. Find a necessary and suﬃcient condition (formulated not in topological terms) for a subset of the arrow to be compact? 15. iﬀ any cover which consists of sets open in X contains a ﬁnite subcover. n=1 § 15 ◦ 5 Compact Sets Versus Closed Sets 15.6. . A subset A of a topological space X is compact. Any compact subset of a Hausdorﬀ space is closed. What is the minimal number of points needed? .F. . Construct a nonclosed compact subset of some topological space.7.
For this. 15.S. 15. Recall that ndimensional cube is the set 15. Prove that a compact Hausdorﬀ space is normal.) Matrix (aij ) with 1 ≤ i ≤ n.12.R. (c) S 1 . § 15 ◦ 7 Compactness in Euclidean Space 15. Which of the following sets are compact: (a) (d) (g) [0.Q. 15.11. but noncompact set of a metric space. 1] for i = 1. (f) ellipsoid.7. A compact Hausdorﬀ space is regular. . 15. Are the metric spaces of Problem 4. Let {Kn } be a decreasing sequence of compact nonempty connected subset of a Hausdorﬀ space. 15. 15. (Cf.g. (Cf. 1] ∩ Q? ray R+ = {x ∈ R  x ≥ 0}. see 15.15. Section § 12.O. Prove that the intersection of any family of compact subsets of a Hausdorﬀ space is compact. The cube I n is compact. Let X be a Hausdorﬀ space.A compact? 15. a point of Rnk .M.Q. Construct a closed and bounded. 1). 15. Prove that there exists an open and closed set V such that K ⊂ V ⊂ U . (b) Sn.§ 15.10. Any compact subset of a metric space is bounded. Prove that U ⊃ ∩α∈A Kα for some ﬁnite A ⊂ Λ. iﬀ it is closed and bounded. any compact subset of a metric space is closed and bounded. This identiﬁes the set L(nk) of all matrices like that with Rnk and endows it with a topological structure. . Therefore. and let U be an open set containing ∩α∈Λ Kα . . A subset of a Euclidean space is compact. lexicographically) its elements by numbers from 1 till nk. COMPACTNESS 96 § 15 ◦ 6 Compactness and Separation Axioms 15.) 15.N. The interval I is compact. .16.K and 15. one needs to enumerate somehow (e. Let K be a connected component of a compact Hausdorﬀ space X and let U be an open set containing K.P. Prove that the intersection ∩∞ Kn is nonempty n=1 and connected. 15. I n = {x ∈ Rn  xi ∈ [0. 1 ≤ j ≤ k with real ai j can be considered as . onesheeted hyperboloid. n}. 15. let {Kα }α∈Λ be a family of its compact subsets. 15.13. (e) [0. Construct a decreasing sequence of connected subsets of the plane with nonconnected intersection.14.
T.U. Prove that if f : I → R is a continuous function then f (I) is an interval. Prove that for any open cover of a compact metric space there exists a number r > 0 such that any open ball of radius r is contained in some element of the cover.22. n)  det A = 0}.§ 15. if X is a compact space and f : X → Y is a continuous map then f (X) is compact.S. F is closed and G compact then ρ(F. Prove that V is pathconnected.. Prove that if F and G are disjoint subsets of a metric space. Then there exists a number δ > 0 such that for any set A ⊂ X with diameter diam(A) < δ the image f (A) is contained in some element of Γ. (d) {A ∈ L(n. COMPACTNESS 97 15.e. b ∈ X such that f (a) ≤ f (x) ≤ f (b) for any x ∈ X.18. (c) O(n) = {A ∈ L(n. Let f : X → Y be a continuous map of a compact metric space X to a topological space Y . . 15. A) < ε} for some ε > 0). G) = inf {ρ(x. n) :  A is an orthogonal matrix}. y)  x ∈ F. Which of the following subsets of L(n. 15. 15. Let A be a closed connected subset of Rn and let V be its closed εneighborhood (i.17. f (y)) < ρ(x. n)  A2 = E}. 15. A continuous image of a compact set is compact.T and 15.e. Prove that if in a compact metric space the closure of any open ball is the closed ball with the same center and radius then any ball of this space is connected. Let X be a compact metric space and f : X → X be a map such that ρ(f (x). (b) SL(n) = {A ∈ L(n. V = {x ∈ Rn  ρ(x.V Lebesgue Lemma. § 15 ◦ 9 Closed Maps A continuous map is said to be closed if the image of any closed set under this map is closed. On a compact set any continuous function is bounded and attains its maximal and minimal values. Prove that any open set containing a compact set A of a metric space X contains an εneighborhood of A.21..) Cf. the set {x ∈ X  ρ(x. y ∈ G} > 0. y ∈ X with x = y. (In other words.23. (i. n)  det A = 1}. if X is a compact space and f : X → R is a continuous function.) 15. (Recall that a ﬁxed point of f is a point x such that f (x) = x. (In other words. 15. 15. here E is the unit matrix? § 15 ◦ 8 Compactness and Continuous Maps 15. n) are compact: (a) GL(n) = {A ∈ L(n.20. then there exist a.19. A) < ε}).24. 15. Prove that f has a unique ﬁxed point. y) for any x. 15. and let Γ be an open cover of Y .) 15.
A continuous map of a compact space to a Hausdorﬀ space is closed. 15. 15. Does there exist a noncompact subspace of Euclidian space such that any its map to a Hausdorﬀ space is closed? (Cf. 15.W. 15.W. 15. 15. Here are two important corollaries of this theorem. deﬁne the same topological structure).27.31. Prove that any norm Rn → R (see Section § 4) is a continuous function (with respect to the standard topology of Rn ).§ 15.Y can be omitted without making the statement false. 15. COMPACTNESS 98 15. See 4. Does the same hold true for metrics in Rn ? . A continuous injection of a compact space to a Hausdorﬀ space is a topological embedding.Y.25. cf.e. Show that none of the hypothesis in 15. 4.) § 15 ◦ 10 Norms in Rn 15.X.29.28. A continuous bijection of a compact space to a Hausdorﬀ space is a homeomorphism.U and 15.27.26. Prove that any two norms in Rn are equivalent (i.
Ω∗ ) is compact.) A topological embedding of a space X into a compact space Y is called a compactiﬁcation of X if the image of X is dense in Y . Prove that the onepoint compactiﬁcation of Rn is homeomorphic to Sn. Which of the following spaces are locally compact: (a) R. Let X ∗ be the set obtained by adding a point to X (of course. 16:1. Ω∗ ) is Hausdorﬀ. Give explicit descriptions of onepoint compactiﬁcations of the following spaces: 99 . In this situation Y is also called a compactiﬁcation of X. Find two locally compact sets on the line such that their union is not locally compact. 16:H. Prove that the inclusion X → X ∗ is a topological embedding (with respect to the original topology of X and Ω∗ ). (b) Q. Prove that the onepoint compactiﬁcation of the plane is homeomorphic to S 2 . Prove that the space (X ∗ .e. where C ⊂ X is a compact set. Prove that if X is locally compact then the space (X ∗ . § 16 ◦ 2 OnePoint Compactiﬁcation Let X be a Hausdorﬀ topological space. 16:E. Prove that if X is a locally compact Hausdorﬀ space and Y is its compactiﬁcation with Y X consisting of a single point then there exists a homeomorphism Y → X ∗ which is the identity on X. (d) a discrete space? 16:2. iﬀ each of its points has a locally compact neighborhood. 16:D. (Recall that X is assumed to be Hausdorﬀ. Let Ω∗ be the collection of subsets of X ∗ consisting of • sets open in X and • sets of the form X ∗ C. 16:J. the point does not belong to X). The space Y of Problem 16:I is called a onepoint compactiﬁcation or Alexandroﬀ compactiﬁcation of X. 16:G. (c) Rn . A closed subset of a locally compact space is locally compact. 16:A. 16:3. 16:4. Local compactness is a local property.. Prove that Ω∗ is a topological structure. i. a space is locally compact. 16:B. Is local compactness hereditary? 16:C. 16:I. 16:F. An open subset of a locally compact Hausdorﬀ space is locally compact.§ 16 Local Compactness and Paracompactness § 16 ◦ 1 Local Compactness A topological space X is called locally compact if each of its points has a neighborhood with compact closure.
Y. 16:6. y) ∈ R2  1 < x2 + y 2 < 2}. 16:N. xy < 1}. 1]}. Problem 16:M is related to Theorem 15. if x ∈ X otherwise. a compact space. Let {Ui }i∈N be a locally ﬁnite open cover of Rn . 16:L. Y ∗ Y. 16:R. Then ∆ is said to be a reﬁnement of Γ if for each A ∈ ∆ there exists B ∈ Γ such that A ⊂ B. y) ∈ R2  x ∈ [0. y) ∈ R2  x. if x = X ∗ X. Prove that there exist an open cover {Vi }i∈N such that Cl Vi ⊂ Ui for each i ∈ N.§ 16. 16:7.X and 15. iﬀ f is proper. Any continuous map f : X → Y is naturally extended to a map X ∗ → Y ∗ deﬁned by the following formula: f ∗ (x) = f (x). 16:K. strip {(x. Let Γ and ∆ be covers of a set X. § 16 ◦ 4 Locally Finite Collections of Subsets A collection Γ of subsets of a space X is said to be locally ﬁnite if each point b ∈ X has a neighborhood U such that A ∩ U = ∅ for all but ﬁnite number of A ∈ Γ. If a collection Γ of subsets of a space X is locally ﬁnite then so is {Cl A  A ∈ Γ}..W. i. 1]. 16:O. Prove that f ∗ is continuous. y ∈ [−1. If a collection Γ of subsets of a space X is locally ﬁnite and Cl A is compact for each A ∈ Γ then each A ∈ Γ intersects only ﬁnite number of elements of Γ.e. Prove that a locally compact Hausdorﬀ space is regular. Any locally ﬁnite cover of a compact space is ﬁnite. Y be Hausdorﬀ spaces. . Any locally ﬁnite cover of a sequentially compact space is ﬁnite. 16:M. Prove that any proper map of a Hausdorﬀ space to a Hausdorﬀ locally compact space is closed. 16:5. § 16 ◦ 3 Proper Maps A continuous map f : X → Y is said to be proper if the preimage of any compact subset of Y is compact. square without vertices {(x. Prove that any open cover of Rn has a locally ﬁnite open reﬁnement. Find an example of an open cover of Rn which does not possess a locally ﬁnite subcover. LOCAL COMPACTNESS AND PARACOMPACTNESS 100 (a) (b) (c) (d) annulus {(x. Extend this analogy: formulate and prove statements corresponding to theorems 15. Let X. 16:P. 16:Q.
Then there exists a locally ﬁnite open cover ∆ such that {Cl V  V ∈ ∆} is a reﬁnement of Γ. . Then X can be embedded in Rk(n+1) . . . Information. where Ui comprise an open cover of a space X. 16:Y. A disjoint union of paracompact spaces is paracompact.§ 16. Any Hausdorﬀ paracompact space is regular. § 16 ◦ 7 Partitions of Unity For a function f : X → R. 16:10. A Hausdorﬀ space is paracompact. § 16 ◦ 6 Paracompactness and Separation Axioms 16:9. Rn is paracompact. the set Cl{x ∈ X  f (x) = 0} is called the support of f and denoted by supp f . 16:T. k. Any metrizable space is paracompact. Let {fα }α∈Λ be a family of continuous functions X → R such that the sets supp(fα ) comprise a locally ﬁnite cover of the space X. Prove that the relation f (x) = α∈Λ fα (x) deﬁnes a continuous function f : X → R. be embeddings. For every normal space X there exists a partition of unity which is subordinate to a given locally ﬁnite open cover of X. Any compact space is paracompact. 16:11. . iﬀ any its open cover admits a partition of unity which is subordinate to this cover. i=1 16:V. Let X be a normal space and Γ its locally ﬁnite open cover. LOCAL COMPACTNESS AND PARACOMPACTNESS 101 § 16 ◦ 5 Paracompact Spaces A space X is said to be paracompact if any its open cover has a locally ﬁnite open reﬁnement. 16:8. Let hi Ui → Rn . A partition of unity {fα } is said to be subordinate to a cover Γ if each supp(fα ) is contained in an element of Γ. Any closed subspace of a paracompact space is paracompact. . i = 1. § 16 ◦ 8 Application: Making Embeddings from Pieces 16:Z. A family of nonnegative functions fα X → R+ is called a partition of unity if the sets supp(fα ) comprise a locally ﬁnite cover of the space X and α∈Λ fα (x) = 1. 16:X. 16:W. 16:S. Any Hausdorﬀ paracompact space is normal. Let X = ∪∞ Xi and Xi are compact sets. 16:U. Then X is paracompact.
§ 16. LOCAL COMPACTNESS AND PARACOMPACTNESS
102
ˆ 16:Z.1. Show that the map x → (fi (x)hi (x)), where fi X → R comrise ˆ a partition of unity, which is subordinate to the given cover and hi (x) = n+1 , is an embedding. (hi (x), 1) ∈ R
§ 17 Sequential Compactness
§ 17 ◦ 1 Sequential Compactness Versus Compactness A topological space is said to be sequentially compact if every sequence of its points contains a convergent subsequence. 17.A. Any compact ﬁrst countable space is sequentially compact.
A point b is called an accumulation point of a set A if every neighborhood of b contains inﬁnitely many points of A. 17.A.1. Prove that in a space satisfying the ﬁrst separation axiom the notions of accumulation point and limit point coincide. 17.A.2. In a compact space any inﬁnite set has an accumulation point. 17.A.3. The space, in which any inﬁnite set has an accumulation point, is sequentially compact.
17.B. A sequentially compact second countable space is compact.
17.B.1. In a sequentially compact space a decreasing sequence of nonempty closed sets has a nonempty intersection. 17.B.2. Prove that in a topological space every decreasing sequence of nonempty closed sets has nonempty intersection, iﬀ any centered countable collection of closed sets has nonempty intersection.
17.C. For second countable spaces compactness and sequential compactness are equivalent. § 17 ◦ 2 In Metric Space A subset A of a metric space X is called an εnet (where ε is a positive number) if ρ(x, A) < ε for each point x ∈ X. 17.D. Prove that in any compact metric space for any ε > 0 there exists a ﬁnite εnet. 17.E. Prove that in any sequentially compact metric space for any ε > 0 there exists a ﬁnite εnet. 17.F. Prove that a subset of a metric space is everywhere dense, iﬀ it is an εnet for any ε > 0. 17.G. Any sequentially compact metric space is separable. 17.H. Any sequentially compact metric space is second countable. 17.I. For metric spaces compactness and sequential compactness are equivalent.
103
§ 17. SEQUENTIAL COMPACTNESS
104
17.1. Prove that a sequentially compact metric space is bounded. (Cf. 17.E and 17.I.) 17.2. Prove that in any metric space for any ε > 0 there exists (a) a discrete εnet and even (b) an εnet such that the distance between any two of its points is greater than ε. § 17 ◦ 3 Completeness and Compactness A sequence {xn }n∈N of points of a metric space is called a Cauchy sequence if for any ε > 0 there exists a number N such that ρ(xn , xm ), ε for any n, m > N . A metric space is said to be complete if each Cauchy sequence in it is convergent. 17:A. A Cauchy sequence, which contains a convergent subsequence, converges. 17:B. Prove that a metric space is complete, iﬀ any decreasing sequence of its closed balls with radii tending to 0 has nonempty intersection. 17:C. Prove that a compact metric space is complete? 17:D. Is any locally compact, but not compact metric space complete? 17:E. Prove that a complete metric space is compact, iﬀ for any ε > 0 it contains a ﬁnite εnet. 17:F. Prove that a complete metric space is compact iﬀ for any ε > 0 it contains a compact εnet.
§ 17 ◦ 4 NonCompact Balls in Inﬁnite Dimension
By l∞ denote the set of all bounded sequences of real numbers. This is a vector space with respect to the componentwise operations. There is a natural norm in it: x = sup{xn  : n ∈ N}. 17.3. Are closed balls of l∞ compact? What about spheres? 17.4. Is the set {x ∈ l∞ : xn  ≤ 2−n , n ∈ N} compact? 17.5. Prove that the set {x ∈ l∞ : xn  = 2−n , n ∈ N} is homeomorphic to the Cantor set K introduced in Section § 2. 17.6*. Does there exist an inﬁnitely dimensional normed space, in which closed balls are compact?
§ 17 ◦ 5 pAdic Numbers
Fix a prime integer p. By Zp denote the set of series of the form a0 + a1 p + · · · + an pn + . . . with 0 ≤ an < p, an ∈ N. For x, y ∈ Zp put ρ(x, y) = 0 if x = y and ρ(x, y) = p−m , if m is the smallest number such that the mth coeﬃcients in the series x and y diﬀer. 17.7. Prove that ρ is a metric in Zp . This metric space is called the space of integer padic numbers. There is an injection Z → Zp assigning to a0 + a1 p + · · · + an pn ∈ Z with 0 ≤ ak < p the series a0 + a1 p + · · · + an pn + 0pn+1 + 0pn+2 + · · · ∈ Zp
§ 17. SEQUENTIAL COMPACTNESS
105
and to −(a0 + a1 p + · · · + an pn ) ∈ Z with 0 ≤ ak < p the series b0 + b1 p + · · · + bn pn + (p − 1)pn+1 + (p − 1)pn+2 + . . . , where b0 + b1 p + · · · + bn pn = pn+1 − (a0 + a1 p + · · · + an pn ). Cf. 4:D. 17.8. Prove that the image of the injection Z → Zp is dense in Zp . 17.9. Is Zp a complete space? 17.10. Is Zp compact?
§ 17 ◦ 6 Induction on Compactness
A function f : X → R is locally bounded if for any point a ∈ X there exists a neighborhood U and a number M > 0 such that f (x) ≤ M for x ∈ U (i.e., each point has a neighborhood such that the restriction of f to this neighborhood is bounded). 17.11. Prove that if a space X is compact and a function f : X → R is locally bounded then f is bounded. This statement is one of the simplest applications of a general principle formulated below in 17.12. This principle may be called induction on compactness (cf. induction on connectedness discussed in Section § 11). Let X be a topological space, C a property of subsets of X. We say that C is additive if the union of any ﬁnite family of sets having C also has C. The space X is said to possess C locally if each point of X has a neighborhood with property C. 17.12. Prove that a compact space which possesses locally an additive property has this property itself. 17.13. Deduce from this principle the statements of problems 15.Q, 17:E, and 17:F.
§ 17 ◦ 7 Spaces of Convex Figures
Let D ⊂ R2 be a closed disc of radius p. Consider the set of all convex polygons P with the following properties: • the perimeter of P is at most p; • P is contained in D; • P has ≤ n vertices (the cases of one and two vertices are not excluded). See 4:E, cf. 4:G. 17.14. Equip this set with a natural topological structure. For instance, deﬁne a natural metric. 17.15. Prove that this space is compact. 17.16. Prove that there exists a polygon belonging to this set and having the maximal area. 17.17. Prove that this is a regular ngon.
§ 17. SEQUENTIAL COMPACTNESS
106
Consider now the set of all convex polygons of perimeter ≤ p contained in D. In other words, consider the union of the sets of ≤ ngons considered above. 17.18. Construct a topological structure in this set such that it induces the structures introduced above in the spaces of ≤ npolygons. 17.19. Prove that the space provided by the solution of Problem 17.18 is not compact. Consider now the set of all convex subsets of the plane of perimeter ≤ p contained in D. 17.20. Construct a topological structure in this set such that it induces the structure introduced above in the spaces of polygons. 17.21. Prove that the space provided by the solution of Problem 17.20 is compact. 17.22. Prove that there exists a convex plane set with perimeter ≤ p having a maximal area. 17.23. Prove that this is a disc of radius
p 2π .
17.24. Consider the set of all bounded subsets of a compact metric space. Prove that this set endowed with the Hausdorﬀ metric (see 4:F) is a compact space.
Problems for Tests Test. If X is: Has Y the same property. Fill Table 1 with pluses and minuses according to your answers to the corresponding questions. 107 . if: Y ⊂X Y is open subset of X Y is closed subset of X X is dense in Y Y is quotient space of X Y = X as sets.1. Let X be a topological space. ΩX ⊂ ΩY Y is open subset of Rn Y is antidiscrete nonnonsecond connected Hausdorﬀ Hausdorﬀ separable compact compact countable Table 1.
Give as many proves as you can for nonexistence of a homeomorphism between (a) S 1 and R1 . Test.3.2. (c) R and RT1 (d) R and R+ = {x ∈ R : x ≥ 0}. If X is: Has Y the same property. . (b) I and I 2 . Let X be a topological space. ΩX ⊃ ΩY Y is closed and bounded subset of Rn Y is discrete nonnonsecond connected Hausdorﬀ Hausdorﬀ separable compact compact countable Table 2. Fill Table 2 with pluses and minuses according to your answers to the corresponding questions.PROBLEMS FOR TESTS 108 Test. if: X =Y ×Z Y =X ×Z Y is open dense in X X is open dense in Y X is quotient space of Y Y = X as sets.
(A1 × B1 ) ∩ (A2 × B2 ) = (A1 ∩ A2 ) × (B1 ∩ B2 ). A2 ⊂ X and B1 . 109 .C. There are natural maps of X × Y onto X and Y deﬁned by formulas (x. Prove that (A × B) ∩ ∆ = ∅. −1 18. y) ∈ X × X : x = y}. 18.5. Prove that for any map f : X → Y and any set A ⊂ X. Write down the corresponding formula for B ⊂ Y . They are denoted by prX and prY and are called (natural) projections. To a map f : X → Y there corresponds a subset Γf of X × Y deﬁned by Γf = {(x. −1 f (A) = prY (Γf ∩ (A × Y )) = prY (Γf ∩ prX (A)) 18. Product of Sets Let X and Y be sets. x). A set Γ ⊂ X × Y is the graph of a map X → Y . y) → x and (x. y) → (y.4. Let T : X × Y → Y × X be the map deﬁned by (x. Prove that Γf −1 = T (Γf ) for any invertible map f : X → Y . 18. B2 ⊂ Y (A1 ∪ A2 ) × (B1 ∪ B2 ) = (A1 × B1 ) ∪ (A1 × B2 ) ∪ (A2 × B1 ) ∪ (A2 × B2 ). 18.A. Prove that the map prX Γf and f −1 (B) = prX (Γ ∩ (X × B)) for any B ⊂ Y . is bijective. 18. f (x)) : x ∈ X} and called the graph of f . The set of ordered pairs (x. Prove that f is injective.3. iﬀ prY is injective. Prove that for any A1 .B. y) → y.CHAPTER 3 Topological Constructions § 18 Multiplication § 18 ◦ 1 SetTheoretic Digression.2.6.1. Prove that prX (A) = A × Y for A ⊂ X. Γf 18. iﬀ A ∩ B = ∅ 18. y) with x ∈ X and y ∈ Y is called a direct product or Cartesian product or just product of X and Y and denoted by X ×Y . iﬀ for each a ∈ X the intersection Γ ∩ (a × Y ) contains exactly one point. If A ⊂ X and B ⊂ Y then A×B ⊂ X ×Y . 18. Let A and B be subsets of X and ∆ = {(x. Sets X × {b} with b ∈ Y and {a} × Y with a ∈ X are called ﬁbers of the product X × Y .
7.§ 18. 18.10. 18. Prove that if A is closed in X and B is closed in Y then A × B is closed in X × Y . then we say that U × V is an elementary set of X × Y . B and Fr B.J.D. Y × X is canonically homeomorphic to X × Y .13. 18. 18. 18. The set of elementary sets of X × Y is a base of a topological structure in X × Y . The natural projections prX : X × Y → X and prY : X × Y → Y are continuous. 18. MULTIPLICATION 110 § 18 ◦ 2 Product of Topologies Let X and Y be topological spaces.F. The topology of product is the coarsest topology with respect to which prX and prY are continuous. If U is an open set of X and B is an open set of Y . Is it true that Fr(A × B) = (Fr A × B) ∪ (A × Fr B)? 18. Prove that for any subspaces A and B of spaces X and Y the topology of the product A × B coincides with the topology induced from X × Y via the natural inclusion A × B ⊂ X × Y . 18. 18. Fr A. so that one may expect that it has additional pleasant properties. Prove that R1 × R1 = R2 .14.9. A ﬁber of a product is canonically homeomorphic to the corresponding factor.G.E.I.12. 18. The word canonically means here that a homeomorphism between X × Y and Y × X which exists according to the statement can be chosen in a nice special (or even obvious?) way. (I)n = I n (recall that I n is the ndimensional cube). 18. (R1 )n = Rn . Is it true that Int(A × B) = Int A × Int B? 18. Prove that Cl(A × B) = Cl A × Cl B for any A ⊂ X and B ⊂ Y . Prove that Fr(A × B) = (Fr A × B) ∪ (A × Fr B) for closed A and B? 18. . X × (Y × Z) is canonically homeomorphic to (X × Y ) × Z. Is it true that Fr(A × B) = Fr A × Fr B? 18.11. Find a formula for Fr(A × B) in terms of A. § 18 ◦ 3 Topological Properties of Projections and Fibers 18. The canonical homeomorphism is the restriction to the ﬁber of the natural projection of the product onto the factor.H.8. The product of topological spaces X and Y is the set X × Y with the topological structure deﬁned by the base consisting of elementary sets.
Prove that a topological space is Hausdorﬀ. f can be recovered from them as a sort of product. Prove that for any maps f1 : Z → X and f2 : Z → Y there exists a unique map f : Z → X × Y with prX ◦ f = f1 and prY ◦ f = f2 18. Let f R → R be a continuous function. x2 ) → (g1 (x1 ). Prove that its graph is: (a) closed. g2 (x2 )). Prove that a map f : X → Y is continuous. To a map f : Z → X × Y one assigns the compositions f1 = prX ◦ f : Z → X and f2 = prY ◦ f : Z → Y .25.28. 18. Prove that a metric ρ : X × X → R is continuous with respect to the topology deﬁned by the metric. (b) connected. § 18 ◦ 5 Properties of Diagonal and Other Graphs 18.20. Prove that f is continuous.§ 18.26.16. Prove that if Y is a Hausdorﬀ space and a map f : X → Y is continuous then the graph Γf is closed in X × Y .22.27. Y . Y .23. Let X.17. Prove that sets U × V with U ∈ ΣX and V ∈ ΣY constitute a base for X × Y . and Z be sets.15. iﬀ f1 and f2 are continuous. 18. For any maps g1 : X1 → Y1 and g2 : X2 → Y2 there is a map X1 × X2 → Y1 × Y2 deﬁned by formula (x1 . Prove that then f is continuous. These compositions are called factors of f .17 states that prX : X × Y → X is an open map.24. Is prX a closed map? 18. A map of a topological space X to a topological space Y is said to be open.27 the hypothesis on compactness is necessary.29. Prove that if W is open in X × Y then prX (W ) is open in X. Prove that the Cartesian product of open maps is open. Indeed. (c) path connected. 18. iﬀ prX Γf is a homeomorphism.19. Prove that for each topological space X and each compact topological space Y the map prX : X × Y → X is closed. . and Z be topological spaces. if the image of any open set under this map is open. 18. 18. Prove that the Cartesian product of continuous maps is continuous. § 18 ◦ 4 Cartesian Products of Maps Let X. 18. iﬀ the set ∆ = {(x. 18. 18. 18. (d) locally connected.21. Prove that in 18.18. Therefore 18. 18. MULTIPLICATION 111 18. Let ΣX and ΣY be bases of topological spaces X and Y . 18. Let Y be a compact space and Γf be closed. This map is called a (Cartesian) product of g1 and g2 and denoted by g1 × g2 . x) : x ∈ X} (which is called the diagonal of X × X) is closed.
Let Γf be closed.1 the subspace {(x. B ⊂ Y the set X × Y A × B is connected. if X × Y has the property then X also has? .36.1.Q. 18. 18. 18. MULTIPLICATION 112 (e) locally compact. (b) f is locally bounded.30.33.36. 18. Prove that the product of locally compact spaces is locally compact. For which of the topological properties studied above.§ 18. 18. y) : x = −y} is closed and discrete.K. Prove that if Γf is connected and locally compact then f is continuous. 18. The product of metrizable spaces is metrizable. 18.R.36. Prove that in the Cartesian square of the space introduced in 18. 18.36. 18. The product of compact spaces is compact.29 imply its continuity? 18.39.O. The product of separable spaces is separable.P.3.37. y) : x = −y} which have no disjoint neighborhoods in the Cartesian square of the space of 18. Then the following assertions are equivalent: (a) f is continuous. 18.31 – 18.2. The product of connected spaces is connected. The product of pathconnected spaces is pathconnected.35. 18. Does any of properties of the graph of a function mentioned in 18.40. (c) the graph Γf of f is connected.36. 18.L. Prove that for connected spaces X and Y and any proper subsets A ⊂ X. 18. First countability of factors implies ﬁrst countability of the product. 18.36. The product of normal spaces is not necessarily normal. 18. Prove that the product of regular spaces is regular. Prove that the set of real numbers with the topology deﬁned by the base which consists of all semiopen intervals [a.1. 18.32. Find two disjoint subsets of {(x. Prove that if Γf is connected and locally connected then f is continuous.33 true for mappings f : R2 → R? § 18 ◦ 6 Topological Properties of Products 18.N. 18. The product of second countable spaces is second countable. If X is a paracompact space and Y compact then X × Y is paracompact. b) is normal. 18. 18.M. The product of Hausdorﬀ spaces is Hausdorﬀ.34. Are some of assertions in problems 18.31. 18.38.
GL+ (n) = {A ∈ L(n. where 18. and S 2 × I into R3 . Construct a topological embedding of the torus to R3 . Prove that Rn homeomorphic to S k−1 × Dn−k+1 . The space S 1 × S 1 is called a torus.V. Prove that S n ∩ {x ∈ Rn+1 : x2 + · · · + x2 ≤ x2 + · · · + x2 } is 1 n+1 k k+1 18. The product S 1 ×· · ·×S 1 of k factors is called the kdimensional torus. . 18. 18. 18.T. Prove that SO(4) is homeomorphic to S 3 × SO(3). Prove that O(n) is homeomorphic to SO(n) × O(1). 18.46. Prove that GL(n) is homeomorphic to SL(n) × GL(1).41.43.U. S 1 × S 1 × I. Find topological embeddings of S 1 × D 2 . Prove that GL+ (n) is homeomorphic to SO(n) × R 18.§ 18. MULTIPLICATION 113 § 18 ◦ 7 Representation of Special Spaces as Products 18. Prove that R2 {0} is homeomorphic to S 1 × R. Rk is homeomorphic to S n−k−1 × Rk+1 . 18.45.44.42.S. Prove that the kdimensional torus can be topologically embedded into Rk+1 . 18. n(n+1) 2 . n) : det A > 0}.
Thus partitions of a set into nonempty subsets and equivalence relations in the set are essentially the same. 19. 19. since S and X/S have the same elements.. Hence. a relation. How is this operation related to division of numbers? Why is there a similarity in terminology and notations? At ﬁrst glance. but in our way of thinking of elements of partitions. there seems to be no need to introduce X/S . 114 . Prove that the saturation of a set A equals pr−1 pr(A) . according to this principle. we speak of a partition. If we think of them as atoms. More precisely. The smallest saturated set containing a subset A of X is called the saturation of A. A subset of X which is a union of elements of a partition is said to be saturated. 19. of their elements). The real sense of the notion of quotient set is not in its literal settheoretic meaning. getting rid of their possible internal structure then we speak on the quotient set. The set X/S is called also the set of equivalence classes for the equivalence relation corresponding to the partition S. iﬀ A = pr−1 (point) where pr : X → X/S is the natural projection.§ 19 Quotient Spaces § 19 ◦ 1 SetTheoretic Digression.2. which is reﬂexive. Enigma. Each partition of a set X gives rise to an equivalence relation (i. Indeed. the deﬁnition of quotient set contradicts one of the very profound principles of the set theory which states that a set is deﬁned by its elements. Vice versa. The set whose elements are members of the partition S (which are subsets of X) is called the quotient set or factor set of X by S and denoted by X/S . each equivalence relation in X gives rise to the partition of X to classes of equivalent elements. Let X be a set. and S be a partition.1.A. symmetric and transitive): two elements of X are said to be equivalent if they belong to the same element of the partition.e. If we remember that they are subsets of the original set and want to keep track of their internal structure (at least. Relations Partitions and Equivalence Recall that a partition of a set is its cover consisting of pairwise disjoint sets. Prove that A ⊂ X is an element of a partition S of X. they are two ways of describing the same phenomenon. The mapping X → X/S that maps x ∈ X to the element of S containing this point is called a (canonical) projection and denoted by pr. X/S = S.
R + is 19.5. R not Hausdorﬀ. 19. The set X/S with this topology is called the quotient space of the space X by partition S.F. The canonical projection pr : X → X/S is continuous. 1] 1 2 by the partition consisting of [0. A quotient space of a separable space is separable. What can you say about a partition S of a topological space X if the quotient space X/S is known to be discrete? 19. 19. 19. A quotient space of a connected space is connected.D. 3 ]. Give an example showing that second countability may get lost when we go over to a quotient space. 19.6. A subset of a quotient space X/S is closed. A quotient space of a compact space is compact. A quotient space of a pathconnected space is pathconnected.C. iﬀ it is equal to its saturation.J.H. 19. ( 1 . 19. Formulate similar necessary and suﬃcient conditions for a quotient space to satisfy other separation axioms and countability axioms. iﬀ any two elements of S possess disjoint saturated neighborhoods.B. ( 3 . QUOTIENT SPACES 115 19. 19. 2 ].E.M. . 19.L. A subset of a quotient space X/S is open. § 19 ◦ 3 Topological Properties of Quotient Spaces 19. 1].§ 19.G. The collection of these sets is a topological structure in the quotient set X/S . The quotient space of a topological space X by a partition S is Hausdorﬀ. Give an explicit description of the quotient space of the segment [0. 19. 19. iﬀ it is the image of a closed saturated set. This topological structure is called the quotient topology. 19. The quotient space of the real line by partition R + . 3 3 19. iﬀ it is the image of an open saturated set under the canonical projection pr. Prove that a set is saturated.I.3. Prove that the quotient topology is the ﬁnest topology in X/S such that the canonical projection pr is continuous with respect to it.K. iﬀ its preimage under pr is closed in X.4. § 19 ◦ 2 Quotient Topology A quotient set X/S of a topological space X with respect to a partition S into nonempty subsets is provided with a natural topology: a set U ⊂ X/S is said to be open in X/S if its preimage pr−1 (U) under the canonical projection pr : X → X/S is open.
This map is called injective factor (or injective quotient) of the map f .N. Let X. Prove that a map f : X → Y is constant on each element of a partition S of X iﬀ there exists a map g : X/S → Y such that the following diagram is commutative: X pr −− Y −→ g f X/S Prove that such a map g coincides with f /S . Y be topological spaces. Let f : X → Y be a map which is constant on each element of S. Prove that the formula f → f /S deﬁnes a bijection of the set of all continuous maps X → Y . A map f : X → Y deﬁnes a partition of the set X into nonempty preimages of the elements of Y . S be a partition of X into nonempty sets. This map is denoted by f /S.7. Then there is a map X/S → Y which assigns to each element A of S the element f (A). Let X. S be a partition of X into nonempty sets. Formulate and prove for f /S. gives rise to a map X/S → Y /T which assigns to an element A of partition S the element of partition T containing f (A). which maps each element of S into an element of T . which are constant on each element of the partition S. T and called the quotient map or factor map of f (with respect to S and T ). 19. . 19. Then the factor f /S of f is continuous. 19. 19. Y be topological spaces.O. onto the set of all continuous maps X/S → Y . T : X/S → Y /T is continuous.R. Quotients and Maps Let S be a partition of a set X into nonempty subsets. if S and T are partitions of sets X and Y then every map f : X → Y . Then the map f /S. Y be topological spaces. More generally.Q.N.§ 19. The map f /S(f ) : X/S(f ) → Y is injective. S and T partitions of X and Y .P. and f : X → Y a continuous map. § 19 ◦ 5 Continuity of Quotient Maps 19. which maps each element of S into an element of T . Let X. This map is denoted by f /S and called the quotient map or factor map of f (by partition S). which is constant on each element of S. QUOTIENT SPACES 116 § 19 ◦ 4 SetTheoretic Digression. This partition is denoted by S(f ). T a statement which generalizes 19. 19. and f : X → Y be a continuous map.
a A binary relation in a set X is called a preorder if it is transitive and reﬂective. although are not its consequences. 19:1. 19:D. 19:B. 19:C. Let be a transitive relation in a set X. The quotient space of a topological space satisfying the ﬁrst separation axiom with respect to a closed partition satisﬁes the ﬁrst separation axiom. . a a for any a. a for each a ∈ X). § 19 ◦ 7 Open Partitions A partition S of a topological space X is called open. if the saturation of each open set is open. then a c. if a b or a = b. 19:3. Prove that if a set A is saturated with respect to an open partition. A set X equipped with a preorder is called preordered. The quotient space of a ﬁrst countable space with respect to an open partition is ﬁrst countable. 19:F. Below we supply a formal justiﬁcation of this feeling by showing that the other conditions are natural companions of transitivity. is closed if this element is a closed set. then Int A and Cl A are also saturated. if the saturation of each closed set is closed. Denote by S × T the partition of X × Y consisting of A × B with A ∈ S and B ∈ T . If a b and b c. iﬀ the canonical projection X → X/S is a closed map. 19:2. which contains only one element consisting of more than one point. The quotient space of a second countable space with respect to an open partition is second countable. 19:4. • Reﬂexivity.§ 19. Prove that a partition is open. Prove that a partition is closed. If a preorder is antisymmetric then this is a nonstrict order. QUOTIENT SPACES 117 § 19 ◦ 6 Closed Partitions A partition S of a topological space X is called closed. i. iﬀ the canonical projection X → X/S is an open map. 19:A. e. 19:E. Then the relation a b. § 19 ◦ 8 SetTheoretic Digression: Splitting a transitive relation into equivalence and partial order In the deﬁnitions of equivalence and partial order relations the condition of transitivity seems to be the most important. deﬁned as follows is also transitive (and furthermore it is certainly reﬂexive. Prove that a partition. The quotient space of a normal topological space with respect to a closed partition is normal. Then the injective factor X × Y /S × T → X/S × Y /T of pr × pr X × Y → X/S × Y /T is a homeomorphism. that is it satisﬁes the following conditions: • Transitivity. Let S be an open partition of a topological space X and T be an open partition of a topological space Y .
By Theorem 9. In the set of all subsets of an arbitrary topological space the relation A is a preorder. B.§ 19. it is partitioned to indiscrete clusters of points. Let X be a ﬁnite topological space. The maximal T0 quotient of a ﬁnite space is a smallest neighborhood space (as a ﬁnite space). 19:K. homeomorphisms between spaces with poset topologies are monotone bijections. How this chain of constructions would degenerate. § 19 ◦ 9 Finite Topological Spaces The results of the preceding subsection provide a key to understanding of structure of ﬁnite topological spaces. By 19:L. This means that we can consider a ﬁnite topological space as its maximal T0 qoutient whose points are equipped with multiplicities. ) be a preordered set and ∼ be an equivalence relation deﬁned in X by according to 19:G. What equivalence relation is deﬁned in Z by the preorder ab? 19:H. QUOTIENT SPACES 118 19:5. Then a ∼ a. if a b and b a 19:7. The quotient space of Theorem 19:K is called the maximal T0 quotient of X. if a ∈ Cl{b}. Prove that any continuous map X → Y induces a continuous map of the maximal T0 quotient of X to the maximal T0 quotient of Y . it is symmetric. 19:L.P. Is the relation ab a preorder in the set Z of integers? 19:G. a b and b ∼ b imply a b and in this way deﬁnes a relation in the set of equivalence classes X/∼. This relation is a nonstrict partial order.N. a ∼ b. 19:I. This preorder deﬁnes the following equivalence relation: sets are equivalent iﬀ they have the same closure. which are natural numbers. if the original relation was (a) an equivalence relation. 19:M. By Theorem 13. 19:6. the numbers of points in the corresponding clusters of the original space. Thus any transitive relation generates an equivalence relation and a partial order in the set of equivalence classes. By Theorem 19:K. A continuous image of an indiscrete topological space is indiscrete. . Let (X. e. and by 19:M induce continuous maps between the maximal T0 quotient spaces. reﬂexive and transitive) in X. The quotient space satisﬁes the Kolmogorov separation axiom T0 . 19:J. The equivalence relation deﬁned by the preorder of Theorem 19:J deﬁnes the partition of the space to maximal (with respect to inclusion) indiscrete subspaces. its topology is deﬁned by a partial order. deﬁned by b. In any topological space the relation a is a preorder. or (b) nonstric partial order. continuous maps between ﬁnite spaces respect these clusters. If (X.. if A ⊂ Cl{B}. ) is a preordered set then the relation ∼ deﬁned by is an equivalence relation (i.
§ 19. QUOTIENT SPACES
119
Thus a ﬁnite topological space is characterised up to a homeomorphism by a ﬁnite poset whose elements are equipped with multiplicities (natural numbers). Two such spaces are homeomorphic, iﬀ there exists a monotone bijection between the corresponding posets which preserve the multiplicities. To recover the topological space from the poset with the multiplicities, one has to equip the poset with the poset topology and then replace each of its elements with an indiscrete cluster of points, the number points in which is the multiplicity of the element. § 19 ◦ 10 Simplicial schemes Let V be a set and Σ be a set of some of its subsets. A pair (V, Σ) is called a simplicial scheme with set of vertices V and set of simplices Σ, if • each subset of any element of Σ belongs to Σ, • the intersection of any collection of elements of Σ belongs to Σ, • each oneelement subset of V belongs to Σ. The set Σ is partially ordered by inclusion. When equipped the poset topology of this partial order, it is called the space of simplices of the simplicial scheme (X, Σ). Each simplicial scheme gives rise also to another topological space. Namely, for a simplicial scheme (V, Σ) consider the set S(V, Σ) of all functions c : V → I (= [0, 1]) such that Supp(c) = {v ∈ V : c(v) = 0} ∈ Σ and v∈V c(v) = 1. Equip S(V, Σ) with the topology generated by metric ρ(c1 , c2 ) = sup c1 (v) − c2 (v).
v∈V
Space S(V, Σ) is called simplicial or triangulated space. It is covered by sets {c ∈ S  Supp(c) = σ}, where σ ∈ Σ, which are called its (open) simplices. 19:8. Which open simplices of a simplicial space are open sets, which are closed, and which are neither closed nor open? 19:N. Find for each σ ∈ Σ a homeomorphism of the space onto an open simplex of dimension equal to the number of vertices belonging to σ minus one (recall that ndimensional open simplex is the set {(x1 , . . . , xn+1 ) ∈ Rn+1  xj > 0 for j = 1, . . . , n + 1 and
n+1 i=1
{c ∈ S  Supp(c) = σ} ⊂ S(V, Σ)
xi = 1, }).
19:O. Prove that for any simplicial scheme (V, Σ) the quotient space of the simplicial space S(V, Σ) by its partition to open simplices is homeomorphic to the space Σ of simplices of the simplicial scheme (V, Σ). § 19 ◦ 11 Baricentric Subdivision of a Poset 19:P. Find a poset, which is not isomorphic to the set of simplices (ordered by inclusion) of whatever simplicial scheme. Let (X, ) be a poset. Consider the set of all nonempty ﬁnite stricly increasing sequences a1 a2 · · · an of elements of X. It can be described also as the set of all nonempty ﬁnite subsets of X in each of which deﬁnes a linear order. Denote this set by X . It is naturally ordered by inclusion. Poset (X , ⊂) is called the baricentric subdivision of (X, ).
§ 19. QUOTIENT SPACES
120
19:Q. For any poset (X, ), pair (X, X ) is a simplicial scheme. There is a natural map X → X which maps an element of X (i. e., nonempty ﬁnite linearly ordered subset of X) to its greatest element. 19:R. Is this map monotone? Strictly monotone? The same questions concerning a similar map which maps a nonempty ﬁnite linearly ordered subset of X to its smallest element. Let (V, Σ) be a simplicial scheme and Σ be the baricentric subdivision of Σ (ordered by inclusion). Simplicial scheme (Σ, Σ ) is called the baricentric subdivision of the simplicial scheme (V, Σ) There is a natural mapping Σ → S(V, Σ) which maps a simplex σ ∈ Σ (i. e., a 1 subset {v0 , v1 , . . . , vn } of V ) to function bσ : V → R with bσ (vi ) = n+1 and bσ (v) = 0 for any v ∈ σ. Deﬁne a map β : S(Σ, Σ ) → S(V, Σ), which maps a function ϕ : Σ → R to function V →R:v→ ϕ(σ)bσ (v).
σ∈Σ
19:S. Prove that the map β : S(Σ, Σ ) → S(V, Σ) is a homeomorphism and constitutes, together with projections S(V, Σ) → Σ and S(Σ, Σ ) → Σ and the natural map Σ → Σ a commutative diagram S(Σ, Σ ) − − → S(V, Σ) −− Σ −−→ −− Σ
β
§ 20 Zoo of Quotient Spaces
§ 20 ◦ 1 Tool for Identifying a Quotient Space with a Known Space 20.A. If f : X → Y is a continuous map of a compact space X onto a Hausdorﬀ space Y then the injective factor f /S(f ) : X/S(f ) → Y is a
homeomorphism.
20.B. The injective factor of a continuous map of a compact space to a Hausdorﬀ one is a topological embedding.
20.1. Describe explicitly partitions of a segment such that the corresponding quotient spaces are all the connected letters of the alphabet. 20.2. Prove that there exists a partition of a segment I with the quotient space homeomorphic to square I × I.
§ 20 ◦ 2 Tools for Describing Partitions Usually an accurate literal description of a partition is cumbersome, but can be shortened and made more understandable. Of course, this requires a more ﬂexible vocabulary with lots of words with almost the same meanings. For instance, the words factorize and pass to a quotient can be replaced by attach, glue, identify, contract, and other words accompanying these ones in everyday life. Some elements of this language are easy to formalize. For instance, factorization of a space X with respect to a partition consisting of a set A and onepoint subsets of the complement of A is called a contraction (of the subset A to a point), and the result is denoted by X/A.
20.3. Let A, B ⊂ X form a fundamental cover of a topological space X. Prove that the quotient map A/A ∩ B → X/B of the inclusion A → X is a homeomorphism.
If A and B are disjoint subspaces of a space X, and f : A → B is a homeomorphism then passing to the quotient of the space X by the partition into onepoint subsets of the set X (A ∪ B) and twopoint sets {x, f (x)}, where x ∈ A, is called gluing or identifying (of sets A and B by homeomorphism f ). Rather convenient and ﬂexible way for describing partitions is to describe the corresponding equivalence relations. The main advantage of this approach is that, due to transitivity, it suﬃces to specify only some pairs of equivalent elements: if one states that x ∼ y and y ∼ z then it is not needed to state x ∼ z, since this follows.
Hence, a partition is represented by a list of statements of the form x ∼ y, which are suﬃcient to recover the equivalence relation. By such
121
§ 20. ZOO OF QUOTIENT SPACES
122
a list enclosed into square brackets, we denote the corresponding partition. For example, the quotient of a space X obtained by identifying subsets A and B by a homeomorphism f : A → B is denoted by X/[a ∼ f (a) for any a ∈ A] or just X/[a ∼ f (a)]. Some partitions are easy to describe by a picture, especially if the original space can be embedded into plane. In such a case, as in the pictures below, one draws arrows on segments to be identiﬁed to show directions which are to be identiﬁed. Below we introduce all these kinds of descriptions for partitions and give examples of their usage, providing simultaneously literal descriptions. The latter are not nice, but they may help to keep the reader conﬁdent about the meaning of the new words and, on the other hand, appreciating the improvement the new words bring in. § 20 ◦ 3 Entrance to the Zoo 20.C. Prove that I/[0 ∼ 1] is homeomorphic to S 1 . In other words, the quotient space of segment I by the partition consisting of {0, 1} and {a} with a ∈ (0, 1) is homeomorphic to a circle.
20.C.1. Find a surjective continuous map I → S 1 such that the corresponding partition into preimages of points consists of onepoint subsets of the interior of the segment and the pair of boundary points of the segment.
20.D. Prove that D n /S n−1 is homeomorphic to S n . In 20.D we deal with the quotient space of ball D n by the partition into S n−1 and onepoint subsets of its interior. Reformulation of 20.D: Contracting the boundary of an ndimensional ball to a point gives rise to an ndimensional sphere.
20.D.1. Find a continuous map of ball D n to the sphere S n that maps the boundary of the ball to a single point, and maps the interior of the ball bijectively onto the complement of this point.
20.E. Prove that I 2 /[(0, t) ∼ (1, t) for t ∈I] is homeomorphic to S 1 × I. Here the partition consisits of pairs of points {(0, t), (1, t)} where t ∈ I, and onepoint subsets of (0, 1) × I. Reformulation of 20.E: If we glue the side edges of a square identifying points on the same hight, we get a cylinder.
§ 20. ZOO OF QUOTIENT SPACES
123
20.F. Let X and Y be topological spaces, S a partition of X. Denote by T the partition of X × Y into sets A × y with A ∈ S, y ∈ Y . Then the natural bijection X/S × Y → X × Y /T is a homeomorphism. 20.G. Enigma. How are the problems 20.C, 20.E and 20.F related? 20.H. S 1 × I/[(z, 0) ∼ (z, 1) for z ∈ S 1 ] is homeomorphic to S 1 × S 1 . Here the partition consists of onepoint subsets of S 1 × (0, 1), and pairs of points of the basis circles lying on the same generatrix of the cylinder. Reformulation of 20.H: If we glue the basis circles of a cylinder identifying points on the same generatrix, then we get a torus. 20.I. I 2 /[(0, t) ∼ (1, t), (t, 0) ∼ (t, 1)] is homeomorphic to S 1 × S 1 . In 20.I the partition consists of • onepoint subsets of the interior (0, 1) × (0, 1) of the square, • pairs of points on the vertical sides, which are the same distance from the bottom side (i.e., pairs {(0, t), (1, t)} with t ∈ (0, 1)), • pairs of points on the horizontal sides which lie on the same vertical line (i.e., pairs {(t, 0), (t, 1)} with t ∈ (0, 1)), • the four vertices of the square Reformulation of 20.I: Identifying the sides of a square according to the picture , we get a torus .
§ 20 ◦ 4 Transitivity of Factorization A solution of Problem 20.I can be based on Problems 20.E and 20.H and the following general theorem. 20.J Transitivity of Factorization. Let S be a partition of a space X, and let S be a partition of the space X/S . Then the quotient space (X/S )/S is canonically homeomorphic to X/T , where T is the partition of the space X into preimages of elements of the partition S under projection X → X/S .
M¨bius strip or M¨bius band is I 2 /[(0. i. +∞) is homeomorphic to Int D2 ∪ {(0. R2 /B . Prove that if f : X → Y is a homeomorphism then the quotient spaces X/A and Y /f (A) are homeomorphic. (c) 20. Figure 1. and [0. where A is a union of several segments with a common end point. 1} is 3 3 3 20. See Figure 1. 1]/[ 1 . Prove that the M¨bius strip is homeomorphic to the surface swept o 3 in R by an interval. . Prove that [0. 1)}. which rotates in a halfplane around the middle point while the halfplane rotates around its boundary line. ZOO OF QUOTIENT SPACES 124 § 20 ◦ 5 M¨bius Strip o this is the quotient space of square I 2 by the partition into pairs of points symmetric with respect to the center of the square and lying on the vertical edges and onepoint set which do not lie on the vertical edges. In other words. 20. 1].. (e) (f) R2 /A.4. In such that the initial point of Ii+1 coincides with the ﬁnal point of Ii . where B is a simple ﬁnite polygonal line. Prove that R2 /[0.6. 1]/{ 1 . The ratio of the angular velocities of these rotations is such that rotation of the halfplane by 360◦ takes the same time as rotation of the interval by 180◦. . (b) R2 / I . a union of a ﬁnite sequence of segments I1 .e. Figuratively speaking. .K. the M¨bius strip is obtained by identifying o the vertical sides of a square in such a way that the directions shown on them by arrows are superimposed . § 20 ◦ 6 Contracting Subsets homeomorphic to letter P.5.7. .§ 20. o o 20. . t) ∼ (1. 1 − t)]. Prove that the following spaces are homeomorphic: (a) R2 . 20. 2 ] is homeomorphic to [0.
y) ∼ (−y. 0) ∼ (t. Prove that S 1 /[z ∼ e2πi/3 z] is homeomorphic to S 1 .8. 20. (b) the M¨bius strip.9 with Dn substituted for D2 .12. 1 − t) symmetric with respect to the center of the square which lie on the vertical edges.15.K and 18. 1 − t)]. In other words. .§ 20.9. Represent the M¨bius strip as a quotient space of cylinder S 1 × I. • pairs of points (t. 0). 20.8 the partition consists of triples of points which are vertices of equilateral inscribed triangles.) ¯ 20. (t. 20. ZOO OF QUOTIENT SPACES 125 § 20 ◦ 7 Further Examples 20. (b) D2 /[(x. t) ∼ (1. Prove that S 1 × S 1 /[(z. t). Prove that the following quotient spaces of disk D2 are homeomorphic to D2 : (a) D2 /[(x. w)] is homeomorphic to the Klein ¯ bottle.14. Describe explicitly the quotient space of line R1 by equivalence relation x ∼ y ⇔ x − y ∈ Z. • pairs of points (0. −y)].13. (0. (c) D2 /[(x. and • the quadruple of vertices. 20. 1) on horizontal edges which lie on the same vertical line.T). y) ∼ (−x. 20.10.11. w) ∼ (−z. (Here w denotes the complex number conjugate to w. Present the Klein bottle as a quotient space of (a) a cylinder. Find a generalization of 20. 20. o § 20 ◦ 8 Klein Bottle this is the quotient space of square I 2 by the partition into • onepoint subsets of its interior. y) ∼ (x. 20. 1). x)]. −y)]. Embed the Klein bottle into R4 so that the image of this embedding under the orthogonal projection R4 → R3 would look as follows: Klein bottle is I 2 /[(t. o 20. In 20.16. (1. Embed the Klein bottle into R4 (cf.
The result is called the projective plane. The sum is denoted by α∈A Xα . 20.§ 20. Y )  y ∈ Y }.L has not appeared yet. Sums of Sets A sum of a family of sets {Xα }α∈A is the set of pairs (xα . Let us consider the ﬁrst operation in details. 20. The map of Xβ (β ∈ A) to α∈A Xα deﬁned by formula x → (x.e. It is a combination of two operations: ﬁrst we must make a single space consisting of disjoint copies of the original spaces. is a topological structure. i. factorize the disk by the partition consisting of onepoint subsets of the interior of the disk and pairs of points on the boundary circle symmetric with respect to the center of the disk.M. Instead.L. The sum α∈A Xα with this topology is called the (disjoint) sum of topological spaces Xα . Thus we are not able to draw it. § 20 ◦ 11 SetTheoretic Digression. § 20 ◦ 12 Sums of Spaces 20. Indeed. we can avoid indices and deﬁne the sum by setting X Y = {(x. ZOO OF QUOTIENT SPACES 126 § 20 ◦ 9 Projective Plane Let us identify each boundary point of the disk D 2 with the antipodal point. The maps inβ : Xβ → α∈A Xα are topological embedding..17. o § 20 ◦ 10 You May Have Been Provoked to Perform an Illegal Operation Solving the previous problem you did something which does not ﬁt into the theory presented above. This space cannot be embedded into R3 . (α ∈ A). β) is an injection and denoted by inβ . and then we factorize this space identifying points of one copy with points of another. the operation with two spaces called gluing in 20. α) such that xα ∈ Xα . and their α∈A images are both open and closed in Xα . X)  x ∈ X} ∪ {(y. If {Xα }α∈A is a collection of topological spaces then the collection of subsets of α∈A Xα whose preimages under all inclusions inα (α ∈ A) are open. A projective plane is the result of gluing of a disk and the M¨bius o strip by homeomorphism between boundary circle of the disk and boundary circle of the M¨bius strip. we present it in other way. If only sets X and Y are involved and they are distinct. too. 20. .N. Topology described in 20.M is the ﬁnest topology with respect to which all inclusions inα are continuous.
22. g : A → X continuous maps.P. Prove that the Klein bottle can be obtained as a result of gluing two copies of the M¨bius strip by the identity map of the boundary circle. Prove that the result of gluing two copies of a cylinder by the identity map of the boundary circles (of one copy to the boundary circles of the other) is homeomorphic to S 1 × S 1 . Prove that the composition of inclusion X → X X Y → X ∪f Y is a topological embedding. Classify up to homeomorphism topological spaces. and f : A → X a continuous map. Y and projection A) 20. which can be obtained from a square by identifying a pair of opposite sides by a homeomorphism. 1} by a homeomorphism. The latter is called the attaching map. . and f. 20.24.O. 20.23. 20.§ 20. Prove that the result of gluing two copies of solid torus S 1 × D2 by the identity map of the boundary torus S 1 × S 1 is homeomorphic to S 1 × S 2 . 20. Prove that D n ∪h D n is homeomorphic to S n for any homeomorphism h : S n−1 → S n−1 . Prove that attaching a ball D n to its copy by the identity map of the boundary sphere S n−1 gives rise to a space homeomorphic to S n . Y be topological spaces. 20. A a subset of Y . Prove that if X is a point then X ∪f Y is Y /A.21. A a subset of Y . ZOO OF QUOTIENT SPACES 127 20. and sets in1 (x) ∪ in2 f −1 (x) with x ∈ f (A). Y be topological spaces. and is said to be the result of attaching or gluing the space Y to the space X by f . Let X. Obtain the Klein bottle by gluing two copies of the cylinder S 1 × I to each other. Prove that if there exists a homeomorphism h : X → X such that h ◦ f = g then X ∪f Y and X ∪g Y are homeomorphic. 20. 20.25. x) of the boundary torus to its copy is homeomorphic to S 3 . Prove that the result of gluing two copies of solid torus S 1 × D2 by the map S 1 × S 1 → S 1 × S 1 : (x.19.Q.27. The quotient space (X Y )/[a ∼ f (a) for a ∈ A] is denoted by X ∪f Y . Here the partition of X Y consists of onepoint subsets of in2 (Y and in1 (X f (A)).26. Which topological properties are inherited from summands Xα by the sum α∈A Xα ? Which are not? § 20 ◦ 13 Attaching Space Let X. 20.20.18. y) → (y. o 20. 20. Classify up to homeomorphism the spaces which can be obtained from two copies of S 1 × I by identifying of the copies of S 1 × {0. 20.
U. 20. orientable connected closed surface of genus p.R. 20. or.e. It is called pantaloons.31. 20.T.29. The result of attaching p copies of a handle to a sphere with p holes by embeddings of the boundary circles of handles onto the boundary circles of the holes (the boundaries of the holes) is called a sphere with p handles. 20.W. Prove that a sphere with q crosscaps is welldeﬁned up to homeomorphism (i. and spheres with crosscaps are called basic surfaces. or nonorientable connected closed surface of genus q. which can be obtained from S 1 × I by identifying S 1 × 0 with S 1 × 1 by a homeomorphism. Sometimes this word denotes also a sphere with more handles.. A sphere with two crosscaps is homeomorphic to the Klein bottle. A sphere with two holes is homeomorphic to cylinder S 1 × I. A sphere with a hole is homeomorphic to disk D 2 . ZOO OF QUOTIENT SPACES 128 20. A sphere with one handle is homeomorphic to torus S 1 × S 1 . § 20 ◦ 14 Basic Surfaces A torus S 1 × S 1 with the interior of an embedded disk deleted is called a handle.V. The space obtained from a sphere with q holes by attaching q copies of the M¨bius strip by embeddings of the boundary circles of the M¨bius o o strips onto the boundary circles of the holes (the boundaries of the holes) is called a sphere with q crosscaps. more ceremonial (and less understandable.. A sphere. Prove that a sphere with p handles is welldeﬁned up to homeomorphism (i. Prove that the topological type of the space resulting in gluing two copies of the M¨bius strip by a homeomorphism of the boundary circle does o not depend on the homeomorphism.S. . A sphere with one crosscap is homeomorphic to the projective plane. Classify up to homeomorphism topological spaces.e. the topological type of the result of gluing does not depend on the attaching embeddings). spheres with handles. 20.30. A sphere with two handles is homeomorphic to the result of gluing two copies of a handle by the identity map of the boundary circle. 20.28. 20. 20. A sphere with three holes has a special name. for a while). 20. A twodimensional sphere with the interior of n disjoint embedded disks deleted is called a sphere with n holes. the topological type of the result of gluing does not depend on the attaching embeddings).§ 20. A sphere with two handles is called a pretzel.
X.32.§ 20. 20. Prove that a sphere with p handles and q crosscaps is homeomorphic to a sphere with 2p + q crosscaps (here q > 0). Classify up to homeomorphisms topological spaces. which can be obtained by attaching to a sphere with 2p holes p copies of S 1 ×I by embeddings of the boundary circles of the cylinders onto the boundary circles of the sphere with holes. ZOO OF QUOTIENT SPACES 129 20. .
Therefore the projective space RP n can be considered as the result of extending of Rn by adjoining “nonproper” or “inﬁnite” points. λx1 . . . 21. . but are not deﬁned by this point: proportional vectors of coordinates (x0 . x1 . where i is the embedding from Problem 21.C. which constitute a projective space RP n−1 . . RP 0 is a point. 21. . and denoted by RP n . 21.H. 21. xn is denoted by (x0 : x1 : · · · : xn ). These numbers are called homogeneous coordinates of the corresponding point of RP n . The quotient spaces described here are of too great importance to consider them just as examples of quotient spaces. Prove that this is really a metric. . xn ) and (λx0 . The space RP n is canonically homeomorphic to the quotient space of Rn+1 {0} by the partition into onedimensional vector subspaces of Rn+1 punctured at 0. 2 21. .G. . . . . .E. The space RP n is homeomorphic to the quotient space of the ball D n by the partition into onepoint subsets of the interior of D n . π ]).G. The space RP n is canonically homeomorphic to the metric space. Prove that the map is a topological embedding.§ 21 Projective Spaces This section can be considered as a continuation of the previous one. xn ) → (1 : x1 : · · · : xn ) . The point with homogeneous coordinates x0 .D. 21. Homogeneous coordinates deﬁne a point of RP n . § 21 ◦ 1 Real Projective Space of Dimension n This space is deﬁned as the quotient space of the sphere S n by the partition into pairs of antipodal points. x1 . . The space RP 2 is homeomorphic to the projective plane deﬁned in the previous section. The space RP 1 is homeomorphic to the circle S 1 . . whose points are lines of Rn+1 passing through the origin 0 = (0. 21.F. What is its image? What is the inverse map of its image onto Rn ? 21.B. . .A. 0) and the metric is deﬁned as the angle between lines (which takes values in [0. Construct a topological embedding RP n−1 → RP n such that its image is RP n i(Rn ). . 130 i : Rn → RP n : (x1 . λxn ) deﬁne the same point of RP n . . . A point of the space Rn+1 {0} is a sequence of real numbers which are not all zeros. . and pairs of antipodal point of the boundary sphere S n−1 .
. whose points are the (complex) lines of the space Cn+1 passing through the origin 0 and the metric is deﬁned to be the angle between lines (which takes values in [0. 21:A. kj = −i ik = −j. j = (0. y3 . i = (0.Hamilton in 1843. 1 is really a unity: (1. The complex projective line CP 1 is homeomorphic to S 2 . . x1 . The rest of multiplication table looks as follows: ij = k. y4 ) = (x1 y1 − x2 y2 − x3 y3 − x4 y4 . ki = j. 1. 21:B. It is bilinear and to describe it in a shorter way it suﬃces to specify the products of the basis vectors. 0) k = (0. following Hamilton. .2. CP n can be viewed as the space of complexproportional nonzero complex sequences (x0 . It can be deﬁned by the formula (x1 . 21. The space CP n is canonically homeomorphic to the quotient space of the space Cn+1 {0} by the partition into complex lines of Cn+1 punctured at 0. x1 y2 + x2 y1 + x3 y4 − x4 y3 . jk = i. 0. 0. Introduce a natural topological structure in the set of all lines on the plane and prove that the resulting space is homeomorphic to a) RP 2 { point }. Prove that the set of all rotations of the space R3 around lines passing through the origin equipped with the natural topology is homeomorphic to RP 3 . 2 § 21 ◦ 3 Quaternionic Projective Spaces Recall that in R4 there is a remarkable multiplication. xn ). π ]). 0. The latter are traditionally denoted in this case. 0. x4 ) × (y1 . 1. ji = −k. 21:C. 0). CP 0 is a point. § 21 ◦ 2 Complex Projective Space of Dimension n This space is deﬁned as the quotient space of unit sphere S 2n+1 of the space Cn+1 by the partition into circles which cut by (complex) lines of Cn+1 passing through the point 0. The space CP 1 is called a complex projective line. . e. 1). b) open M¨bius strip (i. 0). x1 . Notation (x0 : x1 : · · · : xn ) and term homogeneous coordinates introduced for the real case are used in the same way for the complex case. 0. which was discovered by R. In this notation. 0.W. 21:D. x1 y4 + x2 y3 − x3 y2 + x4 y1 ).1. 0. x1 y3 − x2 y4 + x3 y1 + x4 y2 . x3 . 0) × x = x for any x ∈ R4 . PROJECTIVE SPACES 131 21. CP n is homeomorphic to the quotient space of the unit ball D2n of the space Cn by the partition whose elements are onepoint subsets of the interior of D2n and circles cut on the boundary sphere S 2n−1 by (complex) lines of the space Cn passing through the origin 0 ∈ Cn . The space CP n is canonically homeomorphic to the metric space. . y2 . 21:E. M¨bius strip with the boundary circle o o removed). 0.§ 21. It is denoted by CP n . Hence. as follows: 1 = (1.
HP n is homeomorphic to the quotient space of the unit closed ball D4n of the space Hn by the partition into points of the interior of D4n and 3sphere which are intersections of the boundary sphere S 4n−1 with (left quaternionic) lines of Hn . . It is deﬁned by the formula (x1 . x4 ) → (x1 . HP 0 consists of a single point. the (left) quaternionic projective space of dimension n. there are right quaternionic lines. 21:I. 0. . −x2 . e. 21:K. ξan )  ξ ∈ H}. The latter property allows one to deﬁne. 21:N. . Are the right and left quaternionic projective space of the same dimension homeomorphic? The left quaternionic projective space of dimension n is denoted by HP n . : xn ) and term homogenious coordinates introduced above in the real case are used in the same way in the quaternionic situation. the product of any quaternion a by the conjugate quaternion a equals (a2 . quaternions are very similar to complex numbers. e. 21:M. Hence. PROJECTIVE SPACES 132 Together with coordinatewise addition. . the inverse quaternion such that aa−1 = 1. 21:O. but with left quaternionic lines. subsets {(a1 ξ. . as the conjugation of complex numbers. ab = ba for any quaternions a and b.§ 21. for any a ∈ R4 . −x3 . Space HP 1 is called the quaternionic projective line. and similar left quaternionic lines {(ξa1 . g. 21:L. an ξ)  ξ ∈ H}. i. . this multiplication deﬁnes a structure of algebra in R4 . Prove that two right quaternionic lines in Hn either meet only at 0. Each of them is a real 4dimensional subspace of Hn = R4n . aa = a2 . Find a right quaternionic line. by a bar: x → x. Similarly. Its elements are called quaternions. . . As in C. The quotient space of the unit sphere S 4n+3 of the space Hn+1 = R4n+4 by the partition into its intersections with right quaternionic lines is called the (right) quaternionic projective space of dimension n. In the space Hn = R4n . It is denoted. or coincide. . −x4 ) and has the following two remarkable properties: 21:G. . 21:J. . ij = k = −k = ji). who discovered it. that is not a left quaternionic line. Quaternionic projective line HP 1 is homeomorphic to S 4 . 21:H. x3 . . It is denoted by H after Hamilton. a−1 = a−2 a Hence. It is not commutative (e. 0). i. . x2 . Check that the quaternion multiplication is associative. the quaternion algebra is a division algebra or a skew ﬁeld. Notation (x0 : x1 : .. Otherwise. HP n is canonically homeomorphic to the quotient space of Hn+1 {0} by the partition to left quaternionic lines of Hn+1 passing through the origin and punctured at it. . . 0. there is a transformation called conjugation acting in the set of quaternions. 21:F. HP n can be presented as the space of classes of left proportional (in the quaternionic sense) nonzero sequences (x0 . xn ) of quaternions.
PROJECTIVE SPACES 133 21:P. HP n is canonically homeomorphic to the set of (left quaternionic) lines of Hn+1 equipped with the topology induced by the angular metric (which takes values in [0.§ 21. π ]). 2 .
unless the contrary is speciﬁed explicitly. the cardinality card C(X. is homeomorphic Y . Let X be a set of n points with discrete topology. Prove that there exists a bijection between C(X. 22:2. 22:7. × Y (n times). 22:1. {0}. Denote by ∆(pw) the set {W (a. Y ) is greater than or equal to card Y . Prove that the topology of pointwise convergence and compactopen topology of C(X. Enigma. B ⊂ Y . 22:6. Y ). Y ) of C(X. Let X be a discrete space of n points. ∆(pw) is a subbase of a topological structure on C(X. Y ) the space of all continuous mappings X → Y with the compactopen topology. Prove that C(X. I) is not homeomorphic to C (pw) (I. 22:5. which. . U )  C ⊂ X is compact. Y ). Y ) can be identiﬁed with Y × . 22:C CompactOpen Versus Pointwise. Y ). The compactopen topology is ﬁner than the topology of pointwise convergence. Find natural conditions implying C(X. Y ) and the topological structure of X. Y ). Let Y be a set of k points with discrete topology. 22:3. Y }. Hereafter we denote by C(X. Y ) = Y . Denote by Const(X. Y be topological spaces. 1} equipped with topology {∅. Prove that there exists an injection Y → C(X. § 22 ◦ 2 Topologies on Set of Continuous Mappings {f ∈ C(X. The topological structure generated by ∆(pw) is called the topology of pointwise convergency. The topological structure deﬁned by ∆(co) is called the compactopen topology. Prove that C (pw) (X. Prove that C(X. B) the set U is open in Y } . Prove that C(I. U is open in Y } 22:A. In other words. 22:9. Y ) we denote the set of all continuous mappings of a topologival space X to a topological space Y . Y ) consists of a single element iﬀ so does Y . Y ) the set of all constant mappings f : X → Y . Is this true for C(X. U )  a ∈ X. Y ) equipped with this structure is denoted by C (pw) (X. I).§ 22 Spaces of Continuous Maps § 22 ◦ 1 Sets of Continuous Mappings By C(X. Y ). 22:4. . . ∆(co) is a subbase of a topological structures on C(X. Y ) is homeomorphic Y × . 22:8. with this topology. Y ) induce the same topological structure on Const(X. The set C(X. 22:B. A ⊂ X. Y ) contain k 2 elements. Find necessary and suﬃcient condition for the set C(X. . Denote by W (A. × Y (n times). Y )  f (A) ⊂ B}. Let Y = {0. Y )? 134 Let X. and by ∆(co) the set {W (C.
22:11. then d is a metric on the set C(X. Y ) does not coincide with the compactopen topology. . ρ) be a metric space and X a compact space. g ∈ Cb (X. I) compact? § 22 ◦ 4 Metric Case 22:E. . Y ). . I) is not compact. Prove that the space C(R. Let X be a compact space and Y a metric space. Then the topology deﬁned by d on C(X. Y ). g) = sup{ρ(f (x). Prove that if Y is Hausdorﬀ. Prove that C(X. Prove that C (pw) (I. where Xi is compact and Xi ⊂ Int Xi+1 i=1 for i = 1. 22:J d∞ and Uniform Convergence. Y ) is metrizable. Y )? 22:10. g(x))  x ∈ X}. SPACES OF CONTINUOUS MAPS 135 § 22 ◦ 3 Topological Properties of Mapping Spaces 22:D.. 22:I Metric on Bounded Mappings. f (x)) < ε for any n > N and x ∈ X. Y ). I) is metrizable. Y ) coincides with the compactopen topology. Let X be a topological space and Y a metric space with metric ρ. . g) = max{ρ(f (x). Y ) is metrizable. X) is path connected iﬀ X is path connected. This is a metric in Cb (X. Y ) is Hausdorﬀ for any topological space X. A sequence fn of bounded maps X → Y converges to f : X → Y in the topology deﬁned by d∞ . 22:G Metric of Uniform Convergence. iﬀ fn uniformly converges to f . g : X → Y put d(f. Let (Y. Y ) the set of all continuous bounded maps from a topological space X to a metric space Y . A sequence fn of maps X → Y converges to f : X → Y in the topology deﬁned by d. iﬀ fn uniformly converges to f . This is a straightforward generalization of the notion of uniform convergence which is known from Calculus. Is this true for C(X. For maps f. Let Y be a bounded metric space and X a topological space which admits presentation X = ∞ Xi . 22:12. Find X and Y such that the topology deﬁned by d∞ on Cb (X. g(x))  x ∈ X}. then C (pw) (X. A sequence fn of maps X → Y is said to uniformly converge to f : X → Y if for any ε > 0 there exists a natural N such that ρ(fn (x). 22:K When Uniform Is Not CompactOpen. For continuous maps f. Denote by Cb (X. Let X be a compact space and Y a metric space. 22:H Uniform Convergence Versus CompactOpen. If X is a compact space and Y a metric space. Prove that C(I. 22:F This is a Metric. 22:13. put d∞ (f.§ 22. 2. Let X be a topological space and Y a metric space. If Y is metrizable and X is compact then C(X. Is the space C(I.
Let {X1 . Z be topological spaces and Y a Hausdorﬀ locally compact space. 22:O Maps to Product. Let X. y) → F (x)(y) is continuous. 22:M Continuity of Composing. y). Let X. Are the conditions imposed on S and X in 22:Q necessary? 22:R The Evaluation Map. Recall that a partition is called closed. Y and B ⊂ Y the map C(X. Z) be a continuous mapping. if the saturation of each closed set is closed. Let F : X → C(Y. 22:16. Xn } be a fundumental cover of X. g) → g ◦ f is continuous. Are the conditions imposed on X in 22:R necessary? § 22 ◦ 6 Mappings X × Y → Z and X → C(Y. f Xn ) is a topological embedding. Then the map F : X → C(Y. n C(X. . Y ) is a topological embedding. 22:14. Z) → C(X. Y be topological spaces. Y ) : f → iB ◦ f is a topological embedding. is continuous. B) → C(X. Y ) : f → (f X1 . Y ) : f → f A is continuous. Y ) → C(X. . x) → f (x) is continuous. Prove that the map C(X. Prove that the map C(X. 22:15. 1 . Then the mapping f : X × Y → Z : (x. . . Let X. . Z). What if the cover is not fundamental? 22:Q Factorizing Source. Let X. Let S be a closed partition1 of a Hausdorﬀ compact space X. Y ) × X → Y : (f. 22:T. Z) : F (x) : y → f (x. SPACES OF CONTINUOUS MAPS 136 § 22 ◦ 5 Interactions With Other Constructions 22:L Continuity of Restricting. . Prove that for any topological space Y . Y ) × C(X. . Y ) → C(A. 22:P Restricting to Sets Covering Source. Y be topological spaces and A ⊂ X. For any topological spaces X. Z) 22:S. .§ 22. Y ) × C(Y. Z) : (f. Y × Z) is canonically homeomorphic to C(X. Y ) → i=1 C(Xi . Y and Z be topological spaces and f X × Y → Z be a continuous map. Y and Z the space C(X. For any topological spaces X. Is local compactness of Y necessary in 22:M? 22:N Extending Target. Prove that if X is locally compact and Hausdorﬀ then the map C(X. Let X be a topological space and Y a locally compact Hausdorﬀ space. Prove that for any topological space Y the mapping C(X/S.
(b) if Y is locally compact and Hausdorﬀ then Φ is a homeomorphism. y). Z) → C(X. SPACES OF CONTINUOUS MAPS 137 22:U. Let the mapping be deﬁned by the relation Φ : C(X × Y. . Let X. Z)) Φ(f )(x) : y → f (x. Then (a) Φ is continuous. C(Y.§ 22. Y and Z be topological spaces.
the reader will meet groups. Groups are attributed to Algebra. In the second part of the book. which is devoted to algebraic topology. sooner or later. this material provides a greate collection of exercises. It emerges in overwhelming majority of mathematical environments. Algebra studied algebraic operations with numbers and similar objects and contributed to the settheoretic Mathematics various structures based on operations. groups appear in a more profound way. In the frameworks of general topology. So. In the Mathematics built on sets. This interaction is a subject of Topological Algebra. It is also about an interaction of Topology and Algebra. This is a sort of interdisciplinary material. but the group is built to encode topological properties of the space. it is not that important in the frameworks of general topology. Often it appears together with topology and in a nice interaction with it. Topology and metric evolved from geometric considerations. this material can be postponed till the introductory chapters of the mathematical courses (functional analysis. Quite often. partial order.) which really require it. The second part of this book is called Algebraic Topology.CHAPTER 4 A Touch of Topological Algebra In this chapter we are going to study topological spaces strongly related to groups: they either themselves are groups in a nice way (so that all the maps coming from the group theory would be continuous). or groups act on topological spaces and can be thought of as consisting of homeomorphisms. for study of fundamental groups. One of the simplest (and most versatile) of these structures is the structure of group. the main objects are sets with additional structure. spaces and groups. etc. But this is a completely diﬀerent interaction. Lie groups. There structures of topological space and group do not live on he same set. Although it plays important roles in many areas of Mathematics. metric. Above we met a few of the most fundamental of these structures: topology. At latest by the next chapter. 138 .
b). b) = a + b. They provide a suﬃcient bases for most of what follows. Can a group be empty? Can it consist of one element? A group consisting of one element is said to be trivial. 23:A Uniqueness of Neutral Element.§ 23 Digression. § 23 ◦ 1 The Notion of Group Recall that a group is a set G equipped with a group operation. quotient groups. 23:1 Simplest Group. For any a ∈ G there exists b ∈ G such that ω(a. ω(a. b) = ab. b ∈ G there exists uniques elements x. b) = a+b. Let G be a set with associative operation ω : G × G → G. What is its neutral element? How to calculate the element inverse to a given one? • The set G is the set Z of integers. ω(a. etc. x) = b and ω(y. which hadto be acquired for mastering of the material presented above in this book. The reader is assumed to be familiar with groups. the reader can read few deﬁnitions and prove few theorems gathered in this section. Check if in each of the following situations we have a group. Prove that this is a group. iﬀ for any a. b) = a + b. ω(a. The mathematical culture. A group operation in a set G is a map ω : G × G → G satisfying the following three conditions (known as group axioms): • Associativity. • G is the set of all bijection of a set A onto itself. a) = e. • G = R 0. b. b) = ab. 139 . If this is not yet so. c)) = ω(ω(a. • Existence of Inverse. ω(b. c) for any a. e) = a for every a ∈ G. • G = C. 23:B Uniqueness of Inverse. a) = b. homomorphisms. • Existence of Neutral Element. We do not mean to provide a selfcontained outline of the group theory. For any element of a group an element inverse to it is unique. 23:2 Solving Equations. subgroups. the group operation is addition: ω(a. 23:C The First Examples of Groups. Generalities on Groups This section is included mainly to recall the most elementary deﬁnitions and statements concerning groups. b) = ω(b. • The set G is the set Q>0 of positive rational nambers. As a temporary solution. ω(a. There exists e ∈ G such that ω(e. In a group a neutral element is unique. c ∈ G. y ∈ G such that ω(a. would make this an easy and pleasant exercise. a) = ω(a. • G = R. b) = a ◦ b. the group operation is multiplication: ω(a. the group operatio is composition of bijections: ω(a. we recommend to read one of numerous algebraic textbooks covering the elementary group theory.
23:D. . An operation ω : G×G → G is commutative provided that ω(a. say abcde = (ab)(c(de)). if the deﬁnition of group is discussed. In the case of a sequence of three elements this is just associativity (ab)c = a(bc). b) → a + b. (A.) (c) the sets Rn and Cn with coordinatewise addition. The element inverse to a is denoted by a−1 . (j) the set R>0 of positive real numbers with multiplication. . a) for all a. a0 = 1. A group with commutative group operation is called commutative or abelian. For an element a of a group G powers an with n ∈ Z are deﬁned by formulas: = an a. of group of nonzero rational numbers with the usual multiplication. The distribution of brackets does not matter. while the multiplicative notation is used both for commutative and noncommutative cases. b ∈ G. 23:3. (h) the set Zn of classes of natural numbers congruent modulo n with the addition deﬁned by the addition of natural numbers. . (l) the set of translations of a plane with composition. GENERALITIES ON GROUPS 140 § 23 ◦ 2 Additive and Multiplicative Notations The notation above are never used! (The only exception may happen. which can be calculated by a sequence of pairwise multiplications deﬁned by any placement of backets. Associativity implies that in a group any ﬁnite sequence of elements has a welldeﬁned product. as here. Under multiplicative notation the group operation is called multiplication and denoted as multiplication: (a.) Instead. B) → (A ∪ B) (A ∩ B) a n+1 . Below we use mostly the multiplicative notation. (d) the set Homeo(X) of all homeomorphisms of a topological space X with composition. DIGRESSION. (b) the set Sn of bijections of the set {1. (k) S 1 ⊂ C with multiplication. b) = ω(b. Traditionally the additive notation is used only in the case of commutative group. . a−n = (a−1 )n . Check if in each of the following situations we have a group: (a) a singleton {a} with multiplication aa = a. 2. n} of n ﬁrst natural numbers with composition (symmetric group of degree n. The neutral element is called unity and denoted by 1.§ 23. b) → ab. The neutral element is called zero and denoted by 0. (g) the set of all subsets of a set X with symmetric diﬀerence . (f) the set Mn (R) of all real n × nmatrices with addition of matrices. R) with matrix multiplication. These notation are borrowed from the case. The element inverse to a is denoted by −a. (e) the set of invertible real n × nmatrices GL(n. of group of integer numbers with the usual addition. say. These notation are borrowed from the case. say. Under additive notation the group operation is called addition and denoted as addition: (a. Prove that independence of brackets in the product of any length follows from associativity. one uses either multiplicative or aditive notation. (i) the set of complex roots of unity of degree n with multiplication.
R) a subgroup of Mn (R) (for notations see 23:3)? 23:L. ab ∈ A for any a. 23:4. DIGRESSION. with the operation induced by the operation in G. The inverse map to an isomorphism is also an isomorphism. How does this deﬁnition look in additive notation? What if one of the groups is multiplicative. in the deﬁnition of homomorphism. 23:K. Recall that a homomorphism is called an epimorphism. and elements inverse to each other to elements inverse to each other. Prove that an is welldeﬁned and has the properties ap aq = ap+q and (ap )q = apq . 23:9. The one element subset of a group consisting of its neutral element is a subgroup. Let a be an element of a multiplicative group G. and isomorphism. List all subgroups of the additive Z.e. GENERALITIES ON GROUPS 141 23:E. 23:6. b ∈ A) and A. iﬀ AA ⊂ G and A−1 ⊂ A. if it is surjective. § 23 ◦ 4 Subgroups A subset A of a group G is called a subgroup of G if it is invariant under the group operation of G (i. Is GL(n. 23:I.§ 23. A homomorphism maps the neutral element to the neutral element. if there exists an isomorphism of one of them to another one. while the other is additive? 23:5. Let G and H be groups. § 23 ◦ 3 Homomorphisms Recall that a map f : G → H of a group to other one is called a homomorphism. . if it is injective. is a group. monomorphism. Prove that a subset A of a ﬁnite group G is a subgroup. Above. Is the contant map G → H mapping the whole G to the neutral element of H a homomorphism? Is any other constant map G → H a homomorphism? 23:G. y ∈ G. if it is bijective. 23:7. if f (xy) = f (x)f (y) for any x. For subsets A and B of a group G put AB = {ab  a ∈ A. Is a map Z → G deﬁned by formula n → an a homomorphism? 23:F. The identity map of a group is a homomorphism. multiplicative notation is used. The composition of homomorphisms is a homomorphism. Being isomorphic is an equivalence relation. The image of a group homomorphism f : G → H is a subgroup of H. if AA ⊂ A (the condition A−1 ⊂ A is not needed). b ∈ B} and A−1 = {a−1  a ∈ A}. A subset A of a multiplicative group G is a subgroup of G. 23:J. 23:H. 23:8. Groups are called isomorphic.
23:Q. The elements of a set that generates G are called generators of G. The subgroup generated by S is the intersection of all the subgroups of G that contain S..§ 23. 23:S. The preimage of the neutral element under a group homomorphism f : G → H is called the kernel of f and denoted by Ker f . The number of elements of a group G is called the order of G. 23:R.) Any cyclic group is commutative. Kernel of a group homomorphism is a subgroup. . DIGRESSION. Any subgroup of a cyclic group is cyclic. iﬀ its kernel is trivial. Shortly speaking: The preimage of a subgroup under a group homomorphism is a subgroup. if G is a cyclic group and a is its generator then G = {