Everything You Always Wanted To

Know About Mathematics*

(*But didn’t even know to ask)

A Guided Journey Into the World of Abstract

Mathematics and the Writing of Proofs

Brendan W. Sullivan
[email protected]
with Professor John Mackey

Department of Mathematical Sciences

Carnegie Mellon University
Pittsburgh, PA

May 10, 2013

This work is submitted in partial fulfillment of the requirements for the degree
of Doctor of Arts in Mathematical Sciences.
2
Contents

I Learning to Think Mathematically 11

1 What Is Mathematics? 13
1.1 Truths and Proofs . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.1.1 Triangle Tangle . . . . . . . . . . . . . . . . . . . . . . . . 14
1.1.2 Prime Time . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.1.3 Irrational Irreverence . . . . . . . . . . . . . . . . . . . . . 21
1.2 Exposition Exhibition . . . . . . . . . . . . . . . . . . . . . . . . 22
1.2.1 Simply Symbols . . . . . . . . . . . . . . . . . . . . . . . 22
1.2.2 Write Right . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.2.3 Pick Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
1.2.4 Obvious Obfuscation . . . . . . . . . . . . . . . . . . . . . 37
1.3 Review, Redo, Renew . . . . . . . . . . . . . . . . . . . . . . . . 41
1.3.1 Quick Arithmetic . . . . . . . . . . . . . . . . . . . . . . . 42
1.3.2 Algebra Abracadabra . . . . . . . . . . . . . . . . . . . . 43
1.3.3 Polynomnomnomials . . . . . . . . . . . . . . . . . . . . . 49
1.3.4 Let’s Talk About Sets . . . . . . . . . . . . . . . . . . . . 59
1.3.5 Notation Station . . . . . . . . . . . . . . . . . . . . . . . 60
1.4 Quizzical Puzzicles . . . . . . . . . . . . . . . . . . . . . . . . . . 61
1.4.1 Funny Money . . . . . . . . . . . . . . . . . . . . . . . . . 61
1.4.2 Gauss in the House . . . . . . . . . . . . . . . . . . . . . . 65
1.4.3 Some Other Sums . . . . . . . . . . . . . . . . . . . . . . 71
1.4.4 Friend Trends . . . . . . . . . . . . . . . . . . . . . . . . . 77
1.4.5 The Full Monty Hall . . . . . . . . . . . . . . . . . . . . . 86
1.5 It’s Wise To Exercise . . . . . . . . . . . . . . . . . . . . . . . . . 92
1.6 Lookahead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

2 Mathematical Induction 101

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
2.1.1 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
2.1.2 Segue from previous chapter . . . . . . . . . . . . . . . . . 102
2.1.3 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . 102
2.1.4 Goals and Warnings for the Reader . . . . . . . . . . . . . 103
2.2 Examples and Discussion . . . . . . . . . . . . . . . . . . . . . . 104
2.2.1 Turning Cubes Into Bigger Cubes . . . . . . . . . . . . . 104

3
4 CONTENTS

2.2.2 Lines On The Plane . . . . . . . . . . . . . . . . . . . . . 112

2.2.3 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 117
2.3 Defining Induction . . . . . . . . . . . . . . . . . . . . . . . . . . 119
2.3.1 The Domino Analogy . . . . . . . . . . . . . . . . . . . . 119
2.3.2 Other Analogies . . . . . . . . . . . . . . . . . . . . . . . 125
2.3.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
2.3.4 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 127
2.4 Two More (Different) Examples . . . . . . . . . . . . . . . . . . . 129
2.4.1 Dominos and Tilings . . . . . . . . . . . . . . . . . . . . . 129
2.4.2 Winning Strategies . . . . . . . . . . . . . . . . . . . . . . 133
2.4.3 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 137
2.5 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
2.5.1 Recursive Programming . . . . . . . . . . . . . . . . . . . 137
2.5.2 The Tower of Hanoi . . . . . . . . . . . . . . . . . . . . . 139
2.5.3 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 143
2.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
2.7 Chapter Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . 144
2.8 Lookahead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148

3 Sets 149
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
3.1.1 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
3.1.2 Segue from previous chapter . . . . . . . . . . . . . . . . . 150
3.1.3 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . 150
3.1.4 Goals and Warnings for the Reader . . . . . . . . . . . . . 151
3.2 The Idea of a “Set” . . . . . . . . . . . . . . . . . . . . . . . . . . 151
3.3 Definition and Examples . . . . . . . . . . . . . . . . . . . . . . . 153
3.3.1 Definition of “Set” . . . . . . . . . . . . . . . . . . . . . . 153
3.3.2 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
3.3.3 How To Define a Set . . . . . . . . . . . . . . . . . . . . . 154
3.3.4 The Empty Set . . . . . . . . . . . . . . . . . . . . . . . . 158
3.3.5 Russell’s Paradox . . . . . . . . . . . . . . . . . . . . . . . 159
3.3.6 Standard Sets and Their Notation . . . . . . . . . . . . . 162
3.3.7 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 163
3.4 Subsets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
3.4.1 Definition and Examples . . . . . . . . . . . . . . . . . . . 164
3.4.2 The Power Set . . . . . . . . . . . . . . . . . . . . . . . . 167
3.4.3 Set Equality . . . . . . . . . . . . . . . . . . . . . . . . . . 168
3.4.4 The “Bag” Analogy . . . . . . . . . . . . . . . . . . . . . 168
3.4.5 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 170
3.5 Set Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
3.5.1 Intersection . . . . . . . . . . . . . . . . . . . . . . . . . . 172
3.5.2 Union . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
3.5.3 Difference . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
3.5.4 Complement . . . . . . . . . . . . . . . . . . . . . . . . . 175
3.5.5 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 176
CONTENTS 5

3.6 Indexed Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177

3.6.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . 177
3.6.2 Indexed Unions and Intersections . . . . . . . . . . . . . . 181
3.6.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
3.6.4 Partitions . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
3.6.5 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 184
3.7 Cartesian Products . . . . . . . . . . . . . . . . . . . . . . . . . . 186
3.7.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
3.7.2 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
3.7.3 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 189
3.8 Defining the Natural Numbers . . . . . . . . . . . . . . . . . . . 190
3.8.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
3.8.2 Principle of Mathematical Induction . . . . . . . . . . . . 193
3.8.3 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 193
3.9 Proofs Involving Sets . . . . . . . . . . . . . . . . . . . . . . . . . 194
3.9.1 Logic and Rigor: Using Definitions . . . . . . . . . . . . . 194
3.9.2 Proving “⊆” . . . . . . . . . . . . . . . . . . . . . . . . . 195
3.9.3 Proving “=” . . . . . . . . . . . . . . . . . . . . . . . . . 198
3.9.4 Disproving Claims . . . . . . . . . . . . . . . . . . . . . . 203
3.9.5 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 206
3.10 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
3.11 Chapter Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . 208
3.12 Lookahead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213

4 Logic 215
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
4.1.1 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
4.1.2 Segue from previous chapter . . . . . . . . . . . . . . . . . 216
4.1.3 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . 216
4.1.4 Goals and Warnings for the Reader . . . . . . . . . . . . . 216
4.2 Mathematical Statements . . . . . . . . . . . . . . . . . . . . . . 217
4.2.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
4.2.2 Examples and Non-examples . . . . . . . . . . . . . . . . 219
4.2.3 Variable Propositions . . . . . . . . . . . . . . . . . . . . 221
4.2.4 Word Order Matters! . . . . . . . . . . . . . . . . . . . . . 224
4.2.5 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 224
4.3 Quantifiers: Existential and Universal . . . . . . . . . . . . . . . 226
4.3.1 Usage and notation . . . . . . . . . . . . . . . . . . . . . . 226
4.3.2 The phrase “such that”, and the order of quantifiers . . . 229
4.3.3 “Fixed” Variables and Dependence . . . . . . . . . . . . . 230
4.3.4 Specifying a quantification set . . . . . . . . . . . . . . . . 232
4.3.5 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 233
4.4 Logical Negation of Quantified Statements . . . . . . . . . . . . . 235
4.4.1 Negation of a universal quantification . . . . . . . . . . . 235
4.4.2 Negation of an existential quantification . . . . . . . . . . 236
4.4.3 Negation of general quantified statements . . . . . . . . . 237
6 CONTENTS

4.4.4 Method Summary . . . . . . . . . . . . . . . . . . . . . . 239

4.4.5 The Law of the Excluded Middle . . . . . . . . . . . . . . 240
4.4.6 Looking Back: Indexed Set Operations and Quantifiers . 241
4.4.7 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 242
4.5 Logical Connectives . . . . . . . . . . . . . . . . . . . . . . . . . 244
4.5.1 And . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
4.5.2 Or . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246
4.5.3 Conditional Statements . . . . . . . . . . . . . . . . . . . 246
4.5.4 Looking Back: Set Operations and Logical Connectives . 255
4.5.5 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 256
4.6 Logical Equivalence . . . . . . . . . . . . . . . . . . . . . . . . . . 258
4.6.1 Definition and Uses . . . . . . . . . . . . . . . . . . . . . 259
4.6.2 Necessary and Sufficient Conditions . . . . . . . . . . . . 263
4.6.3 Proving Logical Equivalences: Associative Laws . . . . . . 264
4.6.4 Proving Logical Equivalences: Distributive Laws . . . . . 268
4.6.5 Proving Logical Equivalences: De Morgan’s Laws (Logic) 269
4.6.6 Using Logical Equivalences: DeMorgan’s Laws (Sets) . . . 270
4.6.7 Proving Set Containments via Conditional Statements . . 271
4.6.8 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 276
4.7 Negation of Any Mathematical Statement . . . . . . . . . . . . . 278
4.7.1 Negating Conditional Statements . . . . . . . . . . . . . . 278
4.7.2 Negating Any Statement . . . . . . . . . . . . . . . . . . . 280
4.7.3 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 282
4.8 Truth Values and Sets . . . . . . . . . . . . . . . . . . . . . . . . 284
4.9 Writing Proofs: Strategies and Examples . . . . . . . . . . . . . . 286
4.9.1 Proving ∃ Claims . . . . . . . . . . . . . . . . . . . . . . . 287
4.9.2 Proving ∀ Claims . . . . . . . . . . . . . . . . . . . . . . . 291
4.9.3 Proving ∨ Claims . . . . . . . . . . . . . . . . . . . . . . . 293
4.9.4 Proving ∧ Claims . . . . . . . . . . . . . . . . . . . . . . . 295
4.9.5 Proving =⇒ Claims . . . . . . . . . . . . . . . . . . . . . 297
4.9.6 Proving ⇐⇒ Claims . . . . . . . . . . . . . . . . . . . . . 304
4.9.7 Disproving Claims . . . . . . . . . . . . . . . . . . . . . . 307
4.9.8 Using assumptions in proofs . . . . . . . . . . . . . . . . . 309
4.9.9 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 311
4.10 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
4.11 Chapter Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . 313
4.12 Lookahead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319

5 Rigorous Mathematical Induction 321

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
5.1.1 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
5.2 Regular Induction . . . . . . . . . . . . . . . . . . . . . . . . . . 322
5.2.1 Theorem Statement and Proof . . . . . . . . . . . . . . . 322
5.2.2 Using Induction: Proof Template . . . . . . . . . . . . . . 324
5.2.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
5.2.4 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 329
CONTENTS 7

5.3 Other Variants of Induction . . . . . . . . . . . . . . . . . . . . . 331

5.3.1 Starting with a Base Case other than n = 1 . . . . . . . . 331
5.3.2 Inducting Backwards . . . . . . . . . . . . . . . . . . . . . 334
5.3.3 Inducting on the Evens/Odds . . . . . . . . . . . . . . . . 335
5.3.4 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 341
5.4 Strong Induction . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
5.4.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . 342
5.4.2 Theorem Statement and Proof . . . . . . . . . . . . . . . 343
5.4.3 Using Strong Induction: Proof Template . . . . . . . . . . 348
5.4.4 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
5.4.5 Comparing “Regular” and Strong Induction . . . . . . . . 355
5.4.6 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 356
5.5 Variants of Strong Induction . . . . . . . . . . . . . . . . . . . . 357
5.5.1 “Minimal Criminal” Arguments . . . . . . . . . . . . . . . 358
5.5.2 The Well-Ordering Principle of N . . . . . . . . . . . . . . 362
5.5.3 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 364
5.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 366
5.7 Chapter Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . 366
5.8 Lookahead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373

II Learning Mathematical Topics 375

6 Relations and Modular Arithmetic 377
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377
6.1.1 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . 377
6.1.2 Segue from previous chapter . . . . . . . . . . . . . . . . . 378
6.1.3 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . 379
6.1.4 Goals and Warnings for the Reader . . . . . . . . . . . . . 379
6.2 Abstract (Binary) Relations . . . . . . . . . . . . . . . . . . . . . 380
6.2.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . 380
6.2.2 Properties of Relations . . . . . . . . . . . . . . . . . . . . 383
6.2.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . 384
6.2.4 Proving/Disproving Properties of Relations . . . . . . . . 386
6.2.5 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 391
6.3 Order Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
6.3.1 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 398
6.4 Equivalence Relations . . . . . . . . . . . . . . . . . . . . . . . . 399
6.4.1 Definition and Examples . . . . . . . . . . . . . . . . . . . 399
6.4.2 Equivalence Classes . . . . . . . . . . . . . . . . . . . . . 402
6.4.3 More Examples . . . . . . . . . . . . . . . . . . . . . . . . 409
6.4.4 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 412
6.5 Modular Arithmetic . . . . . . . . . . . . . . . . . . . . . . . . . 414
6.5.1 Definition and Examples . . . . . . . . . . . . . . . . . . . 414
6.5.2 Equivalence Classes modulo n . . . . . . . . . . . . . . . . 423
6.5.3 Multiplicative Inverses . . . . . . . . . . . . . . . . . . . . 433
8 CONTENTS

6.5.4 Some Helpful Theorems . . . . . . . . . . . . . . . . . . . 447

6.5.5 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 455
6.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 456
6.7 Chapter Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . 457
6.8 Lookahead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 466

7 Functions and Cardinality 467

7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467
7.1.1 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . 467
7.1.2 Segue from previous chapter . . . . . . . . . . . . . . . . . 468
7.1.3 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . 469
7.1.4 Goals and Warnings for the Reader . . . . . . . . . . . . . 469
7.2 Definition and Examples . . . . . . . . . . . . . . . . . . . . . . . 469
7.2.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . 470
7.2.2 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . 472
7.2.3 Equality of Functions . . . . . . . . . . . . . . . . . . . . 476
7.2.4 Schematics . . . . . . . . . . . . . . . . . . . . . . . . . . 480
7.2.5 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 481
7.3 Images and Pre-images . . . . . . . . . . . . . . . . . . . . . . . . 482
7.3.1 Image: Definition and Examples . . . . . . . . . . . . . . 482
7.3.2 Proofs about Images . . . . . . . . . . . . . . . . . . . . . 490
7.3.3 Pre-Image: Definition and Examples . . . . . . . . . . . . 493
7.3.4 Proofs about Pre-Images . . . . . . . . . . . . . . . . . . . 495
7.3.5 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 496
7.4 Properties of Functions . . . . . . . . . . . . . . . . . . . . . . . 497
7.4.1 Surjective (Onto) Functions . . . . . . . . . . . . . . . . . 497
7.4.2 Injective (1-to-1) Functions . . . . . . . . . . . . . . . . . 502
7.4.3 Proof Techniques for Jections . . . . . . . . . . . . . . . . 506
7.4.4 Bijections . . . . . . . . . . . . . . . . . . . . . . . . . . . 507
7.4.5 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 509
7.5 Compositions and Inverses . . . . . . . . . . . . . . . . . . . . . . 511
7.5.1 Composition of Functions . . . . . . . . . . . . . . . . . . 511
7.5.2 Inverses . . . . . . . . . . . . . . . . . . . . . . . . . . . . 516
7.5.3 Bijective ⇐⇒ Invertible . . . . . . . . . . . . . . . . . . . 519
7.5.4 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 521
7.6 Cardinality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 522
7.6.1 Motivation and Definition . . . . . . . . . . . . . . . . . . 522
7.6.2 Finite Sets . . . . . . . . . . . . . . . . . . . . . . . . . . 528
7.6.3 Countably Infinite Sets . . . . . . . . . . . . . . . . . . . 530
7.6.4 Uncountable Sets . . . . . . . . . . . . . . . . . . . . . . . 549
7.6.5 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 555
7.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 557
7.8 Chapter Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . 558
7.9 Lookahead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 566
CONTENTS 9

8 Combinatorics 567
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 567
8.1.1 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . 567
8.1.2 Segue from previous chapter . . . . . . . . . . . . . . . . . 568
8.1.3 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . 568
8.1.4 Goals and Warnings for the Reader . . . . . . . . . . . . . 569
8.2 Basic Counting Principles . . . . . . . . . . . . . . . . . . . . . . 570
8.2.1 The Rule of Sum . . . . . . . . . . . . . . . . . . . . . . . 570
8.2.2 The Rule of Product . . . . . . . . . . . . . . . . . . . . . 574
8.2.3 Fundamental Counting Objects and Formulas . . . . . . . 580
8.2.4 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 588
8.3 Counting Arguments . . . . . . . . . . . . . . . . . . . . . . . . . 589
8.3.1 Poker Hands . . . . . . . . . . . . . . . . . . . . . . . . . 589
8.3.2 Other Card-Counting Examples . . . . . . . . . . . . . . . 595
8.3.3 Other Counting Objects . . . . . . . . . . . . . . . . . . . 604
8.3.4 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 620
8.4 Counting in Two Ways . . . . . . . . . . . . . . . . . . . . . . . . 623
8.4.1 Method Summary . . . . . . . . . . . . . . . . . . . . . . 623
8.4.2 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . 625
8.4.3 Standard Counting Objects . . . . . . . . . . . . . . . . . 634
8.4.4 Binomial Theorem . . . . . . . . . . . . . . . . . . . . . . 639
8.4.5 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 641
8.5 Selections with Repetition . . . . . . . . . . . . . . . . . . . . . . 643
8.5.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . 643
8.5.2 Formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . 644
8.5.3 Equivalent Forms . . . . . . . . . . . . . . . . . . . . . . . 645
8.5.4 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . 647
8.5.5 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 651
8.6 Pigeonhole Principle . . . . . . . . . . . . . . . . . . . . . . . . . 652
8.6.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . 652
8.6.2 Statement and Proof . . . . . . . . . . . . . . . . . . . . . 653
8.6.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . 654
8.6.4 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 656
8.7 Inclusion/Exclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 657
8.7.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . 657
8.7.2 Statement and Proof . . . . . . . . . . . . . . . . . . . . . 658
8.7.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . 659
8.7.4 Questions & Exercises . . . . . . . . . . . . . . . . . . . . 662
8.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 662
8.9 Chapter Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . 663
8.10 Lookahead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 669
10 CONTENTS

A Definitions and Theorems 671

A.1 Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 671
A.1.1 Standard Sets . . . . . . . . . . . . . . . . . . . . . . . . . 671
A.1.2 Set-Builder Notation . . . . . . . . . . . . . . . . . . . . . 671
A.1.3 Elements and Subsets . . . . . . . . . . . . . . . . . . . . 672
A.1.4 Power Set . . . . . . . . . . . . . . . . . . . . . . . . . . . 672
A.1.5 Set Equality . . . . . . . . . . . . . . . . . . . . . . . . . . 673
A.1.6 Set Operations . . . . . . . . . . . . . . . . . . . . . . . . 673
A.1.7 Indexed Set Operations . . . . . . . . . . . . . . . . . . . 674
A.1.8 Partition . . . . . . . . . . . . . . . . . . . . . . . . . . . 674
A.2 Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 675
A.2.1 Statements and Propositions . . . . . . . . . . . . . . . . 675
A.2.2 Quantifiers . . . . . . . . . . . . . . . . . . . . . . . . . . 675
A.2.3 Connectives . . . . . . . . . . . . . . . . . . . . . . . . . . 676
A.2.4 Logical Negation . . . . . . . . . . . . . . . . . . . . . . . 677
A.2.5 Proof Strategies . . . . . . . . . . . . . . . . . . . . . . . 678
A.3 Induction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 680
A.3.1 Principle of Specific Mathematical Induction . . . . . . . 680
A.3.2 Principle of Strong Mathematical Induction . . . . . . . . 680
A.3.3 “Minimal Criminal” Argument . . . . . . . . . . . . . . . 681
A.4 Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 682
A.4.1 Properties of Relations . . . . . . . . . . . . . . . . . . . . 682
A.4.2 Equivalence Relations . . . . . . . . . . . . . . . . . . . . 682
A.4.3 Modular Arithmetic . . . . . . . . . . . . . . . . . . . . . 684
A.5 Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 686
A.5.1 Images and Pre-Images . . . . . . . . . . . . . . . . . . . 686
A.5.2 Jections . . . . . . . . . . . . . . . . . . . . . . . . . . . . 687
A.5.3 Composition of Functions . . . . . . . . . . . . . . . . . . 687
A.5.4 Inverses . . . . . . . . . . . . . . . . . . . . . . . . . . . . 688
A.5.5 Proof Techniques for Functions . . . . . . . . . . . . . . . 688
A.6 Cardinality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 692
A.6.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . 692
A.6.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 692
A.6.3 Standard Catalog of Cardinalities . . . . . . . . . . . . . . 694
A.7 Combinatorics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 695
A.7.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . 695
A.7.2 Counting Principles . . . . . . . . . . . . . . . . . . . . . 695
A.7.3 Formulas . . . . . . . . . . . . . . . . . . . . . . . . . . . 695
A.7.4 Standard Counting Objects . . . . . . . . . . . . . . . . . 696
A.7.5 Counting In Two Ways . . . . . . . . . . . . . . . . . . . 696
A.7.6 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 696
A.7.7 Inclusion/Exclusion . . . . . . . . . . . . . . . . . . . . . 697
A.7.8 Pigeonhole Principle . . . . . . . . . . . . . . . . . . . . . 697
A.8 Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 698
A.8.1 General Phrases . . . . . . . . . . . . . . . . . . . . . . . 698
A.8.2 Induction . . . . . . . . . . . . . . . . . . . . . . . . . . . 698
 Part I

Learning to Think
Mathematically

11
Chapter 1

What Is Mathematics?

1.1 Truths and Proofs

How do you know whether something is true or not? Surely, you’ve been told
that the angles of a triangle add to 180◦ , for example, but how do you know for
sure? What if you met an alien who had never studied basic geometry? How
could you convince him/her/it that this fact is true? In a way, this is what
mathematics is all about: devising new statements, deciding somehow whether
they are true or false, and explaining these findings to other people (or aliens,
as the case may be). Unfortunately, it seems like many people think math-
ematicians spend their days multiplying large numbers together; in actuality,
though, mathematics is a far more creative and writing-based discipline than
its widely-perceived role as ever-more-complicated arithmetic. One aim of this
book is to convince you of this fact, but that’s merely a bonus. This book’s
main goals are to show you what mathematical thinking, problem-solving, and
proof-writing are really all about, to show you how to do those things, and to
show you how much fun they really are!
As a side note, you might even wonder “What does it mean for something
to be true?” A full discussion of this question would delve into philosophy,
psychology, and maybe linguistics, and we don’t really want to get into that.
The main idea in the context of mathematics, though, is that something is
true only if we can show it to be true always. We know 1 + 1 = 2
always and forever. It doesn’t matter if it’s midnight or noon, we can rest
assured that equation will hold true. (Have you ever thought about how to show
such a fact, though? It’s actually quite difficult! A book called the Principia
Mathematica does this from “first principles” and it takes the authors many,
many pages to even get to 1 + 1 = 2!) This is quite different from, perhaps,
other sciences. If we conduct a physical experiment 10 times and the same result
occurs, do we know that this will always happen? What if we do the experiment
a million times? A billion? At what point have we actually proven anything? In
mathematics, repeated experimentation is not a viable proof! We would need to

13
14 CHAPTER 1. WHAT IS MATHEMATICS?

find an argument that shows why such a phenomenon would always occur. As
an example, there is a famous open problem in mathematics called the Goldbach
Conjecture. It is unknown, as of now, whether it is true or not, even though
it has been verified by computer simulations up until a value of roughly 1018 .
That’s a huge number, but it is still not enough to know whether the conjecture
is True or False. Do you see the difference? We mathematicians like to prove
facts, and checking a bunch of values but not all of them does not constitute a
proof.

1.1.1 Triangle Tangle

We’ve introduced the idea of a proof by talking about what we hope proofs to
accomplish, and why we would care so much about them. You might wonder,
then, how one can define a proof. This is actually a difficult idea to address!
To approach this idea, we are going to present several different mathematical
arguments. We want you to read along with them, and think about whether
they are convincing. Do they prove something? Are they correct? Are they
understandable? How do they make you feel? Think about them on your own
and develop some opinions, and then read along with our discussion.
The mathematical arguments we will present here are all about triangles.
Specifically, they concern the Pythagorean Theorem.

Theorem 1.1.1 (The Pythagorean Theorem). If a right triangle has base

lengths a, b and hypotenuse length c, then these values satisfy a2 + b2 = c2 .

How do we know this? It’s a very useful fact, one that you’ve probably used
many times in your mathematics classes (and in life, without even realizing).
1.1. TRUTHS AND PROOFS 15

Have you ever wondered why it’s true? How would you explain it to a skeptical
friend? This is what a mathematical proof attempts to accomplish: a clear
and concise explanation of a fact. The reasoning behind requiring a proof makes
a lot of sense, too, and is twofold: it’s a relief to know that what we thought was
true is, indeed, true and it’s nice to not have to go through the explanation of
the fact every time we’d like to use it. After proving the Pythagorean Theorem
(satisfactorily), we merely need to refer to the theorem by name whenever a
relevant situation arises; we’ve already done the proof, so there’s no need to
prove it again.
Now, what exactly constitutes a proof? How do we know that an explanation
is sufficiently clear and concise? Answering this question is, in general, rather
difficult and is part of the reason why mathematics can be viewed as an art as
much as it is a science. We deal with cold, hard facts, yes, but being able to
reason with these facts and satisfactorily explain them to others is an art form
in itself.

Examples of “Proofs”
Let’s look at some sample “proofs” and see whether they work well enough.
(We say “proof” for now until we come up with a more precise definition for it,
later on.) Here’s the first one:
“Proof ” 1. Draw a square with side length a + b. Inside this square, draw four
copies of the right triangle, forming a square with side length c inside the larger
square.

The area of the larger square can be computed in two ways: by applying the
area formula to the larger square or by adding the area of the smaller square to
the area of the four triangles. Thus, it must be true that
ab
(a + b)2 = c2 + 4 · = c2 + 2ab
2
16 CHAPTER 1. WHAT IS MATHEMATICS?

Expanding the expression on the left and canceling a common term on both
sides yields
a2 +  + b2 = c2 + 
2ab 2ab


Therefore, a2 + b2 = c2 is true.

Are you convinced? Did each step make sense? Maybe you’re not sure yet,
so let’s look at another “proof” of the theorem.

“Proof ” 2. Suppose the Pythagorean Theorem is true and draw the right tri-
angle with the altitude from the vertex corresponding to the right angle. Label
the points and lengths as in the diagram below:

Since the Pythagorean Theorem is true, we can apply it to all three of the
right triangles in the diagram, namely ABC, BCD, ACD. This tells us (defining
e = c − d)

a2 = d2 + f 2
b2 = f 2 + e2
c2 = a2 + b2

Adding the first two equations together and replacing this sum in the third
equation, we get
c2 = d2 + e2 + 2f 2
Notice that angles ∠ABC and ∠ACD are equal, because they are both comple-
mentary to angle ∠CAB, so we know triangles 4CDB and 4ADC are similar
triangles. (We are now assuming some familiarity with plane geometry.) This
tells us fe = fd , and thus f 2 = ed. We can use this to replace f 2 in the line
above and factor, as follows:

c2 = d2 + e2 + 2de = (d + e)2
1.1. TRUTHS AND PROOFS 17

Taking the square root of both sides (and knowing c, d, e are all positive num-
bers) tells us c = d + e, which is true by the definition of the lengths d and e.
Therefore, our assumption that the Pythagorean Theorem is true was valid.
What about this proof? Was it convincing? Was it clear? Let’s examine one
more “proof” before we decide what a constitutes a “correct” or “good” proof.
“Proof ” 3. Observe that

a c−b
so c+b = a and thus a2 + b2 = c2 .
Did that make any sense to you? Finally, here’s one last “proof” to consider.
“Proof ” 4. The Pythagorean Theorem must be true, otherwise my teachers
have been lying to me.

Discussion
Before reading on, we encourage you to think about these four “proofs” and even
discuss them with another student or a friend. What do you think constitutes
a “correct” proof? Is clarity and ease of reading important? Does it affect the
“correctness” of a proof?
From a historical perspective, mathematical proof-writing has evolved over
the years and there is a good, general consensus as to what constitutes a “cor-
rect” proof:
• It is important that every step in the proof, every logical inference and
claim, is valid, mathematically speaking.
• It is also important that the proof-writer makes (reasonably) clear why a
statement follows from the previous work or from outside knowledge.
18 CHAPTER 1. WHAT IS MATHEMATICS?

What’s nice about the truth requirement is that mathematics has been built up
so that we can read through an argument and verify each claim as True or False.
What’s difficult to define is clear writing. In a way, it is much like Supreme
Court Justice Potter Stewart’s famous definition of obscenity: “I know it when
I see it”.
Given these four arguments for comparison, let’s assess them for clarity and
correctness:
Clarity:
• “Proofs” 1 and 2 are fairly well explained. There are clear statements
about what the writer is doing and why. They indicate where any equa-
tions come from, and even include some pictures to illustrate their ideas
for the reader.
Notice that “proof 1” does rely on some basic prior knowledge, like the
algebraic manipulation of variables and formulae for the area of a triangle
and square, but this is fine.
Likewise, “Proof 2” relies on some understanding of similar triangles and
what this means about the lengths of their sides. At least the proof-writers
pointed this out, so an interested reader could look up some relevant ideas.
If they didn’t say this, a reader might be confused and have no idea how
to figure out what they’re missing!
• “Proof” 3 is very poorly worded! It offers no explanation whatsoever.
This makes it quite difficult to determine whether their claims are actually
correct. Yes, a picture is included, but there is no indication of why they
chose to draw a circle around the triangle, or why the stated equations
follow from the diagram.
• “Proof” 4 is a grammatically correct English sentence, but it doesn’t ex-
plain anything!
Already, we can see that “Proof” 4 is certainly ont a viable candidate for being
a good and proper proof. “Proofs” 1 and 2 are still in the running, since they
are at least written clearly. “Proof” 3, as it is written now, would probably
not be a good candidate; however, maybe it does contain correct ideas that just
require better explanations. Perhaps it could be rewritten as a good and proper
proof.
Let’s analayze the logical correctness of these four arguments:
Correctness:
• “Proof” 1 mostly good. The formulae for the areas of the square and
triangles are correctly applied, and the algebraic manipulation thereof is
correct. But how do we know that the process they described—putting
four copies of the given triangle inside a larger square—creates a square
with side length c on the inside? They merely say it does so without
1.1. TRUTHS AND PROOFS 19

really saying why. Other than this omission, though, this proof is both
well-written and correct.
(Can you prove that fact, that the shape inside is actually a square? Just
look at its angles: can you show why they are all right angles?)

• Unfortunately, “Proof” 2 is completely invalid! Every logical step that

it makes does follow from the previous one. For instance, assuming we
have the triangles set up this way, we can correctly deduce that ∆CDB
and ∆ADC are similar triangles. However, why is it that we can assume
the theorem is True right at the beginning? Isn’t that what we are trying
to accomplish in the proof, overall? This is a crucial flaw. Assuming a
fact and deducing something True from it does not allow us to
conclude the original assumption was valid.
If this method were valid, we could “prove” just about anything we wanted!
Here’s an example: What do you think of the following “proof” that 0 = 1?

“Proof ”. Assume 0 = 1. Then, by the symmetric property of =, it is also

true that 1 = 0. Adding these two equations tells us 1 = 1, which is True.
Therefore, 0 = 1 was a valid assumption, so it must be True.

Do you see the similarity between this and “Proof” 2 above? The same
sort of flawed reasoning was used: we assumed one fact, did some work to
get to something else we know to be True, and then said that the assumed
fact must be True, as well.

• Regarding “Proof” 3, most mathematicians would say it is a “bad proof”,

despite the fact that everything it appears to claim is correct. We say
“appears” because, without any words to explain what’s going on, we
don’t actually know what the proof-writer is trying to say! However, we
will say that the kernel of a perfectly good proof is contained therein.
a
From the diagram, you can show that the stated equation, c+b = c−b
a ,
must follow. (Hint: Use similar triangles!) From there, it is a simple
manipulation to deduce that a2 + b2 = c2 .
Can you write some sentences to go along with the diagram that would
turn this into a proper proof?

• Lastly, just about every reasonably logical person (we hope!) would say
that “Proof” 4 is not even close to being a proof, however convenient it
might be to make such statements.

This discussion shows that “Proof” 1 is actually a good proof. Amongst all
four, it is the most clearly-written, and the one that is logically correct. We can
refer to it now as a proof. “Proof” 2 is outright incorrect, despite how clearly
it is presented. “Proof” 3 contains correct ideas, but is not presented clearly.
“Proof” 4 is so far from a proof that we don’t even want to discuss it.
20 CHAPTER 1. WHAT IS MATHEMATICS?

Question
Before moving on to other topics, we’ll leave you with a question: if we give you
three positive numbers a, b, c that satisfy a2 + b2 = c2 , is it necessarily true that
there is a right triangle with side lengths a, b and hypotenuse length c? If so,
how could you go about constructing it? If not, why not?

1.1.2 Prime Time

While we’re on the topic of proofs, let’s look at another proof, for a different
theorem. As a reminder (or brief introduction), let’s talk about prime numbers.

Definition, Examples, and Uses

Definition 1.1.2. A positive integer p that is larger than 1 is called a prime
number if the only positive divisors of p are 1 and p. A non-prime positive
integer is called a composite number.

Prime numbers have shown to be incredibly important in all branches of

mathematics, not just the study of integers and their properties, which is known
as number theory. One of the most famous conjectures (a guess at a theorem
that has been neither proven nor disproven thus far) in all of mathematics is
the Riemann Hypothesis. Its conclusion has been shown to be closely related to
the distribution of prime numbers throughout the integers. Many books have
been written on this topic. Also, most modern cryptography schemes are based
on multiplying huge prime numbers together, relying on the fact that it’s quite
difficult to undo this process and figure out the two huge prime factors, given
their product. So now you know: every time you buy a song on iTunes with
your credit card, some computer just multiplied two large prime numbers!
The first few prime numbers are 2, 3, 5, 7, 11, 13, 17, 19, 23, . . . (remember, 1
does not fit our definition). How many prime numbers are there? How far apart
are they? Is there a pattern? Answering questions like these can be interesting
and fun, but also difficult (and sometimes, impossible!). Here, we’ll answer one
of the questions: are there an infinite number of prime numbers?

Theorem and Proof

Theorem 1.1.3 (Infinitude of the Primes). There are infinitely-many prime
numbers.

“Proof ”. Assume instead that there are only finitely-many prime numbers, and
list them in ascending order: p1 , p2 , p3 , . . . , pk , so that pk is the largest of these
prime numbers. Define the new number

N = (p1 · p2 · p3 · · · · · pk ) + 1

It must be true that N is divisible by some prime number. However, it cannot

be divisible by p1 or p2 or . . . or pk , because that would leave a remainder of 1,
1.1. TRUTHS AND PROOFS 21

based on how we defined N . Thus, N is divisible by some other prime number

that is not in the list.
If N itself is composite (i.e. not prime), then we have found some new prime
p < N that is not in the list of all primes we presumably had. If N itself is
prime, then we have a new prime N > pk , so pk is not actually the largest prime
number. Either way, we have a new prime guaranteed to not be in the given
list of k primes. Therefore, there must be infinitely-many prime numbers.

What do you think of this “proof”? Are you convinced? It feels a little
different from the other arguments we’ve seen so far, doesn’t it? Try explaining
to a classmate how this one differs from “Proof 1” of the Pythagorean Theorem
from the previous section. We will reveal this, though: this “proof” here is
actually a fully correct proof, sans quotation marks!

1.1.3 Irrational Irreverence

Let’s talk about a different type of number, now: rational numbers. You might
know rational numbers as “fractions” or “quotients” or “ratios”.

Definition and Examples

Here is a precise definition of rational numbers:

Definition 1.1.4. A real number r is a rational number if and only if it can

be expressed as a ratio of two integers r = ab , where a and b are both integers
(and b 6= 0).
A real number that is not rational is called irrational.

Nothing about this definition says that there has to be only one such repre-
sentation of a rational number; it merely requires that a rational number have
at least one such a representation. For instance, 1.5 is a rational number be-
cause 1.5 = 23 = 12 30
8 = 20 and so on. A real number that is not rational is called
an irrational number, and that’s the entire definition: not rational, i.e. there
is no √
such representation of the number as a ratio of integers. You may know
that 2 is an irrational number, but how do you prove such a thing? Try it for
yourself. We will actually reexamine this question later on (see Example
√ 4.9.4).
Other irrational numbers you may know already include e, π, ϕ and n where
n is any positive integer that is not a perfect square.

Questions
Given this definition of rational/irrational, we might wonder how we can com-
bine irrational numbers to produce a rational number. Try to answer the follow-
ing questions on your own. If your answer is “yes”, try to find an example, and
if your answer is “no”, try to explain why the desired situation is not possible.

(1) Are there irrational numbers a and b such that a · b is a rational number?
22 CHAPTER 1. WHAT IS MATHEMATICS?

(2) Are there irrational numbers a and b such that a + b is a rational number?
(3) Are there irrational numbers a and b such that ab is a rational number?
Did you find any examples? It turns out that the answer to all three questions
is “yes”! The first two are not too hard to figure out, but the third one is a
little trickier.
Here, we’ll work through a proof that says the answer to the third is “yes”.
The interesting part about it, though, is that we won’t actually come up with
definitive numbers a and b that make ab a rational number; we’ll just narrow
it down to two possible choices and show that one of those choices must work.
Sounds interesting, right? Let’s try it.
√ √ √2
Proof. We know 2 is an irrational number. Consider the number x = 2 .
There are two possibilities to consider:
√ √
• If x is rational, then we can choose a = 2 and b = 2 and have our
answer.
√ √2 √
• However, if x is irrational, then we can choose a = 2 and b = 2
because then
√ √2 √ √2·√2 √ 2
√

b 2
a = 2 = 2 = 2 =2

and 2 is a rational number.

In either case, we can find irrational numbers a and b such that ab is a rational
number. Thus, such a pair of numbers must exist.
How do you feel about this proof? Is it convincing? It answers the third
question above with a definitive “yes”, but it does not tell us which pair a, b is
actually√the correct one, merely that one of the pairs will work. (It turns out
√ 2
that 2 is also irrational, but that fact takes a lot more work to prove.)
There are plenty of other concrete examples that answer this question,
though. Can you come up with any? (Hint: try using the log10 function...)

1.2 Exposition Exhibition

1.2.1 Simply Symbols
Mathematics is a Language
Despite appearances (and some densely-written textbooks), mathematics is not
just a collection of symbols that we push around on paper. The English lan-
guage is based on a fixed group of symbols (the 26 letters of the alphabet plus
common punctuation like the period and comma and parenthesis) but we put
these symbols together in a specific way, while following some standard and
1.2. EXPOSITION EXHIBITION 23

agreed-upon conventions, to craft meaningful words, phrases, sentences, para-

graphs, and so on; in essence, English, like any language, is a way to convey
meaning via a collection of symbols and a collection of rules governing those
symbols. The same concept applies to the language of mathematics: there is a
collection of symbols and a set of rules that we apply to those symbols.
One difference is that the collection of symbols we use in mathematics can
be rather large, depending on which branch of mathematics currently being
discussed. A big part of the structural versatility of mathematics is that we can
always create and define new symbols to use. Oftentimes, this is even done to
make things shorter and easier to read.
Another main difference between mathematics and other languages is that
we choose carefully how to define our words and the concepts they represent.
Frequently, most of the debates mathematicians have revolve around definitions.
This may be surprising to you; perhaps it would make more sense to think that
mathematicians debate over proofs and conjectures, or maybe it’s a novel idea
that mathematicians even debate at all! Choosing the right definitions and
terms for a newly-discovered concept is a crucial component of mathematical
discovery and exposition since it helps the discoverer/inventor explain his/her
ideas to other, interested people. (Without this process, there is no advancement
in mathematics, just a bunch of isolated people trying to discover truths on their
own.)
The situation is similar with spoken languages, but not as extreme, it seems.
For instance, if you said to your friend, “I’m hungry”, or “I’m feeling a bit
peckish”, or “Oh my god, I’m starving”, they hear essentially the same message
and would respond roughly the same way in each case. In mathematics, though,
our definitions are far more precise and don’t incorporate the types of nuances
that spoken language permits. Of course, there are benefits and disadvantages to
both philosophies, but in mathematics we strive for precision whenever possible,
so we like our definitions to be exact and unwavering. That said, though, we
have control over what those definitions are! This is why debates over definitions
are so prevalent in the mathematical world: choosing the right definitions for
concepts at hand can make future work with those concepts much easier and
more convenient.

Choosing Definitions Properly

As a concrete example, let’s return to Definition 1.1.2 of a prime number that
we saw in the previous subsection. It said:
Definition. A positive integer p that is larger than 1 is called a prime number
if the only positive divisors of p are 1 and p. A non-prime positive integer is
called a composite number.
There doesn’t seem to be anything questionable about this definition, does
there? Perhaps you would have worded it differently or been more concise or
used a different variable letter or what have you, but the ultimate message would
be the same: a prime number is a certain type of number that has a certain
24 CHAPTER 1. WHAT IS MATHEMATICS?

property. However you choose to write out what that specific type of number is
(a positive integer larger than 1) and what that property is (having no positive
divisors except itself and 1), you obtain an equivalent definition.
There are some subtle questions behind this definition, though: Why is it
that particular type of number? Why is it that particular property—only being
divisible by 1 and itself—that we care so much about? What if the definition was
slightly different? Would things really change that much? We’ll address these
questions with another question: What do you think of the following alternative
definition of a prime number?

Definition 1.2.1. An integer p that is less than −1 or greater than 1 is called

a prime number if the only positive divisors of p are 1 and p.

Do you notice the subtle difference? All of the numbers that fit the previous
definition of “prime” still fit this one, but now so do negative numbers! Specif-
ically, given any number p that is prime under the old definition, −p is now
prime under the new definition. Is this a reasonable idea? What’s wrong with
having negative prime numbers?
How about this third definition of prime numbers?

Definition 1.2.2. A positive integer p is called a prime number if the only

positive divisors of p are 1 and p.

(Remember that 0 is neither positive nor negative, by convention.) Now, the

negative numbers are out of bounds, but 1 fits this definition. Is this reasonable?
The only positive divisors of 1 are 1 and . . . itself, right?
This is where a debate could arise: perhaps you don’t mind allowing 1 to be
a prime number, but your friend is vehemently against it. Well, without solid
reasons either way, there’s no way to say that either of you is wrong, really;
you just made different choices of terminology, and neither of them change the
inherent property that the only positive divisors of 1 are 1 and itself. As a
similar idea, consider this: whether you call them sandals or thongs or flip-
flops, the fact remains that those types of shoes are appropriate footwear at the
beach.
With historical hindsight and new desires in mind, though, oftentimes one
particular definition is shown to be more appropriate. In the future, we will
look at prime factorizations, a way of writing every (positive) integer as a
product of only prime numbers. For instance, 15 = 3 · 5 and 12 = 2 · 2 · 3 = 22 · 3
and 142857 = 33 · 11 · 13 · 37 are all prime factorizations.
There is a special property about these factorizations, too: in general, a
prime factorization of a positive integer is unique! That is, there is one and
only one way to write a positive integer as a product of prime numbers (since
we think of different orderings of the factors as the same thing, so 105 = 3 · 5 · 7
and 105 = 7 · 3 · 5 are the same factorization). This is something we will prove
rigorously using the first definition we gave above. What if we use the second
definition, or the third? Is this property of uniqueness still true? Why do you
think this uniqueness property is so important? Ultimately, the lesson here is
1.2. EXPOSITION EXHIBITION 25

that definitions should be driven by both logic and usefulness, and this can
change over time and stir some debate.

Mathematicians Study Patterns

Another benefit of establishing clear and precise definitions is the knowledge
and understanding you gain as a thinker; establishing logical foundations can
be helpful in the future. A major aspect of how human beings learn involves
identifying patterns through everyday experience and then associating ideas,
concepts, words and events with those patterns. Then, one can use those pat-
terns to predict and theorize about abstract ideas, concepts and events.
For instance, it has been studied and shown that human babies initially lack,
but develop over time, the concept of object permanence. If you show a child a
colorful toy that they smile at and enjoy, and then hide it under a cardboard
box, the child doesn’t quite understand that the toy still exists but is just out of
sight. He/she will act as if the object is no longer in existence. At some point,
though, we learn that this isn’t true and that objects that are outside our realm
of vision are still existent. How exactly does this happen? Well, perhaps we
recognize the pattern of many such occurrences where an object “disappears”
and then we find it again later.
Better examples can be found in the natural sciences, and they illustrate
an extra facet of pattern recognition and abstract thinking that is of utmost
importantance, particularly in mathematics and the sciences. One can imagine
that Neanderthals somehow knew that any time they picked up a rock and
held it at arm’s length and then let go, the rock would fall to the ground.
This probably happened over and over and so they “understood” that this
phenomenon is a necessary product of nature. After enough occurrences, it
was likely understood that this would always happen, or, at least, any instance
in which it didn’t happen would cause great confusion and fear. (It is this
type of emotional response which might serve to explain how the infrequent
but powerful occurrences of, say, volcanic eruptions led ancient civilizations to
blame such events on “angry gods”).
None of these observations of events brought these prehistoric human beings
any closer to understanding why the rock would always fall to the ground, or
being able to explain why it would necessarily happen every time. It would
be many milennia before people even began to think to ask why and how this
phenomenon occurred, and even longer before Isaac Newton finally proposed a
model that sought to explain the behavior of gravity (the name given to this type
of phenomenon, eventually). And even now, some say, we still haven’t figured
out precisely how it works. (Go online and Google “loop quantum gravity” and
try to understand that, if you’re curious).
It’s this abstractive leap in thinking—from observations of a pattern to an
epistemological understanding of that pattern—that characterizes a truly inquis-
itive and intellectual thinker, a true scientist, in the best sense of the word.
Whom would you consider the better entomologist: the voracious reader who
has memorized and can list all of the currently-known species of beetle in the
26 CHAPTER 1. WHAT IS MATHEMATICS?

world, or the laboratory scientist who has examined a variety of species and can
take a new specimen and classify it as a beetle or non-beetle? This is somewhat
of a leading question, but the main point is this: it is far more beneficial to
understand a definition and the motivations behind it than it is to simply know
a bunch of instances that satisfy a certain definition.
This is, arguably, even more important in mathematics. Can you imagine a
mathematician who didn’t know what a prime number was but could merely list
the first 100 prime numbers from memory and was content with that? Of course
not! Part of the beauty, versatility, and appeal of the study of mathematics
is that we examine patterns and phenomena and then choose how to make
the appropriate definitions associated with those patterns. We then use our
newfound understanding of those patterns to make rigorously precise predictions
about other patterns and phenomena. Thoroughly understanding a definition
or concept increases the predictive power, and is far more effective than merely
knowing examples of that definition/concept.

1.2.2 Write Right

Another interesting aspect of mathematics is that, as much as it is a language
unto itself, we rely on an external language to convey the mathematical thoughts
and insights we have. Try rewriting any of the definitions and proofs we’ve
looked at before without using any words. It’s tough, isn’t it? Accordingly, we
want the written language we use to convey mathematical ideas to follow the
same types of standards we apply to the mathematical “sentences” we write:
we want them to be precise, logical, and clear.
Now, deciding on a precise, logical, and clear definition for each of these
three words is a difficult task, in itself. However, we can all agree that it would
be ideal for a proof to be:

• precise: no individual statement should be untrue or interpretable in

multiple ways that would make the truth debatable;

• logical: each step should follow from previous steps with proper motiva-
tion and explanation; and,

• clear: steps should be connected and described with proper English gram-
mar and usage, helping the reader to see what’s going on.

Let’s examine a few “proofs” that disregard these standards and somehow fail
to fit the definition of proof that we have so far.

Bad “Proof ” #1
First, we have a “proof” that 1=2, so we know there must be something wrong
with this one. Can you find the error? Which standard does it violate? Preci-
sion, logic, or clarity?
1.2. EXPOSITION EXHIBITION 27

“Proof ”. Suppose we have two real numbers x and y, and consider the following
chain of equalities:

x=y
x2 = xy multiply both sides by x
2 2 2
x − y = xy − y subtract y 2 from both sides
(x + y)(x − y) = y(x − y) factor both sides
x+y =y cancel (x − y) from both sides
y+y =y remembering x = y, from the first line
2y = y
2=1 divide both sides by y

The issue here is precision. After factoring in line four, it seems convenient
and wise to divide by the common factor (x − y) to obtain line five; however,
line one tells us that x = y so x − y = 0, and division by zero is not allowed!
Working with the variables x and y was just a way to throw you off the scent
and disguise the division by zero. (While we’re on the topic, why is division by
zero not allowed? Can you think of a reasonable explanation? Think about it
in terms of multiplication.)

Bad “Proof ” #2
Here’s another proof of a similar “fact”, namely that 0 = 36.

“Proof ”. Consider the equation x2 + y 2 = 25. Rearranging to isolate x tells us

p
x= 25 − y 2

and then adding 3 to both sides and squaring yields

 p 2
(x + 3)2 = 3 + 25 − y 2

Notice that x = −3 and y = 4 is a solution to the original equation, so the final

equation should be true, as well. Plugging in these values for x and y tells us
√
2
0 = (−3 + 3)2 = 3 + 25 − 16 = (3 + 3)2 = 36

Therefore, 0 = 36.

What happened here? Can you spot the illogical step? Perhaps it would
help if we rewrote the steps of the proof using the specific values of the variables
28 CHAPTER 1. WHAT IS MATHEMATICS?

x and y that we chose towards the end:

(−3)2 + 42 = 25
p
−3 = 25 − 42
 p 2
(−3 + 3)2 = 3 + 25 − 42
0 = 36

It’s obvious now, isn’t it? There’s an issue with applying the square root opera-
tion to both sides of an equation, and it’s dependent on the fact that (−x)2 = x2 .
When we are looking to solve an equation like z 2 = x2 , we have to remember
there are two roots of this equation: z = −x and z = x. Accordingly, starting
from an equation and squaring both sides is a completely logical step (the truth
of the resulting equations follows from the truth of the original equation), but
working the other way is an illogical step (the truth of the squared equation
does not necessarily tell us that the square-rooted equation is also true). This
is an issue with conditional statements or logical implications, an idea we
will discuss in detail later on (in Section 4.5.3). For now, we can summarize
this idea with the following line:

If a = b then a2 = b2 , but if a2 = b2 then a = b or a = −b.

p
This shows why moving from x2 + y 2 = 25 to x = 25 − y 2 in the “proof”
above is an illogical step: we are immediately assuming one particular choice
for the square root when there are two possible options. What would have
happened if we had chosen the negative squarep root there? Try rewriting the
proof with the second step reading −x = 25 − y 2 , instead, and then use the
same values for x and y at the end. What happens? What if you use x = 3 and
y = −4 instead? Or x = −5 and y = 0? Can you describe how to determine
when we should use the positive root x and when we should use the negative
root −x?

Mathematics Uses the “Inclusive Or”

Since this word just arose, let’s mention the use of or in the sentence above.
When we say “a = b or a = −b”, we mean that at least one of the two statements
must be true, possibly both. Now, if both a 6= 0 and b 6= 0, then only one of the
concluding statements can be true; that is, in that context, only one of the roots
(positive or negative) will be the correct one and not both. If b = 0, though,
then both of the concluding statements say the same thing, a = 0, so it would
be illogical to dictate that or means only one of the statements can be true and
doesn’t allow both of them to be true, simultaneously. In other situations, this
distinction makes a more marked difference.
For instance, if you order a sandwich at a restaurant and the waiter asks, “Do
you want fries or potato salad on the side?”, it is understood that you can choose
one of those options, but not both. This is an example of the exclusive or since
1.2. EXPOSITION EXHIBITION 29

it excludes you from choosing both options. Alternatively, if you forgot to bring
a writing implement to class and are looking for any old way to take notes and
ask your friend, “Do you have a pencil or pen I can borrow?”, it is understood
that you really don’t care which one of the two options is provided, as long as
at least one is available. Maybe your friend has both, and any one of them will
do. This is an example of the inclusive or, and this is the interpretation that
is assumed in all mathematical examples.

Unclear Arguments
The last two bad “proofs” failed because of issues with precision and logical
correctness. The third condition we require of a good proof is that it be clear :
we want the writing to explain what the proof-writer accomplishes in each step
and why that accomplishment is relevant. In other words, we don’t want the
reader to stop at any point and ask, “What does that sentence mean?” or
“Where did that come from?” or similar questions born from confusion. If it
helps, think about writing a proof in terms of explaining it to your friend in your
class, or the grader who will be reading your homework assignment, or a family
member of similar intelligence. Reread your own writing and try to anticipate
any questions that might arise or any clarifications that might be asked of you,
and then address those issues by rewriting.
There are several ways that a proof could fail this condition and come across
as unclear. For one, the words and sentences might fail to properly explain the
steps and motivations of the proof, and this could actually be because there are
too many words (obscuring the proof by overburdening the reader) or because
there are too few words (not giving the reader enough information to work with)
or because the words chosen are confusing (not properly explaining the proof).
These are issues with the language of the proof.
Mathematically, any number of issues could arise, in terms of clarity. Per-
haps the proof-writer suddenly introduces a variable without stating what type
of number it is (an integer, a real number, etc.) or skips a few steps of arith-
metic/algebra or uses new notation without defining what it means first or . . .
None of these acts is technically wrong or illogical, but they can certainly cause
confusion for a reader. Can you think of any other ways that a proof can be
unclear? Try to think of a language-based one and a mathematical one.

Bad “Proof ” #3
Let’s state a simple fact about a polynomial function and then examine a
“proof” about that fact. Read the argument carefully and try to pinpoint some
sentences or mathematical steps that are unclear.
Fact: Consider the polynomial function f (x) = x4 − 8x2 + 16. This function
satisfies f (x) ≥ 0 for any value of x.

“Proof ”. No matter what the value of x is that we plug into the function f of x
we can write the value that the function puts out by factoring the polynomial,
30 CHAPTER 1. WHAT IS MATHEMATICS?

like this:
f (x) = x4 − 8x2 + 16 = (x − 2)2 (x + 2)2
Now, any number z must be less than −2, or greater than 2, or strictly between
−2 and 2, or equal to one of them. When z > 2 then z − 2 and z + 2 are both
greater than 0 so f (z) > 0. When z < −2 then both terms are negative and a
negative squared is positive so f (z) > 0, too. When −2 < x < 2, a similar thing
happens, and when x = 2 or x = −2 one of the terms is 0 so f = 0. Therefore,
what we were trying to prove has to be true.
What is there to criticize in this proof? First of all, is it correct? Is it
precise? Logical? Clear? Where is it unclear? Try to identify the statements,
both linguistic and mathematical, that are even slightly unclear, and try to
amend them appropriately. Without pointing out any of the individual errors,
we offer below a much better, clearer proof of the fact above.
Proof. We begin by factoring the function f (x) by considering it as a quadratic
function in the variable x2

f (x) = (x2 )2 − 8x2 + 16 = (x2 − 4)2

Next, we can factor x2 − 4 = (x + 2)(x − 2) and rewrite the original function as

2
f (x) = (x + 2)(x − 2) = (x + 2)2 (x − 2)2

Now, for any real number x, (x + 2)2 ≥ 0 and (x − 2)2 ≥ 0, since a squared
quantity is always nonnegative. A product of two nonnegative terms is also
nonnegative, so f (x) = (x + 2)2 (x − 2)2 ≥ 0, for any value of x.
What are the differences between the first “proof” and this second proof? Does
your rewritten proof look like this second one, as well?
One of the critiques of the first “proof” is that it does not fully explain the
situation where −2 < x < 2; rather, it merely says that something “similar”
happens and does not actually carry out any of the details. This is a common
situation in mathematics (where some steps of a proof are “left to the reader”)
and it is a convenient technique that can sometimes avoid tedious arithmetic/al-
gebra and make reading a proof easier, faster, and more enjoyable. However, it
should be used sparingly and with caution. It is important, as a proof-writer, to
make sure that those steps do work, even if you are not going to present them
in your proof; you should consider providing the reader with a short summary
or hint as to how those steps would actually work. Also, a proof-writer should
try not to use this technique on steps that are crucial to the ultimate result of
the proof.
In this particular case, the actual steps of factoring are skipped completely
and the analysis of the case where −2 < x < 2 is only mentioned in passing, yet
these are essential components of the proof! It is such a short proof, anyway,
that showing these steps does not represent a great sacrifice in brevity or clarity.
Again, this brings up the point of proof-writing as an art, as much as a science:
1.2. EXPOSITION EXHIBITION 31

choosing when to leave some of the verification of details to the reader can be
tricky. In this particular instance, showing all of the steps is important.
That being said, though, the second proof we showed here is much clearer.
Moreover, it completely avoids the case analysis that appears in the first “proof”!
There was an issue of clarity with one of the cases in the first “proof”, but rather
than simply expound the details in the amended version, we opted to scrap that
technique altogether and use a shorter, more direct proof. Now, this is not to
say that the technique of the first proof is incorrect. Were we to fill in the
gaps of the argument of the first “proof”, we would obtain a completely correct
proof. However, some of the steps in that technique are redundant. Notice that
the cases where −2 < x < 2 and where x > 2 are actually identical, in a sense:
the factors satisfy (x − 2)2 > 0 and (x + 2)2 > 0 in both cases. In fact, this is
true of the first case, where x < −2, as well! So why separate this argument
into three separate cases when the same ultimate observation is applied to all
three of them? In this case, it is best to combine them into one (also using the
knowledge that when x = 2 or x = −2, one of the factors is 0). Again, using
that expanded technique is certainly not incorrect; rather, it just adds some
unnecessary length to the proof.
We mentioned the term “case” and the phrase “case analysis” in the above
paragraphs without properly defining or explaining what we mean. For now,
we want to postpone a discussion of these terms until we thoroughly discuss
logic in Chapter 4. If you’re itching for immediate gratification regarding this
issue, though, you can skip ahead to Section 1.4.4 and check out the “Hungarian
friends” problem, which contains some intricate case analyses.

1.2.3 Pick Logic

We have used the word “logical”, and its associated forms, quite frequently,
already, without fully explaining what we mean by it. We realize this seems to
go against the precision and clarity that we have been so strongly advocating
thus far, but we also have to admit that, unfortunately, it is extremely difficult
to provide a thorough definition of logic.

Games
If you’re looking for a decent heuristic understanding of logic, try thinking about
it in terms of “logic puzzles” like Sudoku or Kakuro. These puzzles/games
are built around very specific rules that are established and agreed upon from
the very beginning, and then the solver is presented with a starting board and
expected to apply those rules in a rigorous manner until the puzzle is solved. For
instance, in Sudoku, remembering the conditions that each digit from 1 through
9 must appear exactly once in every row, column, and 3 × 3 box allows the
solver to systematically place more and more numbers in the grid, continually
narrowing down the large number of potential “solutions” to find the unique
answer that the starting grid of numbers yields. An important aspect of this
solving process is that at no time is it necessary (or wise, at that) to guess;
32 CHAPTER 1. WHAT IS MATHEMATICS?

every step should be guided by a rational choice given the current situation and
the established rules of the puzzle, and within that framework, the puzzle is
guaranteed to be solvable (given enough time, of course).
Mathematical logic is a little different, in some respects, but the essence is
the same: there are established rules of how to play the game and every move
should be guided by those rules and current knowledge, and nothing else. This is
what we mean when we say that writing mathematical proofs should be governed
by logic: every step, from one truth to another, should follow the agreed-upon
rules and only reference those rules or already-proven facts. The “game” or
“puzzle” that we’re playing in a proof (and in mathematics, in general) is not
as clear-cut as a Sudoku puzzle. Even more confusing, though, is the idea that
sometimes we start playing an unwinnable game and don’t realize it!
This idea of an “unwinnable game” is an astounding, surprising, and down-
right powerful conclusion of the work of mathematician Kurt Gödel, a 20th
century Austrian logician. His Incompleteness Theorems address an inherent
problem with strong logical systems: there can be True statements that aren’t
provable within that system. We are unable to provide a thoroughly detailed
explanation of some terms here (namely, logical system and provable), but hope-
fully you see that there is something weird going on here. How could this be
possible? If something is True in mathematics, can’t we somehow show that it
is true? How else would we know that it is true?

Some Mathematical History

To begin to address these natural questions, let’s step a little further back in
time and discuss the beginnings of logic as a full-fledged branch of mathematics.
One thing to keep in mind throughout this discussion is that we can’t completely
address every topic that comes up, and that this may feel dissatisfying, and we
understand that. Part of the beauty of mathematics is that learning about any
one topic brings up so many other questions and concepts to think about, and
these can be addressed, as well, with more mathematics. Context is important,
though, and for the context of this book, we just don’t have the time and
space to address all of these tangentially-related topics. We are not trying to
hide anything from you or sweep some issues under the rug; rather, we’re just
dealing with the reality of making sure we’re not forcing you to read 10,000
pages on the entire history of mathematics just to get our point across!
You will probably study many of the people that we mention below (and the
work they did) further along in your mathematical careers. At that point, you’ll
have a deeper understanding and appreciation for the subject built by hands-
on experience with the material, and you’ll be better-equipped to tackle the
issues therein. For now, we are merely introducing these people out of interest.
Mathematics has a rich and interesting history, and it helps to be aware of it!
Here, we will try to present a concise yet meaningful interpretation of logic—its
history, motivations, and meaning—that fits with the current context.
The mathematicians and philosophers in the mid- to late-19th century who
first studied the ideas that would evolve into modern logic were interested in
1.2. EXPOSITION EXHIBITION 33

many of the same issues we are trying to investigate here: How do we know
something is True? How can we express that truth? What types of “somethings”
can we even declare to be True or not? Breaking down mathematical language
to its very roots, these mathematicians studied ways to combine a fixed set of
symbols in very specific ways to create more complicated statements, but in the
grand scheme of things, these statements were still rather simple. This is not
meant to be a knock against their efforts; one must start somewhere, after all,
and these people were working from the ground up.
One of the first major efforts was to investigate the foundations of arithmetic,
or the study of the natural numbers (1, 2, 3, 4, . . . ). Much like Euclid sought
to study geometry by establishing a short list of accepted truths, or axioms,
and then derive truths from these given assumptions, Italian mathematician
Giuseppe Peano established a set of axioms for the natural numbers, while
others approached the topic from a slightly different viewpoint. Meanwhile,
this newfound appreciation for being rigorous and decisive about truths and
proving those truths led David Hilbert and others to bring up some issues with
Euclid’s axioms, specifically the parallel postulate.
This work on geometry and arithmetic naturally led into further, intricate
study of other areas of mathematics and fervent attepmts to axiomatize fields
like analysis of the real numbers. Karl Weierstrass, in studying this topic, pro-
duced some mind-blowing examples of functions with strange properties. For
instance, try to define a continuous function that is not differentiable anywhere.
(If you’re unfamiliar with these terms from calculus, don’t worry about it; suf-
fice it to say, it’s difficult.) Finally, Richard Dedekind was able to establish a
rigorous, logical definition of the real numbers, derived entirely from the natural
numbers, and not dependent on some vague, physical notion that a continuum
of numbers must exist.
Later on, this study branched off slightly into the study of sets, or collections
of objects. The groundwork for much of this area was laid by Georg Cantor
towards the end of the 19th century. He was the first to truly study the theory
of infinite sets, establishing the controversial idea that there are different “sizes”
of infinity. That is, he showed that some infinite sets are strictly bigger than
other infinite sets. This idea was so controversial at the time, that he was hated
by many other mathematicians! Nowadays, we realize Cantor was right. (This
also gives you a flavor of what we’ll discuss later in Section 7.6. Take this as an
intriguing example: the set of odd integers and the set of even integers are the
same size, sure, but they are both also the same size as the set of all integers.
However, the set of all real numbers is strictly bigger!)
Indeed, some mathematicians were quite shocked by Cantor’s discoveries,
and even the great Bernhard Riemann thought the development of set theory
would be the scourge of mathematics (at first, anyway). This was not the case,
though, and it has flourished since then, with many mathematicians working on
ways to represent all of mathematics in just the right way and understand the
“foundations” of mathematics. In a way, you can think of set theory as the study
of the basic objects that all mathematicians are working with, ultimately, in a
way similar to the fact that all of chemistry is done by appropriately combining
34 CHAPTER 1. WHAT IS MATHEMATICS?

elements of the periodic table in more and more complicated ways.

A further development from these topics was the study of symbolic logic,
which is a bit more concrete than the abstract ideas we’ve mentioned so far,
and whose basic ideas we will be studying frequently in the beginning chapters
of this book. This area covers how we can combine mathematical equations
and symbols with language-based symbols and connectors to make meaningful
mathematical statements that are able to be confirmed as true via a proof.
This is an incredibly important component of mathematics, in general, and
this book, in particular. Individual viewpoints are certainly more nuanced and
specific than this, but, in general, most mathematicians are of the mindset that
there are many mathematical truths out there waiting to be discovered and we
spend our time learning about the truths we have already uncovered with the
hopes of exposing even more of those truths. It’s like a giant archaelogical dig,
whereby studying the bones and artifacts we’ve already unearthed will help us
to predict what kinds of other treasures we will find and where, which tells us
where to look and how to dig once we get there. In a way, logic is that process
that is abstracted from the digging by one step: logic is the study of the digging
process. It tells us how we can actually take our mathematical knowledge and
learn from it and combine it with other knowledge to prove further truths from
that.
Now, this is not a precise analogy, mind you, and the study of abstract logic
is far more complicated and intricate. For our purposes, in this book, though,
this is a sufficiently reasonable way to think of logic. We will learn about some
of the first principles and basic operations of symbolic logic and apply this
knowledge to our study of writing proofs. It will help us to actually understand
what a proof even is, it will help guide the construction of proofs that we want
to write, it will allow us to critique proofs that may be incorrect, and it will
ultimately help us understand just how mathematics works, as a whole.

Applications of Logic: Theoretical Computer Science

One very important application of the ideas and results of logic is in the develop-
ment and study of computer science, particularly theoretical computer science
and computability theory. This particular branch of mathematics was initially
motivated by one of David Hilbert’s Twenty-Three Problems: this was a list
of famously unsolved conjectures in the world of mathematics at the time of
their publishing, in 1900. Problem number ten dealt with solving Diophan-
tine Equations, which are equations of the form
a1 xp11 + a2 xp22 + a3 xp33 + · · · + an xpnn = c
where a1 , a2 , . . . , an and c are fixed, given constants, p1 , . . . , pn are fixed natural
numbers, and x1 , x2 , . . . , xn are variables that are left to be determined so that
they make the equation true.
Given an equation like this, one might wonder whether there are any solu-
tions at all and, if so, just how many there are. Furthermore, if we’re given
that the fixed constants ai and c are all rational numbers, we might wonder
1.2. EXPOSITION EXHIBITION 35

whether we can ensure that there is a solution where all of the variables xi are
also rational numbers. Some theoretical results have been established regarding
this particular problem, but Hilbert’s tenth problem, as stated in 1900, asked
whether there was “a process according to which it can be determined in a finite
number of operations” whether there is a solution to a given equation where all
of the variables xi are rational numbers. They didn’t have a proper notion or
definition of this term at the time, but what Hilbert was asking for was an algo-
rithm that would take in the values of the constants ai and c and output True
or False depending on whether there exists a solution with the desired property.
An important part of his question was that this “process” takes a finite number
of steps before outputting an answer.
A student at Cambridge in the United Kingdom by the name of Alan Turing
began working on this problem years later by thinking of a physical machine
that would be executing the steps required to output an answer to the posed
problem. Some subsequent publications of his described his invention, what we
now call a Turing Machine, which is an interesting theoretical device that could
be used to answer some problems in formal logic, but also represents many of the
ideas that go into building modern computers. We say it’s a theoretical device
because the nature of its definition ensures that it is not physically feasible to
build and operate, but it handles some theoretical problems quite well, including
the aforementioned tenth problem of Hilbert. More specifically, this machine
gave rise to a proper definition for what we mean when we say that something is
computable, or able to be determined in a finite number of steps, and this helped
to establish a proper notion of an algorithm. It would be unfair of us to discuss
these topics without also mentioning Alonzo Church, who was working on similar
problems at the same time as Turing. Their names, together, are placed on the
Church-Turing thesis which relates the work of the Turing machine to the more
theoretical, formal logic-based notion of computability.

What Will We Do with Logic?

While all of these topics in set theory and logic are inherently interesting and
immensely important to mathematics, in general, we simply don’t have enough
time and space to discuss them in detail. Instead, let’s focus a bit more on
the notions of logic that we’ll be using in writing and critiquing mathematical
proofs.
We will consider: (1) what kinds of “things” we can actually state and prove,
(2) how we can combine “things” that we know to be true to produce more
complex truths, and (3) how we can explain how we arrived at the conclusion
that those “things” are, indeed, True. For lack of a better term, we say “things”
since we don’t yet have a formal definition of mathematical statement, which
is really the type of “thing” that we will be proving. In essence, a mathematical
statement is a combination of symbols and sentences from the languages of
mathematics and English (in this book, at least) that can be verified as either
True or False, but not both or neither. A proof, then, amounts to arranging
a sequence of steps and explanations that use true mathematical statements
36 CHAPTER 1. WHAT IS MATHEMATICS?

and sentences to connect these truths together and yield the desired truth of a
specific statement at the end. Our study of logic will deal with just how we can
combine those steps and guarantee that our proof leads to the correct assesment
of truth at the end.
More specifically, we will examine what a mathematical statement really
is and how we can combine them to produce more complicated statements.
The words and and or will be particularly important there, since those two
words allow us to combine two mathematical statements together in new and
meaningful ways. We will also look at conditional mathematical statements,
which are statements of the form “If A, then B” or “A implies B”. These
are really the bread and butter of mathematical statements and a majority of
important mathematical theorems are of this form. These statements involve
making some assumptions or hypotheses (contained in the statement A), and
using those assumed truths to derive a conclusion (contained in the statement
B). Look back at the statement of the Pythagorean Theorem in Section 1.1.1
and notice how it is in the form of a conditional statement. (Could it be written
another way? Try writing the statement of the theorem in a non-conditional
form and think about whether it is an inherently different statement in that
form. Find another famous mathematical theorem that is in the form of a
conditional statement and try doing the same change of format.)
Another important idea in mathematics, and one that will show up all the
time in proof-writing, is the concept of a variable. Sometimes we want to talk
about a type of mathematical object in generality without assigning it a specific
value and this is accomplished by introducing a variable. You have likely seen
this happen all the time in your previous study of mathematics, and we’ve even
done it already in this book. Look again at the Pythagorean Theorem statement
in Section 1.1.1. What do the letters a, b, c represent? Well, we didn’t state it
explicitly, but we know that these are positive real numbers that represent the
lengths of the three sides of a right triangle. What triangle? We didn’t mention
a specific one or point to a specific drawing or anything like that, but you knew
all along what we were talking about. Moreover, the proofs we examined didn’t
depend on what those values actually are, merely that they are positive real
numbers with certain properties. This is incredibly useful and important and,
in a way, it saves time since we don’t have to individually consider all possible
right triangles in the universe (of which there are infinitely-many!) and can
reduce the whole idea into one compact statement and proof.
One thing we can do with variables is quantify them. This involves making
claims about whether a statement is true for any potential value of a variable, or
maybe for just one specific value. For instance, in the Pythagorean Theorem,
we couldn’t claim that a2 + b2 = c2 for any positive real numbers a, b, c; we
had to impose extra assumptions on the variables to obtain the result we did.
This is an example of universal quantification: “For all numbers a, b, c with
this property and that property, we can guarantee that . . . ” Similarly, we can
quantify existentially: “There exists a number n with this property.”
Can you think of a theorem/fact that we have examined so far that uses
existential quantification? Look again at the proof that there are irrational
1.2. EXPOSITION EXHIBITION 37

numbers a and b such that ab is rational. Notice that this claim we proved is
of the existence type: we claimed that there are two such numbers with the
desired properties, and we then proceeded to show that there must, indeed,
be those numbers. Now, the interesting part of that proof was that it was
nonconstructive; that is, we were able to prove our claim without saying what
the numbers a and b actually are, explicitly. We narrowed it down to two choices
but never made a claim as to which one is the correct choice, merely that one
of the pairs must work.

1.2.4 Obvious Obfuscation

As a preview of these logical concepts that we’ll be examining in mathematical
detail later on, let’s take some real-world, language-based examples of these
ideas.

Conditional Statements
First, let’s investigate conditional statements. Mathematical theorems fre-
quently take the form of a conditional statement, but these types of statements
also appear in everyday language all the time, sometimes implicitly (which can
only add to the confusion). For instance, people talk sometimes about what
they would do with their lottery winnings, saying something like
If I win the lottery, then I will buy a new car.
The idea is that the second part of the statement, after the “then”, is dependent
on the first part of the statement, which is associated with the “if”. When the
conditions outlined in the “if” part are satisfied, the actions outlined in the
“then” part are guaranteed to take place.
The part of a conditional statement associated with the “if” is known as
the hypothesis (or sometimes, more formally, the antecedent). The part
associated with the “then” is known as the conclusion (or, more formally, the
consequent).
Sometimes the conclusion of the conditional is more subtle, or the verb tenses
in the sentence are such that it doesn’t even include the word “if”. Take the
following quote from the film Top Gun, for example:
It’s classified. I could tell you, but then I’d have to kill you.
The idea here is that the first part, “I could tell you”, is a hypothesis in disguise.
The sentence “If I told you, I would have to kill you” would have the same logical
meaning as the actual film quote; however it doesn’t convey the same forceful,
dramatic connotations. It’s quite common to actually not include the word
“then” in the conclusion of a conditional statement; while reading the sentence,
you might even add the word in your mind without realizing. Take the following
lyrics from a song by the band The Barenaked Ladies, say:
If I had $1,000,000, we wouldn’t have to walk to the store.
If I had $1,000,000, we’d take a limousine ’cause it costs more.
38 CHAPTER 1. WHAT IS MATHEMATICS?

Both lines are conditional statements, but neither includes the word “then”; it
is understood to be part of the sentence.
Compare these examples to the following sentence and see what’s different:
I carry an umbrella only if it is raining.
The idea here is that the speaker would hate to carry an umbrella around for
no good reason, preferring to make sure it would be of use. Does this sentence
have the same meaning as the following, similar sentence?
If I am carrying an umbrella, it is raining.
In modern language usage, the notion of conditional can be a little fuzzy. The
first sentence could be interpreted to mean that sometimes it might be raining
but the speaker forgets to bring an umbrella, say. The second sentence is a
clear assertion of a conditional statement: seeing me walking around with an
umbrella lets you necessarily deduce this is because it’s raining. In mathematics,
we associate these two sentences and say they have the same logical meaning.
This motivates the meaning of the phrase “only if” and, subsequently, the
phrase “if and only if”. Consider the following two sentences:
I will buy a new car if I win the lottery.
I will buy a new car only if I win the lottery.
The first one says that winning the lottery guarantees I will buy a new car,
whereas the second one says that the act of buying a new car guarantees that
it is because I just won the lottery. If both of these sentences are true, then the
events “winning the lottery” and “buying a new car” are equivalent, in a sense,
because the occurrence of each one necessarily guarantees the occurrence of the
other.
Accordingly, mathematical definitions commonly use the phrase “if and
only if ”. For example, we might write “An integer is even if and only if it is
divisible by 2.” This indicates that knowing a number has that property allows
us to call it “even”, and knowing a number is even allows us to conclude the
divisibility property. (Sometimes, though, a definition will just use if, with the
only if part left unstated but understood. You may have noticed that we did
this with the definition of prime numbers in Section 1.1.2.)

Creating More Conditional Statements from Others

Starting with a conditional statement, we can modify it slightly to produce
three other conditional statements with the same content but different structure.
Continuing to use the “lottery/car” example, let’s consider the following four
versions of the original sentence:
1. If I win the lottery, then I will buy a new car.
2. If I bought a new car, then I won the lottery.
3. If I don’t win the lottery, then I won’t buy a new car.
1.2. EXPOSITION EXHIBITION 39

4. If I didn’t buy a new car, then I didn’t win the lottery.

How do these sentences compare? Do any of them have the same logical meaning
as each other? Are all of them True, necessarily, assuming the truth of the first
one? We would argue that, in this case, sentence two could be False, even if
the first sentence is True. Perhaps I got a hefty raise at work or inherited some
money and decided to buy a new car. What about sentences three and four,
though? Can they be associated with the others somehow? We will leave this
for you to discuss and explore on your own. It might be interesting to ask the
same questions of some of the other conditional statements we’ve looked at and
see if your answers are different, too.
One final example of a conditional statement we’ll mention comes from a
joke by standup comedian Demetri Martin.

I went into a clothing store and a lady came up to me and said, “If
you need anything, I’m Jill.” I’ve never met anyone with a condi-
tional identity before. “What if I don’t need anything! Who are
you?”

This should give you a flavor for the ways that conditional statements in
modern language can be imprecise or subtle, and sometimes open to interpre-
tation. In mathematics, we want these types of statements to be rigorous,
well-defined, and unambiguous. This is something we will investigate further
later on in Section 4.5.3. For now, though, it might help to think of these types
of statements in the rigorous way in which a computer algorithm would inter-
pret an if . . . then statement. When the conditions of the if part are satisfied,
the subroutine is executed, and they are ignored otherwise. Likewise, a while
loop is merely a sequence of if . . . then statements condensed into one, concise
form.

Quantifiers
Next, let’s examine some examples of quantifiers. We will use quantifiers when
there is an unknown variable meant to be an object drawn from a collection
of possible values or representations. For instance, when we quantified the
variables a, b, c in the statement of the Pythagorean Theorem, they were drawn
from the collection of real numbers that represent the side lengths of right
triangles. For a non-mathematical example, consider the following sentence:

Every person is loved by someone.

What are the variables here? How are they quantified? Be careful because, yes,
there are in fact two quantifications in this sentence, one for each of two separate
variables. In both cases, the variables represent a member of the collection of
all people in the world, and the first variable is quantified universally while the
second one is quantified existentially. That may sound confusing, so let’s try
rewriting the sentence more verbosely:
40 CHAPTER 1. WHAT IS MATHEMATICS?

For every person x in the world, there exists another person y with
the property that person y loves person x.
Do you see how this has the same logical meaning as the first sentence?
Surely, this one is unnecessarily wordy and precise for a conversation, but we
present it here to show you the underlying variables and quantifiers. The key
phrases for the quantifiers are “for every” (universal) and “there exists” (exis-
tential).

The Order of Quantification Matters!

Now, let’s look at a similar sentence as the last example:
Someone is loved by every person.
This sentence is quite similar to the one above; it has all of the same words,
even! What did the change in word order do to the logical meaning of the
sentence? Well, there are still two variables and two quantifiers, one universal
and one existential, but the order in which those quantifiers are applied has
been switched. The verbose version of this sentence reads:
There exists a person x with the property that, for every person y
in the world, person y loves person x.
This has a completely different meaning from the first sentence! The first
one seemed believable but this one is almost outlandish. This should give you
an idea of how important it is to keep the order of quantification straight so
that you are actually saying what you mean to be saying.

Nested Quantifiers
The following examples illustrate how our brains can sometimes process quanti-
fiers in language-based sentences fairly quickly and easily, even when the inter-
connectedness might make it difficult to understand. When quantifiers follow
one after the other, we call them nested.
The ability to analyze and udnerstand such sentences might depend on the
context of the sentence and the message it is trying to convey. If the message
makes sense and we believe it, it can be easier to grasp. The best example we
know of this phenomenon is embodied by the following quote, attributed to the
great presidential orator Abraham Lincoln:
You can fool some of the people all of the time, and all of the people
some of the time, but you can not fool all of the people all of the
time.
There are quantifiers all over the place here! We are talking about the
collection of all people and the collection of instances in which certain people
are fooled, and quantifying on those collections. Try to write this sentence with
a few different wordings to see if it can sound any “simpler” or more concise.
1.3. REVIEW, REDO, RENEW 41

Could there be another way to phrase the sentence that would remove some (or
all) of the quantifiers without altering the meaning?
Finally, out of personal interest and to inject a bit of humor, we’ll mention a
similar quote that comes from Bob Dylan’s song “Talkin’ World War III Blues”,
from his 1963 album The Freewheelin’ Bob Dylan:

Half of the people can be part right all of the time

Some of the people can be all right part of the time
But all of the people can’t be all right all of the time
I think Abraham Lincoln said that

We will discuss these topics in greater detail later on, where we will examine
their mathematical motivations, meanings, and uses. For now, we can’t stress
enough how important these issues are in writing proofs. Stringing together
a bunch of sentences with no way of knowing how they’re connected is not a
proof, but a properly-structured series of logical statements and implications is
exactly what we’re looking for.

1.3 Review, Redo, Renew

Thus far, we have sought to motivate and explain mathematical reasoning and
proof-writing from a logical standpoint but, along the way, we have used some
mathematical concepts and techniques with which you may or may not be fa-
miliar. It is important, of course, to think logically and rationally when doing
mathematics, but this is only part of the bigger picture. We have tried to ex-
plain how to organize mathematical ideas and structure them in a meaningful
way that can convince others of a specific fact, but those ideas must contain
some mathematical concepts related to that fact!
For instance, we couldn’t have looked at any of the proofs of the Pythagorean
Theorem without having a rudimentary understanding of geometry: what a
triangle is, some basic properties of triangles and lines and angles, etc. What
else did we assume the reader would understand? Many of the steps involved
arithmetic, like manipulating multiple equations by multiplying through by the
same factor or subtracting two equations, and so on. Those ideas may be second-
nature to you now, but at some point you had to learn these things and see why
and how they actually worked so that you could safely and appropriately use
them in the future.
Look back over some of the other proofs we looked at in the previous sections.
What mathematical ideas did we use? Try to write down a few and think about
when and how you learned about them. Try to write down some specific facts
that we may have used without explicitly saying so and think about why we
would do that. Also, try to find a few instances where we made a claim but
didn’t necessarily fully explain why it must be True, and try to do that. For
instance, in “Proof 1” of the Pythagorean Theorem, we drew four identical
triangles inside a square and then said that the figure inside would also be a
square. Is this really True? How can we be so sure? Try to prove it!
42 CHAPTER 1. WHAT IS MATHEMATICS?

Presumed Knowledge
The main point is that we can’t actually write proofs without imbuing them with
some meaningful mathematical content. Accordingly, one of the main goals of
this book is to share some interesting mathematical facts with you. Sometimes,
this involves working with objects you already know about and have seen before
(like triangles or prime numbers) and trying to do new things with them. Other
times, we may be introducing you to completely new mathematical objects (like
equivalence relations or binomial coefficients) and working with those. What
we’d like to do now is discuss some mathematical objects and concepts that
we will use rather frequently and that you might have seen before. We aren’t
necessarily assuming that you’ve seen all of these, but none of these ideas are
too hard to learn/relearn quickly, and they will be quite useful throughout the
remainder of this book, and the remainder of your mathematical life, as well!
We’ve included a handful of problems for you to work on to give you some
practice, both throughout this section and at the end of it.

1.3.1 Quick Arithmetic

We won’t be expecting you to multiply six digit numbers in your head or any-
thing like that, but being able to manipulate “small” numbers via addition,
subtraction and multiplication is an important skill. Sure, calculators and com-
puting programs can be helpful, but we hope that it isn’t necessary to run off to
Maple or Mathematica or your TI-89 whenever we need to add a couple of four
digit numbers, say. Technology provides us with many conveniences in the form
of accuracy and time-efficiency but when we rely on these devices too heavily,
we diminish our ability to verify those answers we get (in the event of a typo or
missed keystroke, for instance) and when we use them too frequently, we may
not actually save any time at all!
We encourage you to continually try to perform any arithmetic steps we face
either in your head or on a piece of scrap paper. It will be fairly infrequent that
any problems/puzzles involve arithmetic with “large” numbers and even when
they do, there may be a special trick that can reduce the problem to something
easier. For instance, try to work on the following series of problems and see
what you notice.
Problem 1.3.1. For each of the following multiplications, try to identify the final
digit of the resulting number. If your answer is “0” then try to identify how
many zeroes are at the end of the resulting number.
1. 1 · 2 · 3 · 4 · 5
2. 1 · 2 · 3 · · · · · 10
3. 1 · 2 · 3 · · · · · 25
4. 1 · 2 · 3 · · · · · 100
5. 1 · 2 · 3 · · · · · 1000
1.3. REVIEW, REDO, RENEW 43

6. 1 · 2 · 3 · · · · · 10000

7. 1 · 2 · 3 · · · · · 109

Try to write down a few sentences that would explain to a friend the procedure
you used above. That is, given any number n, explain how to identify the
number of zeroes at the end of the number resulting from multiplying 1·2·3·· · ··n
What did you notice? Did you use your calculator for the first few? Surely
that would work, or you could even do the first two or three by hand, but how
did that help you later on? How did that help you to explain your procedure?
Certainly, you needed to find a more general way of figuring out how to tackle
this problem, and resorting to a calculator or computer might help you in some
cases, but it won’t provide you with any insight into the answer.
If you haven’t figured out a general procedure, we’ll give you this:
Hint: Think about how many multiples of 2 and how many multiples of 5
appear in the multiplication. Try to pair them together. (Why would you want
to do that?)

1.3.2 Algebra Abracadabra

The term algebra has a couple of meanings in the mathematical world with
some different nuances to each. Usually, the term brings to mind manipulating
equations with variables and trying to find a numerical solution for them. This
is probably how you handled word problems in a high school algebra class. More
generally, though, abstract algebra is a branch of mathematics that studies
certain mathematical structures that have specific properties, oftentimes having
no relation to integers or real numbers.
Much of the groundwork for this field was laid by mathematicians before the
19th century who were seeking roots of polynomial equations where the variables
were raised to the third or fourth or even higher powers. For a quadratic equa-
tion (containing a variable raised to the second power), you probably remember
the formula for finding a root of the equation (i.e. a value for the variable
that will make the expression evaluate to zero); this is the famous Quadratic
Formula.
Did you also know there is a procedure for finding the root of an expres-
sion involving a variable that is cubed? Or even raised to the fourth power?
Interestingly enough, the mathematicians working on this general problem dis-
covered that there are no such procedures possible for higher powers! A lot of
the concepts and structures they were working with developed into some inher-
ently interesting mathematics, and people have been studying those objects ever
since, eventually branching off and working with the underlying properties of
those objects, stripping away the numerical context of finding roots of equations.
This is usually what a mathematician means when he/she says “algebra”.
In this particular context, though, we will be using “algebra” in the sense
that you’re likely thinking of it: manipulating multiple equations and variables
44 CHAPTER 1. WHAT IS MATHEMATICS?

to obtain numerical values for the variables that make the expressions evaluate
to numbers that satisfy all of the equations involved. There is actually a rich
and wonderful theory behind solving systems of linear equations, but this type
of in-depth study is better suited for a course on matrix algebra (also called
linear algebra). For now, we’ll look at a couple of handy tricks and then let you
practice them.

Solving Systems of Linear Equations

A system of linear equations is just a collection of equations involving a certain

number of variables (all raised to the first power, hence linear ) multiplied by
coefficients and added together, and set equal to some constants. There are
specific conditions on the coefficients and constants that guarantee whether or
not a solution exists (and whether there are infinitely many or just one, in fact)
but we won’t get into those specific details. Suffice it to say that the systems of
equations we will have to handle in this book will have unique solutions, and this
means that the number of equations we have will be the same as the number
of variables involved. Knowing that ahead of time, how do we manipulate a
system of equations to find that unique solution?
In practice, the most time-efficient way of solving a system depends on the
coefficients and constants, as well, and perhaps spotting a particular way of
applying the methods we are about to explain. That said, simply following
these methods will always work in a short amount of time, anyway, so don’t be
too concerned with finding the absolute shortest method in any given case.
Method 1: The first method involves a system of two equations and two
unknowns. In this case, we can use one of the equations to express one variable
in terms of the other, then substitute this into the second equation, yielding
one equation and one unknown. From that, we can find a specific value for one
variable, and substitute this back into the first equation to find a specific value
for the other variable, thereby obtaining the solution we wanted. Let’s see this
process in action with a particular example. Consider the system of equations
below:

7x + 4y = −2
−2x + 3y = 13

Following the method we just described, we would rearrange the first equation
to write y in terms of x
1
y = (−2 − 7x)
4
then substitute this into the second equation

1
−2x + 3 · (−2 − 7x) = 13
4
1.3. REVIEW, REDO, RENEW 45

and solve that new equation for x:

3 21
−2x − − x = 13
2 4
29 29
− x=
4 2
x = −2

Then, we would use this value in the first equation and solve for y:

7 · (−2) + 4y = −2
4y = −2 + 14 = 12
y=3

Therefore, the solution we sought is (x, y) = (−2, 3).

What if we had used the value of x we found in the second equation instead of
the first? Well, it would produce the same value of y, but maybe the arithmetic
would have been slightly quicker. Or, what if we had done it the other way
around, and expressed x in terms of y, solved for y, then went back and solved
for x? Again, we would have found the same solution, but perhaps the numbers
would work out more “nicely” and save us a few seconds of scratch work. This
is what we mean by not worrying about finding the most “efficient” method.
Sure, there are multiple ways to approach this system of equations, but they
ultimately stem from the same method (substitute and solve) and yield the same
solution.
Method 2: An alternative way to handle a system of two equations and two
unknowns is to multiply both equations through by particular values and then
add them together, choosing those multipliers appropriately so that one of the
variables is eliminated. Using the example from above, we could multiply the
first equation by 2 and the second equation by 7, making the coefficient of x
in both equations equal yet opposite; then, adding the equations reduces the
system to one equation and one unknown, just y. Observe:


2 · 7x + 4y = −2


7 · −2x + 3y = 13
14x + (−14x) + 8y + 21y = −4 + 91
29y = 87
y=3

From there, we can substitute this value into the first or second equation and
solve for x.
You can use either of these approaches to handle any system of two equations
and two unknowns. Perhaps one would be slightly quicker than the other,
depending on the numbers involved, but you won’t be saving more than a minute
either way, so we encourage you to just choose one and work through it.
46 CHAPTER 1. WHAT IS MATHEMATICS?

Method 3: It can sometimes be convenient to interpret these systems of equa-

tions graphically; this is not usually an efficient way of identifying a specific
solution to the system, but it can give an indication of whether a solution exists
and a rough estimate of the magnitude of the values of the solution.
With two unknowns, we can interpret an equation like ax + by = c in terms
of a line in the plane by rearranging: y = − ab x + cb . This is the line with slope
− ab and y-intercept cb . Given two such equations, we can draw two lines in
the plane and locate the point of intersection visually. The (x, y) coordinates
of that point are precisely the solution we would find by solving the system of
equations as we described above.
This visual method also applies to a system of three equations and three
unknowns, but this requires drawing lines in three-dimensional space. This can
be difficult to do in practice, but is technically achievable. These same concepts
also apply to larger numbers of equations and unknowns, but drawing “lines”
in four or more dimensions is impossible for us human beings to visualize!
More than two variables: Reduce! The next part of this method builds
upon the first by taking a system of more than two equations (and unknowns)
and continually reducing it to smaller systems, eventually obtaining a system
of two equations and two unknowns, where we can apply the first part of the
method. We will illustrate the method by considering a system of three equa-
tions and three unknowns, like the one below:

6x − 3y + z = −1
−3x + 4y − 2z = 12
5x + y + 8z = 6

The first goal is to eliminate one of the three variables. In essence, this can
be done in one of two ways, much like the method for two equations and two
unknowns. Let’s say we’re going to try to eliminate z from the system; we can
try to express z in terms of x and y and substitute somehow, or we can multiply
some equations and add them together to cancel the coefficients of z. The only
difference here is that, whichever option we choose, we need to do it twice. Let’s
use the first equation to write

z = −6x + 3y − 1

After substituting this expression for z into both the second and third equations,
we will have a system of two equations and two unknowns.
One way to think about this is that we need information from all three
equations to ultimately arrive at an answer, and in reducing the system to two
equations, we need to somehow retain information from all three of the original
equations. The expression we have for z came from the first equation, so we
need to subtitute it into the other two to retain all of the information we need.
Compare this to the following sequence of steps: rearrange the first equation
to isolate z and substitute this into the second equation, then rearrange the
second equation to isolate z and substitute this into the first equation. What
1.3. REVIEW, REDO, RENEW 47

happens? The intuition is that we have somehow “lost” information from the
third equation and, yes, we will obtain a system of two equations and two
unknowns, but it will have insufficient information to yield a unique solution
for x and y. If you actually perform the steps we just described (try doing this
to check our work), you obtain the following “system” of two equations after
minimal simplification:
9x − 2y = 10
9
x−y =5
2
These are really the same equation! Accordingly, we wouldn’t actually be able
to solve them for unique values of x and y.
Let’s return to where we were and substitute the expression for z above into
the second and third equations
−3x + 4y − 2 · (−6x + 3y − 1) = 12
5x + y + 8 · (−6x + 3y − 1) = 6
and then simplify
9x − 2y = 10
−43x + 25y = 14
Applying one of the methods from the first problem will give us the solution
(x, y) = (2, 4). Having both of these values in hand, we can now return to any
one of the original three equations and solve for z; better yet, we can just use
the isolated expression for z we found already from the first equation:
z = −6x + 3y − 1 = −6 · (2) + 3 · 4 − 1 = −12 + 12 − 1 = −1
More than two variables: Reduce another way! Another way to reduce
a system from three equations to two equations is related to the “multiply and
add” method from before, but we still have to be careful about ensuring that
we retain information from all three equations. Using the same system of three
equations from above, we might notice that after multiplying the first equation
by 8 and the second equation by 4, the coefficient of z in all three equations is
either ±8. This allows us to add/subtract the equations in a convenient way to
reduce the system to two equations and two unknowns. Specifically, let’s do the
multiplication we just mentioned
48x − 24y + 8z = −8
−12x + 16y − 8z = 48
5x + y + 8z = 6
and then add the first equation to the second
(48x − 12x) + (−24y + 16y) + (8z − 8z) = −8 + 48
36x − 8y = 40
48 CHAPTER 1. WHAT IS MATHEMATICS?

and the second equation to the third

(−12x + 5x) + (16y + y) + (−8z + 8z) = 48 + 6

−7x + 17y = 54

This produces two equations involving only x and y; furthermore, we have

combined information from all three original equations to produce these, so we
can be assured that we haven’t “lost” anything. Solving this new system

36x − 8y = 40
−7x + 17y = 54

via any of the previous methods we discussed produces the solution (x, y) =
(2, 4). Substituting these values into any of the three original equations and
solving for z produces the ultimate answer we sought.
We could have performed similar steps to eliminate y from the system of
three equations, too; for instance, we could add 4 times the first equation to
three times the second, and subtract 4 times the third equation from the second.
Any of these methods would produce the same ultimate answer, but some of
them may shorten the arithmetic steps or involve “nicer” numbers (i.e. fewer
fractions, smaller multiplications, what have you). Solving a system with more
equations amounts to the same general procedure: multiply the equations and
add to eliminate one variable from the system and continue doing this until
there are only two equations and two unknowns; then, solve for the values of
those two variables and work backwards, substituting those values to solve for
the values of the variables that had been eliminated.

Algebra Practice
Problem 1.3.2. Solve the following system of equations for (x, y, z):

x + y + z = 15
2x − y + z = 8
x − 2y − z = −2

Now, solve this similar system for (x, y, z):

x + y + z = 15
2x − y + z = 9
x − 2y − z = −2

Compare the changes in the values of x, y, and z between the two systems.
Which variable changed the most? The least? What is the ratio of these
changes?
How large/small can you make this ratio by changing the constant on the right-
hand side of the second equation of the system?
1.3. REVIEW, REDO, RENEW 49

Problem 1.3.3. A father, mother and son were sitting in a restaurant eating
dinner, when they were approached by another family consisting of a father,
mother and son. This second family was struck by their resemblance to the
first family, so the second father asked the first, “How old are the three of you?
I’m guessing we are all about the same age”. The first father happened to be
a mathematician and did not feel like giving away his family member’s ages so
easily, and thus “revealed” them to the others in a tricky way. He said, “Well,
our current ages combine to make 72, and I happen to be six times as old as
my son. However, in the future when I am just twice his age, our combined age
will be twice our current combined age. How old do you think we are?”
How old are the three family members?

1.3.3 Polynomnomnomials
Sometimes we will need to work with variables that are squared or cubed or
raised to even higher powers. In general, a polynomial is the term we use for
a function that has one or more variables raised to integer powers, multiplied
by coefficients, and added together. Here are some examples of polynomials:
1 2
x2 − 7x + 1, 7p6 + 5p4 + 3p2 + 2p, z + 9y 2 z − 2y + z 3 y 2 − 7z
2
These types of functions are quite common and popular in mathematics,
partly due to their convenient properties and partly due to their prevalence in
nature. We will see them appear throughout this book. For now, though, let’s
focus on polynomials that only have one input variable.

Roots of Polynomials
Sometimes, we will define a polynomial function in the context of a puzzle and
wonder whether there are any values for the input variable that make the ouput
value 0. These input values are called roots of the polynomial.
One way to identify roots of a polnyomial is to factor it into linear terms;
that is, we try to express the function as a series of multiplications instead of
additions, since we can declare that (at least) one of the factors must be 0 to
achieve a 0 value. The motivation behind this technique relies on the following
fact:
Fact: If a and b are real numbers and ab = 0, then a = 0 or b = 0
(or possibly both).
Example 1.3.4. Let’s see a specific example. Let’s try to factor the following
polynomial:
p(x) = x2 + 6x + 8
(It is common notation to define polynomials as p(x), where p stands for poly-
nomial, x is the input variable, and p(x) is the output value corresponding to
the input value x. This doesn’t have to be the case, though.)
50 CHAPTER 1. WHAT IS MATHEMATICS?

You might notice that

p(x) = x2 + 6x + 8 = (x + 4) · (x + 2) = (x + 4)(x + 2)

(It is also fairly common to drop the · when there are factors separated by
parentheses, so we will adopt that convention from here on out, as well.)
The reason this factorization works is because we are applying the distribu-
tive property multiple times, in reverse. If we were to expand the factorization
we just found, explicitly showing every step, it would look like:

p(x) = (x + 4)(x + 2)
= x(x + 2) + 4(x + 2)
= (x2 + 2x) + (4x + 8)
= x2 + 2x + 4x + 8 = x2 + 6x + 8

All we really did to write down the factorization was to notice that the terms
+4 and +2 have product +8, which is the constant term, and they have sum +6,
which is the coefficient on the x term. Knowing how the subsequent expansion of
those factors would work out allows us to write down that factorization without
really checking it.

Factoring Quadratics
Let’s take what we did with that specific example and try to generalize to any
quadratic function. If we want to factor a quadratic polynomial like

p(x) = x2 + bx + c

we seek values r and s so that r · s = c and r + s = b. Usually, we can do this

“by inspection”, or by just staring at these two equations and thinking for a
minute to come up with the appropriate values. (That’s what we did with the
last example!)
What do we do if the coefficient of the x2 is not 1 but some other number a?
Well, notice that if we can factor the polynomial p(x) 2 b c
a = x + a x+ a , then we can
find a factorization of the original polynomial p(x), as well, by just multiplying
by a. This won’t affect our ability to find roots of the polynomial (our original
goal), because we’re assuming a 6= 0 (otherwise we didn’t really have a quadratic
polynomial to begin with and wouldn’t need to factor it). Once we’ve found
this factorization, it’s easy to identify the roots of p(x); since we want to figure
out when p(x) = 0, we can just use the factorization and the fact we mentioned
above to conclude that

0 = p(x) = (x + r)(x + s) implies x + r = 0 or x + s = 0

which implies x = −r or x = −s

That is, the roots must be −r and −s.

1.3. REVIEW, REDO, RENEW 51

What if we have a polynomial of the form p(x) = x2 − a2 ? This particular

type of function is known as a difference of squares, and has a quick fac-
torzation trick. This is a quadratic polynomial so, following the method from
above, we would seek values r, s such that rs = −a2 and r + s = 0 (since there
is no x term in p(x)). The second constraint tells us r = −s and using this in
the first constraint tells us r2 = a2 . Thus, using r = a and s = −a achieves the
factorization p(x) = (x − a)(x + a) and so the roots are ±a. (Notice that using
r = −a and s = a also satisfies both constraints, yet it actually yields the same
factorization of p(x).)
Similar tricks can sometimes be applied to polynomials of higher degree (re-
call that “degree” means the highest power of the input variable). For instance,
the following polynomial has degree 4

p(x) = 4x4 − x2 − 3

but we can factor it easily if we define y = x2 and write it as a quadratic

polynomial
p(y) = 4y 2 − y − 3 = (4y + 3)(y − 1)
Notice that you can think about the factorizations of the coefficients of the
y 2 , y, and constant terms to jump right to the factorization we found, or you
can follow the division trick we mentioned. Here, we would want to factor
p(y) 2 1 3 3 1
4 = y − 4 y − 4 , so we need rs = − 4 and r + s = − 4 ; using r = −1 and
3
s = + 4 works, so we obtain the factorization
 
p(x) 3
= (y + (−1)) y +
4 4

which can be simplified as

 
3
p(x) = 4(y − 1) y + = (y − 1)(4y + 3)
4

which is exactly what we had before.

A Root Yields A Factor

Of course, this trick of identifying roots can work in reverse, too: if we can easily
spot a root of a polynomial, that can help us in identifying one of the factors.
As an example, look at the cubic polynomial below and see if you can find a
root “by inspection”; that is, see if you can find an input value for x that will
make p(x) evaluate to zero:

p(x) = x3 − 3x + 2

If you haven’t spotted it yet, you might want to try plugging in some “easy
values”, like the first few integers (both positive and negative) to see what
happens. If you do so, you’ll find that p(1) = 1 − 3 + 2 = 0. Accordingly, we
52 CHAPTER 1. WHAT IS MATHEMATICS?

know that a factorization of the polynomial p should include the factor (x − 1),
since this corresponds to the root x = 1. Knowing this, we want to divide p(x)
by the factor (x − 1) so that we can further factor that quotient and identify all
of the roots of p.

Polynomial “Division”
How do we divide polynomials, though? We seek another polynomial q(x) so
that p(x) = q(x) · (x − 1), or in other words, we need to find p(x)
x−1 . One way to
identify such a function is by using the same principles of long division that
you learned all about in middle school when you were dividing integers. The
same concepts apply to polynomial functions! Think back to how long division
works, and try working through some basic examples— like 22 ÷ 7, say—to
refresh your memory about how it works.
Now, let’s try to apply those same principles to polynomials. Here’s an
3
example of the idea of long division applied to x −3x+2
x−1 :

x2 + x − 2
x3


x−1 − 3x + 2
− x3 + x2
x2 − 3x
− x2 + x
− 2x + 2
2x − 2
0
In each iteration of the method, we try to find the largest “factor” that can
“go into” the larger term. In this case, these are just multiples of powers of
x; we identify the largest power of x that can “go into” the current term in
question. Since the dividend has x3 and the divisor has x, we write x2 above
the division line. Then, we multiply (x−1) by x2 , write this below the dividend,
and subtract to find the remainder.
We repeat the same process until we have a constant term above the division
line (i.e. a multiple of x0 ) and see the remainder. Since the remainder here is
0, we know that we have a factorization with no remainder. We can then factor
the resultant quadratic by noticing that r = 2 and s = −1 satisfy r + s = 1 and
rs = −2, so we can finally write
p(x) = (x − 1)(x − 1)(x + 2) = (x − 1)2 (x + 2)
Accordingly, the roots of p(x) are x = 1 and x = −2. For this function, the
degree of the polynomial is 3 but the function has only 2 roots. Does this strike
you as odd? Can you think of a polynomial of degree 3 that has only 1 root?
How about a polynomial of degree 3 with no roots? What about 4 roots, or 5 or
more? Are any of these possible? Why or why not? What if we were working
with a polynomial of degree 4? Of degree n? What can you say for sure about
the number of roots a polynomial has, relative to its degree?
1.3. REVIEW, REDO, RENEW 53

Expanding Factors
Sometimes, when working on a puzzle, we start from a factorization of a polyno-
mial and want to expand the factors completely so we can identify the coefficient
of a particular term. How can we quickly and easily multiply polynomials to-
gether? In essence, we are trying to apply the distributive law over and over
without having to write out all of the steps (although that thorough, step-by-
step procedure is guaranteed to work, so if you are unsure of your answer, it is
always a good idea to go back and check each step thoroughly).
One particular instance where we can reduce the number of steps involved
is when we need to expand a factorization like (a + b)n , where a and b represent
any constant or variable and n is an integer. In this specific situation, there is
a convenient way to identify the coefficients of the expanded polynomial, and
those values come from Pascal’s Triangle.
This is an arrangement of rows of integers into a triangular shape, where
each row corresponds to a particular value of n in such an expansion. The trick
to generate Pascal’s Triangle is to write the first two rows as all 1s, and the
outside “legs” of the triangle as all 1s. In the interior of the triangle, any entry
is filled in by finding the sum of the two entries immediately above that entry,
to the left and to the right. Try generating the first few rows of the triangle
yourself and compare to the one below to make sure you’ve done the procedure
correctly.
n = 0: 1
n = 1: 1 1
n = 2: 1 2 1
n = 3: 1 3 3 1
n = 4: 1 4 6 4 1

We’ve written the n values on the left side to indicate the correspondence with
the original problem of expanding (a+b)n . In general, any term of the expansion
will be some coefficient (taken from the triangle) times ak bn−k for some value
of k between 0 and n; that is to say, in every term of the expansion, the sum of
the powers of a and b in that term must be n. The numbers in any given row
of the triangle are written in an order corresponding to decreasing powers of a,
so that the first 1 is the coefficient of an , the next number is the coefficient of
an−1 b, and so on.
If we were faced with expanding (a + b)2 , we would read the n = 2 row of
Pascal’s Triangle and see that the coefficients should be 1, 2, 1, and that these
are the coefficients for a2 , ab, b2 , respectively. Thus,

(a + b)2 = a2 + 2ab + b2

which we could have also accomplished fairly easily by just expanding by hand.

4
What if we were faced with expanding x2 + 2 , say? This isn’t done as quickly
by hand, so let’s see what happens if we use Pascal’s Triangle. The n = 4 row
54 CHAPTER 1. WHAT IS MATHEMATICS?

tells us the coefficients of a4 , a3 b, a2 b2 , ab3 , b4 are 1, 4, 6, 4, 1, respectively, where

a = x2 and b = 2. Thus, we can write

4
4
3
2
x2 + 2 = 1 · x2 + 4 · x2 · 2 + 6 · x2 · (2)2
+ 4 · x2 · (2)3 + 1 · (2)4
= x8 + 4 · x6 · 2 + 6 · x4 · 4 + 4 · x2 · 8 + 16
= x8 + 8x6 + 24x4 + 32x2 + 16

Try performing this expansion step-by-step and compare, too. There are actu-
ally some very interesting properties of Pascal’s Triangle that are deeply rooted
in some other mathematical concepts, and these properties are particularly use-
ful in the field of combinatorics. We will, in fact, examine many of these
properties in greater detail later on! For example, you might wonder why it is
the case that this procedure—adding the two entries above—yields entries that
correspond to expanding factors like this. We will prove that it works when
we discuss the Binomial Theorem and its related ideas! (See Section 8.4.4 if
you’re curious.)

Completing the Square

There is one more polynomial-related trick we need to mention before deriving

an important result. Sometimes, it is useful to rewrite a polynomial as a squared
term plus a constant term, so that we can separate the variables and constants
in a convenient way. This amounts to adding and subtracting a particular term
so that, overall, we have added 0 to the polynomial, but the term is chosen in a
way that lets us rewrite the terms of the polynomial conveniently. This process
is known as completing the square, in the sense that we add a term to create
a squared factor, and complete the polynomial by subtracting a corresponding
amount.
Let’s try this procedure with an example and then attempt to generalize.
Start with the following polynomial:

p(x) = x2 + 8x + 9

A factorization isn’t immediately apparent here, so let’s try to complete the

square. We want to see a term like (x + a)2 , where we know the coefficient
of x is 1 since the polynomial has 1 · x2 . Expanding a term like that gives
x2 + 2ax + a2 . Since we need 8x to appear, we should use a = 4. This expansion
gives x2 + 8x + 16, but we really want to see +9 as the constant term, so let’s
add and subtract 7 from the original polynomial:

p(x) = x2 + 8x + 9 + 7 − 7 = (x2 + 8x + 16) − 7 = (x + 4)2 − 7

Does this look familiar? Precisely, it’s a difference of squares, and we know how
1.3. REVIEW, REDO, RENEW 55

to factor that:
√ 2
p(x) = x2 + 8x + 9 = (x + 4)2 − 7 = (x + 4)2 − 7
 √  √ 
= x+4+ 7 x+4− 7

√ √
Accordingly, the roots of this polynomial are x = −4 − 7 and x = −4 + 7.
Let’s generalize! Suppose we start with a quadratic polynomial of the form

p(x) = ax2 + bx + c

and, to complete the square, we want to add and subtract a particular term.
How did we find that term before? Well, the expansion of a term like (rx + s)2
yields r2 x2 + 2rsx + s2 , and to match these coefficients with the coefficients√of
the original polynomial, we see that we need r2 = a, so we should use r = a.
(Notice that this requires a ≥ 0, of course! What should we do if a < 0?) Then,
b
to have 2rs = b, we need s = 2r = 2√b a . Then, when this is expanded we have
2
b
added on s2 = 4a , so we should subtract that from the polynomial.
These steps are performed below, with some extra algebraic cleanup, of sorts,
to make the terms look “nicer”:

b2 b2
p(x) = ax2 + bx + c = ax2 + bx + +c−
4a 4a
2  2
√
 
b b
= ax + √ + c−
2 a 4a
2 
√ b2
  
b
= a· x+ + c−
2a 4a
 2  2

b b
=a x+ + c−
2a 4a

This now tells us how to complete the square, given any quadratic polynomial!

Visualizing Completing the Square

Here’s a helpful way to remember how to do this process. It’s based on a visual
representation of the areas of squares and rectangles.
Let’s suppose that a, b > 0 so that we can geometrically interpret ax2 + bx
as the area of a rectangle. Specifically, let’s√take the ax2 term to be the area of
a square. This means each side has length a · x:
56 CHAPTER 1. WHAT IS MATHEMATICS?

√
a·x

√
a·x

How might we represent the bx term? We want to build this square into a larger
square; this is what completing the square means. So, we should build some
rectangles around this square that will help us proceed towards that goal. Let’s
split the area contributed by the bx term into√ two rectangles, each with area
b
2 x. Since we must have one side with length a · x, and we want the total area
b
to be 2 x, then we see that we need the other length to be 2√b a :

b b
2x
√
2 a

√ b
a·x ax2 2x

√ b
a·x √
2 a

What do we need to add to make this a square? We see that there is just a tiny
square piece left to fill in the upper-right corner. Each side is length 2√b a , so its
b2
area—the term we need to add on—is 4a .

b b b2
2x
√
2 a 4a

√ b
a·x ax2 2x

√ b
a·x √
2 a

Look at that! This is the same term we produced above with our algebraic
derivation. By adding that on, we were able to factor the terms as a perfect
square. We just needed to make sure to subtract it off, as well, so that the net
change to the original expression is zero.
1.3. REVIEW, REDO, RENEW 57

This is a helpful trick to keep in mind. It can remind you about both the
motivating process for completing the square, as well as how to achieve it. One
thing you should ponder, though: Why is it that this visual representation
works? We had to assume a, b > 0 to be able to draw these diagrams, so why is
it that the general formula works no matter what a and b are?

The Quadratic Formula

Let’s returning to the question of identifying the roots of a polynomial. Specifi-
cally, let’s recall the quadratic formula. You may have memorized this formula
as a way to “solve quadratic equations” but do you know why it actually works?
Let’s try to figure it out! In general, we start with a quadratic polynomial of
the form
p(x) = ax2 + bx + c
where a 6= 0 (otherwise, it’s not actually quadratic), and we want to identify
the values of x such that p(x) = 0. (Did you try to answer our questions above
about how many roots this type of polynomial can have? Keep those concepts
in mind throughout the following derivation.) We can’t hope to factor the
polynomial into linear factors too easily, so let’s take advantage of the process
we used above: completing the square. The benefit of that procedure is that we
can set p(x) = 0 and rearrange the terms after completing the square to solve
for x. Observe:
2 
b2
 
b
0 = p(x) = ax2 + bx + c = a x + + c−
2a 4a

simplifies to:
2
b2

b
−c=a x+
4a 2a
Now, we want to to start “undoing” the processes here to solve for x, and this
b2
would require taking the square root of boths sides. But what if 4a − c < 0?
2
b
We couldn’t take that square root at all! Or what if 4a − c = 0? Is that
b2
a problem? Do we have anything to worry about when 4a − c > 0? These
are issues that are related to the questions we had before about the possible
number of roots a polynomial can have. You may have deduced (correctly) that
a quadratic polynomial can have at most two roots, but here we have uncovered
the possibility (and reasons why) that a quadratic polynomial may have one or
zero roots!
2
b
• In the case where 4a − c < 0, then no value of x can possibly satisfy the
last line in the derivation above. Therefore, there would be no roots of
p(x) in the set of real numbers.
2
b
• In the case where 4a − c = 0, then taking the square root of both sides of
the last line above is perfectly valid, but it will produce exactly one value
58 CHAPTER 1. WHAT IS MATHEMATICS?

of x:
2
b2

b
−c=0=a x+
4a 2a
b
0=x+
2a
b
x=−
2a
2
b
The remaining case is when 4a − c > 0. Here, we can expect two roots
of p(x) because taking the square root of both sides introduces two possible
2
solutions. In general, when we √
have a situation
√ like s = t, we can say that the
only possible solutions are s√= t and s = − t but we must consider both (we
usually write this as s = ± t). Solving for x in that case yields
2
b2

b
−c=a x+
4a 2a
r
b2 − 4ac √ √
 
b b
± = a x+ = ax + √
4a 2a 2 a
√
b b − 4ac √
2
− √ ± √ = ax
2 a 4a
√
b b2 − 4ac
− ± √ =x
2a 4a2
Now, we need
√ to be careful about the square root observation we made before.
In general, 4a2 = ±2a, but we already know that the fractional term involving
that square root already has an associated ±1 factor, so this factor won’t change
that. Therefore, we can conclude
√ √
b b2 − 4ac −b ± b2 − 4ac
x=− ± =
2a 2a 2a
Voilà, the quadratic formula!
Remember that the final steps of the derivation were carried out under the
b2 b2
assumption that 4a − c > 0. Does this formula still apply when 4a − c = 0?
Could we have performed the same steps as we did immediately above while
operating under that assumption? Why or why not?

Problems
Problem 1.3.5. Find all possible values of a so that x−a is a factor of x2 +2ax−3.
Problem 1.3.6. Find all possible values of b so that x3 + b is divisible by x + b
with no remainder.
Problem 1.3.7. Factor xn − 1 for any natural number n.
1.3. REVIEW, REDO, RENEW 59

Problem 1.3.8. Determine the value of x defined by

s r
√
q
x = 2 + 2 + 2 + 2 + ···

Hint: try to express the infinitely-nested square roots by using x, itself.

Problem 1.3.9. Use completing the square to prove that the sum of a positive
number n and its reciprocal is always greater than or equal to 2, and that the
only number that makes the sum equal to 2 is n = 1.
Hint: take the sum, add and subtract 2, and rearrange.
Problem 1.3.10. How can we find the roots of a quartic polynomial of the form
ax4 + bx2 + c?

1.3.4 Let’s Talk About Sets

We’ve mentioned some particular types of numbers already, but we want to
specifically define the sets of numbers we will be working with in the future.
Each of these collections of numbers is represented by a particular letter in the
blackboard bold font. The natural numbers (also known as whole numbers
or counting numbers) are so-called because they feel “natural” to say as we start
counting objects. We can write

N = {1, 2, 3, 4, 5, . . . }

(There is a more specific and technical definition that we will explain later on.)
We use N to stand for “natural”.
Using N, we can define a related collection of numbers: the set of all inte-
gers, which combines the natural numbers, 0, and the negative natural numbers.
We can write
Z = {. . . , −3, −2, −1, 0, 1, 2, 3, . . . }
The letter Z comes from the German word Zahlen, meaning “number”.
From this set, we can define the collection of rational numbers. These
numbers can be represented as a ratio of integers, but they don’t seem to have
a natural “listing” like the sets N and Z, so we can’t write this set in the way
that we did above. For this, we use a very common set notation, as follows:
na o
Q= | a, b ∈ Z and b 6= 0
b
We read this as:

“The set of rational numbers is the set of all numbers of the form ab ,
where a and b are both integers and b is nonzero.”

This conveys the necessary information that a rational number is a fraction,

where the numerator and denominator are integers (but the denominator can’t
60 CHAPTER 1. WHAT IS MATHEMATICS?

be 0 because division by 0 is disallowed). The reason we use the letter Q for the
rational numbers is because R was already reserved for the real numbers and Q
was the next previous letter available. Also, Q contains all of the quotients of
integers, so that makes sense, too!
The real numbers R have a very technical definition that we, unfortunately,
cannot delve into completely in this book. (That just goes to show how difficult
it is to mathematically define that set!) For now, one way to think of the real
numbers is via a number line. The real numbers are all numbers that lie on the
line, while the numbers of N, Z and Q are specific numbers that lie on the line,
but they don’t comprise the entirety of the line. In a way, R is the “completion”
of Q, in the sense of “filling in the gaps” between rational numbers.

1.3.5 Notation Station

A popular and convenient way of writing sums and products is to use a short-
ened notation that collects many terms or factors into one common form. For
instance, what if we wanted to talk about the sum of the first 500 natural num-
bers? It would be tedious to write out all 500 terms of the sum, so it is common
to write something like 1 + 2 + 3 + · · · + 499 + 500. (We’ve already used ellipses
like this, in fact. Did you understand what we meant?) This is popular and
does get the point across, but some mathematicians take offense to the unneces-
sary use of ellipses in the middle. We put off talking about this issue until now
because it’s often the case that notation can be difficult to learn and compre-
hend. Rather than bombard you with new symbols right away, we appealed to
our intuitive understanding of what “. . . ” accomplish.
Now, that we’ve brought it up, let’s see how to avoid using ellipses. To write
the sum we mentioned above, we would use the following notation:

500
X
1 + 2 + 3 + · · · + 499 + 500 = i
i=1

P
The large sigma comes from the Greek letter corresponding to S, for “sum”,
and the index i tells us to find the values
P of the individual terms of the sum.
Writing i = 1 below and 500 above the sign means that we let i assume all
of the natural number values between 1 and 500 (inclusive). Using those values,
we substitute into the general expression for the term, which is just i, in this
case. Accordingly, we find that the terms are 1, 2, 3, . . . , 500, as desired. Try
to find a few other ways of writing this sum by altering the expression for the
general term and/or the values of the index. What if we wanted to find the
sum of the first 500 even natural numbers? What about all of the even natural
numbers up to (and including) 500? Try to write those sums in the notation
style above.
Q
Related to this is the notation. If we wanted to look at the product of the
first 500 natural numbers, we would follow the same conventions of identifying
1.4. QUIZZICAL PUZZICLES 61

values for the index and the general term:

500
Y
1 · 2 · 3 · · · · · 499 · 500 = i
i=1
Q
The large pi comes from the Greek letter corresponding to P, for “product”.
Again, try expressing this in a different way by changing the general term and/or
index values. What if we wanted to find the product of the first 500 even natural
numbers? What about all of the even natural numbers up to (and including)
500? Try to write those products in the notation style above.

Problems
Problem 1.3.11. Write an English sentence that describes what the following
equation means:
n
X n(n + 1)(2n + 1)
i2 =
i=1
6
Problem 1.3.12. Express, in appropriate notation, the sum and product of the
first n powers of 2, starting with 20 = 1. Can you prove a formula for the sum?
The product?
Problem 1.3.13. Consider the sum of all the odd numbers between 17 and 33
(inclusive). Write this sum in summation notation where the index starts at 0.
Now try writing it where the index starts at 1. Now try writing it where the
index starts at 8, and again with 9. Which of these feels “more natural”? Why?

1.4 Quizzical Puzzicles

Let’s put into action some of the principles we have discussed so far. Specifi-
cally, let’s examine some interesting mathematical puzzles and explain how to
go about solving them. None of these involves knowing anything beyond ba-
sic algebra and arithmetic, but this does not mean they are “basic” or “easy”,
since they all involve critical thinking skills and keen insight to solve and under-
stand. Along the way, we will be employing some logical ideas we have brought
up already. We might have to work with polynomial functions, or solve some
equations algebraically. We might have to think careful about order and flow
of our arguments, making sure everything follows from previous knowledge or
deductions. Overall, we should also be thinking about what constitutes a good
and valid proof of the facts we discover!

1.4.1 Funny Money

Problem Statement
This classic puzzle is contained in a story about some friends paying for a shared
hotel room:
62 CHAPTER 1. WHAT IS MATHEMATICS?

Three friends are on a road trip and stop at a hotel late one night
looking for a room to catch up on some rest. The clerk on duty
says that there is only one room vacant for that night and it costs
$30 for the three of them, if they want to squeeze in together. The
friends decide they are desperate for sleep and agree to split the
room amongst them, each placing a $10 bill on the counter to pay
up front. The clerk thanks them, hands them the key, and they
head off to grab their bags from the car. Meanwhile, another clerk
shows up to start his shift and realizes that the previous clerk had
made an error and overcharged the three friends for their room: it
should have only been $25. He takes a $5 bill from the register,
hands it to the bellhop on duty and says, “Bring this to Room #29.
The guests there were overcharged.” The bellhop nods and heads
off to their room. When the three friends answer the door, they are
surprised and happy to discover they earned a refund. To split the
money fairly, one friend makes change with five $1 bills and then
each friend takes $1, giving the remaining $2 to the bellhop as a
friendly tip. The bellhop thanks them kindly and heads back to
work.
Now, each of the three friends contributed $9 to the room, plus a $2
tip, making a total of $29. But they originally gave the clerk $30. . .
What happened to the missing dollar?!

Think carefully about this before turning the page and reading our solution.
1.4. QUIZZICAL PUZZICLES 63

Solution: Keeping Careful Track of the Money

Did you figure it out? Did you realize there is really nothing “missing”? This
puzzle is intended to confuse the reader and mislead them into searching for
something that isn’t really there. The numbers involved are chosen so that
the “missing sum” of $1 is so small that the reader believes that something
mysterious happened, but careful and logical analysis of the events should lead
you to realize that the question at the end is not really a fair one; it is based
on a misinterpretation of the situation, and trying to ignore its reasoning is the
key to figuring out the solution to this puzzle. When the numbers are changed
greatly so that the final discrepancy is of a much larger value, the reader no
longer has that emotional investment to seek out that “missing dollar”.
First, let’s analyze what actually happened in this particular case. The key
is careful interpretation of where the money actually goes. It helps to forget
about the individual people involved and think of two distinct entities: the
group of friends, who we’ll call F , and the clerk/bellhop combo from the hotel,
who we’ll call H. Now, let’s replay the steps of the story and describe where
the money comes from and goes to at each step:

(1) F arrives and gives $30 to H (original cost of room)

(2) H gives $5 back to F (refund for overcharge)

(3) F gives $2 back to H (tip to bellhop)

(4) Net change: F gave $30-$5+$2=$27 to H

It makes more sense now, doesn’t it? The refund was $5, so the room actually
cost $25 and it doesn’t make sense to say that the three friends each paid $9
for it, plus the tip to the bellhop. That $27 contribution from the three of
them includes the tip. The question after the story implies that the we should
be adding the tip to the friends’ contribution but, really, it is part of their
contribution. By grouping the friends together and the bellhop/clerk together,
we can actually track down how the money changes hands.

Generalizing: Changing The Numbers

Let’s change the problem in the way that we mentioned above; specifically, let’s
try to change the numbers of the problem to remove that emotional attachment
to the “missing dollar” and make the discrepancy much larger. To begin, we
will define some variables to represent the dollar amounts used in each of the
steps outlined above. We could try to approach this problem by “testing out”
specific values for these dollar amounts and seeing what happens but it will be
more efficient to essentially “try everything at once” by introducing variables
and substituting specific values for them later on.
We will let 3n represent the original cost of the hotel room (the amount paid
by the three friends when they first arrived), for some value of n. We choose
this because we want the cost to be evenly split by the friends. Next, we want
64 CHAPTER 1. WHAT IS MATHEMATICS?

to define a variable to represent the refund they receive. Knowing that they
will want to split this amount evenly amongst the three of them and have some
leftover to tip the bellhop, let’s say that the refund is of the form 3r + 2. The
variable r represents how much each friend individually receives back from the
hotel, and the 2 represents the tip for the bellhop. Now, let’s restate the puzzle
with these variables instead of the original values.

Three friends are on a road trip and stop at a hotel late one night
looking for a room to catch up on some rest. The clerk on duty
says that there is only one room vacant for that night and it costs
$3n for the three of them, if they want to squeeze in together. The
friends decide they are desparate for sleep and agree to split the room
amongst them, each placing $n on the counter to pay up front. The
clerk thanks them, hands them the key, and they head off to grab
their bags from the car. Meanwhile, another clerk shows up to start
his shift and realizes that the previous clerk had made an error and
overcharged the three friends for their room: it should have only
been $3n − (3r + 2). He takes $3r + 2 from the register, hands it to
the bellhop on duty and says, “Bring this to Room #29. The guests
there were overcharged.” The bellhop nods and heads off to their
room. When the three friends answer the door, they are surprised
and happy to discover they earned a refund. To split the money
fairly, each friend takes $r, giving the remaining $2 to the bellhop
as a friendly tip. The bellhop thanks them kindly and heads back
to work.
Now, each of the three friends contributed $n − r to the room, plus a
$2 tip, making a total of $3(n − r) + 2. But they originally gave the
clerk $3n . . . What happened to the missing $3n − [3(n − r) + 2] =
$3r − 2?!

Do you see what happened now? The discrepancy occurs, as we explained

before, because the question considers adding the $2 tip to the refunded cost of
the room and comparing that to the original cost of $3n. The actual comparison
should be between the refunded contribution of the friends, which is $3(n − r) =
$3n − 3r, and the sum of the refunded cost of the room and the tip, which is
$3n − (3r + 2) + 2 = $3n − 3r. No discrepancy there!

Generalizing: Questions For You

In the original statement of the puzzle, the values were n = 10 and r = 1, so
that the “missing amount” was magically $3r − 2 = $1. If we had chosen larger
values—say n = 100 and r = 10—then the $300 room actually should have
cost $268, the bellhop would bring the friends $32, they would each take $10,
he would keep $2, and the discrepancy becomes $28. Would anyone actually
believe that $28 went missing in those transactions? What if we use even larger
values of n and r? How large can you make the discrepancy? How small? Given
1.4. QUIZZICAL PUZZICLES 65

a desired discrepancy, in dollars, can you find values for n and r that achieve
that value? How many ways are there to do that?

Lessons From This Puzzle

Logic and rational thinking are important when solving a puzzle because it is
sometimes easy to be misled by emotions. Had we stated the puzzle originally
as the “missing $28 problem”, would you have reacted the same way? Would
you have been as temporarily confused before trying to backtrack and discover
what really happened?

1.4.2 Gauss in the House

Problem Statement
There is a popular anecdote among mathematicians that may or may not be
apocryphal, but some of us would like to believe that it is true, because it
features one of the greatest mathematicians/physicists of all time, Carl Friedrich
Gauss. He worked in the late 1700s and early- to mid-1800s and proved some
fundamental and powerful results in a broad range of areas. He studied number
theory, and complex analysis, and optics, and geometry, and astronomy, and so
much more! Read the story below, think about what you would have done in
that situation—as a young child and now—and then read on for a discussion.

It was early in the morning in an elementary school classroom, and

the students were acting noisy and rowdy, much to the dismay of
the teacher, who was feeling quite sick and tired, literally, and quite
sick and tired of their behavior. He needed a way to occupy them
for a while so that he could relax at his desk and recuperate. He
bellowed out to the room and told them to take out their slates
and chalk. After asking a few more times, everyone had obliged.
He then told them to add together all of the numbers from 1 to
100, and that the first person to do so would earn the privelege of
being the teacher’s helper for the day. He returned to his desk and
sat down, relieved that they would be occupied for quite some time
performing large sums. After only a minute, though, one boy walked
up to the desk and showed the teacher his slate with the answer. The
teacher was astounded and had to spend a few minutes performing
the arithmetic himself to check the answer, but in the end, the little
boy was correct, and he had accomplished this feat so quickly. How
did he do it?

Think carefully about this before turning the page and reading our solution.
Remember, this story “happened” in the days before calculators, so you should
not be using anything more than your brain and a pencil and paper.
66 CHAPTER 1. WHAT IS MATHEMATICS?

Solution: Reducing Computations

Perhaps you figured this one out. There are actually a number of ways to
approach this problem that are slightly different, but they mostly amount to
the same insight: trying to reduce the number of computations required.
To naively go through and add each of the 100 numbers to the previously
obtained sum would require 99 additions, with ever larger numbers involved.
Certainly, the trick here is not to just do these additions faster than the others,
it is to be more efficient with the computations that are required. Remember
that multiplication can be viewed as repeated addition of one number with itself,
so perhaps we can reduce all of these additions to one multiplication, provided
we find the right number to add to itself over and over.
Another important fact to remember is that addition is associative and
commutative, meaning we can perform the additions in any order and be
assured that we obtain the same answer. Specifically, we can add all the numbers
from 100 down to 1 and get the same result for the sum, call it S. Let’s write
down this fact in a convenient way here:

1 + 2 + 3 + ··· + 98 + 99 + 100 = S
100 + 99 + 98 + · · · + 3 + 2 + 1 = S
101 + 101 + 101 + · · · + 101 + 101 + 101 = 2S

Notice that we have written down the desired sum in two different ways, added
those two sums entry by entry, and obtained an expression for 2S, twice the
desired sum. That new expression can be written as a multiplication because
there are 100 terms, each of which is the number 101. Thus,

2S = 101 · 100 and therefore, S = 101 · 50 = 5050

This is much faster than performing 99 additions, and in fact, if we think care-
fully, we may be able to do the entire process in our heads!

Alternate Solution: Pairing Terms

Now, a very similar way of seeing this problem is to skip adding the two lines
we wrote above and just pair off the numbers in the original sum, as follows:

S = 1 + 2 + 3 + · · · + 98 + 99 + 100
= (1 + 100) + (2 + 99) + (3 + 98) + · · · + (49 + 52) + (50 + 51)
= 101 + 101 + · · · + 101 = 50 · 101 = 5050

This approach is essentially equivalent to the one we described above; it still

takes advantage of the associative property of addition to convert the sum into
a multiplication, it just skips over the intermediate steps where we found an
expression for 2S and then divided by two.
1.4. QUIZZICAL PUZZICLES 67

Generalizing: Even n
What if the teacher had asked his students to add the numbers from 1 to 1000?
Would they have protested? Would Gauss have been able to find the answer just
as quickly? What would you do? We’re not sure about the first two questions,
but we think that you could handle this sum just as easily. The only thing
different here is that the number of pairs we create will be 500 (instead of 50),
and each of those pairs will sum to 1001 (instead of 101), so the result will be

1 + 2 + 3 + · · · + 998 + 999 + 1000 = 1001 · 500 = 500500

Does it look like there’s a pattern there? Do you think you could say what the
sum of all of the numbers from 1 to 1 million is right away without doing the
multiplication?

Generalizing: Odd n
What if the teacher had asked for the sum of the first 99 numbers instead?
Would the pairing process still work? Let’s see:

S = 1 + 2 + 3 + · · · + 97 + 98 + 99
= (1 + 99) + (2 + 98) + (3 + 97) + · · · + (48 + 52) + (49 + 51) + 50
= (49 · 100) + 50 = 4950

Notice that we had an odd number of terms in total, so we couldn’t pair off
every number, and had to add 50 to the result of the multiplication. Could we
have paired the numbers in a different way?

S = 1 + 2 + 3 + · · · + 97 + 98 + 99
= (1 + 98) + (2 + 97) + (3 + 96) + · · · + (48 + 51) + (49 + 50) + 99
= (49 · 99) + 99 = 50 · 99 = 4950

This seems more similar to the result of the original puzzle, because we ulti-
mately performed one multiplication. This may seem like a strange coincidence
now, but try to follow the steps above with some other odd sums. What is the
sum of the first 7 integers? The first 29? The first 999? The first 999999?

Generalizing: Any n
Let’s step back from the individual cases that we have examined here and try
to solve the problem in a more general sense. Let’s pretened that the teacher
had presented the students with the following problem:

Find a formula for the sum of the first n numbers. I want a specific
formula so that if someone tells me what n is, I can find an answer
quickly by plugging in that particular value.
68 CHAPTER 1. WHAT IS MATHEMATICS?

The caveat in the second sentence rules out a solution of the form given by our
investigations above. We have already come up with some simple algorithms for
finding a solution to this problem, but we have now been asked to find a forumla
that will produce a solution. How do we begin to approach this? Well, based on
some observations we made above, it would make sense to tackle this puzzle by
handling the cases where n is even and where n is odd separately. We saw that
the pairing worked out slightly differently in those cases, so let’s investigate one
and then the other. In each case, we are looking for a formula for S(n), the sum
represented by 1 + 2 + 3 + · · · + (n − 2) + (n − 1) + n. We are using this new
notation S(n) to indicate that the sum depends on that particular value of n.
If n is even, we know that we can pair off every number and have no terms
leftover:
n  n n 
S(n) = 1 + 2 + 3 + · · · + −1 + + + 1 + ···
2 2 2
+ (n − 2) + (n − 1) + n
= (1 + n) + (2 + (n − 1)) + (3 + (n − 2)) + · · ·
 n  n   n   n 
+ −1 + +2 + + +1
2 2 2 2
n n2 + n
= (n + 1) · =
2 2
Try this formula with some of the even values of n we examined above (like 100,
1000, 1000000, etc.) It works, doesn’t it? Note that the reason we can write
terms involving n2 and be assured they are part of the sum is that n is even, so
n
2 is also a whole number.
Okay, now what happens if n is odd? We know that we won’t be able to pair
off every number, so we need to be clever about what we do here. Remember our
approach with summing the first 99 numbers? By leaving off the last term of the
sum, we could pair off all of the other terms with no leftover, and futhermore,
each of those pairs summed to the same value as that last number. Let’s try
using that approach here:

n−1 n−1 n−1

   
S(n) = 1 + 2 + 3 + · · · + −1 + + + 1 + ···
2 2 2
+ (n − 2) + (n − 1) + n
n−1 n−1
   
= (1 + (n − 1)) + (2 + (n − 2)) + · · · + + +1 +n
2 2
2n − 2
 
= n + n + ··· + + 1 + n = (n + n + · · · + n) + n
2

This has shown that each of the pairs of terms sums to n, the final number that
we removed before the pairing process. Now, let’s think carefully about how
many pairs we had. Notice that we can number them by looking at the first
number of the pair: the first pair was (1, n − 1), the second pair was (2, n − 2),
and so on, and the first number in the last pair was n−1 2 . Therefore, we had
1.4. QUIZZICAL PUZZICLES 69

exactly that many pairs: n−12 . (Remember that n is odd, so we can rest assured
that n − 1 is even, so n−12 is a whole number. We haven’t been mentioning
that every time, so be sure to go back over what we’ve done so far and convince
yourself that every step and every term we write is valid.) To those pairs, we
tacked on a final number, n, so we can write the multiplication for the sum as
follows:
n−1 n−1 2 n2 + n
   
n+1
S(n) = +1 ·n= + ·n= ·n=
2 2 2 2 2

Wow, this is the exact same formula we found in the case where n is even!
Did this surprise you? It’s not obvious at all that we should end up with the
same formula, even with the similarity of the approaches to the problem. What
does this suggest to you? A mathematician would see such a “coincidence” and
wonder whether there is a much simpler and direct route to this result; that is,
is there a way we could approach this puzzle that would answer both odd and
even cases simulatenously? Since we obtained the same answer, there might be
a way to do it. Think about this for a minute before reading on.

Generalizing: Any n, without separate cases

It turns out that we already hinted at this other method in our previous discus-
sion of this puzzle. Remember when we wrote the sum forwards on one line and
backwards on another line and added them together? Well, when we treated
the odd/even cases here, we decided to avoid that method because it seemed to
add a couple of extra steps; the “pairing terms” process seemed slightly quicker
so we followed that method. What if we went back and reexamined the “add
the sum twice” method? We’d find something like this:

1 + 2 + ··· + (n − 1) + n = S(n)
n + (n − 1) + · · · + 2 + 1 = S(n)
(n + 1) + (n + 1) + · · · + (n + 1) + (n + 1) = 2S(n)

In this case, we have n terms in the sum on the third line, and each term is
(n + 1). Thus,

1 n2 + n
(n + 1) · n = 2S(n) and therefore, S(n) = (n + 1) · n =
2 2
This is the formula we have obtained already, and we found it here in a way
that didn’t depend on whether n is odd or even! (Look back at the steps we
just performed and verify for yourself that the odd/even property of n is really
irrelevant.)

Alternate Solution: Visual Diagram

Before we wrap up this puzzle, we want to mention a geometric approach to the
solution. We will relate the sum S(n) to the area of a square and find a way
70 CHAPTER 1. WHAT IS MATHEMATICS?

to draw the individual terms of the sum (1, 2, 3, . . . , n − 1, n) as portions of the

square’s area. Specifically, let’s consider an n × n square and draw the sum’s
terms as rectangles with increasing height, each with width one unit. See the
picture below:

Now, asking for a formula that yields the sum S(n) is equivalent to asking
what area is covered by all of the rectangles we have drawn inside this square.
Trying to add up the individual areas is just restating the puzzle, so we need
to think of a way to relate this area to the total area of the square. To do this,
let’s think about what is left over; that is, how can we describe the area of the
square that is not covered by the rectangles? Look at the area strictly above
the first 1 × 1 rectangle: it is also a rectangle, with dimensions (n − 1) × 1.
Look at the area above the 2 × 1 rectangle: it is a (n − 2) × 1 rectangle. This
pattern continues! Eventually, we have a 1 × 1 rectangle above the (n − 1) × 1
rectangle and then no area above the last n × 1 rectangle. What is the total
area of all of those rectangles? Well, it looks a lot like the sum S(n) we are
considering, but it is just missing the final term, n. Now, we can add up the
areas of all the rectangles by relating them to S(n) and then to the square’s
area:
n2 = S(n) + (S(n) − n) = 2S(n) − n
Therefore,
n2 + n
S(n) =
2
the same formula we had before!

Lessons From This Puzzle

Sometimes there are several completely reasonable ways to approach a puzzle
and obtain a solution. Some of them might come to mind first but take longer
to execute, some might be trickier to find but lead to a solution much more
easily, or some might just go nowhere at all! It’s often hard to tell beforehand
what’s going to happen with any particular approach, so just start trying to
work through the puzzle and see what happens, keeping track of what you’ve
1.4. QUIZZICAL PUZZICLES 71

tried and what happened so that you can reassess that approach later. This
is a fact we need to keep in mind as we advance in our mathematical careers.
We can’t always know exactly what to do right away. We are bound to get
stuck sometimes, or try roads that end up being dead ends. This shouldn’t be
discouraging; it’s just the way it is!
As a subpuzzle, try redoing the “paired terms” for the odd n case, but
instead of leaving off the last term of the sum, try separating the middle term
and pairing the numbers from the outside in. Does this give you the same
result? Does it seem any easier/faster/different than the approach we used?
Alternatively, what if we had handled the even n case by saying that n = 2k for
some number k? What would we do for odd n? Does this notation change the
procedure at all? Does it make it any easier to handle? Now, can you think of
any completely different methods to attack this puzzle?

1.4.3 Some Other Sums

Summing Odd Numbers: Observing a Pattern
While we’re on the topic of evaluating sums of integers, let’s look at some related
problems. First, we will look at an interesting geometric way of interpreting
sums of odd integers: let’s represent 1 as a 1×1 block, and then each successively
larger odd integer as a right-angled corner of 1 × 1 blocks that fits perfectly
around the previous such figure. Why would we do this? Well, since these are
all odd numbers, successive terms differ in size by two, and lengthening the sides
of our corner pieces by one each time allows us to snugly fit the corners around
each other and build successively larger squares!

Does this pattern continue? If we believe that it does, how could we prove
such a thing? What does this geometric pattern mean in terms of the numerical
sums? This is a good question to answer first, because as pretty as the geometric
pattern is, it’s difficult to work with and manipulate and, ultimately, prove
decidedly. In essence, pointing to the first few terms of the pattern and saying,
“Look, it works!” does not constitute an official, mathematical proof, so we
must find a better way of formulating this problem. This is not to downplay
the meaningfulness and beauty of the pattern we noticed; it is quite interesting
that it works out that way and it did provide us with some valuable insight into
what might be going on, mathematically, but at the end of the day, that’s all it
can do for us.
72 CHAPTER 1. WHAT IS MATHEMATICS?

Summing Odd Numbers: Proving Our Findings

Let’s try to write the sums represented by the figures above in numerical terms.
The corner pieces are made from 1 × 1 blocks, and there are two more blocks
in each corner than the previous one, so each square figure that we saw is
represented by a sum like

1 or 1+3 or 1+3+5 or 1+3+5+7

and so on. What we notice from these terms is that, indeed, they sum to square
numbers:

1 = 12 1 + 3 = 4 = 22 1 + 3 + 5 = 9 = 32 1 + 3 + 5 + 7 = 16 = 42

This is really the pattern that we want to prove; it is equivalent to the geometric
pattern we noticed before, but now it is written in terms we can manipulate.
Let’s think about how can we do this, now. Is this pattern similar to anything
we’ve seen before? Have we proved any results about sums of integers? Of
course! Look back at the previous puzzle; we proved (in a few ways, in fact)
that
n2 + n
1 + 2 + 3 + · · · + (n − 1) + n =
2
How might this be useful in this puzzle? The sum formula we proved involves
all consecutive integers from 1 to n, but for the current desired formula, we only
want to consider consecutive odd integers.
Before, we used the function S(n) to represent the sum of the first n natural
numbers, so let’s define a function T (n) to represent the sum of the first n odd
natural numbers. Now, we need to identify the terms of that sum first, and
then relate those to S(n) somehow. Below, we have written out the sums for
n = 1, 2, 3 and 4. Can you find a way to identify the largest term in the sum
and express that in terms of n?

n=1: 1, n=2: 1 + 3, n=3: 1 + 3 + 5, n=4: 1+3+5+7

Notice that the last term of the sum is always 2n−1. This is related to a general
fact, that any even integer can be represented as 2k, for some particular integer
k, and any odd integer can be represented as 2n − 1, for some particular integer
n. (We can also express an odd integer as 2n + 1 for some integer, as well,
right? In this context, it is more convenient to use the 2n − 1 form, though.)
Accordingly, we want to find a formula for the sum of the first n odd natural
numbers, given by

T (n) = 1 + 3 + 5 + 7 + · · · + (2n − 3) + (2n − 1)

Can we relate this sum to S(n) or something similar? Well, notice that the sum

S(2n) = 1 + 2 + 3 + · · · + (2n − 3) + (2n − 2) + (2n − 1) + 2n

1.4. QUIZZICAL PUZZICLES 73

contains all of the natural numbers from 1 to 2n, whereas T (n) only contains
the odd natural numbers in that range. Perhaps it would make sense to subtract
those two sums and try to find an expression for the sum of the leftover terms:


S(2n) − T (n) = 1 + 2 + 3 + · · · + (2n − 1) + 2n


− 1 + 3 + 5 + · · · + (2n − 3) + (2n − 1)
= 2 + 4 + 6 + · · · + (2n − 2) + 2n

These terms are all of the even natural numbers from 2 to 2n. How can we find
a formula for this sum? Do we need to do any extra work, or can we apply a
previously-proven result? Well, since all of the terms are even, we can divide
everything by 2 and write

1 1

(S(2n) − T (n)) = 2 + 4 + 6 + · · · + (2n − 2) + 2n
2 2
= 1 + 2 + 3 + · · · + (n − 1) + n = S(n)

and we can be assured that all of the terms in the sum on the far right are,
indeed, integers. Not only that, they are all of the consecutive integers from 1
to n, and we have a formula for that sum! Now, everything is written in terms
of formulas we already know, namely S(n) and S(2n), and the one formula that
we are seeking, namely T (n). The last step now is to rearrange the equation to
isolate T (n) and then substitute what we know about the formulas involving S:

1
(S(2n) − T (n)) = S(n)
2
S(2n) − T (n) = 2S(n)
S(2n) − 2S(n) = T (n)
2
(2n) + 2n 2 · (n2 + n)
− = T (n)
2 2
4n2 + 2n − 2n2 − 2n
= T (n)
2
2n2
= T (n)
2
n2 = T (n)

This looks rather nice, doesn’t it? Despite having to muddle through some
algebraic steps, we arrived at one of the conclusions we were hoping to prove:
that the sum of consecutive odd integers is a perfect square. Not only that,
we have managed to prove precisely how that square number is related to the
number of terms in the sum. Specifically, a concise way of summarizing the
result that we just proved is to say that “the sum of the first n odd integers is
n2 .”
74 CHAPTER 1. WHAT IS MATHEMATICS?

Alternate Solution: An Inductive Argument

Could we have proven this in a different way? What if we had not yet proven
the result from the previous section, or if we hadn’t thought to use it in that
way? Could we have somehow taken advantage of the geometric structure of
the sums that we noticed at first?
Let’s go back and think about this in a slightly different way. Specifically,
let’s see why adding one more term to a sum produces another square number.
Suppose we knew already that one of the sums produced a square number; we
know this is true for the first sum (1 = 12 ), but let’s assume that this happens
for some arbitrary number of terms, n. That is, let’s assume that

1 + 3 + 5 + · · · + (2n − 3) + (2n − 1) = n2

for some value of n. Given this as a fact, what can we subsequently deduce
about the next sum? When we add one more term to the sum, we add on the
next odd integer, 2n + 1, so let’s see how this affects the value of the sum:

1 + 3 + 5 + · · · + (2n − 3) + (2n − 1) + (2n + 1) = n2 + 2n + 1 = (n + 1)2

This seems to confirm our belief, doesn’t it? Knowing that one sum behaves in
the way we expect it to (“if the sum of the first n odd integers is n2 . . . ”) allows
us to deduce that the next sum must also behave in the same way (“. . . then
the sum of the first n + 1 odd integers is (n + 1)2 ”). Does this also prove the
result? What do you think? Does it feel strange to essentially assume our result
to prove something further about it? Is that really what we did?
This proof strategy, using one form of the result to prove something about
a “subsequent” form of the result is called mathematical induction. (In
general, the meaning of the term “subsequent” depends on the context; here,
it meanis the next sum with one more term.) We will examine this strategy
in more detail in the next chapter. For now, we will point out that this is a
perfectly sound strategy, but it is highly dependent on the fact that the first
sum behaves appropriately: 1 = 12 . That way, the work we did allows us to
deduce that the second sum behaves that way (1 + 3 = 22 ), which then alllows
us to deduce that the third sum behaves that way (1 + 3 + 5 = 32 ), and so on. . .
What if we had only been able to prove that second part, but the first sum
didn’t work out in the way that we wanted? Would we stil be able to prove the
result? What does this tell you about the induction strategy, in general? We
will address some of these issues in more generality later on.

Generalizing: Arithmetic Series

One final sum problem we want to mention is strongly related to the two that
we’ve seen so far and, in fact, if we had proven this next result first, we wouldn’t
have had to do anything more to prove the first two! In that sense, this next
result is stronger than the first two: the truth of this result implies the truth of
the first two. (This is a common notion in mathematical terminology, to label
results as stronger or weaker than others.)
1.4. QUIZZICAL PUZZICLES 75

For this result, we want to examine a general arithmetic series. This

phrase means that we’re adding a sequence of numbers where the difference in
value between successive terms is a fixed value. Another way of thinking about
this is that each term is obtained from the previous one by adding on a fixed
constant. Notice that the sums we’ve examined in the last two puzzles were
arithmetic series: in the first sum, each term differed by 1 (or, we added 1 to
each term to get the next term), and in the second sum, each term differed by
2 (or, we added 2 to each term to get the next term).
How can we represent a general arithmetic series? Knowing that successive
terms must differ by a fixed constant, let’s assign that value a variable, say c,
for constant. Now, there must be a first term in the sum, as well, so let’s assign
that value a variable, say a, since it’s the first letter. We just need one more
variable to tell us how many terms there are in the sum, so we will use k, since
we’ve used that variable before with the same meaning. Now, we can represent
the entire sum with just these three variables:

A(a, c, k) = a + (a + c) + (a + 2c) + (a + 3c) + · · · + (a + (k − 2)c) + (a + (k − 1)c)

We can use the fact that each pair of successive terms differ by c to express
the second term using the first term, a, and we can use this to express the
third term, and so on, by continually adding c. We wanted k terms in total so,
thinking of the first term as a + 0 · k, the final term will be what we obtain after
adding c to the first term k − 1 times (there are k numbers from 0 to k − 1,
inclusive). Notice, as well, that we introduced the notation A(a, c, k) to mean
“the sum of the arithmetic series with first term a, constant difference c, and k
terms”. Now, how can we figure out this sum?
Let’s employ a strategy that worked before: in the first sum puzzle, we wrote
the terms of the sum forwards and backwards and added them together. This
allowed us to create many pairs of terms that all had the same sum, reducing
the sum to a multiplication. What happens when we do that here? We see that

a + (a + c) + ··· + (a + (k − 1)c) = A(a, c, k)

(a + (k − 1)c) + (a + (k − 2)c) + ··· + a = A(a, c, k)
(2a + (k − 1)c) + (2a + (k − 1)c) + ··· + (2a + (k − 1)c) = 2A(a, c, k)

Again, we find that each pair of terms has the same sum, and in this case that
sum is 2a+(k −1)c. How many such pairs are there? There are exactly k terms,
of course! (That’s why we chose to use that variable, even.) Representing the
sum as a a multiplication, we can now deduce that

2A(a, c, k) = k · (2a + (k − 1)c)

and therefore,
k
A(a, c, k) =· (2a + (k − 1)c)
2
Does this look like what you expected for a result? Did you have any expecta-
tions? It sometimes helps to try to “guess” what might happen, and then see if
and how the results match up with your intuitions.
76 CHAPTER 1. WHAT IS MATHEMATICS?

Applying the General to the Specific

We mentioned before that the sums we examined previously were both arith-
metic series, so does this formula yield the correct value for those sums? In the
first puzzle, the values of the variables were a = 1, c = 1, and k = n; plugging
in those values yields
n n n2 + n
A(1, 1, n) = · (2 + (n − 1)) = · (n + 1) =
2 2 2
which is, indeed, what we derived. What about the second sum? What were
the values of the variables? Is the formula correct? We will leave it to you to
verify that result.

Another Representation
As a final comment for this puzzle, we want to discuss one other way of represent-
ing the formula we just derived. Look at the term in the parentheses and write it
slightly differently: a + (a + (k − 1)c). Do those terms look particularly interest-
ing? Well, they are the first and last terms of the sum, respectively. This gives
us a different way of stating the sum formula we derived: A(a, c, k) = k2 (a + b),
where a is the first term of the sum and b is the final term. This version of the
formula can be more convenient, and lets us verify some sums more quickly.
For instance, if we told you to find the sum of an arithmetic series with first
term 12 and last term 110 and 14 terms in total, you wouldn’t have to bother
figuring out the constand difference c; instead, you could simply find the sum:
14
2 · (12 + 110) = 854. Much faster, right? What is the value of c for that
arithmetic series? Is there an easy way to find c, given a and b and k?

Lessons From This Puzzle

It can be helpful to be aware of previous results, since they can make other proofs
shorter and easier. Sometimes, it’s difficult to recognize when a particular result
would be useful or, even if you recognize its use, it may be difficult to figure
out how to apply it. In this case, we recognized that we had proven a sum
formula before, so it made sense to at least try to figure out how it might be
useful in proving a different sum formula. However, there was a completely
different way to prove the formula for the sums of odd integers that didn’t
depend on our previous result. That hints at a more general result, and a curious
mathematician would try to explore the problem in more generality, which we
did by looking at an arbitrary arithmetic series. In the end, though, we used
multiple strategies for the first two sum formulas, and applied just one of those to
the general series problem. Could we have used the strategies in other settings?
Could we prove the first sum formula by induction, as well? Could we prove the
second sum formula using the forwards/backwards writing technique? Try to
use those strategies and see what happens. It may seem strange or unnecessary
to you, because we already have the results, but seeing how different techniques
work in different settings is a valuable lesson. In mathematics, it is often as
1.4. QUIZZICAL PUZZICLES 77

hard (or harder, even) to figure out which strategy to use in a proof as it is to
figure out the result to be proven. With that in mind, it’s helpful to practice
particular strategies to develop an intuition for when they will work and when
we need to try something else.

1.4.4 Friend Trends

Problem Statement
This puzzle is based on the following anecdote concerning a Hungarian sociolo-
gist and his observations of circles of friends among children.

“In the 1950s, a Hungarian sociologist S. Szalai studied friendship

relationships between children. He observed that in any group of
around 20 children, he was able to find four children who were mu-
tual friends, or four children such that no two of them were friends.
Before drawing any sociological conclusions, Szalai consulted three
eminent mathematicians in Hungary at that time: Erdos, Turan and
Sos. A brief discussion revealed that indeed this is a mathematical
phenomenon rather than a sociological one. For any symmetric re-
lation R on at least 18 elements, there is a subset S of 4 elements
such that R contains either all pairs in S or none of them. This
fact is a special case of Ramsey’s theorem proved in 1930, the foun-
dation of Ramsey theory which developed later into a rich area of
combinatorics.”
(Quoted from lecture notes by MIT Prof. Jacob Fox.)

The puzzle we now present follows the same idea but with some smaller
numbers. Specifically, we are interested in investigating the smallest size of a
group of people that necessitates a subgroup of three people that are all mutually
friends or all mutually enemies.

Assume that amongst a group of people, any two of them are either
friends or enemies, and that these are the only possible relationships
(i.e. no acquaintances or frenemies or anything like that). Take a
group of four people and try to assign a designation of friend/enemy
to each pair so that there are no groups of three people that are all
friends or all enemies. Can you do this with a group of five people?
How about six? Seven? Ten? Twenty? Try to identify a cutoff
number for the size of the group where you can be guaranteed to
find a subgroup of three people that are all friends or all enemies.

Think carefully about this before turning the page and reading our solution.
78 CHAPTER 1. WHAT IS MATHEMATICS?

Representing The Problem Effectively

Did you figure it out? This is a very tricky puzzle, so don’t feel bad if you
struggled with finding a solution. In fact, we think that investigating this puzzle
is just as important as actually finding an answer, because there are several ways
to approach this puzzle and it’s always interesting to see how different people
interpret the puzzle.

Let’s start by discussing how to even write down/draw/talk about this situ-
ation. For any of the questions posed by this puzzle, we are meant to consider a
group of people with a certain size and think about the relationship between any
two people in the group. To tackle this puzzle, we will need a way to represent
all of these relationships in an efficient and easily-interpretable way, so that we
can verify the desired property about the subgroups of size three. Specifically,
we want to easily identify whether or not there are any subgroups of size three
that are homogeneous, in the sense that the three people are all friends or all
enemies. From now on, we will refer to this as the “homegeneity property”.

How could we do this? How could we represent the people and their relatin-
ships? We could number the people in the group, then write out a list of all
pairs of numbers and label each pair with F (friend) or E (enemy). Let’s try
doing that for a group of four people:

12F 13E 14F 23F 24E 34F

Does this friend/enemy group satisfy the homogeneity property? It’s not so
easy to verify, is it? For one, the numbering makes it difficult to find subgroups
of size three, and to verify the property, we need to check all such subgroups
to make sure they are not EEE or F F F . Perhaps we should find a better way
of representing the information of the puzzle before attempting a solution. Can
you think of a more visually pleasing way of representing whether two people
are friends or enemies, for all possible pairs in the group? Specifically, we would
like to have a relatively efficient way of looking for subgroups of size three and
recognizing whether they are all friends or all enemies.

Let’s try representing each person in the group as a single dot and connecting
two people with a type of line dependent on whether those two are friends or
enemies. For example, let’s connect two people that are friends with a blue
line and two people that are enemies with a red line (and remember that any
two people are either friends or enemies and nothing else, so each pair of dots
must have some colored line between them). For example, the following diagram
depicts the relationships we assigned in the line above, with the other notation:
1.4. QUIZZICAL PUZZICLES 79

1 2

3 4

Now, what would we be looking for to verify the homogeneity property?

We want three dots (three people) so that all of the lines between them are
either blue (all friends) or red (all enemies). That’s right—we’re looking for
monochromatic triangles! (Note: we want the vertices of the triangle to be
one of the original dots we drew; that is, we don’t want a vertex to be a place
where two lines cross. Also, monochromatic comes from the Greek words monos
and khroma, meaning “one” and “color”, respectively.) This representation is
much easier to interpret visually and makes checking for a solution much faster.
Based on the diagram above, we have addressed the question regarding four
people: we have found a particular arrangement of friends and enemies so that
there are no subgroups of size three that are either all friends or all enemies.
That is, there are no subgroups of size three with the homogeneity property.
This shows that such a situation can be achieved with four people, so we are
not guaranteed to have a group with the homogeneity property amongst four
people.
Can you find another such arrangement? How can you be sure that it’s a
different arrangement than the one we’ve already seen? How many different ar-
rangements are there that satisfy the homogeneity property? Now, try drawing
an arrangment that does have a subgroup of size three with the homogeneity
property. What does that look like? How many of these arrangements are
there?

Restating the Problem for n = 5

Let’s move on and think about a group of five people. Our diagram changes
because we have five dots now, and this means there are more lines to draw.
Still, we are looking to fill in all of the connections with a blue or red line and
make sure there are no monochromatic triangles. Is this possible? (Hint: try
arranging the dots into a regular pentagon shape and then filling in the lines.)
Try to draw this a few times and see if any of your arrangements work. It may
also help to draw in a few lines randomly and then guide your choices from there
on out by making sure that you don’t create any triangles when you add a new
line.
Did you figure it out? Turn the page to see how we did it . . .
80 CHAPTER 1. WHAT IS MATHEMATICS?

Solution for n = 5
Here is our arrangment of red/blue lines amongst five dots that completely
avoids the homogeneity property:

5 2

4 3

Notice the elegant symmetry of the figure: all of the red lines are on the
outside of the pentagon, and all of the blue lines are in the interior of the shape.
The reason this works is that any triangle using three dots as vertices must use
either two outside lines and one inside line, or two inside lines and one outside
line. (Think about that: why couldn’t we use three inside lines or three outside
lines to make a triangle?) This guarantees that any triangle we draw will use two
differently-colored lines, so this figure does not have the homogeneity property!
Of course, we could look at all possible triangles inside the diagram and make
sure none of them use one color. How many such triangles are there? How
quickly could you check all of them by hand? Is it easier to do that, or to notice
the inside/outside property that we mentioned above?
Perhaps you found a solution that doesn’t look like the drawing we have.
How can you tell whether it’s actually a different figure? How many blue and red
lines are there in your diagram? In ours? Try redrawing your figure by moving
the dots around but maintaining the relationships between the dots (i.e. the
color of line drawn between any two of them). Can you make your figure look
like ours? What do you think this says about the number of solutions to this
puzzle?

What about n = 6?
Okay, now we’re ready to think about what happens when we have six people.
In terms of the dots and lines, we’re looking to draw all possible connections
between six dots with either blue or red lines and ensure that there are no
triangles with the same line type. Before you start drawing, try to think about
the solutions to this puzzle when we were working with four and five dots. What
did those solutions look like? What was the number of lines that we had to fill
in? How many will we have to draw this time? Can we try to make this figure
look like the solution for five dots? It sometimes helps to think about how a
solution to a current puzzle might be similar to previous work. Now, try to
draw this figure and see what happens.
1.4. QUIZZICAL PUZZICLES 81

Did it work? Why not? Where did you run into trouble? How many lines
could you draw before you were forced to make a monochromatic triangle? That
is, how many lines could you fit into the figure before the next one you drew
would make a monochromatic triangle, no matter whether it was blue or red?
These are just tangential questions, in a way, to solving this particular puzzle,
but they’re worth thinking about because they’re interesting in their own right
and they may guide us toward a solution for this puzzle or a generalization
thereof. For illustration’s sake, here is one of our attempts at trying to assing
red and blue lines in a diagram. Why did we stop here? How many more lines
need to be added? Can we add any of them?

6 2

5
3

The situation we face now is interesting because it’s of the opposite nature
to the kind of situations we faced before. With four and five dots, we wanted to
show that it was possible to arrange all of the lines to make no monochromatic
triangles. To show that, we just had to do it! Exhibiting a particular figure
with the desired property was sufficient to show that it was possible to achieve
the property we wanted. With six dots, though, it seems like it is not possible
to arrange the lines so that there are no monochromatic triangles. How can
we prove that this is a fact? It is tempting to say that we should just look
at all possible arrangements of the lines and argue that there is at least one
monochromatic triangle in every single one of them. Is this feasible? How many
arrangements of the lines are there? How could we easily find a monochromatic
triangle in any given diagram? Remember how we did this with the figure with
five dots? We noticed that any triangle would have to use at least one line from
the outside and at least one line from the inside, which guarantees right away
that any triangle has two types of lines. Could we do the same thing here, and
identify some property that guarantees a triangle?
The issue is that there are too many possible arrangments of the lines in the
diagram with six dots for us to check all of them by hand! There are 15 lines
to be drawn, each of which could be either blue or red, so it seems like there
are 215 possible arrangements. This is a big number! (In actuality, there are
slightly fewer possibilities because some of them are equivalent in some sense;
more technically, they are called “isomorphic”.)
82 CHAPTER 1. WHAT IS MATHEMATICS?

Solution: Working with an Arbitrary Diagram

We need to be more clever with our argument so that we can prove a property
about any diagram without drawing a particular one. That is, we need to find
some fact, some property that is true of all of the possible diagrams with six
dots, that will still allow us to deduce that there must be a triangle. One way
to approach this is to think about drawing the lines in one small section of the
diagram. Specifically, let’s take any of the six dots and consider the five lines
coming out from that dot. For example, we might have something like this:

How many are blue and how many are red? This is partially a trick ques-
tion: we’re not really considering any particular diagram (like the one above)
but rather trying to find a fact about all possible diagrams. Thus, we can’t an-
swer that question too specifically. Pretend we are prsented with an arbitrary
diagram, and we have to make an argument that will work no matter what that
diagram was.
Here’s what we can say: There must be at least three blue lines or at least
three red lines. Do you see why this is true? The only way this wouldn’t be true
is if there were two (or fewer) blue lines and two (or fewer) red lines leaving this
particular dot, for a total of four (or fewer) lines. We know that all possible
connections must be drawn, though, so there should be five! (This argument
is an example of a concept known as the Pigeonhole Principle. The idea is
that we can’t place five objects, of two different colors, into two different boxes
without putting three objects of one color into one box. This is an incredibly
useful strategy with these types of problems, and we will examine the principle
in much greater detail later on in Section 8.6.)
So where do we stand? We started with any of the figures with six dots
and all lines filled in, and focused on one particular dot; coming out from this
dot, there must be three blue lines or three red lines. It could be either color,
so we can’t just assume it’s red and follow an argument that way; we can do
that, but we have to come back to this point afterwards and see what would
change if the three lines were blue. So let’s do that: let’s examine all of the
figures where there are three red lines exiting this particular dot. Where can
we go from there? We haven’t yet assumed anything about the other lines in
the figure, so let’s look at what those could be. Examine the picture below to
see what line colors we have assumed exist so far:
1.4. QUIZZICAL PUZZICLES 83

Now, what lines could be added to this diagram that would avoid making
a triangle of one color amongst three dots? We can’t necessarily make any
assumptions about the lines coming out from the two dots that are isolated in
the picture, so let’s focus on the three dots on the bottom. What color could
the lines among those be? Well, if any of them are red, that would form a
monochromatic triangle between the two endpoints of that line and the original
dot we focused on! That would be a problem.

Red
Triangle

Okay, the only way to avoid that is to make all of those lines blue. But
that would make a blue triangle among those three dots! Wow, it looks like we
cannot avoid making a monochromatic triangle no matter what we do!

Blue
Triangle

Let’s go back to our Pigeonhole argument and reassess the situation. What
if the three lines of the same type that were guaranteed by the argument were
blue instead of red? Well, nothing would really change, right? We would still
84 CHAPTER 1. WHAT IS MATHEMATICS?

be stuck, in terms of adding new lines among the three dots on the bottom of
the figure:

Blue
Triangle

Red
Triangle

If we include any blue lines, that forms a monochromatic triangle with the
original dot, and if we make all of them red, that forms a monchromatic triangle
there! In this sense, the two arguments we followed after the Pigeonhole argu-
ment were identical ; if we were to replace every instance of the word “blue” with
“red” and vice versa, we would have the same argument. Sometimes, mathe-
maticians will use this situation to shorten a proof and just say that “without
loss of generality, the three lines are red.” This is usually taken to mean that
if we were to make the other choice instead (i.e. if the lines were blue) then
the further argument would have an identical structure, mathematically, so we
can save space and time by not writing the same words over again. This is so
common, in fact, that you might sometimes see this phrase, “without loss of
generality”, abbreviated as WOLOG or WLOG.

Solution: Summarizing Our Results

What have we accomplished so far? We produced specific diagrams that showed
we could arrange the lines among four and five dots so that we can avoid a
monochromatic triangle, and we argued that any diagram of lines with six
dots must have a monochromatic triangle. In terms of the friends/enemies
formulation of the puzzle, this means that any group of six people must have a
subgroup of three people that are all friends or all enemies.
Notice how helpful it was to recast the puzzle into this dots/lines formulation;
it allowed us to completely forget about the social context of the problem (which
can be distracting, in a way) and let us simplify our terminology and notation
(we went from labeling pairs of people as “friends” or “enemies” to simply
drawing one line between two dots). This is a very useful strategy: extracting
the inherent structure of a puzzle—the underworkings, the relationships between
the elements, how they interact, etc.—and rewriting everything in terms of just
those parts. This can make the puzzle easier to understand and tackle, and it
can guide us into devising better notation. What if we had continued to solve
this puzzle with the 13F, 23E, . . . notation? That may have eventually worked,
but it would have been much harder!
1.4. QUIZZICAL PUZZICLES 85

One of this puzzle’s original questions was to identify a cutoff number so

that any larger group of people would necessarily have this subgroup property.
Do you think we have accomplished that? Have we identified a cutoff? Could
six be that number? Why or why not? In any group of seven people, there’s a
smaller group of six people, and then our work above proves that there must be
three mutual friends/enemies among that group! Certainly, this works for any
group of people of size larger than six, so this must be that precise cutoff point
we were looking for. This is an analogous result to the one mentioned in the
original statement of the puzzle, where the Hungarian sociologist noticed this
phenomenon for subgroups of size four. That problem is much more difficult
to solve, so we handled a smaller, simpler case here. Both of these results are
related to a larger class of problems known as Ramsey Theory. This branch
of combinatorics and graph theory works with identifying these kind of “cutoff
points” where, as the size of some structure (like a group of people) grows
and grows, there is eventually a point where we can be guaranteed to find a
subgroup with a certain property (three mutual friends/enemies). What was
at first thought to be a sociological phenomenon turned out to be a rigorous
mathematical fact. How about that!

Generalizing: Questions for You

Let’s pose some interesting and related questions before moving on. What if
we had been looking for homogeneous groups of a different size, like four or
five or twelve? Certainly, we would have to have a larger group of people,
overall, to guarantee finding such a subgroup. Can we always do this? That
is, given any desired subgroup size, can we identify a cutoff point in the way
that we did here? Can you figure out how to prove that such a cutoff point
must exist, even without finding the particular number? Furthermore, what
if we allowed for a third possibility: friend, enemy, unacquainted. Could we
answer similar questions about homogeneous groups? These are all questions
related to Ramsey Theory and some of them are quite difficult to answer and
took mathematicians many years to address. Many of these kinds of simply-
stated questions are still open, unsolved problems! Don’t be discouraged if you
don’t make any progress on these questions. We believe that even attempting to
answer them and thinking about the issues therein is meaningful and beneficial
enough, in itself.

Lessons From This Puzzle

This puzzle brought up several difficulties. First, we had to find a way to
interpret the puzzle in a meaningful way so that we could even address the
questions it asked, and this involved coming up with appropriate notation to
represent the elements of the puzzle. This is an important part of mathematical
problem-solving, particularly for a puzzle like this that doesn’t incorporate the
notation and visualization as part of the problem statement.
Second, to identify six as the group cutoff size, we had to somehow prove that
86 CHAPTER 1. WHAT IS MATHEMATICS?

something is not possible, but the number of possible configurations to examine

was way too large to examine each one individually. This happens frequently,
particularly in problems related to computer science and algorithms. To address
this, we had to employ a strategy far more clever than mere brute force, and
it’s not always clear what strategy that should be. Here, we essentially started
to try to fill in the lines as if it was going to work out, and then realized that we
reached a point that was impossible to fix. Proving that something is possible
can amount to just showing an example of that phenomenon (which we did with
the groups of size four and five), but proving something is impossible can be
much trickier and require some context-dependent ingenuity.
Lastly, we saw that it can be interesting to think of questions closely-related
to the puzzle at hand that simply tweak one or more of the conditions of the
problem. What if we look for larger subgroups? What if we allow for more types
of lines? How does this change the results? Exploring the boundaries of puzzles
by changing the conditions like this can lead to new mathematical discoveries
and techniques, and keeps mathematicians actively searching for new knowledge
and ways of sharing that knowledge.

1.4.5 The Full Monty Hall

Problem Statement
This puzzle involves only basic probability and arithmetic, yet it has continually
stumped some very smart people over the years. In fact, a debate erupted in
1990 when Marilyn vos Savant published the puzzle and her solution in her col-
umn in Parade magazine, with many people (mathematicians included) writing
her letters to agree or disagree with her (correct, we should say) answer. Let’s
see what you think!

Suppose you’re on a game show and you’re given the choice of three
doors. Behind one door is a car; behind the others, goats. The car
and the goats were placed randomly behind the doors before the
show. The rules of the game show are as follows: After you have
chosen a door, the door remains closed for the time being. The game
show host, Monty Hall, who knows what is behind the doors, now
has to open one of the two remaining doors, and the door he opens
must have a goat behind it. If both remaining doors have goats
behind them, he chooses one at random. After Monty Hall opens
a door with a goat, he will ask you to decide whether you want to
stay with your first choice or to switch to the last remaining door.
Imagine that you chose Door 1 and the host opens Door 3, which has
a goat. He then asks you ”Do you want to switch to Door Number
2?” Is it to your advantage to change your choice?

Of course, we are assuming that you would prefer to win a car over a goat,
and that you want to maximize the chances of winning that one car. Also, we
1.4. QUIZZICAL PUZZICLES 87

should mention that this puzzle takes its name from the host (Monty Hall) of a
TV game show called Let’s Make a Deal.
So what do you think? Imagine yourself standing on the stage in front of a
TV audience, when Monty Hall asks you, “Do you want to switch to the other
door?” What do you do?
Think carefully about this before turning the page and reading our solution.
88 CHAPTER 1. WHAT IS MATHEMATICS?

Solution: Always Switch

We’ll state the answer right away because it might shock you: you should defi-
nitely switch your choice! Reasoning this out and obtaining this solution is the
tricky and confusing part, and establishing the right way to interpret the puzzle
is part of what has confused solvers for so long.

Analyzing an Incorrect Argument

Let’s start by showing you an incorrect “solution” that claims switching is irrel-
evant. Imagine you and your friend heard this puzzle, and he/she gave you this
explanation. How would you respond? Would you agree? Why? If not, how
would you tell them they’re wrong? What is wrong with their explanation?

Well, after I pick a door and Monty Hall shows me a goat behind
another door, then there are only two doors that are unopened. One
of them has a goat and the other one has a car, so there’s a 50/50
chance that the car is behind my door and a 50/50 chance that the
car is behind the other door. Therefore, it doesn’t matter whether
I switch or not, so I might as well stick with the choice I already
made.

Are you convinced by this? Let’s try to figure out what’s wrong with this
argument. The main question we need to address to solve this puzzle involves
figuring out two numbers: the probability of winning the car by sticking with
our first choice, and the probability of winning the car by switching to the other
door. We need to identify these two values and compare them; only then can
we definitively address the question of the puzzle.
Now, the argument above seems to address both of these probabilities by
saying they are both 50%, but there is a problem with how the arguer interprets
the situation. What do you think the chances are of winning the car by staying
with the first choice? In essence, this is equivalent to not even having Monty
Hall show us another door with a goat behind it. If we are going to stay with
our first choice of a door, we might as well not even see a goat behind a different
door since that doesn’t affect the object behind the door we originally chose.
Let’s restate this idea to reiterate its importance:
Since there are three doors, the chances of picking the right one first
is 13 , and seeing a goat behind another door doesn’t change that fact.
This is the problem with the above argument and, in fact, one of the most
common mistakes made in “solving” this puzzle.
The next step is to figure out the probability of winning the car after switch-
ing and compare this to 31 . There are several ways to accomplish this, in fact.
One concise way is to reason that switching results in success (winning the car)
whenever we first choose a door that happens to have a goat. In those cases,
the two unchosen doors conceal a goat and a car, in some order, and the game
show host is forced to show us the goat; thus, the car is concealed behind the
1.4. QUIZZICAL PUZZICLES 89

remaining door, and switching results in a win. Since we will choose a door with
a goat behind it 32 of the time, we conclude that switching results in a win 23 of
the time.

Enumerating the Possibilities

These explanations might seem unsatisfactory to you, so let’s try to actually
enumerate (explicitly count) the possible arrangements of the goats and the car
behind the doors and write down what happens if we switch in each case. The
first thing to notice is that the numbering of the doors is irrelevant since all
choices are made randomly; that is, whether the car lies behind the door with
“#1” printed on it, or “#2” or “#3”, the results will be the same: we still have
a 31 chance of identifying that door with the car. Accordingly, we can assume
WOLOG (remember that this acronym means “without loss of generality”) that
the car lies behind Door #1 and the goats stand behind Doors #2 and #3. Of
course, this is our imposition on the problem, and we can’t say that the game-
player knows this (otherwise he/she would pick Door #1 every time!). With
this arrangement in mind, let’s examine all 3 choices we could make at the
beginning, and see what switching or staying would accomplish in each case:

Door #1 Door #2 Door #3

Car Goat Goat

Our choice Host shows Result of switching Result of staying

Door #1 Door #2 or Door #3 Goat Car
Door #2 Door #3 Car Goat
Door #3 Door #2 Car Goat

One important observation is that when we initially choose the door with the
car, the host could choose either of the remaining doors to show us a goat, and
he makes that choice randomly. For either choice, though, we lose by switching
and win by staying. Still, those situations only occur 31 of the time, i.e. after
we chose the door with the car behind it, initially. Since each of the rows of
the table above is equally likely, we can conclude that 23 of the time we win by
switching, while 31 of the time we win by staying.
Does this puzzle make more sense now? Try posing this puzzle to some of
your friends and family and gauge their reactions. How many produced the
correct answer? How many could correctly explain it? How many erroneously
said “it doesn’t matter”? How many had already heard the puzzle before?

Generalizing to Many Doors and Cars

Let’s look at a generalization of this game show situation and try to prove
whether or not switching is also a good idea there. Specifically, let’s say there
are n doors and m cars in total, so there are n − m goats. To analyze this, we
need to specify that m ≤ n − 2. Think about why this is necessary:
90 CHAPTER 1. WHAT IS MATHEMATICS?

• If m = n were true, then we always win all the time, whether we switch
or not. Hence, there’s nothing to prove in this case.
• If m = n − 1 were true, then whenever we happen to choose the door with
the only goat behind it at first, the host is unable to show us a door with
a goat. Hence, the game is ruined and the question of switching is moot.
Now, with these variables in place, here are the new rules of the game: We
choose one of the n doors. The host identifies all other doors that conceal a
goat and randomly chooses one of those doors and opens it. We then have the
option to stick with our original choice or switch to any of the other doors, of
our choosing. What is the strategy now? Should we switch? Should we stay?
Does the answer depend on m and n at all? How?
We will approach this modified puzzle in much the same way as the first
approach to the version above. We can’t possibly enumerate all situations in
this version because m and n are unknown variables. Instead, we need to
apply logical reasoning to deduce the probabilities associated with winning when
staying and when switching. The first key observation is exactly the same as
one we made before: the probability of winning when staying is exactly the
probability of choosing a door concealing a car at first. When we choose a door
with a car behind it first, no matter what other door the host reveals, staying
on our current choice results in a win. Furthermore, when we chose a door
concealing a goat at first, staying results in a loss. Thus, the only way to win
while sticking with our first choice is by choosing one of the m doors with a car
behind it out of the n total doors. This probability is precisely m n.
To identify the probability of winning after switching, we need to think
carefully about the probabilities associated with each step. Notice that since
m ≥ 2 is a possibility, it may be that we chose a door with a car at first, switched
our choice, and still won. With that in mind, we should examine two different
cases, here: (a) what happens when we choose a door with a goat first, and
(b) what happens when we choose a door with a car first. Each case will leave
a different number of options for the host and, subsequently, a different number
of ways for us to switch and win, so we should handle them separately.
(a) Let’s say we first chose a door with a goat. There are now n − m − 1 doors
remaining that conceal goats, and the host randomly picks one of those to
open. From our perspective, switching leaves us with n−2 options (we can’t
switch to the opened door or our first choice), m of which are cars. Thus,
m
the probability of winning after switching, in this particular case, is n−2 .
Since there were n − m goats originally, this case occurs with a probability
of n−m
n . Therefore, the contribution of this case to the total chances of
winning after switching is
n−m m nm − m2
· =
n n−2 n(n − 2)
(Think about why we multiplied these probabilities together. Why did we
need to do that at all? Why didn’t we add them together? What will we do
1.4. QUIZZICAL PUZZICLES 91

to combine this probability with the probability associated with the next
case?)
(b) Next, let’s say we first chose a door with a car. There are now n − m doors
remaining that conceal goats, and the host randomly picks one of those to
open. From our perspective, switching leaves us with n − 2 options, m − 1
of which are cars. Thus, the probability of winning after switching, in this
particular case, is m−1
n−2 .
Since there were m cars originally, this case occurs with a probability of mn.
Therefore, the contribution of this case to the total chances of winning after
switching is
m−1 m m2 − m
· =
n−2 n n(n − 2)
Since these two cases occur separately (i.e. they both can’t occur simultane-
ously) we should add these probabilities together. This will tell us the totaly
probability of winning a car after switching from our original choice to another
random door:
nm − m2 m2 − m nm − m2 + m2 − m
+ =
n(n − 2) n(n − 2) n(n − 2)
nm − m
=
n(n − 2)
m(n − 1)
=
n(n − 2)
m n−1
= ·
n n−2
Now, there’s a reason we chose to write the fraction the way we did here. We
want to compare this probability to the chances of winning after staying with
our first choice of door, which was m n . We see that the probability of winning
after switching is, in fact, a multiple of that other probability, and the factor
n−1
n−2 > 1 because n − 1 > n − 2. Written in inequality form:

m m n−1
< ·
n n n−2
| {z }
>1

Therefore, the chances of winning after switching are strictly better (i.e. not
equal to, always better) than the chances of winning after staying. We should
always switch to another random door!

Applying the General to the Specific

The original version of this puzzle is the specific case where n = 3 and m = 1,
so we can check that our result makes sense. The formulas we derived tell us
the chances of winning after switching are 31 , as we found previously, and the
chances of winning after staying are 1(3−1) 2
3(1) = 3 , as we found previously. Neat!
92 CHAPTER 1. WHAT IS MATHEMATICS?

Generalizing: Questions for You

What happens for other values of m and n? Can you make the “always switch”
strategy significantly better than the “always stay” strategy? That is, how large
of a difference can we get between the probability of winning associated with
the two strategies? How small can we make it? Is it possible to make them
equal?
Another modified version of this puzzle is based on the host opening more
than 1 door, revealing multiple goats. Specifically, suppose there are n total
doors, m of which have cars, and that, after your first choice, the host randomly
identifies p doors with goats and opens all of those, after which you may choose
to switch to any of the remaining n − p − 1 doors, or stick with your first pick.
What is the best strategy in this game? What sorts of conditions do you need
to impose on m, n and p to ensure we can even play the game at all? Should
you always switch, or does it depend on p? How large/small can we make the
difference in the chances of winning for the “always switch” and “always stay”
strategies?

Lessons From This Puzzle

Intuition and quick decisions are sometimes helpful in guiding us to a solution,
but it is always important to check those snap judgments to make sure they
are based on sound, rational arguments. In this puzzle, saying that the chances
were “50/50” maybe made sense at first, but after thinking about it more care-
fully and reassessing the situation, we realized there was a flaw in the argument.
Specifically, that flaw had to do with interpreting the puzzle correctly and fol-
lowing the steps of the game show in the appropriate order. It was best to assess
the probabilities in the order in which they occur as the game is played, rather
than starting from an ending position and looking backwards.
In general, puzzles involving probability are quite tricky and require careful
analysis, so it’s important to keep that in mind. A larger lesson here, as well,
is that oftentimes the most simply stated puzzles are the trickiest to figure out.
Never be fooled into thinking that a puzzle will be easy to solve because the
statement is short or easy to understand!
For more information about the Monty Hall problem and the psychology in-
volved, check out this link to the following paper: Krauss, Stefan and Wang, X.
T. (2003). “The Psychology of the Monty Hall Problem: Discovering Psycholog-
ical Mechanisms for Solving a Tenacious Brain Teaser”, Journal of Experimental
Psychology: General 132(1).

1.5 It’s Wise To Exercise

We’ll conclude this chapter with a handful of exercises that incorporate some of
the ideas we’ve discussed so far, or give you a chance to practice your previous
knowledge, or just keep you on your mental toes. Attempt as many as you can,
and discuss potential solutions with some friends to see what they think. At the
1.5. IT’S WISE TO EXERCISE 93

end of the day, though, just think of this as a way to keep your brain muscles
limber!
Problem 1.5.1. A fly is resting on the front of a train that is hurtling forward
at 60 kilometers per hour. On the same track, 300 kilometers straight ahead,
another train is hurtling towards the first train at 60 kilometers per hour. At
that moment, when the trains are 300 km apart, the fly takes off at 90 km per
hour. He continually flies back and forth between the trains, flying just above
the track and instantaneously turning around when he reaches a train. What
is the total distance traveled by the fly before the two trains crash together,
squishing the fly between them in the process? How did you figure this out?
Try to generalize the situation to when one train travels at a km/hr, the other
at b km/hr, and the fly at c km/hr.
Problem 1.5.2. A government mint is commissioned to produce gold coins. The
mint has 20 machines, each of which is producing coins weighing 5 grams apiece.
One day, the foreman of the mint notices that some coins are light, and after
assessing the machines, he finds that one of them is making 4 gram coins, while
the other 19 machines are working perfectly. He decides to use the situation
to his advantage and identify the smartest employee, to be promoted next. He
tells the workers that exactly one machine is producing coins that are 4 grams,
and that they need to determine which machine is broken. You, as an employee,
are allowed to take one, and only one, reading on a scale. You can place any
number of coins from any machines, of your choosing, but you must pool them
together and will only see the total weight of all the coins, as a number in grams.
How do you do this so that you can determine precisely which machine is the
broken one?
Problem 1.5.3. In a game of chess, the Queen is allowed to move vertically,
horizontally, or diagonally, in any direction, for any number of spaces. Try to
place 8 Queens on a standard 8 × 8 chessboard so that none of the Queens is
attacking any of the others (that is, no Queen can make one move and capture
another one immediately). Exhibit a way to do this or show that it is not
possible. If you found a way, how many different ways do you think there are
to do so? If you showed it isn’t possible, find a smaller number of Queens to
place so that it is possible. What is the maximum number of Queens you can
possibly place on the board in this way?
Problem 1.5.4. Start with a standard 8 × 8 chessboard and remove 2 squares
at opposite corners, say the top-right and bottom-left. Can you cover all of
the remaining squares with 2 × 1 dominos so that none of the dominos overlap?
(Note: this is known as a tiling of the board.) Why or why not? What about in
the general case, with an n × n board? Does your answer depend on n at all?
How so?
Problem 1.5.5. Consider a standard 8 × 8 chessboard, with two adjacent squares
removed from each corner. (See picture.) Can you tile this board with T-shaped
tetrominoes? (See picture.) If so, how? If not, why not?
94 CHAPTER 1. WHAT IS MATHEMATICS?

What about in the general case, with an n × n board? Does your answer depend
on n at all? How so?
Problem 1.5.6. Given a real number x, we let bxc denote the largest integer
smaller than (or equal to) x and we let dxe denote the smallest integer greater
than (or equal to) x. For instance, b6.02c = 6 and b6.99999c = 6 and b6c = 6
and b−6.5c = −7.
Determine a more specific and concise representation for the value of the
following expressions, whenever possible. (You might have to find a few different
expressions, depending on x.)

1. bxc + b1 − xc

2. dxe + d1 − xe

3. bxc + dxe
bxc
4. x

5. bx2 c − bxc2

6. dx2 e − dxe2

Problem 1.5.7. Find three natural numbers a, b, c so that no subset has a sum
that is divisible by 3. That is, find a, b, c so that none of the following sums is
divisible by 3: a, b, c, a + b, a + c, b + c, a + b + c. Is this possible? Why or why
not?
Try to do the same thing with 4 numbers: find a, b, c, d so that no subset
has a sum that is divisible by 4. Is this possible? Why or why not?
Try to generalize. Can you say anything about finding n natural numbers
so that no subset has a sum divisible by n?
Problem 1.5.8. Recall the way that we solved the Friend Trends problem, by
working with dots and colored lines. For this problem, we want to address a
similar situation where we have a certain number of dots and need to draw all
possible lines so that any two dots are connected by exactly one line, but in this
case we don’t care about the color, so we can say all lines are drawn in black,
say. Can you draw this figure with 3 dots so that none of the lines cross? What
about with 4 dots? Or 5? Or 6? Why or why not? Try to explain why any of
those figures are impossible to achieve. If you can’t achieve 0 crossings, what is
the minimum number you could possibly achieve?
1.5. IT’S WISE TO EXERCISE 95

Problem 1.5.9. Let’s use the same assumptions from the Friend Trends problem:
any two people are either friends or enemies, and there are no other relationships.
Take a group of n people and assume that no person has more than k friends.
We want to separate the n people into some number of clubs so that each person
is in exactly one club. How many clubs do we need, in total, so that we can
distribute the n people so that nobody is in a club with a friend? Given the
relationships between all the people (i.e. given two people, we know whether
they are friends or enemies) and that number of clubs, how would we go about
separating the people to achieve this property?
Problem 1.5.10. Draw a circle. Place 3 dots along the circumference of the circle.
We want to color the sections between the dots (each section gets exactly one
color) so that no dot is touched by two sections of the same color. How many
colors do we need? What about if we place 4 dots around the circumference
of the circle? Or 5? Try to generalize to n dots. What can you say about the
number of colors required?
Problem 1.5.11. Suppose you have a drawer full of socks. It contains 2 pairs of
blue socks, 3 pairs of red socks, and 4 pairs of green socks. (Also assume that
left and right socks are indistinguishable.) One morning, you are in a great rush
and start grabbing socks randomly, one at a time, and holding on to all of them
until you have a pair. How many socks do you need to take out of the drawer
to guarantee that you have a pair on your hands?
How does your answer change if we had twice as many pairs, of each color,
as before? What if we had 3 pairs of socks in each of red, green, blue, yellow
and brown? What if we had n pairs of socks in each of m colors?
Problem 1.5.12. A group of four friends find themselves on one side of a bridge
late at night while walking home. The bridge is old and rickety and unsafe to
cross as a group. They only have one flashlight between them and the light is
only strong enough to light the way for two friends at a time. Each friend has a
different level of comfort with the bridge, so they will cross at different speeds.
One friend will take 5 minutes to cross, one will take 10, one will take 15, and
one will take 20. If two friends cross together, they cross at the pace of the
slower friend. How long will it take all 4 friends to get across the bridge? Can
you find the method that produces the absolute shortest time?
Problem 1.5.13. Consider the usual denominations of coins for the American
dollar: pennies (1 cent), nickels (5 cents), dimes (10 cents), and quarters (25
cents). What coins would you need to carry around in your pocket to guarantee
that you can pay any price between 0 cents and 100 cents with exact change?
Are there several possible sets of coins that would achieve this? What is the
smallest total value of coins with this property? Are there several possible sets
of coins with this same minimum total value?
Problem 1.5.14. Let a, b, c be real numbers, with a 6= 0. What is wrong with
b
the following “spoof” that − 2a is a solution to the equation ax2 + bx + c = 0,
if anything?
“Spoof ”: Let x and y be solutions to the equation. Subtracting ay 2 +by+c = 0
96 CHAPTER 1. WHAT IS MATHEMATICS?

from ax2 +bx+c = 0 yields a(x+y)(x−y)+b(x−y) = 0. Hence, a(x+y)+b = 0,

and so x + y = − ab . Since x and y were any solutions, we may repeat this
computation with x = y. Thus, 2x = − ab , and therefore x = − 2a
b
is a solution.
“”

Problem 1.5.15. Explain why (−1)(−1) = 1. Pretend you are writing for a
fellow classmate of similar intelligence who is skeptical of this fact and needs to
be convinced. It is not enough to just say “Because it is!” Try to come up with
a helpful geometric or physical explanation, some kind of memorable argument.
Problem 1.5.16. For each of the following proposed equations, identify all of the
real numbers that satisfy them:

(1) |x − 2| = |x − 3|

(2) |2x − 1| = |2x − 3|

(3) |2x − 2| = |3x − 3|

(4) |x + 1| = |x − 5|

(5) |x − 1| + |x − 2| = |x − 3|

Problem 1.5.17. The First Rule Of Logic Club Is . . . : To join Logic Club,
you must decide to always tell the truth or always lie. Members of Logic Club
know who lies and who soothsays. I do not belong to Logic Club, but I encounter
three members on the street who make the following statements:

• Jack: “All three of us are liars.”

• Tyler: “Exactly two of us are liars.”

• Chuck: “Jack and Tyler are liars.”

Who should I believe, if anyone?

√
Problem 1.5.18. Solve x − 1 = x − 3 for all real values of x that satisfy the
equation. Explain your work, and try to indicate why your answer/s is/are the
only answer/s.
Problem 1.5.19. You have two strings of fuse. Each will burn for exactly one
hour. The fuses are not necessarily identical, though, and do not burn at a
constant rate. All you have with you is a lighter and these two fuses. Can you
measure exactly 45 minutes? If so, explain how. If not, explain why.
Problem 1.5.20. This problem is a variation of a standard puzzle, a form of
which first appeared in the Saturday Evening Post way back in 1926!
Three friends pitched in and bought a big bag of M&M candies. They brought
the box back to their apartment and decided to wait and share the bag the next
day at a party.
During the night, the first friend woke up and felt like having a snack. He
1.5. IT’S WISE TO EXERCISE 97

decided to just eat his share of the candies now and not have any the next day.
He opened the bag, divided the M&Ms into three equal piles, but realized there
was one left over. He figured that one extra couldn’t hurt, and ate his share
and the extra, then put the rest back in the bag.
Later in the night, the second friend did the exact same thing. He woke up
feeling hungry, divded what was left in the bag into three piles, and ate his
share plus the extra one that was left over.
Even later in the night, the third friend did the exact same thing, including the
extra M&M left over.
The next day at their party, the friends split the bag of remaining candies into
three equal shares and enjoyed them. (No one acknowledged what they had all
done, of course).
How many M&Ms were in the bag to begin with? What is the smallest possible
number?
Problem 1.5.21. Given a list of real numbers, their arithmetic mean is defined
to be their sum divided by the number of terms, and their geometric mean
is defined to be their product raised to the power of one over the number of
terms. That is, supposing we have x1 , x2 , . . . , xn that are real numbers, then
the arithmetic mean is
x1 + x2 + · · · + xn
n
and the geometric mean is √n
x1 · x2 · · · xn
(Note: The n-th root of a number is the same as that number raised to the
power of n1 .)
Can you identify two numbers so that the arithemtic and geometric means are
equal? Can you identify two numbers so that the arithmetic mean is strictly
greater than the geometric mean? How about the other way around?
Repeat this with three numbers, four numbers, etc. Can you identify a general
pattern?
Problem 1.5.22. Consider the variable equation 6x + 15y = 93. We want to find
some integral solutions; that is, we want to find values of x and y that are both
integers (natural numbers, zero, and negative natural numbers) that satisfy the
equation.
1. Find a solution where both x and y are positive integers. Decribe, with a
few sentences, how you came up with this solution.
2. Find a solution where one of the values, x or y, is positive and the other
is negative. Again, describe how you came up with this solution.
3. How many solutions do you think there are? Try to write down a charac-
terization of all possible solutions, or describe how you might find them
all.
98 CHAPTER 1. WHAT IS MATHEMATICS?

Problem 1.5.23. A magic square is an n × n array that contains each of the

numbers from 1 to n2 exactly once and has the property that every row and
column (and both of the main diagonals) sums to the same number.
For example, here is a 3 × 3 magic square:

8 1 6
3 5 7
4 9 2

Notice that the so-called magic sum of each row/column/diagonal is 15 in this

case.
Can you find a formula for what the magic sum of an n × n magic square must
be?
(Hint: There is a result we discovered in this chapter that will be useful.)
Problem 1.5.24. How many whole numbers less than or equal to 1000 have a 1
as at least one of their digits? For instance, we want to count 1 and 12 and 511
once each.
Problem 1.5.25. We have several piles of koala bears. In an attempt to disperse
them, we remove exactly one koala bear from each pile and place all of those
koalas into one new pile. For example, if we started with koala piles of sizes 1,
4, and 4, we would then end up with koala piles of sizes 3, 3, and 3; or, if we
started with piles of size 3 and 4, we would end up with piles of size 2 and 3
and 2.
It is possible that we do this operation exactly once and end up with the exact
same pile sizes as we started with (the order of them is irrelevant; only the sizes
matter).
Identify all the collections of piles that have this property and explain why they
are the only ones.
Hint: An example of a starting situation with this property is when we have
just one pile of size 1. We do the operation and again obtain one pile of size 1.
Bingo.
Hint 2: Be sure to also explain why your situations are the only ones that work.
How can we be sure you didn’t miss some answers?

1.6 Lookahead
This introductory chapter is meant to get you thinking about what mathe-
matics is, how we solve problems, and what it means to write a proof.
Throughout the rest of the book, we will be discussing all three of these ideas
in more and more detail. In doing so, we will explore several different areas of
the mathematical universe. We have an overarching plan for our journey, so
don’t think that we are just stumbling randomly through the forest. Our major
1.6. LOOKAHEAD 99

goals are to (a) formalize some of the intuitive ideas we have about mathemat-
ical objects, (2) see many examples of good proofs and develop the ability to
create and write good proofs, (3) develop problem-solving skills and the ability
to apply mathematical knowledge, and (4) cultivate an appreciation for both
the art and science of mathematics.
Scan the table of contents at the beginning of the book to get a sense for
where our journey is headed. The phrases and terminology might be foreign to
you now, but by the end of the book, we will all be speaking the same language:
mathematics.
100 CHAPTER 1. WHAT IS MATHEMATICS?
Chapter 2

Mathematical Induction:
“And so on . . . ”

2.1 Introduction
This chapter marks our first big step toward investigating mathematical proofs
more throughly and learning to construct our own. It is also an introduction
to the first significant proof technique we will see. As we describe below,
this chapter is meant to be an appetizer, a first taste, of what mathematical
induction is and how to use it. A couple of chapters from now, we will we be
able to rigorously define induction and prove that this technique is mathemati-
cally valid. That’s right, we’ll actually prove how and why it works! For now,
though, we’ll continue our investigation of some interesting mathematical puz-
zles, with these particular problems hand-picked by us for their use of inductive
techniques.

2.1.1 Objectives
The following short sections in this introduction will show you how this chapter
fits into the scheme of the book. They will describe how our previous work
will be helpful, they will motivate why we would care to investigate the topics
that appear in this chapter, and they will tell you our goals and what you
should keep in mind while reading along to achieve those goals. Right now,
we will summarize the main objectives of this chapter for you via a series of
statements. These describe the skills and knowledge you should have gained by
the conclusion of this chapter. The following sections will reiterate these ideas
in more detail, but this will provide you with a brief list for future reference.
When you finish working through this chapter, return to this list and see if you
understand all of these objectives. Do you see why we outlined them here as
being important? Can you define all the terminology we use? Can you apply
the techniques we describe?

101
102 CHAPTER 2. MATHEMATICAL INDUCTION

By the end of this chapter, you should be able to . . .

• Define what an inductive argument is, as well as classify a presented ar-
gument as an inductive one or not.

• Decide when to use an inductive argument, depending on the structure of

the problem you are solving.

• Heuristically describe mathematical induction via an analogy.

• Identify and describe different kinds of inductive arguments by comparing

and contrasting them, as well as identify the underlying structures of the
corresponding problems that would yield these similarities and differences.

2.1.2 Segue from previous chapter

As in the previous chapter, we won’t assume any familiarity with more advanced
mathematics beyond basic algebra and arithmetic, and perhaps some visual,
geometric intuition. We will, however, make use of summation and product
notation fairly often, so if you feel like your notational skills are, go back and
review Section 1.3.5.

2.1.3 Motivation
Look back at the Puzzle in Section 1.4.3, where we proved that the sum of
the first n odd natural numbers is exactly n2 . We first observed this pattern
geometrically, by arranging the terms of the sums (odd integers) as successively
larger “corner pieces” of a square. The first way we proved the result, though,
didn’t seem to depend on that observation and merely utilized a previous result
(about sums of even and odd integers) in an algebraic way; that is, we did some
tricky manipulation of some equations (mutliplying and subtracting and what
have you) and then–voilà!–out popped the result we expected. What did you
think about that approach? Did it feel satisfying? In a way, it didn’t quite match
the geometric interpretation we had, at first, so it might be surprising that it
worked out so nicely. (Perhaps there is a different geometric interpretation of
this approach. Can you find one?)
Our second approach was to model that initial geometric observation. We
transformed visual pieces into algebraic pieces; specifically, a sum was related to
the area of a square, and the terms of the sum were related to particular pieces of
that square. We established a correspondence between different intepretations
of the same problem, finding a way to relate one to the other so that we could
work with either interpretation and know that we were proving something about
the overall result. The benefit of the visual interpretation is that it allowed us to
take advantage of a general proof strategy known as mathematical induction,
or sometimes just induction, for short. (The word induction has some non-
mathematical meanings, as well, such as in electromagnetism or in philosophical
arguments, but within the context of this book, when we say induction, we mean
2.1. INTRODUCTION 103

mathematical induction.) What exactly is induction? How does it work? When

can we use this strategy? How do we adapt the strategy to a particular puzzle?
Are there variations of the strategy that are more useful in certain situations?
These are all questions that we hope to answer in this chapter.
The first topic we’d like to address is a question that we didn’t just ask in
the last paragraph, namely, “Why induction? Why bother with it?” Based on
that puzzle in Section 1.4.3, it would seem that mathematical induction isn’t
entirely necessary since there might be other ways of proving something, instead
of by induction. Depending on the context, this very well may be true, but the
point we’d like to make clear from the beginning is that induction is incredibly
useful! There are many situations where a proof by induction is the most concise
and clear approach, and it is a well-known general strategy that can be applied
in a variety of such situations. Furthermore, applying induction to a problem
requires there to be a certain structure to the problem, a dependence of one
“part” of the result on a “previous part”. (The “parts” and the “dependence”
will depend on the context, of course.) Recognizing that induction applies,
and actually going through the subsequent proof process, will usually teach us
something about the inherent structure of the problem. This is true even when
induction fails! Perhaps there’s a particular part of a problem that “ruins”
the induction process, and identifying that particular part can be helpful and
insightful.
We hope to motivate these points through some illustrative examples first,
after which we will provide a reasonably thorough definition of mathematical
induction that will show how the method works, in generality. (A completely
rigorous definition will have to be put off until a little bit later, after we have
defined and investigated some relevant concepts, like set theory and logical
statements and implications. For now, though, the definition we give will suffice
to work on some interesting puzzles and allow us to discuss induction as a general
proof strategy.)

2.1.4 Goals and Warnings for the Reader

Do keep in mind that we are still building towards our goal of mathematical
rigor, or as much as is possible within the scope and timing of this book and
course. Some of the claims we make in this Chapter will be clarified and tech-
nically proven later on, once we have properly discussed the natural numbers
and some basic mathematical logic. All in due time!
That said, this chapter is still very important, since we are continuing to
introduce you to the process of solving mathematical problems, applying our
existing knowledge and techniques to discover new facts and explain them to
others. In addition, mathematical induction is a fundamental proof technique
that will likely appear in every other mathematics course you take! This is
because of its usefulness and the prevalence of inductive properties throughout
the mathematical world.
104 CHAPTER 2. MATHEMATICAL INDUCTION

2.2 Examples and Discussion

2.2.1 Turning Cubes Into Bigger Cubes
To motivate the overall method of mathematical induction, let’s examine a ge-
ometric puzzle and solve it together. This example has been chosen carefully
to illustrate how mathematical induction is relevant when a puzzle has a
particular type of structure; specifically, some truth or fact or observation de-
pends or relies or can be derived from a “previous” fact. This dependence on a
previous case (or cases) is what makes a process inductive, and when we observe
this phenomenon, applying induction is almost always a good idea.

1-Cube into a 2-Cube

Let’s examine cubic numbers and, specifically, let’s try to describe a cubic num-
ber in terms of the previous cubic number. Imagine a 1 × 1 × 1 cube, just one
building block. How can we build the “next biggest” cube, of size 2 × 2 × 2, by
adding 1 × 1 × 1 building blocks? How many do we need to add? Arithmetically,
we know the answer: 23 = 8 and 13 = 1, so we need to add 7 blocks to have
the correct volume. Okay, that’s a specific answer, but it doesn’t quite tell us
how to arrange those 7 blocks to make a cube, nor does it give us any insight
into how to answer this question for larger cubes. Ultimately, we would like
to say how many blocks are required to build a 100 × 100 × 100 cube into a
101 × 101 × 101 cube without having to perform a lot of tedious arithmetic; that
is, we are hoping to eventually find an answer to the question: given an n×n×n
cube, how many blocks must we add to build it into a (n + 1) × (n + 1) × (n + 1)
cube? With that in mind, let’s think carefully about this initial case and try to
answer it with a general argument.
Given that single building block, and knowing we have to add 7 blocks to
it, let’s try to identify exactly where those 7 blocks should be placed to make
a 2 × 2 × 2 cube. (For simplicity, we will refer to a cube of size n × n × n
as an n-cube, for any value of n. We will only need to use values of n that
are natural numbers, i.e. non-negative whole numbers, in this example.) Look
at the pictures of the 1-cube and 2-cube below and try to come up with an
explanation of constructing one from the other.

Here’s one reasonable explanation that we want to use because it will guide us in
the general explanation of building an (n+1)-cube from an n-cube, and because
2.2. EXAMPLES AND DISCUSSION 105

it is a mathematically elegant and simple explanation. Start with the 1-cube

positioned as it is above and “enlarge” the 3 exposed faces by the appropriate
amount, in this case by one block. This accounts for 3 of the 7 blocks, thus far:
23 = 13 + 3 + . Where are there “holes” now?

The blocks we just added have created “gaps” between each pair of them, and
each of those “gaps” can be filled with one block. This accounts for 3 more of
the 7 total blocks: 23 = 13 + 3 + 3 + . Now what?

There is just one block left to be filled, and it’s the very top corner. Adding
this block completes the 2-cube and tells us how to mathematically describe our
construction process with the following picture and equation:

23 = 1 + 3 + 3 + 1
2-Cube into a 3-Cube
Okay, we might now have a better idea of how to describe this process in general,
but let’s examine another case or two just to make sure we have the full idea.
106 CHAPTER 2. MATHEMATICAL INDUCTION

Let’s start with a 2-cube and construct a 3-cube from it. (You can even try
this out by hand if you happen to own various sizes of Rubik’s Cubes!) We
can follow a process similar to the steps we used in the previous case and just
change the numbers appropriately. Starting with a similar picture

we see that we need to “enlarge” the three exposed faces of the 2-cube but, in
this case, the amount by which we need to enlarge them is different than before
(with the 1-cube) since we are working with a larger initial cube. Specifically,
each face must be enlarged by a 2 × 2 square of blocks (whereas, in the previous
case, we added a 1 × 1 square of blocks). Thus, an equation to account for this
addition is
33 = 23 + 3 · 22 +

After we do this, we see that we need to fill in the gaps between those enlarged
faces with 2 × 1 of blocks (whereas, in the previous case, we added 1 × 1 rows
of blocks). An equation to account for the additions thus far is

33 = 23 + 3 · 22 + 3 · 2 +
2.2. EXAMPLES AND DISCUSSION 107

After we do this, we see that there is only the top corner left to fill in. Accord-
ingly, we can depict our construction process and its corresponding equation:

33 = 23 + 3 · 22 + 3 · 2 + 1

n-Cube into an (n + 1)-Cube

Do you see now how this process will generalize? What if we started with an
n-cube? How could we construct an (n + 1)-cube? Let’s follow the same steps
we used in the previous two cases. First, we would enlarge the three exposed
faces by adding three squares of blocks. How big is each square? Well, we want
each square to be the same size as the exposed faces, so they will be n × n
squares, accounting for n2 blocks for each face:
108 CHAPTER 2. MATHEMATICAL INDUCTION

(n + 1)3 = n3 + 3n2 +
Next, we would fill in the gaps between these enlarged faces with rows of blocks.
How long are those rows? Well, they each lie along the edges of the squares of
blocks we just added, so they will each be of size n × 1, accounting for n blocks
for each gap:

(n + 1)3 = n3 + 3n2 + 3n +
Finally, there will only be the top corner left to fill in! Therefore,

(n + 1)3 = n3 + 3n2 + 3n + 1

“Wait a minute!” you might say, abruptly. “We already knew that, right?” In
a way, yes; the equation above is an algebraic identity that we can also easily
see by just expanding the product on the left and collecting terms:

(n + 1)3 = (n + 1) · (n + 1)2
= (n + 1) · (n2 + 2n + 1)
= (n3 + 2n2 + n) + (n2 + 2n + 1)
= n3 + 3n2 + 3n + 1

So what have we really accomplished? Well, the main point behind deriving this
identity in this geometric and visual way is that it exhibits how this identity
represents some kind of inductive process. We sought to explain how to derive
one “fact” (a cubic number, (n + 1)3 ) from a previously known “fact” (the
next smallest cubic number, n3 ) and properly explained how to do just that.
Compare this to one of the methods we used to investigate the fact that the
sums of odd integers yield perfect squares, too. That observation also belies an
2.2. EXAMPLES AND DISCUSSION 109

inductive process and, although we didn’t describe it as such at the time, we

encourage you to think about that now. Look back at our discussion and try
to write out how you could write (n + 1)2 in terms of n2 by looking at squares
of blocks. Does it look anything like an “obvious” algebraic identity? (If you’re
feeling ambitious, think about what happens with writing (n + 1)4 in terms of
n4 . Is there any geometric intuition behind this? What about higher powers?)
The benefit of the method we’ve used is that we now know how to describe
cubic numbers in terms of smaller cubic numbers, all the way down to 1; that
is, any time we see a cubic number in an expression, we know precisely how to
write that value in terms of a smaller cubic number and some leftover terms.
Furthermore, each of those expressions and leftover terms have an inherent
structure to them that depends on the cubic number in question. Thus, by
iteratively replacing any cubic number, like (n + 1)3 , with an expression like
the one we derived above, and continuing until we can’t replace any more,
should produce an equation that has some built in symmetry. This idea is best
illustrated by actually doing it, so let’s see what happens. Let’s start with the
expression we derived, for some arbitrary value of n,

(n + 1)3 = n3 + 3n2 + 3n + 1

and then recognize that we now know a similar expression

n3 = (n − 1)3 + 3(n − 1)2 + 3(n − 1) + 1

We proved that this equation holds when we gave a general argument for the
expression above for n3 , since that only relied on the fact that n ≥ 1. We can
follow the same logical steps, throughout replacing n with n − 1, and end up
with the second expression above, for (n − 1)3 . (Does this keep going, for any
value of n? Think about this for a minute. Does our argument make any sense
when n ≤ 0? Would it make physical sense to talk about, say, constructing a
(−2) × (−2) × (−2) cube from a different cube?)
Therefore, we can replace the n3 term in the line above

(n + 1)3 = 3
n + 3n2 + 3n + 1
+ (n − 1)3 + 3(n − 1)2 + 3(n − 1) + 1

This is also an algebraic identity, but it’s certainly not one that we would easily
think to write down just by expanding the product on the left-hand side and
grouping terms. Here, we are taking advantage of the structure of our result
to apply it over and over and obtain new expressions that we wouldn’t have
otherwise thought to write down. Let’s continue with this substitution process
and see where it takes us! Next, we replace (n − 1)3 with the corresponding
expression and find

(n + 1)3 = 3n2 + 3n + 1
(n− 1) + 3(n − 1)2
3
+ 3(n − 1) + 1
 

+ (n − 2)3 + 3(n − 2)2 + 3(n − 2) + 1
110 CHAPTER 2. MATHEMATICAL INDUCTION

Perhaps you see where this is going? We can do this subsitution process over and
over, and the columns that we’ve arranged above will continue to grow, showing
us that there is something deep and mathematically symmetric going on here.
But where does this process stop? We want to write down a concise version of
this iterative process and be able to explain all of the terms that arise, so we
need to know where it ends. Remember the very first step in our investigation
of the cubic numbers? We figured out how to write 23 = 13 + 3 + 3 + 1. Since
this was our first step in building this inductive process, it should be the last
step we apply when building backwards, as we are now. Accordingly, we can
write
(n + 1)3 = 3n2 + 3n + 1
+ 3(n − 1)2 + 3(n − 1) + 1
+ 3(n − 2)2 + 3(n − 2) + 1
+ 3(n − 3)2 + 3(n − 3) + 1
.. .. ..
. + . + .
+ 3 · 22 + 3·2 + 1
3
+ 1 + 3 · 12 + 3·1 + 1

This is definitely an identity we wouldn’t have come up with off the top of our
heads! In addition to being relatively pretty-looking on the page like this, it also
allows us to apply some of our previous knowledge and simplify the expression.
To see how we can do that, let’s apply summation notation to the columns
above and collect a bunch of terms into some simple expressions:

n
X n
X n
X
3 3 2
(n + 1) = 1 + 3 · k +3· k+ 1
k=1 k=1 k=1

In the last chapter, we saw a couple of different proofs that told us

n
X n(n + 1)
k=
2
k=1

Let’s use that fact in the line above, and also simplify the term on the far right,
to write
n
X 3n(n + 1)
(n + 1)3 = 1 + 3 · k2 + +n
2
k=1

What does this tell us? What have we accomplished after all this algebraic
manipulation? Well, we previously proved a result about the sum of the first
n natural numbers, so a natural question to ask after that is: What is the sum
of the first n natural numbers squared ? How could we begin to answer that?
That’s a trick question, because we already have! Let’s do one or two more
algebraic steps with the equation above by isolating the summation term and
2.2. EXAMPLES AND DISCUSSION 111

then dividing:
n
3n(n + 1) X
(n + 1)3 − 1 − n − =3· k2
2
k=1
n
1 1 n(n + 1) X 2
(n + 1)3 − (n + 1) − = k
3 3 2
k=1

This is what we’ve accomplished: we’ve derived a formula for the sum of the first
n square natural numbers! Of course, the expression on the left in the line above
isn’t particularly nice looking and we could perform some further simplification,
and we will leave it to you to verify that this yields the expression below:
n
X 1
k2 = n(n + 1)(2n + 1)
6
k=1

“And so on” is not rigorous!

There are a couple of “morals” that we’d like to point out, based on all of this
work. The first moral is that generalizing an argument is a good method for
discovering new and interesting mathematical ideas and results. Did you think
about how this puzzle is related to the sums of odd natural numbers? If not,
we encourage you strongly to try that now, as well as think about generalizing
this even further to four or five dimensional “cubes” and so on. In addition to
giving you some other interesting results, it will also be incredibly instructive for
learning to think abstractly and apply inductive processes. The second moral
is more like an admission: we have not technically proven the formula above
for the sum of the first n square natural numbers. It seems like our derivation
is valid and tells us the “correct answer” but there is a glaring issue: ellipses!
In expanding the equation for (n + 1)3 and obtaining those columns of terms
.
that we collected into particular sums, writing .. in the middle of those columns
was helpful in guiding our intuition, but this is not a mathematically rigorous
technique. How do we know that all of the terms in the middle are exactly what
we’d expect them to be? How can we be so sure that all of our pictures of cubes
translate perfectly into the mathematical expressions we wrote down? What do
we really mean by “and keep going all the way down to 1”?
As an example, consider this:
1, 2, 3, 4, . . . , 100
What is that list of numbers? You probably interpreted it as “all the natural
numbers between 1 and 100, inclusive”. That seems reasonable. But what if we
actually meant this list?
1, 2, 3, 4, 7, 10, 11, 12, 14, . . . , 100
Why, of course, we meant the list of natural numbers from 1 to 100 that don’t
have an “i” in their English spelling! Wasn’t it obvious?
112 CHAPTER 2. MATHEMATICAL INDUCTION

The point is this: when talking with a friend, and verbalizing some ideas,
it might be okay to write “1, 2, 3, . . . , 100” and ensure that whoever is listening
knows exactly what you mean. In general, though, we can’t assume that a reader
would just naturally intuit whatever we were trying to convey; we should be as
explicit and rigorous as possible.
It may seem to you now like we’re nit-picking, but the larger point is that
there is a mathematical way of making this argument more precise, so that it
constitutes a completely valid proof. Everything we have done so far is useful in
guiding our intuition, but we will have to do a little more work to be sure our
arguments are completely convincing. There are a few other concepts required
to make this type of argument rigorous, in general, and we will investigate those
in the next chapter and return to this subject immediately after that. However,
in the meantime, let’s examine one more example to practice this intuitive
argument style and recognizing when induction is an applicable technique.

2.2.2 Lines On The Plane

Take a clean sheet of paper and a pen and a ruler. How many regions are on
your sheet? Just one, right? Draw a line all the way across the paper. Now
there are two regions. Draw another straight line that intersects your first line.
How many regions are there? You should count four in total. Draw a third line
that intersects both of the first two, but not at the point where the first two
intersect. (That is, there should be three intersection points, in total.) How
many regions are there? Can you predict the answer before counting? What
happens when there are 4 lines? Or 5? Or 100? How do we approach this
puzzle and, ultimately, solve it? Let’s give a more formal statement to be sure
we’re thinking the same way:
Consider n lines on an infinite plane (two-dimensional surface) such that no
two lines are parallel and no more than two lines intersect at one point. How
many distinct regions do the lines create?
We can draw a few examples by hand when n is small (up to, say, n = 5
is reasonable), and let’s use this to guide our intuition into making a general
argument for an arbitrary value of n. (Notice that this strategy is very similar to
what we did in the previous puzzle: identify a pattern with small cases, identify
the relevant components of those cases that can generalize, then abstract to an
aribtrary case.) Specifically, we want to attempt to identify how the number
of regions in one drawing depends on the number of regions in a drawing with
fewer lines. What happens when we draw a new line? Can we determine how
this changes the already existing regions? Can we somehow count how many
regions this creates? Do some investigation of this puzzle on your own before
reading on. If you figure out some results, compare your work to the steps we
follow below.
Let’s start with a small case, say n = 2. We know one line divides a plane
into 2 regions; what happens when we add a second line? We know we get 4
regions, because we can just look and count them:
2.2. EXAMPLES AND DISCUSSION 113

2
1
3
4

However, we are only looking at one specific case of two intersecting lines. How
do we know that we will always find four regions, no matter how we draw those
two lines appropriately? That is, can we describe how this happens in a way
that somehow incorporates the fact that the number of liens is n = 2? Think
about it!
Here’s our approach. Notice that each of the already existing regions is split
into two when we add a second line, and that this is true no matter how you
choose to draw the lines; as long as we make sure the two lines aren’t parallel,
they will always behave this way. That is, if we take one line that splits the
plane into two regions,

2
1

then adding a new line will split each of those existing regions in two. This adds
two new regions to the whole plane, giving four regions in total:

2 n2 split
Regio
1
3
it
n 1 spl 4
Regio

What about when n = 3? In this case, we want to think about adding a

third line to a diagram with two lines and four regions. We want to make
an argument that doesn’t depend on a particular arrangement of the lines, so
eventually the only facts we should use are that no lines are parallel and any
point of intersection only lies on two lines (not three or more). For now, though,
114 CHAPTER 2. MATHEMATICAL INDUCTION

it helps to look at a particular arrangement of lines so that we are talking about

the same diagram; we can use our observations of this specific diagram to guide
our general argument. Let’s start with a two-line diagram, on the left below,
and add a third line, but let’s choose the third line so that all of the intersection
points are “nearby” or within the scope of the diagram, so that we don’t have
to rescale the picture:

2
5 Region 2 split 7
Region 1 split
6 Region 3 split
1
3
4

We certainly have 7 regions now, but we made the third line a separate color
so that we can identify where the “new” regions appear: one region (the lower
quadrant, Region 4) remains unchanged, but the three other regions are split in
two and each of those “splits” adds 1 to our count (where there was 1 region,
now there are 2). What if we had placed the line differently?

6 2
Region 2
split
Region
1 split
5 3
1
4
7 Region
4 split

The same phenomenon occurs, where one quadrant remains untouched but
the other three are split in two. (How do we know there aren’t any other
regions not depicted within the scale of our diagram? This is not as easy a
question to answer as you might think at first blush, and it’s worth thinking
about.) Experiment with other arrangements of the three lines and try to
convince yourself that this always happens; also, think about why this is the
case and how we could explain that this must happen. Before giving a general
explanation, though, let’s examine another small case.
When n = 4, we start with three lines and 7 regions and add a fourth line
that is not parallel to any of the existing lines and doesn’t pass through any
existing intersection points. Again, we will want to make an argument that isn’t
tied to a specific arrangement of the lines, but looking at the following specific
diagram will help guide our intuition into making that argument:
2.2. EXAMPLES AND DISCUSSION 115

5 9 2
7
11
6
1 8 3

4
10

Notice that three of the original regions remain unchanged (Regions 3 and
5 and 7), and the other four are split in two. Do you notice a pattern here?
It seems like for every n we’ve examined, adding the n-th line leaves exactly
n − 1 regions unchanged while the rest are split in two. Let’s try to explain why
this happens. Remember that we’re trying to identify how many regions appear
when we draw n lines, so let’s assign that value a “name” so we can refer to it;
let’s say R(n) represents the number of regions created by drawing n lines on
the plane so that no two lines are parallel and no intersection point belongs to
more than two lines. In these examples we’ve considered for small values of n,
we’ve looked at what changes when we add a new line; that is, we’ve figured out
what R(n) is by already knowing R(n − 1). Let’s try to adapt our observations
so that they apply to any arbitrary value of n.
Assume that we know R(n) already. (Why can we do this? Do we know any
particular value of R(n) for sure, for some specific n? Which? How?) Say we
have an arbitrary diagram of n lines on the plane that satisfy the two conditions
given in the puzzle statement above. How many regions do these lines create?
Yes, exactly R(n). Now, what happens when we add the (n + 1)-th line? What
can we say for sure about this line and how it alters the diagram? Well, the
only information we really have is that (a) this new line is not parallel to any of
the existing n lines and (b) this new line does not intersect any of the already
existing intersection points. Now, condition (a) tells us that this new line must
intersect all of the exisiting n lines; parallel lines don’t intersect, and non-parallel
lines must intersect somewhere. Thus, we must create n new intersection points
on the diagram. Can any of those intersection points coincide with any existing
intersection points? No! This is precisely what condition (b) tells us. These
two pieces of information together tell us that, no matter how we draw this new
line, as long as it satisfies the requirements of the puzzle, we must be able to
identify n “special” points along that line. Those special points are precisely
the points where the new line intersects an existing line.
We’d now like to take these special points and use them to identify new
regions in the diagram. Look back to the cases we examined above: identify
the new intersection points and see if you can associate them with new regions.
Perhaps it would help to label those intersections with a large dot and mark
the new regions with an X to make them all stand out. We’ll show you one
example below, where n = 4. What do you notice? Can you use these dots to
116 CHAPTER 2. MATHEMATICAL INDUCTION

help identify how many new regions are created with the addition of that n-th
line? Think about this for a minute and then read on.

Exactly! Between any two of the new intersection points, we have a line
segment that splits a region in two! All that remains is to identify how many
new such segments we’ve created. Since each one corresponds to exactly one
existing region split in two, this will tell us exactly how many new regions
we’ve created. We’ve already figured out that this (n + 1)-th line creates n new
intersection points. Think about how these points are arranged on the line. Any
two “consecutive” points create a segment, but the extreme points also create
infinite segments (that coninue past those extreme points forever). How many
are there in total? Exactly n + 1. (Look at the diagram above, for n = 3. We
see that there are 3 new intersection points and 4 new segments, with two of
them being infinite rays.) This means there are n + 1 line segments that split
regions in two, so we have created exactly n + 1 new regions! This allows us to
say that
R(n + 1) = R(n) + n + 1
Phew, what an observation! It took a bit of playing around with examples
and making some geometric arguments, but here we are. We’ve identified some
inductive structure to this puzzle; we’ve found how one case depends on another
one. That is, we’ve found how R(n+1) depends on R(n). This hasn’t completely
solved the puzzle, but we are now much closer. All that remains is to replace
R(n) with a similar expression, and continually do this until we reach a value
we know, R(1) = 2. Observe:
R(n + 1) = R(n)
 + n+1
= R(n
 −1) + n + n+1
=  − 2) + (n − 1) +
R(n n + n+1
 

..
.
= R(2)

+ 3
 + ··· + n + n+1
= R(1) + 2 + 3 + ··· + n + n+1
Since we know R(1) = 2, we can say
n+1 n+1
!
X X
R(n + 1) = 2 + (2 + 3 + · · · + n + (n + 1)) = 2 + k −1=1+ k
k=1 k=1
2.2. EXAMPLES AND DISCUSSION 117

and this is a sum we have investigated before! (Also notice that we had to
subract 1 because of the missing first term of the sum in parentheses.) Recall
Pn
that k=1 k = n(n+1)
2 , and to represent the sum we have in the equation above,
we just replace n with n + 1. Therefore,
(n + 1)(n + 2)
R(n + 1) = 1 +
2
One final simplification we would like to make is to replace n+1 with n through-
out the equation, because it makes more sense to have an expression for R(n)
(For what values of n is this valid?)

n(n + 1)
R(n) = 1 +
2
Finally, we have arrived at an answer to the originally-posed puzzle! In so doing,
we employed an inductive technique: we explained how one “fact”, namely the
value of R(n + 1), depends on the value of a “previous fact”, namely R(n),
and used these iterative dependencies to work backwards until we reached a
particular, known value, namely R(1).
We want to point out, again, that the derivation we followed and the obser-
vations we made in this section have guided our intuition into an answer, but
this has not rigorously proven anything. The issue is with the “· · · ”, which is
not a concrete, “officially” mathematical way of capturing the inductive pro-
cess underlying our technique. Furthermore, our method with the “lines in the
plane” problem had us starting with a diagram of n − 1 lines and building a
new diagram with n lines; is this okay? Why does this actually tell us anything
about an arbitrary diagram of n lines? Do all such diagrams come from a smaller
diagram with one fewer line?
We will, in the next two chapters, learn the necessary tools to fully describe
a rigorous way of doing what we have done so far, and in the chapter after
that, we will employ those tools to make mathematical induction officially
rigorous. For now, though, we want to give a heuristic definition of induction and
continue to examine interesting puzzles and observations that rely on inductive
techniques. Practicing with these types of puzzles–learning when to recognize
an inductive process, how to work with it, how to use that structure to solve
a problem, and so on–will be extremely helpful in the future, and we have no
need to delve into technical mathematical detail. (At least, not just yet!)

2.2.3 Questions & Exercises

Remind Yourself
Answering the following questions briefly, either out loud or in writing. These
are all based on the section you just read, so if you can’t recall a specific defi-
nition or concept or example, go back and reread that part. Making sure you
can confidently answer these before moving on will help your understanding and
memory!
118 CHAPTER 2. MATHEMATICAL INDUCTION

(1) What properties characterize an inductive process?

n
X n(n + 1)
(2) How did we prove that k= is correct? How was our method
2
k=1
inductive? (Reread Section 1.4.2 if you forget!)
(3) Why can we take the sum formula mentioned in the previous question and
“replace” n with n+1 and know that it still holds true? Can we also replace
n with n − 1?
(4) Work through the algebraic steps to obtain our final expression for the sum
of the first n squares; that is, verify that
1 1 n(n + 1) 1
(n + 1)3 − (n + 1) − = n(n + 1)(2n + 1)
3 3 2 6

(5) Try to recall the argument that adding the (n + 1)-th line on the plane
created exactly n + 1 new regions. Can you work through the argument for
a friend and convince him/her that it is valid?
(6) To find the sum of the first n squares, why couldn’t we just square the
formula for the sum of the first n numbers? Why is that wrong?

Try It
Try answering the following short-answer questions. They require you to actu-
ally write something down, or describe something out loud (to a friend/class-
mate, perhaps). The goal is to get you to practice working with new concepts,
definitions, and notation. They are meant to be easy, though; making sure you
can work through them will help you!
(1) Draw 5 lines on the plane (that satisfy the two conditions of the puzzle)
and verify that there are 16 regions. Can you also verify that 6 lines yield
22 regions?
(2) Come up with another description of a sequence that goes 1, 2, 3, 4, . . . , 100
that is not simply all of the numbers from 1 to 100. (Recall the example we
gave: all the numbers from 1 to 100 with no “i” in their English spelling.)
(3) Come up with an algebraic expression that relates (n + 1)4 to n4 , like we
did with cubes.
(Challenge: Can you come up with a geometric interpretation for the
expression you just derived?)
(4) Challenge: Let’s bump the “lines in the plane” puzzle up one dimension!
Think about having n planes in three-dimensional space. How many regions
are created? Assume that no two planes are parallel, and no three of them
intersect in one line. (Think about how these two conditions are directly
analogous to the specified conditions for the “lines” puzzle.)
2.3. DEFINING INDUCTION 119

2.3 Defining Induction

To properly motivate the forthcoming definition of mathematical induction
as a proof technique, we want to emphasize that the above examples used some
intuitive notions of the structure of the puzzle to develop a “solution”, where
we use quotation marks around solution to indicate that we haven’t officially
proven it yet. In that sense, we ask the following question: What if we had been
given the formula that we derived and asked to verify it? What if we had not
gone through any intuitive steps to derive the formula and someone just told
us that it is correct? How could we check their claim? The reason we ask this
is because we are really facing that situation now, except the person telling us
the formula is . . . the very same intuitive argument we discovered above!
Pretend you have a skeptical friend who says, “Hey, I heard about this
formula for the sum of the first n natural numbers squared. Somebody told me
that they add up to 61 n(n + 1)(2n + 1). I checked the first two natural numbers,
and it worked, so it’s gotta be right. Pass it on!” Being a logical thinker, but
also a good friend, you nod along and say, “I did hear that, but let’s make sure
it’s correct for every number.” How would you proceed? Your friend is right
that the first few values “work out” nicely:
1
12 = 1 = (1)(2)(3)
6
1
12 + 22 = 5 = (2)(3)(5)
6
1
12 + 22 + 32 = 14 = (3)(4)(7)
6
1
12 + 22 + 32 + 42 = 30 = (4)(5)(9)
6
and so on. We could even check, by hand, a large value of n, if we wanted to:
1
12 + 22 + 32 + 42 + 52 + 62 + 72 + 82 + 92 + 102 = 385 = (10)(11)(21)
6
Remember, though, that this formula is claimed to be valid for any value of
n. Checking individual results by hand would take forever, because there are
an infinite number of natural numbers. No matter how many individual values
of n we check, there will always be larger values, and how do we know that
the formula doesn’t break down for some large value? We need a far more
efficient procedure, mathematically and temporally speaking, to somehow verify
the formula for all values of n in just a few steps. We have an idea in mind,
of course (it’s the upcoming rigorous version of mathematical induction), and
here we will explain how the procedure works, in a broad sense.

2.3.1 The Domino Analogy

Pretend that we have a set of dominoes. This is a special set of dominoes because
we have an infinite number of them (!) and we can imagine anything we want
120 CHAPTER 2. MATHEMATICAL INDUCTION

written on them, instead of the standard array of dots. Let’s also pretend that
they are set up in an infinite line along an infinite tabletop, and we are viewing
the dominos from the side and we can see a label under each one so that we
know where we are in the line:

n=1 n=2 n=3 n=4 n=5

For this particular example, to verify the formula

n
X 1
k2 = n(n + 1)(2n + 1)
6
k=1

we will imagine a particular “fact” written on each domino. Specifically, we will

imagine that the 1st domino has the expression
1
X 1
k2 = (1)(2)(3)
6
k=1

written on it, and the 2nd domino has the expression

2
X 1
k2 = (2)(3)(5)
6
k=1

written on it. In general, we imagine that the n-th domino in the infinite line
has the following “fact” written on it:
n
X 1
k2 = n(n + 1)(2n + 1)
6
k=1

Since we’re dealing with dominos that are meant to fall into each other and
knock each other over, let’s pretend that whenever a domino falls, that means the
corresponding “fact” written on it is a true statement. This is how we will relate
our physical interpretation of the dominos to the mathematical interpretation
of the validity of the formula we derived.
We did check the sum for n = 1 by hand: 12 = 61 (1)(2)(3). Thus, the fact
written on the first domino is a true statement, so we know that the first domino
will, indeed, fall over. We also checked the sum for n = 2 by hand, so we know
that the second domino will fall over:
2.3. DEFINING INDUCTION 121

n=1 n=2 n=3 n=4 n=5

However, continuing like this brings us back to the same problem as before: we
don’t want to check every individual domino to make sure it falls. We would
really like to encapsulate our physical notion of the line of dominos—namely,
that when a domino falls it will topple into the next one and knock that over, and
so on—and somehow relate the “facts” that are written on adjacent dominos.
Let’s look at this situation for the first two dominos. Knowing that Domino
1 falls, can we guarantee that Domino 2 falls without rewriting all of the terms
of the sum? How are the statements written on the two dominos related? Each
statement is a sum of squared natural numbers, and the one on the second
domino has exactly one more term. Thus, knowing already that Domino 1 has
fallen, we can use the true statement written on Domino 1 to verify the truth
of the statement written on Domino 2:

2
X 1
k 2 = 12 + 22 = 1 + 22 = 5 = (2)(3)(5)
6
k=1

Now, this may seem a little silly because the only “work” we have saved is not
having to “do the arithmetic” to write 12 = 1. Let’s use this procedure on a
case with larger numbers so we can more convincingly illustrate the benefit of
this method. Let’s assume that Domino 10 has fallen. (In case you are worried
about this assumption, we wrote the full sum a few paragraphs ago and you can
verify it there.) This means we know that

10
X 1
k2 = (10)(11)(21) = 385
6
k=1

is a true statement. Let’s use this to verify the statement written on Domino
11, which is
11
X 1
k2 = (11)(12)(23)
6
k=1

The sum written on Domino 11 has 11 terms, and the first 10 are exactly the
sum written on Domino 10! Since we know something about that sum, let’s just
separate that 11th term from the sum and apply our knowledge of the other
122 CHAPTER 2. MATHEMATICAL INDUCTION

terms:
11
X
k 2 = 12 + 22 + · · · + 102 + 112

k=1
10
!
X
2
= k + 112
k=1
= 385 + 121
= 506
1 1
= 3036 = (11)(12)(23)
6 6
Look at all of the effort we saved! Why bother reading the first 10 terms of the
sum if we know something about them already?
Now, imagine if we could do this procedure for all values of n, simultaneously!
That is, imagine that we could prove that any time Domino n falls, we are
guaranteed that Domino (n + 1) falls. What would this tell us? Well, think
about the infinite line of dominos again. We know Domino 1 will fall, because
we checked that value by hand. Then, because we verified the “Domino n knocks
over Domino (n + 1)” step for all values of n, we know Domino 1 will fall into
Domino 2, which in turn falls into Domino 3, which in turns falls into Domino
4, which . . . The entire line of dominos will fall! In essence, we could collapse
the whole line of dominos falling down into just two steps:
(a) Make sure the first domino topples;
(b) Make sure every domino knocks over the one immediately after it in line.
With only these two steps, we can guarantee every domino falls and, therefore,
prove that all of the facts written on them are true. This will prove that the
formula we derived is valid for every natural number n.
We have already accomplished step (a), so now we have to complete step
(b). We have done this for specific cases in the previous paragraphs (Domino
1 topples Domino 2, and Domino 10 topples Domino 11), so let’s try to follow
along with the steps of those cases and generalize to an arbitrary value of n. We
assume, for some specific but arbitrary value of n, that Domino n falls, which
tells us that the equation
n
X 1
k2 = n(n + 1)(2n + 1)
6
k=1

is a true statement. Now, we want to relate this to the statement written on

Domino (n + 1) and apply the knowledge given in the equation above. Let’s do
what we did before and write a sum of n + 1 terms as a sum of n terms plus the
last term:
n+1 n
!
X X
2 2 2 2 2
k = 1 + 2 + · · · + n + (n + 1) = k + (n + 1)2
2

k=1 k=1
2.3. DEFINING INDUCTION 123

Next, we can apply our assumption that Domino n has fallen (which tells us
that the fact written on it is true) and write
n+1
X 1
k2 = n(n + 1)(2n + 1) + (n + 1)2
6
k=1

Is this the same as the fact written on Domino (n + 1)? Let’s look at what that
is, first, and then compare. The “fact” on Domino (n + 1) is similar to the fact
on Domino n, except everywhere we see “n” we replace it with “n + 1”:
n+1
X 1 1
k2 = (n + 1)((n + 1) + 1)(2(n + 1) + 1) = (n + 1)(n + 2)(2n + 3)
6 6
k=1

It is not clear yet whether the expression we have derived thus far is actually
equal to this. We could attempt to simplify the expression we’ve derived and
factor it to make it “look like” this new expression, but it might be easier to just
expand both expressions and compare all the terms. (This is motivated by the
general idea that expanding a factored polynomial is far easier than recognizing
a polynomial can be factored.) For the first expression, we get

1 1
n(n + 1)(2n + 1) + (n + 1)2 = n(2n2 + 3n + 1) + (n2 + 2n + 1)
6 6
1 1 1
= n3 + n2 + n + n2 + 2n + 1
3 2 6
1 3 3 2 13
= n + n + n+1
3 2 6
and for the second expression, we get
1 1
(n + 1)(n + 2)(2n + 3) = (n + 1)(2n2 + 7n + 6)
6 6
1 3
(2n + 7n2 + 6n) + (2n2 + 7n + 6)

=
6
1 3 13
= n3 + n2 + n + 1
3 2 6
Look at that; they’re identical! Also, notice how much easier this was than try-
ing to rearrange one of the expressions and “morph” it into the other. We proved
they were identical by manipulating them both and finding the same expression,
ultimately. Now, let’s look back and assess what we have accomplished:

1. We likened proving the validity of the formula

n
X 1
k2 = n(n + 1)(2n + 1)
6
k=1

for all values of n to knocking over an infinite line of dominos.

124 CHAPTER 2. MATHEMATICAL INDUCTION

2. We verified that Domino 1 will fall by checking the formula corresponding

to that case by hand.
3. We proved that Domino n will fall into Domino (n+1) and knock it over by
assuming the fact written on Domino n is true and using that knowledge
to show that the fact written on Domino (n + 1) must also be true.
4. This guarantees that all dominos will fall, so the formula is true for all
values of n!
Are you convinced by this technique? Do you think we’ve rigorously proven
that the formula is valid for all natural numbers n? What if there were a value
of n for which the formula didn’t hold? What would that mean in terms of our
domino scheme?
Remember that this domino analogy is a good intuitive guide for how in-
duction works, but it is not built on mathematically rigorous foundations. That
will be the goal of the next couple of chapters. For now, let’s revisit the other
example we’ve examined in this section: lines in the plane. Again, the use of
ellipses in our derivation of the formula R(n) is troublesome and we want to
avoid it. Let’s try to follow along with the domino scheme in the context of this
puzzle.
Imagine that we have defined the expression R(n) to represent the number
of distinct regions in the plane created by n lines, where no two lines are parallel
and no three intersect at one point. Also, imagine that on Domino n we have
written the “fact” that “R(n) = 1 + n(n+1) 2 ”. Can we follow the same steps as
above and verify that all the dominos will fall?
First, we need to check that Domino 1 does, indeed fall. This amounts to
verifying the statement: “R(1) = 1+ 1(2) 2 = 1+1 = 2”. Is this a true statement?
Yes, of course, we saw this before; one line divides the plane into two regions.
Second, we need to prove that Domino n will topple into Domino (n + 1) for
any arbitrary value of n. That is, let’s assume that “R(n) = 1 + n(n+1) 2 ” is a
(n+1)(n+2)
true statement for some value of n and show that “R(n + 1) = 1 + 2 ”
must also be a true statement. How can we do this? Well, let’s follow along
with the argument we used before to relate R(n + 1) to R(n). By considering
the geometric consequences of adding an extra line to any diagram with n lines
(that also fit our rules about the lines) we proved that R(n + 1) = R(n) + n + 1.
Using this knowledge and our assumption about Domino n falling, we can say
that
n(n + 1)
R(n + 1) = R(n) + n + 1 = 1 + +n+1
2
Is this the same expression as what is written on Domino (n + 1)? Again, let’s
simplify both expressions to verify they are the same. We have
n(n + 1) n2 + n 1 3
1+ +n+1=2+n+ = n2 + n + 2
2 2 2 2
and
(n + 1)(n + 2) n2 + 3n + 2 1 3
1+ =1+ = n2 + n + 2
2 2 2 2
2.3. DEFINING INDUCTION 125

Look at that; they’re identical! Thus, we have shown that Domino n is guaran-
teed to fall into Domino (n + 1), for any value of n. Accordingly, we can declare
that all dominos will fall!
Think about the differences between what we have done with this “domino
technique” and what we did before to derive the expressions we just proved.
Did we use any ellipses in this section? Why is it better to prove a formula
this way? Could we have used the domino induction technique to derive the
formulas themselves?

2.3.2 Other Analogies

The Domino Analogy is quite popular, but it’s not the only description of how
induction works. Depending on what you read, or who you talk to, you might
learn of a different analogy, or some other kind of description altogether. Here,
we’ll describe a couple that we’ve heard of before. It will help solidify your
understanding of induction (at least as far as we’ve developed it) to think about
how these analogies are all equivalent, fundamentally.

Mojo the Magical, Mathematical Monkey

Imagine an infinite ladder, heading straight upwards into the sky. There are
infinitely-many rungs on this ladder, numbered in order: 1, 2, 3, and so on
going upwards. Our friend Mojo happens to be standing next to this ladder.
He is an intelligent monkey, very interested in mathematics but also a little bit
magical, because he can actually climb up this infinite ladder!
If Mojo makes it to a certain rung on the ladder, that means the fact cor-
responding to that number is True. How can we make sure he climbs up the
entire ladder? It would be inefficient to check each rung individually. Imagine
that: we would have to stand on the ground and make sure he got to Rung 1,
then we would have to look up a bit and make sure he got to Rung 2, and then
Rung 3, and so on . . . Instead, let’s just confirm two details with Mojo before he
starts climbing. Is he going to start climbing? That is, is he going to make it to
Rung 1? If so, great! Also, are the rungs close enough together so that he can
always reach the next one, no matter where he is? If so, even greater! These
are exactly like the conditions established in our Domino Analogy. To ensure
that Mojo gets to every rung, we just need to know he gets to the first one and
that he can always get to the next one.

Doug the Induction Duck

Meet Doug. He’s a duck. He also loves bread, and he’s going to go searching
through everyone’s yards to find more bread. These yards are all along Induction
Street in Math Town, where the houses are numbered 1, 2, 3, and so on, all in
a row.
Doug starts in the yard of house #1, looking for bread. He doesn’t find any,
so he’s still hungry. Where else can he look? The house next door, #2, has a
126 CHAPTER 2. MATHEMATICAL INDUCTION

backyard, too! Doug heads that way, his tummy rumbling. He doesn’t find any
bread there, either, so he has to keep looking. He already knows house #1 has
no bread, so the only place to go is next door to house #3. We think you see
where this is going . . .
If we were keeping track of Doug’s progress, we might wonder whether he
eventually gets to every yard. Let’s say we also knew ahead of time that no-
body has any bread. This means that whenever he’s in someone’s yard, he will
definitely go to the next house, still searching for a meal. This means that he
will definitely get to every house! That is, no matter which house we live in, no
matter how large the number on our front door might be, at some point we will
see Doug wandering around our backyard. (Unfortunately, he will go hungry all
this time, though! Poor Doug.)

2.3.3 Summary
Let’s reconsider what we’ve accomplished with the two example puzzles we’ve
seen thus far, and the analogies we’ve given. In our initial consideration of each
puzzle, we identified some aspect of the structure of the puzzle where a “fact”
depended on a “previous fact”. In the case of the cubic numbers, we found a
way to express (n + 1)3 in terms of n3 ; in the case of the lines in the plane,
we described how many regions were added when an extra line was added to
a diagram with n lines. From these observations, we applied this encapsulated
knowledge over and over until we arrived at a “fact” that we knew, for a “small”
value of n (in both cases, here, n = 1). This allowed us to derive a formula or
equation or expression for a general fact that should hold for any value of n.
This work was interesting and essential for deriving these expressions, but
it was not enough to prove the validity of the expressions. In doing the work
described above, we identified the presence of an inductive process and utilized
its structure to derive the expressions in question. This was beneficial in two
ways, really; we actually found the expressions we wanted to prove and, by
recognizing the inductive behavior of the puzzle, we realized that proving the
expressions by mathematical induction would be prudent and efficient.
For the actual “proof by induction”, we followed two main steps. First,
we identified a “starting value” for which we could check the formula/equation
by hand. Second, we assumed that one particular value of n made the cor-
responding formula hold true, and then used this knowledge to show that the
corresponding formula for the value n + 1 must also hold true. Between those
two steps, we could rest assured that “all dominos will fall” and, therefore, the
formulas would hold true for all relevant values of n.

One Concern: What’s at the “top” of the ladder?

You might be worried about something, and we’ll try to anticipate your question
here. (We only bring this up because it’s a not uncommon observation to make.
If you weren’t thinking about this, try to imagine where the idea would come
from.) You might say, “Hey now, I think I see how Mojo is climbing the ladder,
2.3. DEFINING INDUCTION 127

but how can he actually get all the way to the top? It’s an infinite ladder, right?
And he never gets there . . . right?”
In a way, you would be right. Since this magical ladder really does go
on forever, then there is truly no end to it and Mojo will never get “there”.
However, that isn’t the point; we don’t care about any “end” of the ladder (and
not just because there isn’t one). We just need to know that Mojo actually gets
to every possible rung. He doesn’t have to surpass all of them and stand at the
top of the ladder, looking down at where he came from. That wasn’t the goal!
Think of it this way: pretend you have a vested interest in some particular
fact that we’re proving. Let’s say it’s Fact #18,458,789,572,311,000,574,003.
(Some huge number. It doesn’t matter, really.) Its corresponding rung is
waaayyyyyy up there on the ladder, and all you care about is whether or not
Mojo gets there on his journey. Does he? . . . You bet he does! It might take a
long time (how many steps will it take?), but in this magical world of monkeys
and ladders, who cares about time anyway! You know that he’ll eventually
get there, and that makes you happy. Now, just imagine that for each fact,
there’s somebody out there in that magical world that cares about only that
fact. Surely, everyone will be happy with the knowledge that Mojo will get to
their rung on his journey. Nobody cares about whether he gets to the top; that
isn’t their concern. Meanwhile, out here in our regular, non-magical world, we
are extremely happy with the fact that everyone in that world will eventually be
happy. That entire infinite process of ladder-climbing was condensed into just
two steps, and with only those two steps, we can rest assured that every rung
on that ladder will be touched. Every numbered fact is true.
Think about this in terms of the Domino Analogy, as well. Do we care
whether or not there is some “ending point” of the line of dominoes, so that
they all fall into a wall somewhere? Of course not; the line goes on forever.
Every domino will eventually fall over, and we don’t even care how “long” that
takes. Likewise, we know Doug will get to everyone’s yard; we don’t care “when”
he gets to any individual yard, just that he gets to all of them.

2.3.4 Questions & Exercises

Remind Yourself
Answering the following questions briefly, either out loud or in writing. These
are all based on the section you just read, so if you can’t recall a specific defi-
nition or concept or example, go back and reread that part. Making sure you
can confidently answer these before moving on will help your understanding and
memory!

(1) How are the Domino, Mojo, and Doug analogies all equivalent? Can you
come up with some “function” that describes their relationship, that con-
verts one analogy into another?

(2) Find a friend who hasn’t studied mathematical induction before, and try to
describe it. Do you find yourself using one of the analogies? Was it helpful?
128 CHAPTER 2. MATHEMATICAL INDUCTION

(3) Why is it the case that our work with the cubes didn’t prove the summation
formula? Why did we still need to go through all that work?

(4) Think about the Domino Analogy. Is it a problem that the line of dominoes
goes on forever? Does this mean that there are some dominoes that will
never fall down? Try to describe what this means in terms of the analogy.

Try It

Try answering the following short-answer questions. They require you to actu-
ally write something down, or describe something out loud (to a friend/class-
mate, perhaps). The goal is to get you to practice working with new concepts,
definitions, and notation. They are meant to be easy, though; making sure you
can work through them will help you!

(1) Work through the inductive steps to prove the formula

n
X n(n + 1)
k=
2
k=1

(2) Work through the inductive steps to prove the formula

n
X
(2k − 1) = n2
k=1

(3) Work through the inductive steps to prove the formula

n  2
X n(n + 1)
k3 =
2
k=1

(4) Suppose we have a series of facts that are indexed by natural numbers. Let’s
use the expression “P (n)” to represent the n-th fact.

(a) If we want to prove every instance is True, for every natural number n,
how can we do this?
(b) What if we want to prove that only every even value of n makes a True
statement? Can we do this? Can you come up with a modification of
one of the analogies we gave that would describe your method?
(c) What if we want to prove that only every value of n greater than or
equal to 4 makes a True statement? Can we do this? Can you come up
with a modification of one of the analogies we gave that would describe
your method?
2.4. TWO MORE (DIFFERENT) EXAMPLES 129

2.4 Two More (Different) Examples

This short section serves a few purposes. For one, we don’t want you to get the
idea, right away, that induction is all about proving a numerical formula with
numbers and polynomials. Induction is so much more useful than that! One
of the following examples, in particular, will be about proving some abstract
property is true for any “size” of the given situation. You will see how it still
falls under the umbrella of “induction”, but you will also notice how it is dif-
ferent from the previous examples. Furthermore, these examples will illustrate
that sometimes we need to know “more information” to knock over some domi-
noes. In the previous examples, we only needed to know that Domino n fell to
guarantee that Domino n + 1 will fall. Here, though, we might have to know
about several previous dominoes. After these two examples, we will summarize
how this differs from the domino definition given above, and preview a broader
definition of the technique of induction, as it applies to these examples.

2.4.1 Dominos and Tilings

This next example is a little more complicated than the first two. We will still
end up proving a certain numerical formula, but the problem is decidedly more
visual than just manipulating algebraic expressions. Furthermore, we’ll notice
an interesting “kink” in the starting steps, where we have to solve a couple of
“small cases” before being able to generalize our approach. This will be our first
consideration of how the technique of induction can be generalized and adapted
to other situations.
The question we want to answer is nicely stated as follows:

Given a 2 × n array of squares, how many different ways can we tile

the array with dominoes? A tiling must have every square covered
by one–and only one–domino.

For example, the following are proper tilings

whereas the following are not proper tilings

130 CHAPTER 2. MATHEMATICAL INDUCTION

As before, let’s examine the first few cases–where n = 1, 2, 3, and so on–and

see if we notice any patterns. Try working with the problem yourself, before
reading on, even!
When n = 1, we have an array that is exactly the shape of one domino, so
surely there is only one way to do this. Let’s use the notation T (n) to represent
the number of tilings on a 2 × n array. Thus, T (1) = 1.

When n = 2, we have a 2 × 2 array. Since the orientation of the array matters,

we have each of the following distinct tilings. Thus, T (2) = 2.

What about when n = 3? Again, we can enumerate these tilings by hand and
be sure that we aren’t missing any. We see that T (3) = 3.

Okay, one more case, when n = 4. We see that T (4) = 5.

2.4. TWO MORE (DIFFERENT) EXAMPLES 131

Can we start to find a pattern now? Writing out larger arrays will just
be tiresome! Let’s think about how we could have used the fact that T (1) =
1 to deduce something about T (2) . . . Well, wait a minute . . . We couldn’t,
right? There was something fundamentally different about those two cases.
Specifically, because dominoes are 2 × 1 in size, the fact that we only added one
row to the array didn’t help us.

Alright, let’s consider n = 3, then. Could we use the fact that T (2) = 2 at
all? In this case, yes! Knowing there were two tilings of the 2 × 2 array, we
could immediately build two tilings of the 2 × 3 array without much thought.
Specifically, we can just append a vertical domino to each of those previous
tilings. But we know now that T (3) = 3. Where did the third tiling come
from? Look at that tiling again and how it compares to the other two. In that
third tiling, the dominoes on the right side are horizontal, as opposed to the
vertical one in the other two tilings. If we remove those two parallel, horizontal
dominoes, we are left with precisely the situation when n = 1. Put another way,
we can build a tiling of a 2 × 3 array by appending a square of two horizontal
dominoes to the right side. In total, then, we have described all of the tilings of
a 2 × 3 board in terms of boards of smaller sizes, namely 2 × 2 and 2 × 1:

T (2) = 2

T (1) = 1

T (3) = 3 = 2 + 1 = T (2) + T (1)

Now you might see how the pattern develops! Let’s show you what happens
when n = 4, how we can construct all of the tillings that make up T (4) by
appending a vertical domino to each of the tilings that make up T (3), or by
appending two horizontal dominoes to each of the tilings that make up T (2):
132 CHAPTER 2. MATHEMATICAL INDUCTION

T (3) = 3

T (2) = 2

T (4) = 5 = 3 + 2 = T (3) + T (2)

Notice, as well, that no tiling for the 2 × 4 array was produced twice in this
way. (Think carefully about why this is true. How can we characterize the two
types of tilings in a way that will guarantee they don’t coincide at all?) With
this information, we can immediately conclude that T (4) = T (3) + T (2).
Furthermore, we can generalize this argument; nothing was special about
n = 4, right? For any particular n, we can just consider all possible tilings, and
look at what happens on the far right-hand side of the array: either we have
one vertical domino (which means the tiling came from a 2 × (n − 1) array) or
two horizontal dominoes (which means the tiling came from a 2 × (n − 2) array).
With confidence in this argument, we can conclude that
T (n) = T (n − 1) + T (n − 2)
for all of the values of n for which this expression makes sense. What values are
those? Remember that we had to identify T (1) and T (2) separately; this argu-
ment doesn’t apply to those values. Accordingly, we have to add the restriction
n ≥ 3 for the equation above to hold true.
With this information, we can then easily figure out T (n) for any value of
n, given enough time. We could write a computer program fairly easily, even.
It was this inductive argument, though—the pattern that we noticed and our
thorough description of why it occurs—that allowed us to make the conclusion
in the first place. In this case, too, it just so happens that the value of every
term, T (n), depends on the value of two previous terms, T (n − 1) and T (n − 2).
This did not happen in our previous examples in this chapter, and it hints at
something deeper going on here. Do you see how our previous definition of
induction, and the domino analogy, doesn’t exactly apply here anymore? How
might you try to amend our analogy to explain this kind of situation? Think
2.4. TWO MORE (DIFFERENT) EXAMPLES 133

about these issues for a bit and then read on. We’ll talk about them more
in-depth after the next example.
By the way, did you notice something interesting about our solution to this
example? Do you know any other sequences of numbers that behave similarly?
Think about it . . .

2.4.2 Winning Strategies

This example will be our first induction puzzle that doesn’t prove a numerical
formula! It might seem strange to think about that, but it’s true, as you’ll see.
This is actually more common in mathematics than you might think, too: a
problem or mathematical object might have some underlying inductive structure
without depending on something algebraic or arithmetic.
In fact, we will be discussing a game. It’s a game in the usual sense—there
are rules to be followed by two players and there is a clear winner and loser—but
it’s also a game in the mathematical sense, where we can formulate the rules and
playing situations using mathematical notation and discuss formal strategies in
an abstract way. We can even solve the game. This is very different than say,
the game of baseball.
Let’s discuss the rules for this game, which we shall call “Takeaway”, for
now. There are two players, called P1 and P2 . The player P1 goes first every
time. The players start with two piles of stones in the middle of a table, each pile
containing exactly n stones, where n is some natural number. (To distinguish
the different versions of the game, we will say the players are “playing Tn ” when
there are n stones per pile.) On each player’s move, they are allowed to remove
any number of stones from either pile. It is illegal, though, to remove stones
from both piles at once. The player who removes the final stone from the piles
is the winner.
Try playing Takeaway with some friends. Use pennies or candies or penny
candy as stones. Try it for different values of n. Try switching roles so you
are P1 and then P2 . Try to come up with a winning strategy, a method of
playing that maximizes your chances of winning. Try to make a conjecture for
what happens for different values of n. Who is “supposed” to win? Can you
prove your claim? Seriously, play around with this game and attempt to prove
something before reading on for our analysis thereof. You might be surprised
by what you can accomplish!
134 CHAPTER 2. MATHEMATICAL INDUCTION

As with the other examples, let’s use some small values of n to figure out
what’s really going on, then try to generalize. When n = 1, this game is rather
silly. P1 must empty one pile of its only stone, then P2 gets the only remaining
stone in the other pile. Thus, P2 wins. (Notice that it doesn’t matter which
of the two piles P1 picks from, P2 will always get the other one. We might say
that P1 picks the pile on the left “without loss of generality” because it doesn’t
matter either way; the situations are equivalent, so we might as well say it’s the
left pile to have something concrete to say. We will explore this idea of “without
loss of generality” later on when we discuss mathematical logic, too.)

P1’s turn P2’s turn P2 wins!

When n = 2, we now have a few cases that might appear. Think about
P1 ’s possible moves. Again, they might act on either the left or right pile, but
because they’re ultimately identical and we can switch the two piles, let’s just
say (without loss of generality) that P1 removes some stones from the left pile.
How many? It could be one or two stones. Let’s examine each case separately.

P2 ’s turn
P2 wins!

P1 ’s turn
P2 ’s turn
???

If P1 removes both stones, how should P2 react? Duh, they should take the
other pile, so P1 probably shouldn’t have made that move in the first place.
However, P1 might not be thinking straight or something and, besides, we need
to consider all possible situations here to fully analyze this game. Thus, in this
case (the top line in the above diagram) P2 wins. Okay, that’s the easy situation.
What if P1 removes just one stone from the left pile (the bottom line above)?
How should P2 react? We now have some options:

• If P2 removes the other stone from the left pile . . . well, P1 takes the other
pile and P1 wins.

• If P2 removes both stones from the right pile . . . well, P1 takes the last
stone from the left pile and P1 wins.
2.4. TWO MORE (DIFFERENT) EXAMPLES 135

• However, if P2 removes just one stone from the right pile . . .

P1’s turn P2’s turn

Now we have exactly the same situation presented by T1 , which we already

analyzed! It is, again, P1 ’s move first, so we know what will happen: P2 wins
no matter what. If you are player P2 , this is obviously the best move: you win
no matter how P1 responds!
Stepping back for a second, let’s think about what this has shown: no matter
what P1 does first (remove one or two stones from either pile), there is some
possible response that P2 can make that will guarantee a win for P2 , regardless
of P1 ’s subsequent response. Wow, P2 is sitting pretty! Let’s see if this happens
for other values of n.
When n = 3, we will again assume (without loss of generality) that player
P1 acted on the left pile. They could remove one, two, or three stones:

• If P1 removes all three, P2 responds by taking the other pile completely

and wins.

• If P2 removes two stones . . . well, what should player P2 do?

Finishing off that left pile is stupid (because P1 can take the whole right pile
and win), and pulling the entire right pile is similarly stupid (because P1 can
take the whole left pile and win), so something in between is required. Now,
if P2 removes just one stone from the right pile, notice that P1 can respond
with the same action; this leaves exactly one stone in both piles, but the roles
reversed. With P2 going first in such a situation, they are now bound to lose,
per our previous analysis. Bad move, P2 !

P1 wins!

P1’s turn P2’s turn P1’s turn

Let’s try again. If P2 removes two stones from the right pile instead . . . look
at that! We now have exactly one stone in each pile, with P1 up first, so we
know P1 is going to lose. P2 strikes again!
136 CHAPTER 2. MATHEMATICAL INDUCTION

P2 wins!

P1’s turn P2’s turn

Think about the case where n = 4 for a minute, and you’ll find the exact
same analysis occurring. You’ll another possibility to consider: player P1 can
remove one, or two, or three, or four stones from the left pile. Whatever they
do, though, you’ll find that P2 can just mimic that action on the other pile,
reducing the whole game to a previous, smaller version of the game, where P2
was shown to be guaranteed a win! It looks like P2 is in the driver’s seat the
whole time, since they can respond to whatever P1 does, making an identical
move on the other pile. No matter what P1 does, there is always a response for
P2 that means they win, regardless of P1 ’s subsequent moves. In this sense, we
say “P2 has a winning strategy”. There is a clear and describable method for
P2 to assess the game situation and choose a specific move to guarantee a win.
How might we prove this? How does this even fit into this chapter on
induction? It might be hard to see, at the moment. What are we really even
proving here? What are the dominoes or rungs in our analogy for this problem?
In wrapping your brain around this example, you should hopefully realize the
following: induction is not about algebraic formulas all the time; induction
represents some kind of “building-up” structure, where larger situations depend
on smaller ones; we have to prove some initial fact, and then argue how an
arbitrary, larger fact can be reduced so that it depends on a previous fact. This
is really what the dominoes analogy is meant to accomplish. It just so happens
that this analogy is particularly illustrative for certain induction problems (but
not all) and is visualizable and memorable. It does not perfectly apply to all
situations, though.
Read back through these four examples from this chapter and think about
how they are similar and how they are different. Try to come up with a more
precise, mathematical description of mathematical induction using some better
terminology, perhaps of your own invention. (By this, we mean something
better than our intuitive analogy. You’d be surprised at how well you might be
able to describe induction without really knowing what you “ought” to say, and
you’ll actually learn a lot, in the process!) In due time, we will have a rigorous
statement to make, and prove, about mathematical induction and its various
forms. In the meantime, we need to take a trip through some other areas of
mathematics to build up the necessary language, notation, and knowledge to
come back and tackle this problem. Before we go, though, we should mention
a few useful applications of mathematical induction.
2.5. APPLICATIONS 137

2.4.3 Questions & Exercises

Remind Yourself
Answering the following questions briefly, either out loud or in writing. These
are all based on the section you just read, so if you can’t recall a specific defi-
nition or concept or example, go back and reread that part. Making sure you
can confidently answer these before moving on will help your understanding and
memory!
(1) How are these two examples inductive? In what ways are they similar to
the previous examples, with the cubes and lines? In what ways are they
different?
(2) With the domino tilings, how many previous values did we need to know to
compute T (n)?
(3) What is the difference between writing T (n) = T (n − 1) + T (n − 2) and
T (n + 2) = T (n + 1) + T (n)?
(4) What is the winning strategy in the Takeaway game? Try playing with a
friend who doesn’t know the game, and use that strategy as player P2 . How
frustrated do they get every time you win? Do they start to catch on?

Try It
Try answering the following short-answer questions. They require you to actu-
ally write something down, or describe something out loud (to a friend/class-
mate, perhaps). The goal is to get you to practice working with new concepts,
definitions, and notation. They are meant to be easy, though; making sure you
can work through them will help you!
(1) What is T (5)? Can you draw all of those tilings?
(2) Work through the possibilities for takeaway with two piles of 4 stones. Can
you make sure that player P2 always has a winning move?
(3) Challenge: What happens if you play Takeaway with three piles of equal
sizes? Can you find a winning strategy for either player? Try playing with
a friend and see what happens!
(4) Look up the Fibonacci numbers. How are they related to the sequence of
numbers T (n) we found in the domino tiling example?

2.5 Applications
2.5.1 Recursive Programming
The concepts behind mathematical induction are employed heavily in computer
Pn
science, as well. Think back to how we first derived the formula for k=1 k 2 .
138 CHAPTER 2. MATHEMATICAL INDUCTION

Once we had a way to represent a cubic number in terms of a smaller cube and
some leftover terms, we repeated this subsitution process over and over until
we arrived at the “simplest” case, namely, the one that we first observed when
starting the problem: 23 = 1+3+3+1. Recursive programming takes advantage
of this technique: to solve a “large” problem, identify how the problem depends
on “smaller” cases, and reduce the problem until one reaches a simple, known
case.
A classical example of this type of technique is seen in writing code to com-
pute the factorial function, n!, which is defined as the product of the first n
natural numbers:
n! = 1 · 2 · 3 · · · (n − 1) · n
This is a simple definition that we, as humans, intuitively understand, but
telling a computer how to perform this product doesn’t work quite the same
way. (Try it! How do you say “and just keep going until you reach n” in
computer code?) A more efficient way to program the function, and one that
models the mathematically inductive definition, in fact, is to have one program
recursively call itself until it reaches that “simple” case. With the factorial
function, that case is 1! = 1. For any other value of n, we can simply apply the
knowledge that
n! = (n − 1)! · n
over and over to compute n!. Consider the following pseudocode that represents
this idea:
factorial(n):

if n = 1
return 1
else
return n * factorial(n-1)
end

We know that 1! = 1, so if the program is asked to compute that, the correct

value is returned right away. For any larger value of n, the program refers to
itself and says, “Go back and compute (n−1)! for me, then I’ll add a factor of n
at the end, and we’ll know the answer.” To compute (n − 1)!, the program asks,
again, if the input is 1; if not, it calls itself and says, “Go back and compute
(n−2)! for me, then I’ll add a factor of n−1 at the end.” This process continues
until the program returns 1! = 1. From there, it knows how to find 2! = 1 × 2,
and then 3! = 2! × 3, and so on, until n! = (n − 1)! × n.
Another example involving recursive programming arises with the Fibonacci
numbers. You may have seen this sequence of numbers before in a mathematics
course. (In fact, we even mentioned them in the last section, with the domino
tilings!) You also might have heard about how they appear in nature in some in-
teresting and strange ways. (The sequence was first “discovered” by the Italian
mathematician Leonardo of Pisa while studying the growth of rabbit popula-
tions.) The first two numbers in the sequence are specified to be 1, and any
2.5. APPLICATIONS 139

number in the sequence is defined as the sum of the previous two. That is, if
we say F (n) represents the n-th Fibonacci number, then

F (1) = 1 and F (2) = 1 and F (n) = F (n − 1) + F (n − 2) for every n ≥ 3

Now, what is F (5)? Or F (100)? Or F (10000)? This can be handled quite easily
by a recursive program. The idea is the same: if the program refers to either one
of the “simple cases”, i.e. F (1) or F (2), then it will know to return the correct
value of 1 immediately. Otherwise, it will call itself to compute the previous two
numbers and then add those together. Look at the pseudocode below and think
about how it works. What would happen if we used this program to compute
F (10)? How would it figure out the answer?

Fibonacci(n):

if n = 1 or n = 2
return 1
else
return Fibonacci(n-1) + Fibonacci(n-2)
end

This follows the same idea as the factorial program above (let the program
call itself to compute values for “smaller” cases of the function until we reach
a known value) but there’s something a little deeper going on here. If we
were to input n = 10 into the program, it would recognize that it does not
know the output value yet, and it will call itself to compute Fibonacci(9) and
Fibonacci(8). In each of those calls to the program, it would again recognize
the value is as yet unknown. Thus, it would call upon itself again to compute
Fibonacci(8) and Fibonacci(7), but also Fibonacci(7) and Fibonacci(6).
That’s right, the program calls itself multiple times with the same input value.
To compute F (9), we need to know F (8) and F (7), but meanwhile, to compute
F (8), we also need to know F (7) and F (6). In this way, we end up calling the
program Fibonacci many times.
Try to compare the programs Fibonacci and factorial, especially in re-
gards to the inductive processes we have been investigating in this chapter. Do
they use similar ideas? How do they relate to the “domino” analogy of math-
ematical induction that we outlined? Think of the “fact” written on Domino
n as being the computation of the correct value of n! or F (n). How does the
analogy work in each case? Will all the dominos fall? Keep these questions in
mind as you read on. There is some very powerful mathematics underlying all
of these ideas.

2.5.2 The Tower of Hanoi

Let’s take a short break and play a game. Well, it’s not exactly a break because
this is, in a sense, an inductive game, so it’s completely relevant. But it is a
140 CHAPTER 2. MATHEMATICAL INDUCTION

game, nevertheless! The Tower of Hanoi is a very popular puzzle, partly be-
cause it involves such simple equipment and rules. Solving it is another matter,
though!
Imagine that we have three vertical rods and three disks of three different
sizes (colored blue, green, and red) stacked upon each other like so:

The goal is to move all three disks to another rod (either the middle or the right
one, it doesn’t matter) by following these rules:
1. A single move consists of moving one (and only one) disk from the top of
the stack on any rod and moving it to the top of the stack on another rod.
2. A disk cannot be placed on top of a smaller disk.
That’s it! Two simple rules, but a difficult game to play. Try modeling the
game with a few coins or playing cards or whatever you have handy. (You can
even buy Tower of Hanoi sets at some games stores.) Can you solve it? How
many moves did it take you? Is your solution the “best” one? Why or why not?
We mentioned that this is an inductive game, so let’s explore that idea
now. We want to consider how many moves it takes to solve the puzzle (where
one move accounts for moving one disk from one rod to another) and, more
specifically, identify the smallest possible number of moves it would take to
solve the puzzle. To solve the puzzle with three disks, we could keep moving
the smallest disk back and forth between two rods and generate 100 moves, if
we wanted to, and then solve it, but that’s certainly not the best way to do
it, right? Let’s say we found a way to solve the puzzle in a certain number of
moves; how could we show that the number of moves we used is the smallest
possible number of moves?
To address this question, we want to break down the method of solving the
puzzle recursively. In doing so, we are actually going to answer a far more
general question: What is the smallest number of moves required to solve the
Tower of Hanoi puzzle with n disks on 3 rods? We posed the puzzle above with
just 3 disks to give you a concrete version to think about and work with, but
we can answer this more general question by thinking carefully. To make sure
we are on the same page, we will show you how we solved the version with 3
disks:
2.5. APPLICATIONS 141

Start

Move 1

Move 2

Move 3

Move 4

Move 5

Move 6

Move 7
142 CHAPTER 2. MATHEMATICAL INDUCTION

Notice that the largest disk is essentially “irrelevant” for most of the solution.
Since we are allowed to place any other disk on top of it, all we need to do is
“uncover” that disk by moving the other disks onto a different rod, move the
largest disk to the only empty rod, then move the other disks on top of the large
one. In essence, we perform the same procedure (shifting the two smaller disks
from one rod to another) twice and, in between those, we move the large disk
from one rod to another. If the largest disk hadn’t been there at all, what we
actually did was solve the version of the puzzle with 2 disks, but twice! (Think
carefully about this and make sure you see why this is true. Follow along with
the moves in the diagrams above and pretend the large, blue disk isn’t there.)
This shows that the way to solve the 3-disk puzzle involves two iterations of
solving the 2-disk puzzle, with one extra move in between (moving the largest
disk). This indicates a recursive procedure to solve the puzzle, in general. To
optimally solve the n-disk puzzle, we would simply follow the procedure to
optimally solve the (n − 1)-disk puzzle, use one move to shift the largest, n-th
disk, then solve the (n − 1)-disk puzzle again.
Now that we have some insight into how to optimally solve the puzzle, let’s
identify how many moves that procedure requires. Recognizing that solving this
puzzle uses a recursive algorithm, we realize that proving anything about the
optimal solution will require induction. Accordingly, we would need to identify
a “starting point” for our line of dominos, and it should correspond to the
“smallest” or “simplest” version of the puzzle. For the Tower of Hanoi, this is
the 1-disk puzzle. Of course, this is hardly a “puzzle” because we can solve it in
one move, by simply shifting the only disk from one rod to any other rod. If we
let M (n) represent the number of moves required to optimally solve the n-disk
puzzle, then we’ve just identified M (1) = 1. To identify M (2), we can use our
observation from the previous paragraph and say that
M (2) = M (1) + 1
|{z} + M (1) =1+1+1=3
| {z } | {z } | {z }
solve 2-disk solve 1-disk shift largest disk solve 1-disk

and then it must be that

M (3) = M (2) + 1 + M (2) = 3 + 1 + 3 = 7
and
M (4) = M (3) + 1 + M (3) = 7 + 1 + 7 = 15
and so on. Do you notice a pattern yet? Each of these numbers is one less
than a power of 2, and specifically, we notice that M (n) = 2n − 1, for each of
the cases we have seen thus far. It’s important to point out that observing this
pattern doesn’t prove the pattern; just because it works for the first 4 cases
does not mean the trend will continue, but that’s exactly what an induction
proof would accomplish. Also, recognizing that pattern and “observing” that
M (n) = 2n − 1 is a non-trivial matter, itself. We happened to know the answer
and had no problem identifying the formula for you. You should probably try,
on your own, to “solve” the following relationship
M (n) = 2M (n − 1) + 1 and M (1) = 1
2.5. APPLICATIONS 143

and see if you can derive the formula M (n) = 2n − 1. The reason such a formula
is nicer than the above relationship is that, now, M (n) depends only on n, and
not on previous terms (like M (n−1), for example). This relationship and others
like it are known as recurrence relations, and they can be rather difficult to solve,
in general!
We know how to solve this one, though, and it yields M (n) = 2n − 1. We
will leave it to you to verify this. You can do so by checking a few values in the
equation above, but we all know that isn’t a proof. Try working through the
inductive steps to actually prove it! We have already done most of the work, but
it will be up to you to arrange everything carefully and clearly. Remember that
you should identify what the “fact” on each domino is, ensure that Domino 1
falls, and then make a general argument about Domino n toppling into Domino
(n + 1). Try to write that proof. Do the details make sense to you? Try showing
your proof to a friend and see if they understand it. Did you need to tell them
anything else or guide them through it? Think about the best way to explain
your method and steps so that the written version suffices and you don’t have
to add any verbal explanations.

2.5.3 Questions & Exercises

Remind Yourself
Answering the following questions briefly, either out loud or in writing. These
are all based on the section you just read, so if you can’t recall a specific defi-
nition or concept or example, go back and reread that part. Making sure you
can confidently answer these before moving on will help your understanding and
memory!

(1) How is a recursive program inductive?

(2) What is the inductive structure of the Tower of Hanoi? Where did we solve
the 2-disk puzzle while solving the 3-disk puzzle?

Try It
Try answering the following short-answer questions. They require you to actu-
ally write something down, or describe something out loud (to a friend/class-
mate, perhaps). The goal is to get you to practice working with new concepts,
definitions, and notation. They are meant to be easy, though; making sure you
can work through them will help you!

(1) Follow the steps of the psuedocode factorial to compute 5!.

(2) Follow the steps of the psuedocode Fibonacci to compute F (5).

(3) Solve the Tower of Hanoi puzzle with 4 disks. Make sure that you can do
it in the optimal number of moves, 24 − 1 = 15.
144 CHAPTER 2. MATHEMATICAL INDUCTION

2.6 Summary
We have now seen some examples of inductive arguments. We realized that
some of the puzzles we were solving used similar argument styles, and explored
several examples to get a flavor for the different issues that might come up in
such arguments. Specfically, we saw how inductive arguments are not always
about proving a summation formula or an equation: inductive arguments can
apply to any situation where a fact depends on a “previous instance” of that fact.
This led us into developing an analogy for how induction works, mathematically
speaking. We are comfortbal with thinking of induction in terms of the “Domino
Analogy” for now, but one of our main goals in moving forwards is rigorously
stating and proving a principle of induction. For now, let’s get lots of practice
working with these kinds of arguments. This is what this chapter’s exercises are
meant to achieve. Later on, once we’ve formalized induction, we’ll be better off
for it, and we’ll have a thorough understanding of the concept!

2.7 Chapter Exercises

Here are some problems to get you comfortable working with inductive-style
arguments. We aren’t looking for fully rigorous proofs here, just a good de-
scription of what is going on and a write-up of your steps. We’ll come back
to some of these later and rigorously prove them, once we’ve established the
Principle of Mathematical Induction (PMI) and a corresponding proof strategy.
Problem 2.7.1. Prove the following summation formula holds for every natural
number, and for n = 0, as well:
n
X
2i = 2n+1 − 1
i=0

Follow-up question: use this result to state how many games are required to
determine a winner in a single-elimination bracket tournament with 2n teams.
(For example, the NCAA March Madness Tournament uses this format, with
n = 6.)
Problem 2.7.2. Prove that 3n ≥ 2n+1 for every natural number n that is greater
than or equal to 2.
Problem 2.7.3. For which natural numbers n do the following inequalities hold
true? State a claim and then prove it.

1. 2n ≥ (n + 1)2

2. 2n ≥ n!

3. 3n+1 > n4

4. n3 + (n + 1)3 > (n + 2)3

2.7. CHAPTER EXERCISES 145

Problem 2.7.4. The December 31 Game: Two players take turns naming
dates from a calendar. On each turn, a player can increase the month or date
but not both. The starting position is January 1, and the winner is the person
who says December 31. Determine a winning strategy for the first player.
For example, a sequence of moves that yields Player 1 winning is as follows:
(1) January 10, (2) March 10, (1) August 10, (2) August 25, (1)
August 28, (2) November 28, (1) November 30, (2) December 30,
(1) December 31
By winning strategy we mean a method of play that Player 1 follows that guar-
antees a win, no matter what Player 2 does.
Problem 2.7.5. Find and prove a formula for the sum of a geometric series,
which is a series of the form
n−1
X
qi
i=0
for some real number q and some natural number n. (Hint: be careful when
q = 1.)
Problem 2.7.6. Write a sentence that depends on n such that the sentence is
true for all values of n from 1 to 99 (inclusive), but such that the sentence is
false when n = 100.
Problem 2.7.7. What is wrong with the following “spoof” of the claim that
an = 1 for every n?
“Spoof ”: Let a be a nonzero real number. Notice that a0 = 1. Also, notice
that we can inductively write
an 1
an+1 = an · a = an · n−1
=1· =1
a 1
“”

Problem 2.7.8. In a futuristic society, there are only two different denominations
of currency: a coin worth 3 Brendans, and a coin worth 8 Brendans. There is
also a nation-wide law that says shopkeepers can only charge prices that can be
paid in exact change using these two coins.

What are the legal costs that a shopkeeper could charge you for a cup of coffee?
Hint: Try a bunch of small values and see what happens.
Problem 2.7.9. Consider a chessboard of size 2n × 2n , for some arbitrary nat-
ural number n. Remove any square from the board. Is it possible to tile the
remaining squares with L-shaped triominoes?
If your answer is Yes, prove it.

If your answer is No, provide a counterexample argument. (That is, find an n

such that no possible way of tiling the board will work, and show why this is
the case.)
146 CHAPTER 2. MATHEMATICAL INDUCTION

Problem 2.7.10. Consider an n × n grid of squares. How many sub-squares, of

any size, exist within this grid? For example, when n = 2, the answer is 5: there
are 4 1 × 1 squares and 1 2 × 2 square. Find a formula for your answer and try
to prove it is correct.
Problem 2.7.11. Prove that, in a line of at least 2 people, if the 1st person is a
woman and the last person is a man, then somewhere in the line there is a man
standing immediately behind a woman.
Problem 2.7.12. Prove that n3 − n is a multiple of 3, for every natural number
n.
Problem 2.7.13. A binary n-tuple is an ordered string of 0s and 1s, with n
total numbers in the string. Provide an inductive argument to explain why there
are 2n possible binary n-tuples.
Problem 2.7.14. Recall that the Fibonacci Numbers are defined by setting
f0 = 0 and f1 = 1 and then, for every n ≥ 2, setting fn = fn−1 + fn−2 . This
produces the sequence 0, 1, 1, 2, 3, 5, 8, 13, 21, 34, . . .
You might not know that the Fibonacci Numbers also have a closed form;
that is, there is a specific formula that defines them, in addition to the usual
recursive definition given above. Here it is:
" √ !n √ !n #
1 1+ 5 1− 5
fn = √ −
5 2 2

Prove that this formula is correct for all values of n ≥ 0.

Problem 2.7.15. Again, considering the Fibonacci Numbers, fn , prove the fol-
lowing:
n
X
1. fi = fn+2 − 1
i=0
n
X
2. fi2 = fn · fn+1
i=0

3. fn−1 · fn+1 − fn2 = (−1)n

4. fm+n = fn · fn+1 + fm−1 · fn
5. fn2 + fn+1
2
= f2n+1
Problem 2.7.16. Try to provide an inductive argument that explains why every
natural number n ≥ 2 can be written as a product of prime numbers. Can you
also show that this product is unique? That is, can you also explain why there
is exactly one way to factor a natural number into primes?
Problem 2.7.17. Prove that
n
X
k · k! = 1 · 1! + 2 · 2! + 3 · 3! + · · · + n · n! = (n + 1)! − 1
k=1
2.7. CHAPTER EXERCISES 147

Problem 2.7.18. What is wrong with the following “spoof” that all pens have
the same color.
“Spoof ”: Consider a group of pens with size 1. Since there is only 1 pen, it
certainly has the same color as itself.
Assume that any group of n pens has only one color represented inside the
group. (Note: we explained why this assumption is valid for n = 1 already,
so we can make this assumption.) Take any group of n + 1 pens. Line them
up on a table and number them from 1 to n + 1, left to right. Look at the
first n of them, i.e. look at pens 1, 2, 3, . . . , n. This is a group of n pens so, by
assumption, there is only one color represented in the group. (We don’t know
what color that is yet.) Then, look at the last n of the pens; i.e. look at pens
2, 3, . . . , n + 1. This is also a group of n pens so, by assumption, there is only
one color represented in this group, too. Now, pen #2 happens to belong to
both of these groups. Thus, whatever color pen #2 is, that is also the color of
every pen in both groups. Thus, all n + 1 pens have the same color.
By induction, this shows that any group of pens, of any size, has only one
color represented. Looking at the finite collection of pens in the world, then, we
should only find one color. “”

Problem 2.7.19. ? This problem is extremely difficult to analyze, and is taken

from the blog of the famous mathematician Terence Tao (link here).
There is an island upon which a tribe resides. The tribe consists of 1000
people, with various eye colours. Yet, their religion forbids them to know their
own eye color, or even to discuss the topic; thus, each resident can (and does)
see the eye colors of all other residents, but has no way of discovering his or her
own (there are no reflective surfaces). If a tribesperson does discover his or her
own eye color, then their religion compels them to commit ritual suicide at noon
the following day in the village square for all to witness. All the tribespeople
are highly logical and devout, and they all know that each other is also highly
logical and devout (and they all know that they all know that each other is
highly logical and devout, and so forth).
(For the purposes of this logic puzzle, ”highly logical” means that any conclu-
sion that can logically deduced from the information and observations available
to an islander, will automatically be known to that islander.)
Of the 1000 islanders, it turns out that 100 of them have blue eyes and 900
of them have brown eyes, although the islanders are not initially aware of these
statistics (each of them can of course only see 999 of the 1000 tribespeople).
One day, a blue-eyed foreigner visits to the island and wins the complete
trust of the tribe.
One evening, he addresses the entire tribe to thank them for their hospitality.
However, not knowing the customs, the foreigner makes the mistake of men-
tioning eye color in his address, remarking how unusual it is to see another
blue-eyed person like myself in this region of the world.
What effect, if anything, does this faux pas have on the tribe?
148 CHAPTER 2. MATHEMATICAL INDUCTION

2.8 Lookahead
In this chapter, we have introduced you to the concept of mathematical in-
duction. We looked at a few examples of puzzles where an inductive process
guided our solution, and then we described how a proof by induction would fol-
low to rigorously verify that solution. With the mathematical techniques and
concepts we have at hand thus far, we had to rely on a non-technical analogy
to describe this process to you. Thinking of an infinite line of dominos with
“facts” written on them knocking into each other is a perfectly reasonable in-
terpretation of this process, but it fails to represent the full mathematical extent
of induction. In a way, it’s like having a friend describe to you how to swing
a golf club, even though you’ve never played golf before. Sure, they can pro-
vide you with some mental imagery of what a swing “feels like”, but without
getting out there and practicing yourself, how will you truly understand the
mechanics of the golf swing? How will you learn how to adapt your swing, or
tell the differences between using a driver and a five iron and a sand wedge? By
investigating the underlying mechanics and practicing with those concepts, we
hope to gain a better understanding of mathematical induction so that, in the
future, we can use it appropriately, identify situations where it would be useful,
and, eventually, learn how to adapt it to other situations. Of course, it will help
to have that domino analogy in mind to guide our intuition, but we should also
remember that it is not rigorous mathematics. It also doesn’t perfectly describe
the other examples we discussed, where a falling domino depended on not only
the one immediately behind it, but several others before it.
In the next chapter, we will explore some relevant concepts needed to rig-
orously state and prove mathematical induction as a proof technique. Specif-
ically, we will study some ideas of mathematical logic and investigate how to
break down complicated mathematical statements and theorems into their con-
stituent parts, and also how to build interesting and complex statements out
of basic building blocks. Along the way, we will introduce some new notation
and shorthand that will let us condense some of the wordy statements we make
into concise (and precise) mathematical language. With that in hand, we will
explore some more fundamental proof strategies, that we will then apply to ev-
erything else we do in this course, including the induction technique, itself! We
will also study some of the ideas of set theory, a branch of mathematics that
forms the foundation for all other branches. This will be extremely useful for
organizing our ideas in the future, but it will also help us define the natural
numbers in a rigorous manner. With some concepts and knowledge from these
two branches of mathematics under our collective belts, we will be able to build
mathematical induction on a solid foundation and continue to use it properly.
Chapter 3

Sets: Mathematical
Foundations

3.1 Introduction
It’s now time for sets education! This might seem like a weird jump to make,
after the last chapter. You’ll have to trust us when we say that this is actually
quite natural and, ultimately, essential. Everything we do in mathematics is
built upon the foundation of sets, so we better get started talking about them
and getting used to them.

3.1.1 Objectives
The following short sections in this introduction will show you how this chapter
fits into the scheme of the book. They will describe how our previous work
will be helpful, they will motivate why we would care to investigate the topics
that appear in this chapter, and they will tell you our goals and what you
should keep in mind while reading along to achieve those goals. Right now,
we will summarize the main objectives of this chapter for you via a series of
statements. These describe the skills and knowledge you should have gained by
the conclusion of this chapter. The following sections will reiterate these ideas
in more detail, but this will provide you with a brief list for future reference.
When you finish working through this chapter, return to this list and see if you
understand all of these objectives. Do you see why we outlined them here as
being important? Can you define all the terminology we use? Can you apply
the techniques we describe?

By the end of this chapter, you should be able to . . .

• Define what a set is, and identify several common examples.

• Use proper notation to define a set and refer to its elements.

149
150 CHAPTER 3. SETS

• Define and describe several ways to operate on sets; i.e. identify the ways
to take two or more sets and create new sets from them.

• Describe how two sets might be compared, as well as apply a proper

technique to prove such claims.

• Explain how the natural numbers are related to sets, and relate this to
mathematical induction.

3.1.2 Segue from previous chapter

We are building towards a formal statement of mathematical induction as a
theorem, which we will then prove. To get there, we need to have some funda-
mental objects to work with and talk about, logically. Sets are those objects!
Historically speaking, mathematics was placed on the set theory foundation rela-
tively recently, around the turn of the 20th century. Until then, mathematicians
tended to “wave their hands” a bit about what was really going on underneath
their work; they made a lot of “intuitive” assumptions and hadn’t yet tried
to rigorously and axiomatically describe everything they did. After the work
of the mathematician Georg Cantor showed everyone some surprising and
counter-intuitive results that were perfectly correct and consistent with our as-
sumptions . . . well, we realized that we better decide what we’ve been talking
about all along. This is not meant to discredit the work of mathematicians
before 1900, of course! We just mean that they were playing a game all along
where they hadn’t really agreed on a set of rules yet. That’s what the axioms
of set theory are.

3.1.3 Motivation
We are, of course, motivated by our ongoing desire to learn about proofs,
discover what they are and how they work, and, in particular, rigorize mathe-
matical induction. More generally, though, we are interested in learning about
what mathematicians really do, and we are sure that any mathematician in the
world can tell you how important sets are in their work. They might do so
begrudgingly, and say that they themselves could never work in pure set theory,
but we doubt you’ll find anyone who will deny the importance of sets.
Everything we do later on will involve making some claim about a set of
objects; that is, we will attempt to say (and subsequently prove) that some fact
is True about some particular objects. The way we specify those objects involves
sets. The way in which we express such facts will involve mathematical logic,
and we will get to that soon enough. For now, we need to learn how to express
many types of mathematical objects in the first place, before we can even make
claims about them.
3.2. THE IDEA OF A “SET” 151

3.1.4 Goals and Warnings for the Reader

This chapter will likely involve handling some mathematical ideas that are new
to you, as opposed to the previous chapters where we focused on puzzles that
only relied on some knowledge of numbers and algebra and arithmetic and
critical thinking. These new ideas will require careful reading and thinking. As
we introduce these concepts and results, we expect you to read through them
carefully and do some thinking on the side. Mathematical exposition requires
more of the reader than, say, a newspaper article; it expects an engaged reader,
one who will think carefully about every sentence and sometimes have to pause
for a few minutes to ensure full understanding of what has been said so far.
Keep this in mind as you read on: reading mathematics can be difficult, but
this is to be expected! Don’t let it be discouraging; just think of every sentence
as a single jigsaw piece in a larger puzzle to be solved.
In particular, don’t be surprised if this chapter takes as long (if not longer) to
read through (and work through in class) as the previous two chapters combined!
The most baffling part of this, as we have observed over the years, is in the
notation of sets. This is likely the first time in your mathematical careers
where you are expected to be as precise and rigorous as possible with your
writing. It is no longer okay to just “have the right idea” in your written work;
we really care that you say exactly what you meant to say, and nothing else.
As you read what we have written, ask yourself, “Why does that make sense,
as opposed to . . . something else?” After you have written down an answer to
a question or homework problem, read it again and ask yourself, “Does this
actually make sense? Does it say what I meant to say, what was in my head?
Is someone else guaranteed to read it in the same way I wrote it?”
Also, this chapter will involve some more abstract thinking than your typ-
ical mathematics course. This might be a shock for you, or maybe not. Either
way, this is certainly not material that you can just skim through and expect
to pick up on first glance. Now, more than ever, you should take the time and
effort to internalize this material. Read some pages and then go think about the
material while you eat dinner or shower or play basketball. Try to find examples
in real life. Talk with your friends about sets. This may sound silly now but,
ultimately, it will help you. Trust us.

3.2 The Idea of a “Set”

A“Collection of Objects” with a Common Property
The intuitive notion of a set is probably not entirely new to you. If you’re
a baseball card collector, having a “complete set” means owning every single
card from a particular printing run by a card manufacturer. If you play board
games with friends, you agree on a “set of rules” before playing so there are
no unresolved disputes later on. If you performed a laboratory experiment in
biology or chemistry or physics class, you collect information into a “data set”
and analyze those results to test a hypothesis.
152 CHAPTER 3. SETS

Those are three different situations that each refer to the word set, so what
is it about that word that relates those contexts and gives it a proper meaning?
Essentially, a set refers to a collection of objects of some kind that are grouped
together on the basis of having some common property. In the first example,
a copy of every baseball card produced by Topps in 1995 would belong to that
particular set. In the second example, any agreed-upon convention would belong
to your set of rules. In the third example, any data point gathered in your
experiment would belong to your data set. In each case, there is a common
property that lets us associate particular objects with each other and refer to
them as one set.

Sets in Mathematics
Sets are very common, popular, useful and fundamental in mathematics. Be-
cause mathematicians work with abstract objects and relationships between
those objects, it can be quite difficult to describe exactly what one is thinking
about without being able to refer to sets of mathematical objects. We have, in
fact, already done so!
For instance, when investigating polynomials and the quadratic formula, we
mentioned that a quadratic polynomial p(x) = ax2 + bx + c with a negative
b2
discriminant (when 4a − c < 0) will have no roots in the set of real numbers.
What did we mean? Did you understand that sentence? We were trying to
convey the idea that no matter what real number x we choose from among
the collection of all real numbers, it would be guaranteed that p(x) 6= 0. But
what exactly is the set of real numbers? How is it defined? How can we be so
sure it even exists? These are actually rather difficult questions to answer, and
attempting to do so would take us far off course into the world of set theory.
In the language of mathematics, we aim to be precise and unambiguous
with our sentences and statements, and we seek to establish truths based on
certain fundamental assumptions. We need to make those assumptions as a
starting point, or else we would have nothing to base our truths upon. These
assumptions, that everyone agrees to be part of the “set of rules” before “playing
the game” of mathematics, are known as axioms.
Perhaps, if you have studied some geometry or read about the Greek math-
ematician Euclid and his treatise, the Elements, then you have heard the word
“axiom” before. All of the results of basic geometry that Euclid proved were
founded upon some basic assumptions: that any two points can be joined by a
line segment, that a circle with a given center point and radius must exist, that
non-parallel lines intersect, and so on. These statements are simply agreed upon
to be True at the outset.
Another place we find axioms is in the branch of mathematics known as set
theory. The axioms of this branch place all of the results involving sets on
firm foundations, and using those axioms and results derived from them, we
can continue to discover new truths in the mathematical universe. Investigating
these axioms and their consequences is better suited for a course devoted to set
theory, though, and we will take many of the consequences of the set theoretic
3.3. DEFINITION AND EXAMPLES 153

axioms for granted without rigorously proving them. This is not because such
proofs are impossible, but merely because they would take too much time and
space in this book to accomplish.
What we will do is provide a definition of “set” that is satisfactory for the
contexts in which we will be using sets in this book. We will also define some
basic properties of sets, share some illustrative examples, and dicuss different
operations we can perform on sets to create new ones.

3.3 Definition and Examples

3.3.1 Definition of “Set”
Let’s start with a definition. As we started to explain above, we often think
of sets as being characterized by the objects that are grouped together into
that set and the property that makes that grouping make sense. The following
definition attempts to make that notion as precise as we possibly can, while also
introducing some relevant notation and terminology.

Definition 3.3.1. A set is a collection of all objects that have a common, well-
defined property. The objects contained in a set are called elements of the set.
The mathematical symbol “∈” represents the phrase “is an element of ” (and
“∈”
/ represents “is not an element of ”).

3.3.2 Examples
Let’s dive right in with some specific examples of sets (and non-sets, even) to
illustrate this definition. It is common in mathematics to use capital letters
to refer to sets and lowercase letters to refer to elements of sets, and we will
frequently follow this convention (but not always). To define or describe a
set, we need to identify that common, well-defined property that associates the
elements of the set with each other. For instance, we could define B to be the
set of all teams in Major League Baseball. Is this a well-defined property? If
we present you with an object, is ther a definite Yes/No answer to the question
of “Does this object have this defining property?” Yes, this is the case here,
so this is a property that characterizes a set. (To avoid confusion for readers
in the future, let us be more specific and say B refers to MLB teams from the
2012 season.) In the language of mathematics, we would write

B = {Major League Baseball teams from the 2012 season}

The “curly braces”—{ and }—indicate that the description between them will
identify a set, and the text inside is a description of the objects and their
common, well-defined property. It now makes sense to say Pittsburgh Pirates ∈
B and Pittsburgh Penguins ∈ / B.
Common ways to read the mathematical symbol ∈ in English are “is an
element of ” or “is a member of” or “belongs to” or “is in”. We will mostly
154 CHAPTER 3. SETS

use “is an element of” because it is the least ambiguous of them, and uses
the mathematical term element appropriately. Any of these other, equivalent,
phrases may be used, depending on the context, but are less preferable. (In
particular, “is in” can be confused with other set relationships, so we will avoid
it entirely, and encourage you to do the same.)
We’ve also already seen some commonly-used sets of numbers. You know
what they are from previous work with these numbers, but you might not usually
think of them as sets, which is what they are!

N = { natural numbers } = {1, 2, 3, . . .}

Z = { integers } = {. . . , −2, −1, 0, 1, 2, . . .}
Q = { rational numbers }
a
= {numbers of the form , where a, b ∈ Z and b 6= 0 }
b
R = {real numbers}

Think about how the second definition of Q above makes sense. We will see,
quite soon, a more condensed way to write out a phrase like “numbers of the
form . . . blah blah . . . with the additional information that . . . blah blah”. Also,
notice that we can’t really define R except to just say they’re the real numbers.
How do you even define what a real number is? Have you ever tried?

3.3.3 How To Define a Set

Another way of defining or describing a set is simply listing all of its elements.
This is convenient when the number of elements in the set is small. For instance,
the following definitions of the set V are all equivalent:

V = {A,E,I,O,U}
V = {vowels in the English language}
V = {U,E,I,A,O}

By “equivalent”, we mean that each line above defines the same set V , using
different terms. (Note: we have adopted the convention that y is a consonant, so
y∈/ V .) The common, well-defined property that associated the objects A, E, I,
O, and U is the fact that they are all vowels (exhibited in the second definition)
and since there are only five such objects, it is simple and convenient to list
them all (as in the first definition).

Order and Repetition Don’t Matter

Why do you think the third definition is the same as the others? It refers to
the same collection of objects, and any set is completely characterized by its
elements, so the order in which we write the elements does not matter. Is U∈ V ?
The answer to this question is “Yes”, regardless of whether U is written first or
last in the list of elements.
3.3. DEFINITION AND EXAMPLES 155

Not only does the order of elements not matter within a set, the repetition
of elements does not matter! That is, the set A = {a, a, a} and the set A = {a}
are exactly the same. Again, remember that a set is completely characterized
by its elements; we only care about what is “in” a set. (We will mention this
again in Section 3.4.4, when we talk about the “bag analogy” for sets.) Writing
A = {a, a, a} is just a triply-redundant way of saying a ∈ A and that only a is
an element of A. Thus, A = {a} is the most concise way of stating the same
information.

The Common Property Might Be Being an Element of That Set

Now, still following the idea that we can define a set by writing all of the
elements, consider the following definition of a set A:
A = {2, 7, 12, 888}
Surely, this is a set because we just defined it by listing its elements. What is the
common, well-defined property that associates its elements, though? With the
set V of vowels, we could list the elements and provide a linguistic definition, but
for this set A, it seems as though we are relegated to listing the elements without
knowing a way of describing their common property. Mathematically speaking,
though, a common property uniting 2, 7, 12, 888 is that they are all elements
of this set A! In the mathematical universe, we have a license for freedom of
abstract thought, and merely by discussing this set A and its elements, we have
given them that common property. Does this seem satisfactory to you? Can you
come up with another common, well-defined property that would yield exactly
the elements of A? (Hint: identify a polynomial p(x) whose roots are exactly
2, 7, 12, and 888.) If the elements of a set have more than one property that
associates them together, do you think it matters which property we have in
mind when referring to the set? And what do you think about the set S := {2, 7,
M, Boston Red Sox}? Could there possibly be a common property other than
the fact that we have listed them here?

Ellipses Are Sometimes Okay, But Informal

Sometimes, when there is no confusion about the set in question, or it has
been defined in another way and we wish to list a few elements as illustrative
examples, then it is convenient to use ellipses to condense the listing of elements
of a set. For instance, we might write
E = {even natural numbers} = {2, 4, 6, 8, 10, . . . }
This set is infinitely large, in fact, so we could not even list all of its elements
if we tried, but it is clear enough from the first few elements listed that we
are referring to even numbers, especially because we have already referred to E
as “the set of even natural numbers”. However, we cannot stress enough that
this is not a precise definition of the set in question. It suffices in an informal
context, but it is not mathematically rigorous, and this will become clear as we
discuss the following proper way of defining sets.
156 CHAPTER 3. SETS

Set-Builder Notation
The best way to define or describing a set is to identify its elements as particular
objects of another set that have a specific property. For instance, if we wished
to refer to the set S of all natural numbers between 1 and 100 (inclusive), we
could list all of those elements, but this requires a lot of unnecessary writing.
We could also use the ellipsis notation S = {1, 2, 3, . . . , 100} but, again, this is
not precise without already having a formal definition of S. (Someone might
misinterpret the ellipses in a different way.) It is much more precise and concise
to write
S = {x ∈ N | 1 ≤ x ≤ 100}
We read this as “S is the set of all objects x in the set N of natural numbers
such that 1 ≤ x ≤ 100”.
The bar symbol | is read as “such that” and indicates that the information
to the left tells us what “larger set” the objects come from, while the information
to the right tells us the specific property that those objects should have.
(Caution: do not use | in other contexts to mean “such that”. This is only
acceptable in the context of defining sets. It is just used as a place-holder to
separate the left side—the set we use to draw elements—from the right side—a
description of the property those elements should have.)
This is an example of the very popular and useful set-builder notation.
We call it this because we are building a set by drawing elements from a “larger”
set of possibilities, and only including those that have a particular property. To
do this, we need to tell the reader (1) what the larger set is, and (2) what the
common property is. Let’s illustrate this idea with a few examples:
S = {x ∈ N | 1 ≤ x ≤ 100} = {1, 2, 3, . . . , 100}
T = {z ∈ Z | we can find some k ∈ Z such that z = 2k}
= {. . . , −4, −2, 0, 2, 4, . . .}
n √ √ o
U = x ∈ R | x2 − 2 = 0 = − 2, 2


V = x ∈ N | x2 − 2 = 0 = { }


The last two examples show how the context is extremely important. The
same common property (satisfying x2 − 2 = 0) can be satisfied by a different set
of elements when we change the larger set from which we draw elements. Two
real numbers satisfy that property, but no natural numbers satisfy it! Do any
rational numbers satisfy that property? What do you think?
This is 
why it is absolutely essential to specify the larger set. A definition
like “U = x | x2 − 2 = 0 is meaningless because it is ambiguous and could

yield completely different interpretations.

Reading Notation Aloud

We are really learning a new language here, and these are some of the basic
words and rules of grammar. We’ll need some practice translating these sen-
3.3. DEFINITION AND EXAMPLES 157

tences into English (in our heads and out loud) and vice-versa. For example,
we can say the definition for S above as any of the following, reasonably:
S is the set of all natural numbers x such that x is between 1 and
100, inclusive.
S is the set of all natural numbers between 1 and 100, inclusive.
S is the set of all natural numbers x that satisfy the inequality
1 ≤ x ≤ 100.
S is the set of natural numbers x with the property that 1 ≤ x ≤ 100.
Notice that all of them identified the larger set and the common property; the
only differences between them are linguistic/grammatical, and they do not alter
the mathematical meanings.
Try to write similar sentences for the other definitions. Try getting a verbal
definition of a set from a friend and writing down what they said in mathematical
symbols.
Consider a definition of the rational numbers Q that we saw before, and
notice that we can rewrite it as:
na o
Q= , where a, b ∈ Z and b 6= 0
nb a o
= x ∈ R | we can find a, b ∈ Z such that = x and b 6= 0
b
Notice the subtle difference between the two definitions. The upper one tells us
that all rational numbers are of the form ab , and then tells us the particular
assumptions of a and b that must be satisfied. The lower one tells us that all
rational numbers are real numbers with a particular property, namely that we
can express that real number as a ratio of integers. We strongly prefer the lower
definition, because it tells us more information.
In general, if P (x) represents a sentence (involving English and/or mathe-
matical language) that describes a specific, well-defined property, and X is a
given set, then the notation
S = {x ∈ X | P (x)}
is read as
“S is the set of all elements x of the set X such that the property
P (x) is true”.
In the notation P (x), the letter x represents a variable object, and depending
on the particular object we use as x, the property P (x) may hold (i.e. P (x) is
true) or may fail (i.e. P (x) is false). If the property holds, then we include x in
S (so x ∈ S), and if it fails, we do not include x in S (so x ∈ / S).
Returning to our example of the set E of even natural numbers, it is more
precise to write
E = {even natural numbers}
= {x ∈ N | there is a natural number n such that x = 2n}
158 CHAPTER 3. SETS

Notice that there are two “layers” of properties here. A natural number is
included in our set E if we can find another natural number n with the additional
property that x = 2n. Try to write down a similar definition for the set of odd
numbers, or the set of perfect squares. What about the set of primes? The set
of palindromic numbers? The set of perfect numbers? Can you write definitions
for these sets using set-builder notation?

3.3.4 The Empty Set

What if no elements satisfy a property P (x)? What happens then? For instance,
consider the definition

S = x ∈ N | x2 − 2 = 0


√
We √ number x we are “looking for” with that property is 2 (and
√ know that the
− 2, too) but 2 ∈ / N. Thus, no matter what element of N we let x represent,
the property P (x)—given by “x2 − 2 = 0”—actually fails. Thus, there are no
elements of this set. Is this really a set?
Remember, a set is completely characterized by its elements, and a set having
no elements, such as this one, is characterized by that fact. If we attempted to
list its elements, we would end up writing {}. This set is so special, in fact, that
we give it a name and symbol:
Definition 3.3.2. The empty set is the set which has no elements. It is
denoted by the symbol ∅.
There are many ways to define the empty set using set-builder notation.
(And yes, we do mean the empty set; there is only one set with no elements!)
We saw one example above, and we’re sure you can come up with many others.
Consider for example, the following sets:

{a ∈ N | a < 0}
{r ∈ R | r2 < 0}
{q ∈ Q | q 2 ∈
/ Q}

Do you see why these all define the same set, the one with no elements?

Context Matters
We should also note again the significance of specifying the larger set X from
which we draw our variable element x in a set-builder definition like the one
above. For instance, consider the following two sets:

S1 = {x ∈ N | |x| = 5} = {5}
S2 = {x ∈ R | |x| = 5} = {−5, 5}

(Note: It is also quite common to index sets with the subscript notation you
see above, allowing us to use the same letter many times.)
3.3. DEFINITION AND EXAMPLES 159

This specification is clearly important, in this case, because it yields two

entirely different sets! For this reason, we must be precise and clear when
defining a set in this way. A definition like S = {x | |x| = 5} should be regarded
as ambiguous and undesirable, since it leads to issues like the one above.

3.3.5 Russell’s Paradox

Perhaps it seems like we are picking nits here, but the reasoning behind our
vehemence is rooted in some fundamental ideas of set theory. We wish to avoid
some complex issues and paradoxes that might arise without this policy in place.
There is a particularly famous example of a paradox involving sets that illustraes
why we have the requirement described in the above paragraph, namely that
we must specify a larger set when we use set-builder notation. This example is
known as Russell’s Paradox (after the British mathematician Bertrand Russell),
and we we will present and discuss it in this section.

Sets Whose Elements Are Sets

First, we should point out that this discussion will introduce the notion that sets
can also be elements of other sets. This may seem like a strange and far-fetchedly
abstract idea right now, but it is a fundamental concept in mathematics.
For a concrete example, think back to our set B of all Major League Baseball
Teams. We could also regard each team as a set, where its elements are the
players on the team. Thus, it would make sense to say

Derek Jeter ∈ New York Yankees ∈ B

since Derek Jeter is an element of the set New York Yankees, which is itself an
element of the set B. (Notice, however, that Derek Jeter ∈/ B. The relationship
signified by “∈” is not transitive. We will hold off on defining this term until
much later. For now, we will point out that the relationship signified by “≤” on
the set of real numbers is transitive. If we know x ≤ y ≤ z, then we can deduce
x ≤ z. This is not the case with the “∈” relationship, though.)
Another example is S = {1, 2, 3, {10}, ∅}. Yes, the empty set itself can be an
element of another set, as can the set {10}. Why couldn’t they? As a thought
exercise, we suggest thinking about the difference between ∅, {∅}, {{∅}}, and
so on. Why are they different sets?
One final example involves the natural numbers N. Let’s use O and E to
represent the odd and even natural numbers, respectively. What, then, is the set
S = {O, E}, and how does it differ, if at all, from N? This is a subtle question,
so think about it carefully.

The Paradoxical “Set”

Spend some time on the side thinking about this notion of sets whose elements
are sets. For now, though, let us forge ahead with our explanation of Russell’s
Paradox. Consider the following definition of a “set”. We say “set” because it
160 CHAPTER 3. SETS

is actually not a properly-defined set, but it remains to be seen exactly why this
is the case. When it becomes clear this is not a set, this will be an argument for
requiring the specification of a larger set to draw from when we use set-builder
notation; this is because the definition below does not specify a larger set. Here
is that definition:
R = {x | x ∈/ x}

That’s it! Is this a set? What are the elements of R? Think about what the
definition above says: the elements of R are sets that also happen to not have
themselves as elements. Can you identify any elements of R? Can you identify
any objects that are not elements of R?
The first question is far easier to answer: any of the sets we have discussed so
far would be elements of R. For instance, the empty set ∅ contains no elements,
so it certainly doesn’t have itself as an element. Thus, ∅ ∈ R. Also, notice
thatN ∈ / N (because the set of natural numbers is not a natural number, itself)
and thus N ∈ R.
Identifying objects that are not elements of R is a very tricky matter, and
we will help you by asking the following question: Is R an element of itself? Is
it true or false that R ∈ R? Think about this carefully before reading on. We
will walk you through the appropriate considerations.

• Suppose R ∈ R is True.
The defining property of R tells us that any of its elements is a set that
does not have itself as an element. Thus, we can deduce that R ∈ / R.
Wait a minute! Knowing that R ∈ R led us to deduce that, in fact,
R∈ / R. Surely, these contradictory facts cannot both hold simultaneously.
Accordingly, it must be that our original assumption was bad, so it must
be the case that R ∈/ R, instead.

• Now, suppose R ∈
/ R is True.
The defining property of R tells us that any object that is not an element
of R must be an element of itself. (Otherwise, it would have been included
as an element of R.) Thus, we can deduce that R ∈ R.
Wait a minute! Knowing that R ∈ R led us to deduce that, in fact,
R∈/ R. This is also contradictory.

No matter which option we choose—R ∈ R or R ∈ / R—we find that the other

must also be true, but certainly these contradictory options cannot both be
true.
Therein lies the paradox. This is not a properly-defined set. If it were, we
would find ourselves stuck in the two cases we just saw, and neither of them can
be true. It is also not the case that R is simply ∅; no, it must be that R does
not exist as a set.
3.3. DEFINITION AND EXAMPLES 161

The “Set of all Sets” is Not a Set

Could we amend the definition of R somehow to produce the “set” that the
definition is trying to convey? What “larger set” should we draw our objects x
from so that the definition makes sense and properly identifies a set?
Look back at the English-language interpretation we wrote after the defini-
tion: “the elements of R are sets that also happen to not have themselves as
elements.” The objects x that we wish to test for the desired property (x ∈/ x)
are really all sets. Perhaps, then, we should just define X to be the set of all
sets, and use the phrase “x ∈ X” as part of our definition of R. That would fix
it, right?
R = {x ∈ X | x ∈/ x}
Well, no, not at all! The “set of all sets” is, itself, not a set. If it were,
this would lead us to exactly the same paradox as before! Nothing would be
different, except we would have explicitly stated the “larger set” from which we
draw objects x that was previously left implicitly-specified.
The main issue is that not specifying a “larger set” from which to draw
objects, or implicitly referring to the “set of all sets”, results in this type of
undesirable paradox. Accordigly, we must not allow such definitions. Any
attempt to define a set that draws objects x from the “set of all sets”, whether
implicitly or explicitly, is not a proper definition of a set.

Further Discussion
There is nothing inherently wrong with the property P (x) given by “x ∈ / x”,
though. The issue is with that “larger set” we use. For instance, take the set
   
1 3 5 
S= x∈ , , x ∈
/ x
2 4 2 

What are its elements? The only possibilities are those elements drawn from
the larger set 21 , 34 , 52 . Notice that none of those numbers are sets that
 contain
themselves as elements. Thus, this is a proper definition of the set 12 , 34 , 25 ,

itself! With the previous definition for R, the object we were attempting to
define was allowed as one of the variable objects x in its own definition, and
that is where the issue arose.
We hope that we haven’t led your thoughts too far astray from the origi-
nal discussion of examples of sets, but we felt it was important to point out
that it is possible to construct ill-defined “collections” that are not sets in the
mathematical sense of the word. For the most part, we will not face any such
issues with the sets we work with in this book, but to gloss over these issues or
simply not mention them at all would be unfair to you, as a student. If you find
yourself interested in these issues, seek out an introductory book on set theory.
There are other ways that definitions of “sets” can be ill-formed, as well,
but the ensuing examples are based on (English) linguistic issues and not any
mathematical underpinnings, as in Russell’s Paradox. For instance, we could
162 CHAPTER 3. SETS

say “Let N be the set of all classic novels from the 20th century.” Being a
“classic novel” is not a well-defined property, and cannot be used to identify
elements of such a set. The notion of a “classic” is subjective and not rigorously
precise. Also, we could say “Let B be the set of people who will be born
tomorrow” but this temporal dependence in the definition ensures that we will
never actually know what the elements of B are. When tomorrow arrives, the
definition will then refer to the next day, and so on. Can you come up with
some other examples of ill-formed “collections” of elements? Can you come up
with any paradoxes like the one above?
In general, the following statement is the most important idea to take away
from this discussion of Russel’s Paradox:

Under the agreed-upon rules of sets (the axioms of set theory), there
is no set of all sets.

3.3.6 Standard Sets and Their Notation

We have referred to and used some common sets of numbers already, so we will
list now some sets and their standard symbols:

• The natural numbers: N = {1, 2, 3, 4, . . . }

• The first n natural numbers: [n] := {1, 2, 3, . . . , n − 1, n}

• The integers: Z = {. . . , −3, −2, −1, 0, 1, 2, 3, . . . }

• The rational numbers: Q = m


n | m, n ∈ Z and n 6= 0

• The real numbers: R

• The complex numbers: C

We have used N and Z a few times already. The rational numbers Q (we use
Q since R was already taken, and rational numbers are quotients) are all of the
fractions, or ratios of integers, both positive and negative. The real numbers
are harder to describe. Why could we not list a few elements, like we did with
N and Z? Why is it that R 6= Q? For now, we essentially take for granted
our collective knowledge of these sets of numbers, but think for a minute about
that. (We mention the complex numbers C because you might be familiar with
them, but we will not have occasion to use them in this book.)
How do we know that a set like N exists? Why is it that we think of R as
a number line? How many “more” elements are there in Z, as compared to N?
How many “more” elements are there in R, as compared to Q? Can we even
answer these questions? In the very near future, we will rigorously derive the
set N and prove that it exists as the only set with a particular property. This
will be essential when we return to our investigation of mathematical induction.
(Remember our goals from that chapter?)
3.3. DEFINITION AND EXAMPLES 163

3.3.7 Questions & Exercises

Remind Yourself
Answering the following questions briefly, either out loud or in writing. These
are all based on the section you just read, so if you can’t recall a specific defi-
nition or concept or example, go back and reread that part. Making sure you
can confidently answer these before moving on will help your understanding and
memory!
(1) What does the symbol “∈” mean?
(2) How would you read the statement “x ∈ S” out loud?
(3) Is it possible for a set to be an element of another set? If so, give an example.
Is it possible for a set to be an element of itself?
(4) How would read the statement “{x ∈ N | x ≤ 5}” out loud? Can you list
the elements of this set?
(5) What is this set: {z ∈ Z | z ∈ N}?
(6) What is this set: {x ∈ [10] | x ≥ 7}?
(7) For each of the following sets, state how many elements they have:
(a) ∅
(b) {1, 2, 10}
(c) {1, ∅}
(d) { ∅ }
(8) Is x ∈ { 1, 2, {x} }? Is {x} ∈ { 1, 2, {x} }?
(9) Let A = {a, b, c} and B = {b, c, a} and C = {a, a, b, c, a, b}. Are these sets
equal or not?
(10) Is Z = Q? Why or why not?

Try It
Try answering the following short-answer questions. They require you to actu-
ally write something down, or describe something out loud (to a friend/class-
mate, perhaps). The goal is to get you to practice working with new concepts,
definitions, and notation. They are meant to be easy, though; making sure you
can work through them will help you!
(1) Write a definition of the set of natural numbers that are multiples of 4 using
set-builder notation.
(2) Consider the set S = {3, 4, 5, 6}. Define S in two different ways using set-
builder notation.
164 CHAPTER 3. SETS

(3) Give an example of a set X that satisfies N ∈ X but Z ∈

/ X.

(4) Give an example of a set with 100 elements.

(5) Give an example of a sets A, B, C such that A ∈ B and B ∈ C but A ∈

/ C.

(6) Write a definition of the set of odd integers using set-builder notation.

(7) Write a definition of the set of integers that are not natural numbers using
set-builder notation.

(8) Consider the following sets:

A = x ∈ R | x2 − 3x + 2 ≥ 0


B = {y ∈ R | y ≤ 1 or y ≥ 2}

Explain why A = B.

(9) Consider the following sets:

C = x ∈ R | x2 − 4 ≥ 0


D = {y ∈ R | y ≥ 2}

Is C = D? Why or why not? Write your explanation with good mathemat-

ical notation, using ∈ and ∈.
/

(10) Try explaining Russell’s Paradox to a friend, even one who does not study
mathematics. What do they understand about it? Do they object? Do
their ideas make sense? Have a discussion!

3.4 Subsets
3.4.1 Definition and Examples
Let’s discuss a topic whose basic idea we have already been using. Specifically,
let’s investigate the notion of subsets.

Definition 3.4.1. Given two sets A and B, if every element of A is also an

element of B, then we say A is a subset of B.
The mathematical symbol for subset is ⊆, so we would write A ⊆ B.
If we want to indicate that A is a subset of B but is also not equal to B, we
would write A ⊂ B and say that A is a proper subset of B.
We can also write these relationships as B ⊇ A, or B ⊃ A, respectively. In
these cases, we would say B is a superset of A or B is a proper superset of
A, respectively.
3.4. SUBSETS 165

Notice the similarities between these symbols and the inequality symbols we
use to compare real numbers. We write inequalities like x ≤ 2 or 5 > z > 0
and understand what those mean based on the “direction” of the symbol and
whether we put a bar underneath it. The symbols ⊆, ⊂, ⊇, ⊃ work exactly
the same way, except they refer to “containment of elements” as opposed to
“magnitude of a number”.

Standard Sets of Numbers

The standard sets of numbers we mentioned in the last section are related via
the subset relation quite nicely. Specifically, we can say
N⊂Z⊂Q⊂R⊂C
Again, we take for granted our collective knowledge of these sets of numbers to
allow us to make these claims. However, there are some profound and intricate
mathematical concepts involved in describing exactly why, say, the set R exists
and is a proper superset of Q. For now, though, we use these sets to illustrate
the subset relationship.
Since we know the subset relationships above are proper, we used that
corresponding symbol, “⊂”. In general, it is common in mathematical writing
to simply use the “⊆” symbol, even if it is known that “⊂” would apply. We
might only resort to using the “⊂” symbol when it is important, in context, to
indicate that the two sets are not equal. If that information is not essential to
the context at hand, then we might just use the “⊆” symbol.

Set-Builder Notation Creates Subsets

One way that we have already “used” the idea of a subset was in our use of
set-builder notation. This is used to define a set to be all of the elements of a
“larger” set that satisfy a certain property. We define a property P (x), draw a
variable object x from a larger set X, and include any elements x that satisfy
the property P (x). Notice that any element of this new set must be an element
of X, simply based on the way we defined it. Thus, the following relationship
holds
{x ∈ X | P (x)} ⊆ X
regardless of the set X and the property P (x). Depending on the set X and
the property P (x), it may be that the proper subset symbol ⊂ applies but, in
general, we can say for sure that ⊆ applies.
Try to come up with some examples of sets X and properties P (x) so that
⊆ applies, then try to come up with some examples where ⊂ applies. Try to
come up with one set X and two different properties P1 (x) and P2 (x) so that
⊂ applies for P1 (x) and ⊆ applies for P2 (x). Try to identify two different sets
X1 and X2 and two different properties P1 (x) and P2 (x) so that
{x ∈ X1 | P1 (x)} = {x ∈ X2 | P2 (x)}
Can you do it?
166 CHAPTER 3. SETS

Examples

A set is a subset of another set if and only if every element of the first one is an
element of the second one. For instance, this means that the following claims
all hold:

{142, 857} ⊆ N
n√ o
3, −π, 8.2 ⊆ R
x ∈ R | x2 = 1 ⊆ Z


Do you see why these are True?

For a subset relationship to fail, then, we must be able to find an element of
the first set that is not an element of the second set. For instance, this means
that the following claims all hold:

{142, −857} 6⊆ N
n√ o
3, −π, 8.2 6⊆ Q
x ∈ R | x2 = 5 6⊆ Z


Finding All Subsets of a Set

Let’s work with a specific set for a little while. Define A = {1, 2, 3}. Can we
identify all of the subsets of A? Sure, why not?

{1} ⊆ A
{2} ⊆ A
{3} ⊆ A
{1, 2} ⊆ A
{1, 3} ⊆ A
{2, 3} ⊆ A
A = {1, 2, 3} ⊆ A
∅⊆A

Identifying the first six subsets was fairly straightforward, but it’s important to
remember that A and ∅ are subsets, as well. (Notice: it is true, in general, that
for any set S, S ⊆ S and ∅ ⊆ S. Think about this!)
Consider the set B whose elements are all of the sets we listed above:

B= {1} , {2} , {3} , {1, 2} , {1, 3} , {2, 3} , A, ∅

It is true that any element X ∈ B satisfies X ⊆ A. Do you see why?

3.4. SUBSETS 167

3.4.2 The Power Set

This process of identifying all subsets of a given set is common and useful, so
we identify the resulting set with a special name.

Definition 3.4.2. Given a set A, the power set of A is defined to be the set
whose elements are all of the subsets of A. It is denoted by P(A).

Our parenthetical observation from the above paragraph tells us that S ∈

P(S) and ∅ ∈ P(S), for any set S.
Look back at our example set A = {1, 2, 3} above. What do you notice
about the number of elements in P(A)? How does it relate to the number of
elements in A? Do you think there is a general relationship between the number
of elements in S and P(S), for an arbitrary set S?
Example 3.4.3. Let’s find P(∅). What are the subsets of the empty set? There
is only one, the empty set itself! (That is, ∅ ⊆ ∅, but no other set satisfies
this.) Accordingly, the power set P(∅) has one element, the empty set:

P(∅) = { ∅ }

Notice that this is different from the empty set itself:

∅ 6= { ∅ }

Why is this true? Compare the elements! The empty set has no elements, but
the set on the right has exactly one element. (In general, this can be a helpful
way to compare two sets.) To give you some practice, we’ll point out that would
read the above line aloud as:

“The empty set and the set containing the empty set are two different
sets.”

Example 3.4.4. Let’s try this process with another set, say A = { ∅, {1, ∅} }.
We can list the elements of P(A) as

P(A) = { {∅} , {{1, ∅}} , {∅, {1, ∅}} , ∅ , }

This may look strange, with all of the empty sets and curly braces, but it is
important to keep the subset relationships straight. It is true, in this example,
that
∅ ∈ A, {∅} ⊆ A, {∅} ∈ P(A), {∅} ⊆ P(A)

Why are these relationships true? Think carefully about them, and try to
write a few more on your own. The distinction between “∈” and “⊆” is very
important!
168 CHAPTER 3. SETS

3.4.3 Set Equality

When are two sets equal? The main idea is that two sets are equal if they
contain “the same elements”, but this is not a precise definition of equality.
How can we describe that property more explicitly and rigorously? To say that
two sets, A and B, have “the same elements” means that any element of A is
also an element of B, and every element of B is also an element of A. If both
of these properties hold, then we can be guaranteed that the two sets contain
precisely the same elements and are, thus, equal. If you think about it, you’ll
notice that we can phrase this in terms of subsets. How convenient!

Definition 3.4.5. We say two sets, A and B, are equal, and write A = B, if
and only if A ⊆ B and B ⊆ A.

(What happens if we use the ⊂ symbol in the definition, instead of ⊆? Is this

the same notion of set equality? Why or why not?)
This definition will be very useful in the future when we learn how to prove
two sets are equal and we can’t simply list the elements of each and compare
them. By constructing two arguments and proving the subset relationship in
“both directions”, we can show that two sets are equal. For now, let’s see this
definition applied to a straightforward example.
Example 3.4.6. How can we use this to definition to observe that the following
equality holds?
{x ∈ Z | x ≥ 1} = N
We just need to see that the ⊆ and ⊇ relationship applies between the two sides.
First, is it true that every integer that is at least 1 is a natural number? Yes!
This explains why
{x ∈ Z | x ≥ 1} ⊆ N
Second, is it true that every natural number is a positive integer that is at least
1? Yes! This explains why

{x ∈ Z | x ≥ 1} ⊇ N

Together, this shows that the equality stated above is correct.

3.4.4 The “Bag” Analogy

It has been our experience that sets are a rather difficult notion to grasp when
they are introduced. Specifically, the notation associated with sets throws
students for a loop, and they end up writing down things that make no sense!
It is essential to keep straight the differences between the symbols ∈ and ⊆.
Here is a helpful analogy to keep in mind: a set is like a bag with some stuff
in it. The bag itself is irrelevant; we just care about what kind of stuff inside
(i.e. what the elements are). Think of the bag as a non-descript plastic bag
you’d get at the grocery store, even. All those bags are identical; to distinguish
between any two bags, we need to know what kind of things are inside them.
3.4. SUBSETS 169

If I put an apple and an orange in a grocery bag, surely it doesn’t matter

in what order I put them in. All you would need to know is that I have some
apples and oranges. It also doesn’t matter how many apples or oranges I have
in the bag, because we only care about what kind of stuff is in there. Think of
it as answering questions of the form “Are there any s in the bag? Yes or
no?” It doesn’t matter if I have two identical apples or seven or just one in my
bag; if you ask me whether I have any apples, I’ll just say “Yes”. This is related
to the notion that the order and repetition of elements in a set don’t matter. A
set is completely characterized by what its elements are.
This also helps when we think about sets as elements of other sets, them-
selves. Who’s to stop me from just putting a whole bag inside another bag¿
Look at the set A we defined in the example above:
A = { ∅ , {1, ∅} }
This set A is a bag. What’s inside the bag? There are two objects inside the bag
(i.e. there are two elements of A). They both happen to be bags, themselves!
One of them is a plain-old empty bag, with nothing inside it. (That’s the empty
set.) Okay, that’s cool. The other one has two objects inside it. One of those
objects is the number 1. Cool. The other such object is another empty bag.

Distinguishing “∈” and “⊆”

This analogy also helps with understanding the differences between “∈” and
“⊆”. Keep the example A in mind again. When we write x ∈ A, we mean x is
an object inside the bag A. If we peeked into A, we would see an x sitting there
at the bottom amongst the stuff. Let’s use this idea to compare two examples.
• We see that ∅ ∈ A is true here. If we look inside the bag A, we see an
empty bag amongst the stuff (the elements).
• We also see that {∅} ∈ / A is true here. If we look inside the bag A, we
don’t see a bag containing only an empty bag. (This is what {∅} is, mind
you: an empty bag inside another bag.)
Do you see such an object? Where? I defy you to show me, amongst the
stuff inside the bag A, a bag containing only an empty bag.
What do I see inside the bag A? Well, I see two things: an empty bag,
and a bag that has two objects inside it (an empty bag, and the number
1). Neither of those objects is what we were looking for!
When we write X ⊆ A, we mean that the two bags, X and A, are somehow
comparable. Specifically, we are saying that all of the stuff inside X is also stuff
inside A. We are really rooting through all of the objects inside X, taking them
out one by one, and making sure we also see that object inside A. Let’s use this
idea to compare two examples.
• We see that {∅} ⊆ A is true here. We are comparing the bag on the
left with the bag on the right. What stuff is inside the bag on the left?
170 CHAPTER 3. SETS

There’s just one object in there, and it is an empty bag, itself. Now, we
peek inside A. Do we also see an empty bag in there? Yes we do! Thus,
the “⊆” symbol applies here.

• We also see that {1} 6⊆ A is true here. To compare these two bags, we’ll
pull out an object from the bag on the left and see if it is also in the bag
A. Here, we only have one object to pull out: the number 1. Now, let’s
peek inside the bag A. Do we see a 1 sitting in there amongst the stuff?
No, we don’t!
We would have to peek inside something at the bottom of the bag A to
find the number 1; that number isn’t just sitting out in plain sight. Thus,
{1} 6⊆ A.

Look back over some examples we have discussed already with this new analogy
in mind. Does it help you understand the definitions and examples? Does it
help you understand the difference between “∈” and “⊆” and “⊇”? If not, can
you think of another analogy that would help you?

3.4.5 Questions & Exercises

Remind Yourself
Answering the following questions briefly, either out loud or in writing. These
are all based on the section you just read, so if you can’t recall a specific defi-
nition or concept or example, go back and reread that part. Making sure you
can confidently answer these before moving on will help your understanding and
memory!

(1) Is N ⊆ R? Is R ⊆ N? Is Q ⊆ Z? Why or why not?

(2) What is the difference between ⊂ and ⊆? Give an example of sets A, B

such that A ⊆ B is True but A ⊂ B is False.

(3) What is the difference between ∈ and ⊆? Give an example of sets C, D such
that C ⊆ D but C ∈ / D.

(4) Let S be any set. What is the power set of S? What type of mathematical
object is it? How is it defined?

(5) Suppose S ⊆ T . Does this mean S = T ? Why or why not?

(6) Explain why ∅ ⊆ S and ∅ ∈ P(S) for any set S.

(7) Suppose X ∈ P(A). How are X and A related, then?

(8) Is it possible for A = P(A) to be true? (This one is rather tricky, but think
about it!)
3.4. SUBSETS 171

Try It
Try answering the following short-answer questions. They require you to actu-
ally write something down, or describe something out loud (to a friend/class-
mate, perhaps). The goal is to get you to practice working with new concepts,
definitions, and notation. They are meant to be easy, though; making sure you
can work through them will help you!
(1) Write out the elements of the set P(P(∅)).
(2) Write out the elements of the sets P([1]) and P([2]) and P([3]). Can you
make a conjecture about how many elements P([n]) has? (Can you prove
it? We don’t expect you to now, but soon enough; think about it!)
(3) Let A = {x, ♥, {4} , ∅}. For each of the following statements, decide
whether it is True or False and briefly explain why.
(a) x∈A
(b) x⊆A
(c) {x, ♥} ⊆ A
(d) {x, ∅} ⊂ A
(e) {x, ♥, z, 7} ⊇ A
(f) {x} ∈ P(A)
(g) {x} ⊆ P(A)
(h) {♥, x} ∈ P(A)
(i) {4} ∈ P(A)
(j) {∅} ∈ P(A)
(k) {∅} ⊆ P(A)
Hint: 7 of these are True, and 4 are False.
(4) Give an example of sets A, B such that A ∈ B and A ⊆ B are both true.
(5) Is {1, 2, 12} ⊆ R?
(6) Is {−5, 8, 12} ⊆ N?
(7) Is {1, 3, 7} ∈ P(N)?
(8) Is N ∈ P(Z)?
(9) Is P(N) ⊆ P(Z)? Are they equal sets? Why or why not?
(10) Give an example of an infinite set T such that T ∈ P(Z) but T ∈
/ P(N).
(11) Suppose G, H are sets and they satisfy P(G) = P(H). Can we conclude
that G = H? Why or why not? (Don’t try to formally prove this; just
think about it and try to talk it out.)
(12) Give an example of a set W such that W ⊆ P(N) but W ∈
/ P(N).
172 CHAPTER 3. SETS

3.5 Set Operations

When you first learned about numbers, a natural next step was to learn about
how to combine them: multiplication, addition, and so on. Thus, a natural next
step for us now is to investigate how we can take two sets and operate on them
to produce other sets. How can we combine sets in interesting ways? There
are several such operations that have standard, notational symbols and we will
introduce you to those operations now.
Throughout this section, we assume that we are given two sets A and B that
are each subsets of a larger universal set U . That is, we assume A ⊆ U and B ⊆
U . The reason we make this assumption is that each operation involves defining
another set by identifying elements of a larger set with a specific property, so we
must have some set U that is guaranteed to contain all of the elements of A and
B so we can even work with those elements. (Again, ensuring this may seem
nit-picky, but it is meant to avoid nasty paradoxes like the one we investigated
before.) Assuming those sets A, B, U exist, though, we can forge ahead with
our definitions.

3.5.1 Intersection
This operation collects the elements common to two sets and includes them in
a new set, called the intersection.

Definition 3.5.1. Let A and B be any sets. The intersection of A and B

is the set of elements that belong to both A and B, and is denoted by A ∩ B.
Symbolically, we define

A ∩ B = {x ∈ U | x ∈ A and x ∈ B}

Example 3.5.2. Define the following sets:

S1 = {1, 2, 3, 4, 5}
S2 = {1, 3, 7}
S3 = {2, 4, 7}
U =N

Then, we see that, for example,

S1 ∩ S2 = {1, 3}
S1 ∩ S3 = {2, 4}
S2 ∩ S3 = {7}

Also, since S1 ∩ S2 is, itself, a set, it makes sense to consider (S1 ∩ S2 ) ∩ S3 .

However, those two sets share no elements, so we write

(S1 ∩ S2 ) ∩ S3 = ∅
3.5. SET OPERATIONS 173

The situation where two sets have no common elements, as seen in the above
example, is common enough that we have a specific term to describe such sets:
Definition 3.5.3. If A ∩ B = ∅, then we say A and B are disjoint.

Intersections and Subsets

You might have observed already that we can say A ∩ B ⊆ A and A ∩ B ⊆ B,
no matter what A and B are. Let’s prove this fact!
Proposition 3.5.4. Let A and B be any sets. Then,

A∩B ⊆A

and
A∩B ⊆B
By the way, a proposition like this is just a “mini result”. It’s not difficult
or important enough to be called a theorem, but it does require a little proof.
Proof. Let’s say we have two sets, A and B. To prove a subset relationship, like
A ∩ B ⊆ A, we need to show that every element of the set on the left (A ∩ B)
is also an element of the set on the right (A).
Let’s consider an arbitrary element x ∈ A ∩ B. By the definition of A ∩ B, we
know that x ∈ A and x ∈ B. Thus, we know that x ∈ A. This was our goal, so
we have shown that A ∩ B ⊆ A.
Also, we know that x ∈ B, so we have also shown that A ∩ B ⊆ B.
This might seem like a simple observation and an easy proof, but we still
need to go through these logical steps to rigorously explain why those subset
relationships hold true. Also, notice the type of proof structure we used here.
To prove a subset relationship holds true, we need to consider an arbitrary
element of one set and deduce that it is also an element of the other set. This
will be our method for proving any claim about subsets.
What if A ⊆ B? What can we say about A ∩ B, in relation to A and B?
Try to prove a statement about this!

3.5.2 Union
This operation collects the elements of either of two sets and includes them in
a new set, called the union.
Definition 3.5.5. Let A and B be any sets. The union of A and B is the set
of elements that belong to either A or B, and is denoted by A ∪ B. Symbolically,
we define
A ∪ B = {x ∈ U | x ∈ A or x ∈ B}
Notice that the “or” in the definition is an inclusive “or”, meaning that
A ∪ B includes any element that belongs to A or B or possibly both sets.
174 CHAPTER 3. SETS

Example 3.5.6. Returning to the sets S1 , S2 , S3 we defined above in Example

3.5.2, we can say

S1 ∪ S2 = {1, 2, 3, 4, 5, 7}
S1 ∪ S3 = {1, 2, 3, 4, 5, 7}
S2 ∪ S3 = {1, 2, 3, 4, 7}

Also, since each of these unions are sets themselves, we could find their union
with another set. For example,

(S1 ∪ S2 ) ∪ S3 = {1, 2, 3, 4, 5, 7} ∪ {2, 4, 7} = {1, 2, 3, 4, 5, 7}

Unions and Subsets

Notice that A ⊆ (A ∪ B) and B ⊆ (A ∪ B), no matter what A and B are. Let’s
prove that!

Proposition 3.5.7. Let A and B be any sets. Then,

A ⊆ (A ∪ B)

and
B ⊆ (A ∪ B)

Proof. Let’s say we have two sets, A and B. To prove A ⊆ (A ∪ B), we need to
show that every element of A is also an element of A ∪ B.
Let x ∈ A be arbitrary and fixed. Then it is certainly that x ∈ A or x ∈ B
(since x ∈ A). This shows x ∈ A ∪ B. Since x was arbitrary, we have shown
A ⊆ A ∪ B.
Let y ∈ B be arbitrary and fixed. Then it is certainly true that y ∈ A or y ∈ B
(since we already know y ∈ B). This shows y ∈ A ∪ B. Since y was arbitrary,
we have shown B ⊆ A ∪ B.

What can you say about the relationship between A ∩ B and A ∪ B? If

A ⊆ B, can we say anything about the relationship between B and A ∪ B? Try
to prove your observations!
We should emphasize that claims like this—that A ⊆ A ∪ B for any sets A
and B—need to be proven; they do not hold by definition. The definition of
the union of two sets is given above. Notice it says nothing about how A and
A ∪ B are related; it just tells us what the object A ∪ B actually is. When
you invoke or cite a definition and use it, be sure to do so; but also, be sure
to explain any claim that isn’t exactly a definition. Since we have proven these
two little lemmas, we get to use them in the future by referencing them; had we
not done so, we would have to re-explain these little facts every time we try to
invoke them!
3.5. SET OPERATIONS 175

3.5.3 Difference
This operation takes the elements of one set and removes the elements that also
belong to another set.

Definition 3.5.8. The difference between A and B, denoted by A − B, is the

set of all elements of A that are not elements of B. Symbolically, we define

A − B := {x ∈ U | x ∈ A and x ∈
/ B}

Example 3.5.9. Returning to the sets S1 , S2 , S3 we defined above in Example

3.5.2, we can say, for example, that

S1 − S2 = {2, 4, 5}
S2 − S1 = {7}
S2 − S3 = {1, 3}

Set Difference Is Not Symmetric

Notice that S1 − S2 6= S2 − S1 in the example above. In general, the operation
“−”, in the context of sets, is not symmetric, and this example here shows
that. Can you find two sets A, B so that A − B = B − A? Can you find two
sets A, B so that A − B = B − A 6= ∅?
Each of the other operations we have defined thus far is, in fact, symmetric.
That is, A ∩ B = B ∩ A and A ∪ B = B ∪ A. Look back at the definitions for
these operations and see why this makes sense. What is it about the language
used in the property definition of the operation that makes this true?

Notation Notes
One more comment on this set difference notation. Notice that we use the
standard subtraction symbol, “−”, but this has nothing to do with “subtraction”
in the way we usually think of it, like with numbers. This might be the first
time you’ve encountered this ambiguity, or perhaps not, but there is a larger
point that is relevant to mathematical notation and terminology: many symbols
have different meanings depending on the context.
When we write 7 − 5, we clearly mean subtraction, i.e. 7 − 5 = 2. However,
when we write A − A where A has been identified as a set, we mean the set
difference operation, i.e. A − A = ∅. Be sure to check the context of any
statement to ensure that the symbols therein do mean what you think they
mean!

3.5.4 Complement
This operation identifies all of the elements that lie “outside” a set. This oper-
ation depends on the context of the universal set U . You’ll notice that this is
evident in the deifnition, and we will illustrate this through examples, as well.
176 CHAPTER 3. SETS

Definition 3.5.10. The complement of A is the set of all elements that are
not elements of A, and is denoted by Ā. Symbolically, we define
Ā = {x ∈ U | x ∈
/ A}
Remember that we assumed A, B, U are given sets that satisfy A ⊆ U and
B ⊆ U . Within this context, the set Ā is well-defined, but this set certainly
depends on A and U !
Example 3.5.11. For instance, let’s return to the sets S1 , S2 , S3 we defined above
in Example 3.5.2. There, we used the context U = Z. In that case,
S̄1 = {6, 7, 8, 9, . . . }
However, what if we had used U = {1, 2, 3, 4, 5, 6, 7}? In that case,
S̄1 = {6, 7}
Since the symbolic notation Ā makes no indication of the set U that it
depends on, it is important to make this set clear in whatever context we are
working. Try identifying some sets A, U1 , U2 so that Ā in U1 is different from Ā
in U2 , and try identifying some sets so that Ā is the same in both cases.

3.5.5 Questions & Exercises

Remind Yourself
Answering the following questions briefly, either out loud or in writing. These
are all based on the section you just read, so if you can’t recall a specific defi-
nition or concept or example, go back and reread that part. Making sure you
can confidently answer these before moving on will help your understanding and
memory!
(1) What is the difference between a union and intersection of two sets?
(2) What does it mean for two sets to be disjoint?
(3) What is Z ∩ N? What is Z ∪ N? What is Z − N?
(4) Is it possible for A − B = B − A to be true? How?
(5) What is [3] in the context of N? What about in the context of Z? Of R?
Try writing your answers using good mathematical notation, and set-buidler
notation, perhaps.
(6) Is (A ∩ B) ∩ C = (A ∩ B) ∩ C always true? Why or why not? What about
with ∪ instead of ∩?
(7) What is the difference between the statements “7 − 5” and “[7] − [5]”?
(8) Suppose x ∈ A. Does A − x make sense, notationally? How can you change
it to make sense?
(9) What is (Z − N) ∪ R?
3.6. INDEXED SETS 177

Try It
Try answering the following short-answer questions. They require you to actu-
ally write something down, or describe something out loud (to a friend/class-
mate, perhaps). The goal is to get you to practice working with new concepts,
definitions, and notation. They are meant to be easy, though; making sure you
can work through them will help you!
(1) List the elements of the following sets:
(a) [7] ∪ [10]
(b) [10] ∩ [7]
(c) [10] − [7]
(d) ([12] − [3]) ∩ [8]
(e) (N − [3]) ∩ [7]
(f) (Z − N) ∩ N
(g) N ∩ {0}, in the context of Z
(2) Find an example of sets A, B, C such that (A − B) − C = A − (B − C).
Then, find an example where they are not equal.
(3) State and prove a relationship between A and U − A.
(4) Let A = [12], let E be the set of even integers, and let P be the set of prime
natural numbers. What is A ∩ E? What is A ∩ P ? What is (A ∩ E) ∩ P ?
Is it the same as A ∩ (E ∩ P )?
Suppose the context is U = N. What are A ∩ E and A ∩ E?
(5) What is {1} ∩ P({1})?
(6) Consider the sets {1} and {2, 3}. Compare the sets P({1} ∪ {2, 3}) and
P({1}) ∪ P({2, 3}). What do you notice?
Repeat this exercise with “∩” instead of “∪”. What do you notice?
(7) Let A, U be sets, and suppose A ⊆ U . Let B = A, in the context of U .
What do you think B is? Why?

3.6 Indexed Sets

3.6.1 Motivation
Let’s discuss a notion that we referenced briefly before and have been using
already: the concept of indexing a collection of sets. This type of notation is
convenient when we wish to define or refer to a large number of sets without
writing out all of them explicitly. Using similar notation with the set operations
we have defined already, we will be able to “combine” and “operate on” a large
178 CHAPTER 3. SETS

number of sets “all at once”. There is really no new mathematical content in this
section, but the notation involved in these ideas can be confusing and difficult
to work with, at first, so we want to guide you through these ideas carefully.

Relation to Summation Notation

We’ll start with a related concept that we have seen before. Remember when
we investigated sums of natural numbers in Chapter 1? We mentioned some
special notation that allowed
Pus to condense a long string of terms in a sum into
one concise form, using the symbol. For instance, we could write an informal
sum
P (“informal” meaning “not rigorous” because of the use of ellipses) in the
notation as follows:
n
X
1 + 2 + 3 + 4 + · · · + (n − 1) + n = i
i=1

Why does this notation work and make sense? The index variable i is
the key componentP of condensing the sum into this form. Writing “i = 1”
underneath the symbol means that the value of the variable i should start
at
P 1 and increase by 1 until it reaches the terminal value, n, written above the
symbol. For each allowable value of i in that range (from P 1 to n), we include
a term in the sum of the form written to the right of the symbol; in this
case, that term is i. Thus, we should have the terms 1, 2, 3, . . . , n with a + sign
between each.
We should point out that it is implicitly understood that writing i = 1 and n
as the limits on the index variable i means i assumes all values that are natural
numbers between 1 and n.

Example
Let’s see the process of defining indexed sets via an example first. We will also
see how to apply set operations to several sets by using an index variable.
Example 3.6.1. We can similarly condense some set operation notation. For
instance, let’s define the sets A1 , A2 , A3 , . . . , A10 by

A1 = {1, 2}
A2 = {2, 4}
A3 = {3, 6}
..
.
Ai = {i, 2i}
..
.
A10 = {10, 20}
3.6. INDEXED SETS 179

We included the definition of Ai for an arbitrary value i to give these sets a

rigorous definition. Without definining that set—which defines Ai for any rele-
vant value of i—we would be leaving it up to the reader to interpret the pattern
among the sets A1 , A2 , A3 , A10 , and there could be multiple ways of interpreting
that. By defining the term Ai explicitly like this, there is no confusion as to
what we want these ten sets to be.
Furthermore, we can more easily express the union of all of these sets, for
instance. Remember that the union of two sets is the set containing all elements
of both sets (i.e. an element is included in the union if it is in the first set or
the second set, or possibly both). What is the union of more than two sets? It
follows the same idea as the definition for just two sets; we want to include an
element in the union if it is in any of the constituent sets we are combining via
the union operation.
How can we write
P this union concisely and precisely? Let’s follow the same
motivation of the notation. The index of these sets runs from 1 to 10, so we
should write i = 1 below a “∪” symbol and 10 above it. Each term in the union
is of the form {i, 2i}, so we should write that to the right of the “∪” S symbol.
For indexed unions like this one, though, we use a slightly larger “ ” symbol,
like so:
10
[ 10
[
A1 ∪ A2 ∪ A3 ∪ · · · ∪ A10 = Ai = {i, 2i}
i=1 i=1

This is far more concise than writing the elements of all 10 sets, so you can see
how useful this notation is. We will keep reminding you of the imprecision of the
ellipses in the union on the left and tell you that, in fact, an expression like the
one on the right is a truly rigorous mathematical statement about this union.
The expression on the left is more of an intuitive, heuristic way of describing
the union operation applied to these ten sets.

When The Index Set Is Not a Range of Numbers

Let’s examine a more difficult example to motivate the next development in
this notation technique. What if we asked you to write the following sum in
summation notation: the sum of the squared reciprocals of all prime numbers.
How can we accomplish this? (Note: We just want to express all the terms of
the sum without evaluating the sum. That’s a difficult endeavor left for another
time!)
Unfortunately, we cannot use the exact same notation as above, because we
don’t want to sum over a range of index values between two natural numbers;
rather, we want to only include terms in the sum corresponding to prime num-
bers. The solution to this is to define an index set I that will describe the
allowable values of the index that we will then “plug into” the arbitrary term
written to the right of the sum.
In this case, if we have a prime number i, we would like to include Pthe term
1
i2 in our sum, so this expression will be written to the right of the symbol.
We would like to express in our notation that the value i should be a prime
180 CHAPTER 3. SETS

number and that all possible prime numbers should be included. The index set
I of allowable values should, therefore, be the set of all prime numbers. That
is, we can write this sum as
X 1
, where I = {i ∈ N | i is prime}
i2
i∈I

Look at what this notation accomplishes! Not only have we condensed an

infinite number of terms into one expression, we have indicated that the values
of the arbitrary index i should be restricted to prime numbers,
Pn which do not
behave in the “usual” and convenient way as with a sum like i=1 i.
Example 3.6.2. This concept of an index set is extremely useful and extends to
arbitrary sets and even non-mathematical objects. For instance, in our discus-
sion of sets above we used the set B of all Major League Baseball teams. How
can we use this set to express the set P of all Major League Baseball players?
Each team is, itself, a set whose elements are the players on that team, so a
union of all of the teams (i.e. a union of all sets in B) should produce exactly
this set of all players! In this case, our index set is B, and for each element
b ∈ B, we want to include b as a set in our union. Thus, we would write
[
P = b
b∈B

The individual terms in this union are not even dependent on natural num-
bers, so there would have been no way to express this union without the use of
an index set, like this. Also, this union is dependent on the fact that the terms
of the union are elements of the index set B, but they are also sets themselves;
thus, applying the union operation to them makes mathematical sense. This
might still seem like an odd idea, so be sure to think carefully about this idea
of sets having elements that are sets, themselves.

Reading Indexed Expressions Aloud

To verbalize these types of expressions, and to help you think of them in your
head, let’s give you an example. We might read the expression up above as

“The sum, over all i that are prime, of i12 .”

or
“The sum of i12 , where i ranges over all prime numbers.”

Likewise, we might read the other expression above as

“The union, over all b that are MLB teams in the 2012 season, of
those b.”
or
“The union of all sets b, where b ranges over all MLB teams from
the 2012 season.”
3.6. INDEXED SETS 181

3.6.2 Indexed Unions and Intersections

Let’s give a precise definition of this union operation for more than one set,
since we have only rigorously defined the union of two sets.
Definition 3.6.3. The union of a collection of sets Ai indexed by the set I is
[
Ai = {x ∈ U | x ∈ Ai for some (i.e. at least one) i ∈ I}
i∈I

where we assume there is a set U such that Ai ⊆ U for every i ∈ I.

In mathematical language, the phrase “for some i ∈ I” means that we want
there to be at least one i ∈ I with the specified property. If an element x
satisfies x ∈
/ Ai for every i ∈ I, then this says x is not in any of the sets in our
collection, so it should not be included in the union.
Following this idea, we can make a similar definition for set intersections.
Definition 3.6.4. The intersection of a collection of sets Ai indexed by the set
I is \
Ai = {x ∈ U | x ∈ Ai for every i ∈ I}
i∈I

where we assume there is a set U such that Ai ⊆ U for every i ∈ I.

3.6.3 Examples
Let’s return to a previous example and make these ideas clearer.
Example 3.6.5. Previously, in Example 3.6.1, we defined

Ai = {i, 2i}

for every natural number i between 1 and 10. Another way of defining this
collection is to consider the index set I = [10] (recall the notation [n] =
{i ∈ N | 1 ≤ i ≤ n}) and define A to be the set

A = {Ai | i ∈ I} , where Ai = {i, 2i} for every i ∈ I

This defines every set Ai , dependent on the index value i chosen from the index
set I, and it “collects” all of these sets into one set A. Then, another way of
writing that union we wrote before would be
[
Ai
i∈I

with the given definitions of I and Ai .

(Think carefully about how this union differs from the set A. Also, what exactly
is this union? How can we express its elements conveniently? Do we need to
list every element? What if we change the index set I to be N? What is the
union in that case?)
182 CHAPTER 3. SETS

Example 3.6.6. Let I = {1, 2, 3} and, for every i ∈ I, define

Ai = {i − 2, i − 1, i, i + 1, i + 2}
Let’s identify and write out the elements of the following sets:
[ \
Ai and Ai
i∈I i∈I

Notice that we can write out the elements of each Ai set, as follows:
A1 = {−1, 0, 1, 2, 3}
A2 = {0, 1, 2, 3, 4}
A3 = {1, 2, 3, 4, 5}
Thus, [
Ai = A1 ∪ A2 ∪ A3 = {−1, 0, 1, 2, 3, 4, 5}
i∈I
and \
Ai = A1 ∩ A2 ∩ A3 = {1, 2, 3}
i∈I
Now, consider J = {−1, 0, 1}, with Aj defined in the same way as before. Let’s
identify the elements of the sets
[ \
Aj and Aj
j∈J j∈J

Writing out the elements of each set, we can determine that

[
Aj = A−1 ∪ A0 ∪ A1 = {−3, −2, −1, 0, 1, 2, 3}
j∈J

and \
Aj = A−1 ∩ A0 ∩ A1 = {−1, 0, 1}
j∈J

Try answering the same questions with different index sets.

For instance, consider K = {1, 2, 3, 4, 5} or L = {−3, −2, −1, 0, 1, 2, 3}.
Example 3.6.7. Define the index set I = N. For every i ∈ I, define the set
 
i+1
Ci = x ∈ R | 1 ≤ x ≤
i
Then we claim that
[ \
Ci = {y ∈ R | 1 ≤ y ≤ 2} and Ci = {1}
i∈I i∈I

Can you see why these are true? We will discuss the techniques required to
prove such equalities later on. For now, we ask you to just think about why
these are true. Can you explain them to a classmate or a friend? What sort of
techniques might you use to prove these claims?
3.6. INDEXED SETS 183

Example 3.6.8. Let S be the set of students taking this course. For every s ∈ S,
let Cs be the set of courses that student s is taking this semester. What do the
following expressions represent?
[ \
Cs and Cs
s∈S s∈S

We bet you can identify at least one element of the set on the right!

3.6.4 Partitions
Now that we have a way of writing down a union of many sets, we can define
a helpful notion: that of a partition. Linguistically speaking, a partition is a
way of “breaking something apart into pieces”, and that’s pretty much what
this word means, mathematically speaking.
To wit, a partition is just a collection of subsets of a set that do not overlap
and whose union is the entire set. Let’s write down that definition here and
then see some examples and non-examples. We will have occasion to use this
definition many times in the future, so let’s get a handle on it now while we’re
talking about sets and indexed unions.

Definition 3.6.9. Let A be a set. A partition of A is a collection of sets that

are pairwise disjoint and whose union is A.
That is, a partition is formed by an index set I and non-empty sets Si (defined
for every i ∈ I) that satisfy the following conditions:

(1) For every i ∈ I, Si ⊆ A.

(2) For every i, j ∈ I with i 6= j, we have Si ∩ Sj = ∅.

[
(3) Si = A
i∈I

The sets Si are called parts of the partition.

The idea here is that the sets Si “carve up” the set A into non-overlapping,
nonempty pieces.
Example 3.6.10. Let’s see a couple of examples.

(1) Consider the set N. Let O be the set of odd natural numbers, and let E be
the set of even natural numbers. Then {O, E} is a partition of N. This is
because

• E, O 6= ∅, and
• E, O ⊆ N, and
• E ∩ O = ∅, and
• E∪O =N
184 CHAPTER 3. SETS

(2) Consider the set R. For every z ∈ Z, define the set Sz by

Sz = {r ∈ R | z ≤ r < z + 1}

We claim {. . . , S−2 , S−1 , S0 , S1 , S2 , . . . } is a partition of R. Can you see

why? Try to write out the conditions required for this collection of sets to
be a partition, and see if you can understand why they hold.
Specifically, remember that we need these sets to be pairwise disjoint. This
means that any two sets must be disjoint. Notice that this is quite different
from requiring the the intersection of all the sets together to be empty.
For instance, consider the collection of sets

{1, 2}, {2, 3}, {3, 4}

This collection is not pairwise disjoint because, for example,

{1, 2} ∩ {2, 3} = {2} =

6 ∅

However, the intersection of all three sets is empty, because no element is

common to all three of them together.
Example 3.6.11. Now, let’s see a couple of non-examples.
(1) Consider the set R. Let P be the set of positive real numbers and let N be
the set of negative real numbers. Then {N, P } is not a partition because
0∈/ N ∪ P.
Can you modify the choices we made here to identify a partition of R into
two parts?
(2) Consider the set Z. Let A2 be the set of integers that are multiples of 2,
let A3 be the set of integers that are multiples of 3, and let A5 be the set
of integers that are multiples of 5. The collection {A2 , A3 , A5 } is not a
partition for two reasons.
First, these sets are not pairwise disjoint. Notice that 6 ∈ A2 and 6 ∈ A3 ,
since 6 = 2 · 3. Second, these sets do not “cover” all of Z. Notice that 7 ∈ Z
but 7 ∈/ A2 ∪ A3 ∪ A5 .
As we mentioned, we will have occasion to use this definition frequently in
the future, so keep it in mind.

3.6.5 Questions & Exercises

Remind Yourself
Answering the following questions briefly, either out loud or in writing. These
are all based on the section you just read, so if you can’t recall a specific defi-
nition or concept or example, go back and reread that part. Making sure you
can confidently answer these before moving on will help your understanding and
memory!
3.6. INDEXED SETS 185

(1) What is an index set?

(2) Let I = N, and for every i ∈ I, let Ai = {i, −i}. Why are the following sets
all the same set? [ [ [
Ai Ax Aj
i∈I x∈N j∈I

By the way, what are the elements of this set?

(3) List the elements of the following sets:

[
(a) {x}
x∈N
\
(b) {x}
x∈N
[
(c) {x, 0, −x}
x∈N

(4) Why do you think we didn’t talk about an “indexed difference” or an “in-
dexed complementation”, and only talked about unions and intersections?

(5) What is a partition? What conditions does a collection of sets have to

satisfy to be a partition of a set?

Try It
Try answering the following short-answer questions. They require you to actu-
ally write something down, or describe something out loud (to a friend/class-
mate, perhaps). The goal is to get you to practice working with new concepts,
definitions, and notation. They are meant to be easy, though; making sure you
can work through them will help you!

(1) Let A = {−2, −1, 0, 1, 2}. Let B = {1, 3, 5}.

For every i ∈ Z, let Si = {i − 2, i, i + 2, i + 4}.
[ \
What is Si ? What is Sx ?
i∈A x∈B

\ [
(2) For every n ∈ N, let An = [n]. What is An ? What about An ?
n∈N n∈N

(3) Find a way to write the set of all integers between -10 and 10 (inclusive)
using set-builder notation. Then, define the same set using an indexed
union. Can you do this in a way so that the sets in your union are pairwise-
disjoint (meaning that no two of them have any elements in common)?
(Hint: Yes, you can.)
186 CHAPTER 3. SETS

(4) For every n ∈ N, let Mn be the set of all multiples of n. (For example,
M3 = {3, 6, 9, . . . }.) Write a definition for Mn using set-builder notation.
Then, express N as a union, using these sets.
(Challenge: Can you use these sets to define a partition of N?)
[ \
(5) Let X be any set. What is S? What about S?
S∈P(X) S∈P(X)

(It might help to try this with a specific set, like X = {1, 2}, to see what
happens, first.)

(6) Identify an index set and define some sets that allow you to express Q as
an indexed union.
Can you do this so that there are infinitely many elements in the index set?
(Challenge: Can you make this collection a partition of Q?)

3.7 Cartesian Products

There is one more way of “combining” sets to produce other sets that we want
to investigate. This method is based on the idea of order. When we define
sets by listing the elements, the order is irrelevant; that is, the sets {1, 2, 3}
and {3, 1, 2} and {2, 1, 3} are all equal because they contain the same elements.
(More specifically, they are all subsets of each other in both directions). Looking
at mathematical objects where order is relevant, though, allows us to combine
sets in new ways and produce new sets.
You are likely already familiar with the idea of the real plane, R2 (also known
as the Cartesian plane after the French matheamtician René Descartes). Each
“point” on the plane is described by two values, an x-coordinate and a y-
coordinate, and the order in which we write those coordinates is important.
We usually think of the x-coordinate as first and the y-coordinate as second,
and this helps to distinguish two points based on this order. For instance, the
point (1, 0) lies on the x-axis but the point (0, 1) lies on the y-axis. They are
not the same point.
There is a deeper, mathematical idea underlying the Cartesian plane. Given
any two sets, A and B, we can look at the set of all ordered pairs of elements
from A and B. By pair we mean an expression (a, b) where a and b are elements
of A and B, respectively. By ordered we mean that writing a first and b second
is important. In the case of the real plane, it is especially important because any
real number could appear as the x-coordinate or the y-coordinate of a point,
but the point (x, y) is generally different from the point (y, x). (When are they
equal? Think carefully about this.)

3.7.1 Definition
Let’s give an explicit definition of this new set before examining some examples.
3.7. CARTESIAN PRODUCTS 187

Definition 3.7.1. Given two sets, A and B, the Cartesian product of A and
B is written as A × B and defined to be

A × B = {(a, b) | a ∈ A and b ∈ B}

This definition tells us that the Cartesian product A × B collects into a new
set all of the ordered pairs (a, b), where a is allowed to be any element of A and
b is allowed to be any element of B.

Some Technicalities
Notice that we have dropped the assumption of a universal set U . We have
discussed some of the issues that arise when we do not specify a universal set,
but from now on the sets we work with will not address any of these issues.
Accordingly, we will only specify a universal set when not doing so would lead
to ambiguity.
In the case of this definition, we could specify a universal set by defining the
ordered pair (a, b) as a set.Specifically, we could define

(a, b) = { {a}, {a, b} }

This definition incorporates the order of the pair, as well, in the sense that

(a, b) = (c, d) if and only if a = c and b = d

Checking the singleton element in the set tells us the first coordinate, and check-
ing the other element in the set with two elements tells us the second coordinate.
If we have the ordered pair (a, a), then the set reduces to {{a}}, which tells us
a appears in both coordinates.
Using this definition, we could use the universal set U = P(P(A ∪ B)).
We won’t delve into the technical details of these sets and definitions, but we
thought it prudent to point out that such definitions exist. The important point
to remember from this section is given above:

Two ordered pairs are equal if and only if both of their coordinates
are equal.

This is why we call them ordered pairs.

3.7.2 Examples
The Cartesian plane is R × R, which is why we sometimes see this written as R2 .
Indeed, if A = B, then we sometimes write the Cartesian product as A×A = A2 ,
if there is no confusion about the fact that A is a set (and not a number). Let’s
see some more examples where the two sets in the Cartesian product are not
the same.
188 CHAPTER 3. SETS

Example 3.7.2. Define the sets A = {a, b, c} and B = {6, 7} and C = {b, c, d}.
Then we can list the elements of the following Cartesian products:

A × B = {(a, 6), (a, 7), (b, 6), (b, 7), (c, 6), (c, 7)}
B × C = {(6, b), (6, c), (6, d), (7, b), (7, c), (7, d)}
A × C = {(a, b), (a, c), (a, d), (b, b), (b, c), (b, d), (c, b), (c, c), (c, d)}
C × B = {(b, 6), (b, 7), (c, 6), (c, 7), (d, 6), (d, 7)}

Notice that, in general, B × C 6= C × B, as this example shows.

(Can you identify the situations where A × B = B × A? What conditions must
we impose on the sets A and B to make this equality true?)

Ordered Triples and Beyond

This idea also extends to Cartesian products of three or more sets. We simply
write ordered triples for a Cartesian product of three sets and, in general, for
the Cartesian product of n sets, we write ordered n-tuples. (Again, we point
out that there are set-theoretic ways of defining these ordered n-tuples, but we
will not investigate those details.)

Example 3.7.3. The Cartesian product N × N × N (sometimes written as N3 )

is the set of all ordered triples of natural numbers. For instance, (1, 2, 3) ∈ N3
and (7, 7, 100) ∈ N3 , but (0, 1, 2) ∈
/ N3 and (1, 2, 3, 4) ∈
/ N3 .
3
Notice the subtle distinction between N and (N × N) × N. A typical element
of N3 is an ordered triple whose coordinates are each a natural number. A
typical element of (N × N) × N is an ordered pair, the first coordinate of which
is also an ordered pair (of natural numbers) and the second coordinate of which
/ N3 . This
is a natural number. That is, ((1, 2), 3) ∈ (N × N) × N but ((1, 2), 3) ∈
shows the two are different sets.
There is, however, a natural way of relating the two sets which essentially
“drops the parentheses” around the first coordinate (the ordered pair). We
will discuss this later on when we examine functions and bijections. For now,
though, we just want you to notice the subtle differences between the two sets
and remember that a Cartesian product of two sets is a set of ordered pairs,
where each coordinate is drawn from the corresponding constituent set.

Example 3.7.4. What happens if B = ∅, say? Look back at the definition of

A × B. There are actually no elements of B to write as the second “coordinate”
of the ordered pair, so we actually have no elements of A × B to include! Thus,

A×∅=∅

for any set A. Similarly, ∅ × B = ∅, for any set B.

3.7. CARTESIAN PRODUCTS 189

3.7.3 Questions & Exercises

Remind Yourself
Answering the following questions briefly, either out loud or in writing. These
are all based on the section you just read, so if you can’t recall a specific defi-
nition or concept or example, go back and reread that part. Making sure you
can confidently answer these before moving on will help your understanding and
memory!

(1) What is the difference between R × N and N × R? Give an example of an

ordered pair that is an element of one set, but not the other. Then, give an
example of an ordered pair that is actually an element of both sets.

(2) What is ∅ × Z?

(3) Write out all the elements of the set {♥, ♦} × {,, , ♥}.

(4) What is the difference between (N × N) × N and N × (N × N)? Why aren’t

they technically the same set? Can you explain why they are “essentially”
the same set?

(5) Let A, B, C be sets. Suppose A ⊆ B. Do you think A × C ⊆ B × C is true?

Why or why not?

(6) Give an example of a set S such that 12 , −1 ∈ S.

Try It
Try answering the following short-answer questions. They require you to actu-
ally write something down, or describe something out loud (to a friend/class-
mate, perhaps). The goal is to get you to practice working with new concepts,
definitions, and notation. They are meant to be easy, though; making sure you
can work through them will help you!

(1) Write out the elements of [3] × [2].

Can you make a conjecture about how many elements [m] × [n] has, for any
m, n ∈ N? (How might you try to prove your conjecture?)

(2) Give an example of an element of the set N × P(Z).

(3) Give an example of an element of the set (R × N) × Q ∩ (Q × Z) × N .

(4) Give an example of sets C, D such that C × D = D × C.

Follow-up challenge: Can you characterize all possible situations like this
one. What must be true about C and D? Can you prove it?


(5) Write out the elements of P [1] × [2] .
190 CHAPTER 3. SETS

(6) For every n ∈ N, let An = [n] × [n]. Consider the set

[
B= An
n∈N

Is B = N × N or not? Explain, with examples.

(7) If you know some simple computer programming, try writing code (in your
favorite language) that will input m, n ∈ N and print out all the elements
of [m] × [n]. (Use some pseudocode if you’re not totally comfortable with
programming.) How long do you think this takes to run, depending on m
and n?

3.8 [Optional Reading] Defining the Set of Nat-

ural Numbers
Our goal in this section is to put the natural numbers N on a rigorous, mathe-
matical foundation. Specifically, we will prove that the natural numbers exist,
by defining and deriving them from the axioms and principles of set theory. We
will then discuss a few of their properties. We will use some of those properties
to define and prove the Principle of Mathematical Induction in Chapter 5, after
discussing some basic principles and results of mathematical logic.

3.8.1 Definition
How do we define the natural numbers in terms of sets? We intuitively know
what they are. We start with 1 and repeatedly add 1, obtaining all of the other
natural numbers. Thus, we have to identify what we mean by “1” and what
we mean by “add 1”, in terms of sets. To do this, let’s start by thinking about
0. We stated before that we will not include 0 in the set N, but some authors
do, and it will aid us in deriving N, right now, to consider it. We know of
exactly one set that contains no elements, the empty set. Thus, it makes sense
to associate 0 with the empty set; in fact, we define 0 = ∅. Next, we wish to
define 1, and following our definition of 0, it makes sense to choose a set that
has exactly one element. (A set with one element is also known as a singleton.)
There are several such sets:

{∅}, {{∅}} , {{∅, {∅}}}

How do we choose a representative singleton to represent 1? Keeping in mind

that we want to continue this process and eventually define 2 (and 3, and so
on) in terms of previous numbers, it makes sense now to define 1 in terms of
the only object we have at our disposal: 0. Thus, let us choose to define

1 = {0} = {∅}

This guarantees 0 6= 1.
3.8. DEFINING THE NATURAL NUMBERS 191

Next to define 2, we consider sets containing two elements, like

{∅, {∅}} , {∅, {{∅}}} , {{∅} , {{∅}}}
and so on. We seek a natural representative, and we notice that the first set
listed above contains the two objects, 0 and 1, that we have already defined!
Thus, defining 2 = {0, 1} is a natural choice and, again, we know 2 6= 0 and
2 6= 1.

Successors
This gives us an intuitive idea of how to continue this process and define any
natural number: for any n ∈ N, we define
n = {0, 1, 2, . . . , n − 2, n − 1}
However, given a set, it would be quite difficult to verify, using this definition,
whether or not that set represents a natural number. We would like a better
definition of the elements of N; we want to know, given any set, whether it
belongs in N. Look back at the element n above; we could also write
n = {0, 1, 2, . . . , n − 2, n − 1} = {0, 1, 2, . . . , n − 2} ∪ {n − 1} = (n − 1) ∪ {n − 1}
Look at that! We have a natural way of defining an element of N in terms of the
previous element and in terms of set operations. This motivates the following
definition.
Definition 3.8.1. Given any set X, the successor of X, denoted by S(X), is
defined to be S(X) = X ∪ {X}.
This definition applies to all sets, but in the context of natural numbers,
it means that the successor of n is precisely that natural number we “know”
intuitively to be one larger, namely n + 1.

Inductive Sets
This brings us closer to our definition of N. We certainly want 1 ∈ N and we
also want S(n) ∈ N for any element n ∈ N. To codify this symbolically, we
make the following definitions:
Definition 3.8.2. A set I is called inductive provided
1. 1 ∈ I
2. If n ∈ I, then S(n) ∈ I, as well.
Certainly, N itself (as we hope to define it) should be an inductive set. Are
there other inductive sets? Think about this. What properties would they
have? Would they contain elements that are not natural numbers? We don’t
want to address these questions in depth, but for the sake of our discussion here,
we will point out that there are, indeed, other inductive sets. We don’t want
any of those sets to be N, so we make this definition:
192 CHAPTER 3. SETS

Definition 3.8.3. The set of all natural numbers is the set

N := {x | for every inductive set I, x ∈ I}

Put another way, N is the smallest inductive set, in the sense of set inclusion:
\
N= I
I∈{S|S is inductive}

This dictates that N is a subset of every inductive set.

This gives us the “checking property” we desired. Any set x is a natural

number (i.e. x ∈ N) if and only if it is an element of every inductive set (i.e.
x ∈ I for every inductive set I). This also tells us that N ⊆ I for every inductive
set I.
There are some other set theoretic discussions that could be made here:
How do we know that such an infinite set exists? (In actuality, we need to
make this an axiom of set theory! Assuming these types of sets exist, how do
we characterize those other inductive sets that are not N? Addressing these
questions lies outside the scope and goals of this course, so we will not address
them. We will, however, mention a few properties of N now, specifically ones
that will be useful in setting mathematical induction on a rigorous foundation.
(In case you’re wondering, think about the set of integers, Z. Try to explain
why this set is, indeed, inductive. What about R? What about Z − N?)

Properties of N

Before we define the principle of induction, let’s think about some of the common
properties and uses of natural numbers: orderings and arithmetic. Given any
two natural numbers, we can compare them and decide which one is larger and
which one is smaller (or if they are equal). We usually write this with symbols
like 1 < 3, 1 ≤ 5, 4 6< 2, 3 = 3, etc.
Can we phrase these comparisons in terms of sets, knowing that we have
now defined the elements of N as sets, themselves? Yes, we can! Look back at
the definition of successor. Built into that definition is the fact that X ∈ S(X)!
This observation gives us the following definition:

Definition 3.8.4. Given two natural numbers m, n ∈ N, we write m < n if and

only if m ∈ n.

This defines an order relation on the set N. We will discuss the concepts of
relations and orders later on in the book (in Section 6.3).
What about arithmetic? What is m + n in terms of the sets m and n? How
do we define this operation and its output? How do we know m + n is another
natural number? Can we be sure that m + n = n + m? These are questions we
can address later on after discussing functions and relations.
3.8. DEFINING THE NATURAL NUMBERS 193

3.8.2 Principle of Mathematical Induction

For now, let us present a more rigorous version of induction:
Theorem 3.8.5 (Principle of Mathematical Induction). Let P (n) be some
“fact” or “observation” that depends on the natural number n. Assume that
1. P (1) is a true statement.
2. Given any k ∈ N, if P (k) is true, then we can conclude necessarily that
P (k + 1) is true.
Then the statement P (n) must be true for every natural number n ∈ N.
Let us first prove this theorem before discussing its assumptions and conse-
quences in detail.
Proof. Define the set S to be the natural numbers for which the statement P is
true. That is, define S = {n ∈ N | P (n) is true}. By definition, S ⊆ N.
Furthermore, the assumptions of the theorem guarantee that 1 ∈ S and that
whenever k ∈ S, we know k + 1 ∈ S, as well. This means S is an inductive set.
By the observation we made after defining N, we know that N ⊆ S.
Therefore S = N, so the statement P (n) is true for every natural number n.
This is pretty slick, right? It seems that all of the desired conclusions “fell
out of” our definitions! In this sense, the definitions and axioms are natural
choices, because they accomplish what our intuition already “knew” about the
set N and its properties.
There are a few minor issues that we have left undiscussed. Specifically, what
do we mean by a “fact” or “observation” that depends on a natural number n?
What does it mean to necessarily conclude that P (k + 1) is true when P (k)
is true? What do we even mean by true? These are all deep mathematical
questions and involve a thorough study of logic, and we will discuss these issues
in the next chapter! Onward! Huzzah!

3.8.3 Questions & Exercises

Remind Yourself
Answering the following questions briefly, either out loud or in writing. These
are all based on the section you just read, so if you can’t recall a specific defi-
nition or concept or example, go back and reread that part. Making sure you
can confidently answer these before moving on will help your understanding and
memory!
(1) What is an inductive set? Give an example of one that is not N or Z.
(2) We defined S = {n ∈ N | P (n) is true} in the proof of the Principle of
Mathematical Induction. What does this mean? Describe this set in words.
(3) Come up with your own analogy for how Induction works.
194 CHAPTER 3. SETS

Try It
Try answering the following short-answer questions. They require you to actu-
ally write something down, or describe something out loud (to a friend/class-
mate, perhaps). The goal is to get you to practice working with new concepts,
definitions, and notation. They are meant to be easy, though; making sure you
can work through them will help you!

(1) What if we changed the definition of successor to be S(X) = {X}. Using

0 = ∅, what are 1, 2, 3, and 4 in terms of sets? Do they still satisfy the
equality n = {0, 1, . . . , n − 1}? If not, do they satisfy some other relation-
ship? Explore!

(2) Have a debate with a friend about whether or not infinite sets exist. Why
do we need to assume the existence of an inductive set to define N? Does
this seem valid to you? Does it make sense, physically? Mathematically?

(3) Consider a simple arithmetic statement, like 1 + 2 = 3. Write out the

numbers 1, 2, and 3 in terms of sets, and see how this equation might make
sense. What does “+” mean, in this context?

(4) Investigate how one might define Z, using N. Do some exploring online or
in books, or make up an idea on your own.

3.9 Proofs Involving Sets

Now that we’ve gone through many definitions and examples, to introduce what
sets are and how to manipulate them, let’s actually write up some rigorous,
mathematically correct, and well-written proofs about sets. All of the propo-
sitions/lemmas contained here are useful facts that we can cite later on, and we
would expect you to be able to prove claims like these. (Note: A lemma is just
a small result that requires some proof, and can be cited later to prove more
significant theorems.) Furthermore, all of these proofs are of the type of quality
and rigor that we will expect from you in the future, the very near future . . . Use
these as guidelines, if you’d like!

3.9.1 Logic and Rigor: Using Definitions

The main point we’d like to emphasize here—as we transition from descrip-
tive, “wordy”, and intuitive proofs into more rigorous, mathematically correct,
and formally-written ones—is that formal definitions are very important.
Fundamentally, they’re essential because when we say, for example, “A ∪ B”,
we need to know that you know exactly what that symbol means and how it
operates on the sets A and B.
As another example, when we say “Prove A = B”, we have a very specific
goal in mind, and you need to be on the same page. It always helps to have an
intuitive understanding of the main concepts—“Oh, the statement A = B just
3.9. PROOFS INVOLVING SETS 195

means that A and B have the same elements in them”—but this is not the type
of language/ideas we want to use in a rigorous proof. To prove a statement like
A = B, we need you to appeal to the definition of “=” in the context of sets:
A = B if and only if A ⊆ B and B ⊆ A.
This is what we mean when we say “satisfy a definition” or “appeal to a
definition”: to prove that some mathematical object has a certain property,
you must demonstrate that the object satisfies the formal definition of that
property. If you aren’t familiar with that definition, or have forgotten how to
state it precisely . . . by all means, go look it up! We realize this is a lot of
new information to ingest, and there’s nothing wrong with forgetting something
when it’s still new to you. By doing this, you’ll start to internalize these ideas
more quickly and more solidly.
You’ll see how we use the definitions of, for instance, “⊆” and “=” and “∩”
and so on in the examples below. For each proposition/lemma, we will end up
writing a formal proof, but we will also write a little bit about how we would
approach coming up with such a proof. Oftentimes this is the difficult part! We
think you’ll notice that many of those explanations will amount to just recalling
a relevant definition and thinking about what it means and how it applies in the
given situation. In a way, that’s what a lot of mathematics is. We just allow
our definitions we use to get more and more complicated.

3.9.2 Proving “⊆”

Recall the definition of subset, because we will use it frequently here:
Definition 3.9.1. Given two sets A and B, if every element of A is also an
element of B, then we say A is a subset of B.
Say we are presented with the following problem:
Let A be the set . . . and let B be the set . . . Prove that A ⊆ B.
How can we satisfy the definition of A ⊆ B to prove this claim? Yes, the intuitive
idea is that “every element of A is also an element of B”, but we shouldn’t just
try to dance around the issue and try to make that sentence our conclusion.
Rather, we need to verify that every element of A is also necesarily an element
of B. This is where the wonderful phrase “arbitrary and fixed” will come in
handy!

The Phrase “Arbitrary and Fixed”

How can we talk about all possible elements of A all at once? Of course, we
might not need to do this if A has only, say, 3 elements; then, we can just work
with them one by one. But what if A has 100 elements? Or 1 million? Or
infinitely many? How can we prove something about all of them at once in a
reasonable way?
What we will do is introduce an arbitrary and fixed element of A so we
have something to work with. This element will be arbitrary in the sense that
196 CHAPTER 3. SETS

we make no extra assumptions about what it is or what properties it has, only

that it is an element of A. This element will be fixed in the sense that we will
assign it some variable name (usually a letter, like a or x or t or something)
and this letter will represent the same object throughout the remainder of our
proof. As long as we can prove our goal for this element, then we will have
simultaneously proven something about all elements of A. Voilà!

Examples
Let’s see this process in action to really get the point across. We’ll begin with
the statement to be proven, then describe our thought processes in coming up
with a proof, and then present our formal, written proof.
Lemma 3.9.2. Let A, B, X be any sets.
If X ⊆ A and X ⊆ B, then X ⊆ A ∩ B.
Intuition: Consider drawing a Venn diagram to represent this situation. To
make sure the assumptions X ⊆ A and X ⊆ B both hold true, we need to
make the set X “lie inside” both A and B. Accordingly, this means X must
lie entirely “inside” where A and B overlap, which is what A ∩ B represents.
This helps us realize that this statement is, indeed, True. But it doesn’t prove
anything!
To prove the statement, we will introduce an arbitrary and fixed element x ∈ X.
What do we know about it? Well, we assumed that X ⊆ A. The definition of
“⊆” means that any element of X is also an element of A. But we know x is an
element of X; this means it is also an element of A. How convenient! We can
make some similar statements about x and X and B that will tell us x ∈ B.
What does this mean, overall? Oh hey, the definition of “∩” applies, and tells
us x ∈ A ∩ B. Brilliant! Now, let’s write it up.
Proof. Let x ∈ X be arbitrary and fixed.
By assumption X ⊆ A, so x ∈ A, as well, by the definition of ⊆.
Similarly, by assumption X ⊆ B, so x ∈ B, as well.
Since x ∈ A and x ∈ B, this means that x ∈ A ∩ B, by the definition of ∩.
Overall, we have shown that whenever x ∈ X, it is also true that x ∈ A ∩ B.
Since x ∈ X was arbitrary, we conclude that X ⊆ A ∩ B.
Not so bad, right? Let’s try another one, a slightly harder one, even.
Proposition 3.9.3. Let A and B be any sets. Then, P(A) ∩ P(B) ⊆ P(A ∩ B).
Whoa, is this really true? Look back at Problem 6 in Section 3.5, and you’ll
see an example. This claim states that it is true, in general, and not just for
that example. Let’s figure out why, and then prove it.
Intuition: There are several layers of definitions at work here. In particular,
the power set operation might be confusing to you. The key is to just remember
3.9. PROOFS INVOLVING SETS 197

that definition: P(A) is the set of all subsets of A. Now, the main claim here is
one of a subset relationship: whatever the set P(A) ∩ P(B) is (we’ll analyze it
later, but it’s important that you recognize immediately what type of object it
is: a set), it is supposed to be a subset of whatever the set P(A ∩ B) is. That’s
it, and it’s important to notice that this is really motivates the overarching form
of the forthcoming proof.
Without even having to think about what P(A) ∩ P(B) means, we can be
sure that our proof will start with “Let X ∈ P(A) ∩ P(B) be arbitrary and
fixed”. This is because we need to satisfy the definition of “⊆” by taking any
old element of the set on the left and deducing that it is also an element of the
set on the right. This is what we mean by the structure of the proof.
What does an element X ∈ P(A) ∩ P(B) “look like”? It’s a set, and it’s
an element of both P(A) and P(B). This means . . . well, we’re actually going
to skip ahead and jump right into the formal proof now, because we’ll just find
ourselves repeating the same words down below anyway. But before going ahead
to read ours, we think you should go off and try to write your own proof. Then,
when you’re done, you can compare and see whether you are correct, whether
it has the same steps as ours, whether it’s written clearly, and so on. See what
you can do!

Proof. Let X ∈ P(A) ∩ P(B) be arbitrary and fixed.

By the definition of ∩, this means X ∈ P(A) and X ∈ P(B).
Since X ∈ P(A), we know X ⊆ A, by the definition of power set.
Similarly, since X ∈ P(B), we know X ⊆ B.
Since X ⊆ A and X ⊆ B, we know that X ⊆ A ∩ B by Lemma 3.9.2 that we
just proved.
Now, since X ⊆ A ∩ B, we know X ∈ P(A ∩ B), by the definition of power set.
Since X was arbitrary, we conclude that P(A) ∩ P(B) ⊆ P(A ∩ B).

Did you do what we did? Did you also cite the previous lemma? Did you
instead prove that result all over again without realizing it? Consider that a
lesson learned! One of the major benefits of proving results is that we get to
use them in later proofs! There’s nothing technically wrong with proving the
previous result again in the middle of this proof; it just might save a little bit
of time and writing to not do so. If you find yourself working on a problem and
thinking, “Hey, this feels familiar . . . ”, go back and look for related theorems
or lemmas or examples. Maybe you can use some already-acquired knowledge
to your advantage.
198 CHAPTER 3. SETS

3.9.3 Proving “=”

Double-Containment Proofs
Again, we will need to recall the definition of “=” (in the context of sets), since
we will be using it frequently here.

Definition 3.9.4. We say two sets, A and B, are equal, and write A = B, if
and only if A ⊆ B and B ⊆ A.

That’s it! It’s completely built up from a previous definition, that of “⊆”
(since that of “⊇” is completely equivalent). Thus, this isn’t really a new tech-
nique, per se, because it’s really a repeated application of a previous technique.
That is to say, to prove A = B, we just need to use the technique used in the
last subsection and prove A ⊆ B and then prove B ⊆ A.
This technique is so common, in fact, that it is given a name: double-
containment. When we prove two sets are subsets of each other, both ways,
and conclude that they are equal, we call this a double-containment proof.

Examples
Let’s see an example of this double-containment technique in action.

Lemma 3.9.5. Let A and B be any sets. Then, A − (A ∩ B) = A − B.

Intuition: As usual, we could draw a Venn diagram to convince ourselves of

this truth, but that doesn’t prove anything. Instead, we will follow a double-
containment proof. If we take an element x ∈ A − (A ∩ B), we can apply the
definition of “−” first, and then “∩”, to deduce something about x. Hopefully,
it will tell us that x ∈ A − B. Then, if we take an element y ∈ A − B, we can
apply some definitions to hopefully deduce that x ∈ A − (A ∩ B). Maybe we
aren’t sure yet exactly how to do so, but by looking at that Venn diagram and
using the definitions, we can surely figure it out. Why don’t you try to do it
first, then read our proof!

Proof. We will show A − (A ∩ B) = A − B by a double-containment proof.

(“⊆”) First, let x ∈ A − (A ∩ B) be arbitrary and fixed. By the definition of

“−”, we know that x ∈ A and x ∈/ A ∩ B. This means it is not true that x is an
element of both A and B. We already know x ∈ A, so we deduce that x ∈ / B.
Thus, x ∈ A and x ∈
/ B, so by the definition of “−”, we deduce that x ∈ A − B.
This shows A − (A ∩ B) ⊆ A − B.

(“⊇”) Second, let y ∈ A − B be arbitrary and fixed. By the definition of “−”,

this means y ∈ A and y ∈ / B. Now, since y is not an element of B, this means
that certainly y is not an element of both A and B. That is, y ∈
/ A ∩ B, by the
definition of “∩”.
Since we know y ∈ A and y ∈
/ A ∩ B, we deduce that y ∈ A − (A ∩ B). This
3.9. PROOFS INVOLVING SETS 199

shows A − B ⊆ A − (A ∩ B).

Overall, a double-containment proof has shown that A − (A ∩ B) = A − B.

Look at the overall structure of this proof. We knew there would be two
parts to it, since it is a double-containment proof, but we were also kind enough
to point this out ahead of time for our intrepid reader, and separate those two
sections appropriately. It would be technically correct to ignore this and just
dive right in to the proof, but this might leave a reader confused. The whole
point of a proof is to convince someone else of a truth that you have already
figured out, so you might as well make it as easy as possible for them to follow
what you’re doing.
Let’s see another example of proving two sets are equal. This one will be
a little different, because one of the parts of the double-containment will make
use of the complement operation. As a preview, spend a minute now thinking
about why the statements A ⊆ B and B ⊆ A are equivalent (supposing there is
some universal set A, B ⊆ U ). Draw a Venn diagram and try some examples.
Try to prove it, even!
Proposition 3.9.6.
 
8
x∈N|x+ ≤6 = {2, 3, 4}
x
8

Proof. Let’s define A = x ∈ N | x + x ≤ 6 , and B = {2, 3, 4}.
To show A = B, we will show A ⊆ B and B ⊆ A.
First, we will show B ⊆ A. We can consider each of the three elements sepa-
rately, and verify that they satisfy the defining inequality of B:
8
2+ =6≤6
2
8 17
3+ = ≤6
3 3
8
4+ =6≤6
4
Since 2, 3, 4 ∈ N, we deduce that 2 ∈ A and 3 ∈ A and 4 ∈ A, so B ⊆ A.
Next, to show A ⊆ B, we will show that B ⊆ A, where the complement is
taken in the context of N. That is, we will show that all of the natural numbers
1, 5, 6, 7, . . . are not elements of A.
To show this, we will verify that the defining inequality of A is not satisfied by
any of those elements.
The first two cases can be considered easily: 1 + 81 = 9 6≤ 6 and 5 + 58 = 33
5 6≤ 6.
To consider the other cases, we can take an arbitrary and fixed x ∈ N with
x ≥ 6. Notice, then, that x + x8 ≥ 6 + x8 > 6, since x8 > 0.
200 CHAPTER 3. SETS

This shows that only 2,3,4 satisfy the defining inequality of A.

Overall, by a double-containment argument, we have proven that A = B.

Think carefully, again, about why the method employed in the second half
of the proof is valid. (It is actually an instance of using the contrapositive of
a conditional statement, but we haven’t yet defined any of those terms; we will
do so in the next chapter on logic.)
Let’s see another example of proving set equality. This one is only slightly
different in that we are proving some set is actually the empty set and, to do
so, we will prove that it has no elements.

Proposition 3.9.7. For every n ∈ N, define Sn = N − [n]. Then

\
Sn = ∅
n∈N

We suggest you play around with this statement first, if it doesn’t make
sense. For instance, try identifying the element of the sets S1 , and S1 ∩ S2 , and
S1 ∩ S2 ∩ S3 , and so on. Try to come up with a candidate element of the big
intersection on the left, and then figure out why it actually is not an element
of that set. After that, try to figure out a formal proof and write it out; then,
look at ours below!
\
Proof. Let’s define T = Sn , so we can refer to it later.
n∈N

To prove T = ∅, we will show that T does not contain any elements. Notice
that T is formed by an intersection of many sets of natural numbers, so it’s clear
that the only possibilities for elements of T are natural numbers.
Consider an arbitrary and fixed x ∈ N. We want to show that x ∈
/ T.
Observe that x ∈ [x] = {1, 2, . . . , x}. Thus, x ∈
/ N − [x], by the definition of
“−”.
By definition, T contains the elements that belong to all of the sets of the form
N − [n]. We have identified (at least) one set, N − [x], of the intersection such
that x does not belong to that set. Accordingly, x cannot be an element of T ,
since it does not belong to all such sets. Therefore, x ∈
/ T.
Since x ∈ N was arbitrary, we have proven that T contains no natural numbers
as elements, and therefore it has no elements at all.

Summary: Let’s make one more statement about why this technique works.
We showed that there are no elements of T , i.e. that T ⊆ ∅. This completes
the entire process, because it is trivially true, as well, that ∅ ⊆ T . (This claim
holds for any set.) Thus, one of the parts of the double-containment argument
is already achieved, and we can conclude T = ∅.
Alright, one more example. We want to include this one because it gives
us further practice in working with indexed set operations. You will find many
3.9. PROOFS INVOLVING SETS 201

similar problems in the exercises for this section. We encourage you to work on
as many of them as you can!

Proposition 3.9.8. For every n ∈ N, define An = x ∈ R | 0 ≤ x < n1 . Then,



\
An = {0}
n∈N

Think about what this claim T means. Draw a picture of the An sets on a
number line. What does the “ ” intersection accomplish? Why does it work
out that 0 is an element
T of that intersection? Why is it the only element?
The definition of will be crucial in this proof, so let’s recall the definition
here. The key phrase is for every:

Definition 3.9.9. The intersection of a collection of sets Ai indexed by the set

I is \
Ai = {x ∈ U | x ∈ Ai for every i ∈ I}
i∈I

where we assume there is a set U such that Ai ⊆ U for every i ∈ I.

That is, remember that the indexed intersection of several sets collects to-
gether the elements that belong to all of the constituent sets. Thus, in our proof
below, you will see that we need to prove that (1) 0 is, indeed, an element of
all of the An sets, and (2) no other number is an element of all of them, i.e. for
every nonzero real number, we can identify at least one of the An sets such that
the number is not an element of that set.

Proof. First, we will prove that

\
{0} ⊆ An
n∈N

This requires us to show that 0 ∈ An for every n ∈ N.

1
Let n ∈ N be arbitrary and fixed. Notice that the inequality 0 ≤ 0 < n does,
indeed hold.
(Note: You might be worried because “in the limit” 0 is not less than every
fraction n1 “all at once”, but that is not the point! Think of it this way: Is
0 ∈ A1 ? Yes, 0 ≤ 0 < 1. Is 0 ∈ A2 ? Yes, 0 ≤ 0 < 12 . Is 0 ∈ A3 ? Yes, 0 ≤ 0 < 13 .
And so on. The inequality holds for every n ∈ N individually, so 0 is an element
of every such set. If you weren’t worried about this, never mind! Move right
along!)
\ T
Thus, 0 ∈ An for every n ∈ N, and so 0 ∈ An , by the definition of “ ”.
n∈N
\
This shows that {0} ⊆ An .
n∈N
202 CHAPTER 3. SETS

Second, we will prove that \

An ⊆ {0}
n∈N

We will do this by considering the complements of these sets, in the context of

R. Specifically, we will show that
\
{0} ⊆ An
n∈N

which means we will show that every nonzero real number is not an element of
every An .
Let x ∈ R be arbitrary and fixed, with the property that x 6= 0. Either x > 0
or x < 0, then, so let’s consider each case separately.
Case 1: Suppose x > 0. Consider the number x1 ∈ R. Since the natural numbers
are infinite and unbounded in R, we can choose a natural number M that is
bigger than that real number. That is, we can choose M ∈ N such that M > x1 .
(Note: Think about why this works. We haven’t proven that N is infinite, or
that the numbers “go on forever” along the number line of R, but we hope these
ideas make sense to you, intuitively.)
Take such an M ∈ N with M > x1 . Since x > 0, we can multiply the inequality
1 1
on both sides by x; since M > 0 (so M > 0) we can then multiply again by M .
1 1 1
This yields x > M . Accordingly x ∈
/ AM , because − M < x < M is False.
Since\x ∈
/ AM , then surely x is not an element of all such sets. Therefore,
x∈/ An .
n∈N

Case 2: Next, suppose that x < 0. We will make a similar argument as the
previous case; this time, we will just consider −x, since −x > 0. Using the same
logic as above, we can surely identify a natural number M ∈ N that satisfies
1
M > −x = − x1 . Manipulating the inequality tells us that x < − M 1
. Thus,
\
x∈/ AM , and so x ∈ / An .
n∈N

Therefore, we have shown that any x ∈ R with x 6= 0 is not an element of at

least one of the
\An sets, so any such x is not an element of their intersection.
Thus, {0} ⊆ An , and we have proven the claim by a double-containment
n∈N
argument.
This proof is harder than the other ones, we think, so make sure to read it
a couple times to make sure you see what happens in every step. In particular,
think about how we came up with the step where we chose M ∈ N that satisfies
M > x1 . Do you think we magically intuited that choice? Or do we think we
1
recognized that we wanted x < M to be true for some M , and manipulated the
inequality backwards to figure out how to make that happen?
3.9. PROOFS INVOLVING SETS 203

3.9.4 Disproving Claims

Motivating Example
Consider the following proposed claim:

For any sets F, G, H, if F ⊆ G ∪ H, then either F ⊆ G or F ⊆ H.

Is this claim True? If so, how would we prove it? Well, we’d take an arbitrary
and fixed element x ∈ F . Since F ⊆ G ∪ H, this would tell us x ∈ G ∪ H, as
well. Accordingly, either x ∈ G or x ∈ H. Is that it? Are we done with the
proof?
We hope you recognize that this does not work! In particular, we have not
satisfied the definition of “⊆” at the end. If our goal is to prove “either F ⊆ G
or F ⊆ H”, then we should conclude that one or the other of those claims holds:
that every element of F is also an element of G, or else every element of F is
also an element of H.
What we found was that every element of F is itself either an element of
G or H, but we cannot decide collectively that all elements of F are elements
of one or the other, G or H. Read through these last two paragraphs again to
make sure you follow that logical observation. It might be easy to actually write
up a “proof” for this claim and not realize that you’ve made a false step!

Identifying Errors
This recognition of an error is one of the skills we are developing, and it will be
helpful in several ways. You’ll notice that many exercises (some thus far, but
many more as we move onwards) ask you to find the flaw in a proposed “proof”
of some claim. By pointing out that there exists a flaw, we are perhaps helping
you to find it (or them, as the case may be). Reading a proposed proof for
logical, factual, and clarity errors is an essential skill. What’s more, this careful
reading of others’ writing will necessarily make you a more critical reader of
your own writing, and it will help you to catch potential errors like the one in
the preceding paragraphs. Do not worry if you didn’t catch it; now that you’ve
seen it, you’ll be on the lookout for similar mistakes in the future! Like we said,
as well, this skill is ongoing development, and by the end of this book, you will
be a great reader of mathematical proofs, as well as a great writer.

Counterexamples
So, now what do we do? We just recognized that our “proof” above did not
work. Does this mean the claim is actually False? Actually, all this means (so
far) is that our attempt at a proof failed. Maybe some other logical route will
magically take us to the elusive conclusion.
Or, maybe the claim really is False. How could we show that? Think about
the logical form of the claim: it says some statement holds true for any sets
F, G, H. It says that the assumption F ⊂ G ∪ H will always imply, necessarily,
204 CHAPTER 3. SETS

that F ⊆ G or F ⊆ H. To show that this does not always happen, we need to

find what’s called a counterexample.
We will discuss all of these ideas again in the next chapter, when we formalize
logic, but what you need to know for now is this: a counterexample is a
specific, detailed, and described example that illustrates how a statement about
“every . . . ” or “any . . . ” or “all possible . . . ” does not actually hold for every
case. A counterexample amounts to disproving a statement that a whole class
of objects has a certain property, by exhibiting one object in that class without
that property.

Example
Let’s see how the process of finding and stating a counterexample would work
for our example above.
Example 3.9.10. Recall the claim:

For any sets F, G, H, if F ⊆ G ∪ H, then either F ⊆ G or F ⊆ H.

This claim is supposed to work for any sets F, G, H, so when we describe our
counterexample, we better describe exactly what those three sets are going to
be. We can’t just explain our way around the issue and argue about how there
might exist three sets out there with a certain property. Nope, we have to tell a
reader exactly what they are by explicitly defining them. This is what the first
line of our disproof of the claim will be, but we can’t just jump right into that,
because we don’t know how to define them yet!
This is where the fun/work is: we need to play around with the desired
properties of these sets to help us come up with an example. Recall that we
want these sets to satisfy some properties: we should make sure the assumption
F ⊆ G ∪ H holds True, but we want the conclusion—that either F ⊆ G or
F ⊆ H—to be False.
What does this mean? Well, we think you’ll agree that, logically speaking,
the “opposite” or “negation” of a statement like that would be “both F 6⊆ G and
F 6⊆ H”. (This concept of logical negation will return in the next chapter;
for now, we think you can understand it by applying the logical principles that
guide your daily life. Soon, we will formalize this idea.)
We now have a specific goal: to find three sets F, G, H that satisfy all three
of the following:

F ⊆G∪H
F 6⊆ G
F 6⊆ H

One thing left to consider is what “6⊆00 means. We have a definition of “⊆”,
so what is the “opposite” or “negation” of that? For F ⊆ G to be true, we
require that every element of F is also an element of G; so, if this fails, then
we must have at least one element of F that is not an element of G. The same
3.9. PROOFS INVOLVING SETS 205

observation applies to F 6⊆ H. Now, we can restate our goals in a helpful way,

by applying definitions:

Every element of F is an element of either G or H

There is at least one element of F that is not an element of G
There is at least one element of F that is not an element of H

This will be incredibly helpful in finally finding our counterexample! We’ve

boiled down all the essential parts of the claim and have restated the properties
in a more intuitive way. The rest of the work is to just play around on some
scratch paper and see what we can come up with. One approach is to draw a sort
of “empty” Venn diagram, for F and G and H and their potential “overlaps”,
and then fill in enough elements so that the three above properties are satisfied.
The first condition requires the set F to “lie inside” both G and H, entirely;
but, the second and third conditions require the existence of two elements of F ,
one of whom is not an element of G and the other of whom is not an element of
H. That’s all we need! A simple example, you might say, but an effective one,
we say. Let’s jump in and write up our disproof now:
Proof. The following claim is False:
For any sets F, G, H, if F ⊆ G ∪ H, then either F ⊆ G or F ⊆ H.
We will disprove it with a counterexample.
Define F = {1, 2} and G = {1} and H = {2}.
Notice that G ∪ H = {1, 2}. Since F = G ∪ H, then certainly F ⊆ G ∪ H. Thus,
the hypothesis of the claim holds true.
However, notice that 2 ∈ F but 2 ∈
/ G. Thus, F 6⊆ G.
Likewise, notice that 1 ∈ F but 1 ∈
/ H. Thus, F 6⊆ H.
Therefore, the claim is False.
One important lesson from this example is the following:
Tour counterexample does not have to be the most interesting or
complicated one, nor do you somehow need to characterize all possi-
ble counterexamples. We just need to see one counterexample, and
we need to see how it works.
That’s it! This is exactly what we did in the above proof: we defined all of
the important objects (the three sets F, G, H), and then we pointed out and
described all the relevant properties they had. We did not leave it to the reader
to check that the counterexample works; we showed them the details. We did
not argue that some such sets exist somewhere out there in the universe; we
defined them explicitly.
This is important, and we expect your counterexamples to have similar proof
structure to ours above. Most of the work will go on “behind the scenes”, before
206 CHAPTER 3. SETS

the proof starts, when you try to come up with your examples. Once you have
it, though, just write it up much like we did.

3.9.5 Questions & Exercises

Remind Yourself
Answering the following questions briefly, either out loud or in writing. These
are all based on the section you just read, so if you can’t recall a specific defi-
nition or concept or example, go back and reread that part. Making sure you
can confidently answer these before moving on will help your understanding and
memory!

(1) What is the definition of ⊆? How do we use it to prove A ⊆ B?

(2) What does it mean for two sets to be equal?

(3) What is a double-containment proof?

(4) What is a counterexample?

(5) Suppose A, B, U are sets and A, B ⊆ U . Why can we prove B ⊆ A to prove

that A ⊆ B? Try to convince a friend that this is a valid technique.

Try It
Try answering the following short-answer questions. They require you to actu-
ally write something down, or describe something out loud (to a friend/class-
mate, perhaps). The goal is to get you to practice working with new concepts,
definitions, and notation. They are meant to be easy, though; making sure you
can work through them will help you!

(1) First, prove the following claim:

For any sets A, B, C, the subset relationship A − (B − C) ⊆ (A −

B) ∪ C holds.

Second, find a counterexample to the claim that those sets are actually
always equal.

(2) Suppose A, B, C are sets and A ⊆ B. Prove that A × C ⊆ B × C.

(3) Suppose A ⊆ C and B ⊆ D. Prove that A × B ⊆ C × D.

(4) Let A = {x ∈ R | x2 > 2x + 8} and B = {x ∈ R | x > 4}. For each of

the following claims, either prove it is correct or provide a counterexample
to show it is False.

(a) A ⊆ B
(b) B ⊆ A
3.10. SUMMARY 207

(5) Let A, B, U be sets with A, B ⊆ U . Prove that A − B = A ∩ B by a

double-containment argument.
(6) Let S = {x ∈ R | −2 < x < 5} and T = {x ∈ R | −4 ≤ x ≤ 3}. What is
S ∩ T , in the context of R? Identify a set and then prove it is correct, using
a double-containment argument.
(7) Prove the following claim: If A ⊆ B, then P(A) ⊆ P(B).
(8) For every n ∈ N let Sn = x ∈ R | − n1 < x < n1 . Prove that


\
Sn = {0}
n∈N

(9) Let I = {x ∈ R | 0 < x < 1}. For every x ∈ I, define Sx = {y ∈ R | x <

y < x + 1}. Prove that
[
Sx = {z ∈ R | 0 < z < 2}
x∈I

(10) For every n ∈ N, define the sets An and Bn by

n−1
 
An = x ∈ R | 0 ≤ x <
n
 
1
Bn = y ∈ R | − < y < 1
n

Prove the following set equality by a double-containment argument:

[ \
An = Bn
n∈N n∈N

3.10 Summary
This was our first foray into some abstract conepts and results. We introduced
the notion of a set, motivating it via several examples. We discussed the key
relationships of being an element and being a subset, and pointed out how
important is to distinguish the two! (Keeping the “Bag Analogy” in mind
might help you in this regard.) We also discussed some notation, included set-
builder notation. As we continue to move into more abstract mathematics, using
correct, formal notation will be more important than ever to ensure that we are
properly expressing our ideas. One key idea that came up is the notion of the
power set, which represents a place where the element and subset relationships
are both at work.
A discussion of set operations showed us how to combine sets and create new
ones. All of these operations will be used throughout the remainder of our work
in this book. We also showed how these operations can be indexed. This allows
208 CHAPTER 3. SETS

us to use shorthand to write a union of several sets using just a few definitions
and symbols. Again, these ideas will re-appear quite often throughout our work,
so we will present many exercises relating to these ideas; we encourage you to
attempt and work through as many as you can!
We saw a proof technique relating to sets: namely, double-containment
arguments. This is a fundamental proof technique in mathematics. You will
see us use it often, and you will find it appearing in other courses and studies,
as well.
A couple of discussions came up that allowed us to touch upon some profound
ideas in abstract set theory, although we couldn’t completely dive into them.
For one, Russell’s Paradox showed us that there is no “set of all sets”. For
another, we talked about how the natural numbers can be formally defined in
terms of sets. In practice, we won’t use this definition, and will continue to rely
on our intuitions about N. However, we hope it was interesting and somehow
informative to read such a discussion.

3.11 Chapter Exercises

These problems incorporate all of the material covered in this chapter, as well
as any previous material we’ve seen, and possibly some assumed mathematical
knowledge. We don’t expect you to work through all of these, of course, but
the more you work on, the more you will learn! Remember that you can’t truly
learn mathematics without doing mathematics. Get your hands dirty working
on a problem. Read a few statements and walk around thinking about them.
Try to write a proof and show it to a friend, and see if they’re convinced. Keep
practicing your ability to take your thoughts and write them out in a clear,
precise, and logical way. Write a proof and then edit it, to make it better. Most
of all, just keep doing mathematics!
Short-answer problems, that only require an explanation or stated answer with-
out a rigorous proof, have been marked with a .
Particularly challenging problems have been marked with a ?.

Problem 3.11.1.  For each of the following statements about elements and
subsets, state whether it is True or False. Be prepared to defend your choice to
a skeptical friend!
Throughout this problem, we will use the following definitions:

A = {x ∈ Z | −3 ≤ x ≤ 3}
B = {y ∈ Z | −5 < y < 6}
C = x ∈ R | x2 ≥ 9


D = {x ∈ R | x < −3}
E = {n ∈ N | n is even }
3.11. CHAPTER EXERCISES 209

(a) A ⊆ B

(b) C ∩ D = ∅

(c) 4 ∈ E ∩ B

(d) {4} ⊆ A ∩ E

(e) 10 ∈ C − D

(f) A ∪ B ⊇ C

(g) 3 ∈ A ∩ C

(h) 0 ∈ (A − B) ∪ D

(i) E ∩ C ⊆ Z

(j) 0 ∈
/ B−C

Problem 3.11.2.  Let m, n ∈ N. Suppose m ≤ n. Explain why P([m]) ⊆

P([n]).
Problem 3.11.3. Look back at Problem 7 in Section 3.9. We proved that when-
ever two sets satisfy A ⊆ B, then they must also satisfy P(A) ⊆ P(B). Read
through that proof, too, to remind yourself of the details.
Now, does this claim “work the other way”? That is, suppose P(A) ⊆ P(B).
Can you prove that A ⊆ B is also true? Or can you find an example where this
is not true?
Problem 3.11.4. Rewrite the following sentences using the “set-builder notation”
to define a set. Then, if possible, write out all the elements of the set, using set
braces; if not possible, explain why not and write out three example elements
of the set.

(a) Let A be the set of all natural numbers whose squares are less than 39.

(b) Let B be the set of all real numbers that are roots of the equation x2 − 3x −
10 = 0.

(c) Let C be the set of pairs of integers whose sum is non-negative.

(d) Let D be the set of pairs of real numbers whose first coordinate is positive
and whose second coordinate is negative and whose sum is positive.

Problem 3.11.5. Define the following sets:

A = x ∈ R | x2 − x − 12 > 0


B = {y ∈ R | −3 < y < 4}

Prove that A = B.
210 CHAPTER 3. SETS

Problem 3.11.6. Let X be the set of students at your school.

Identify a property P (x) such that A := {x ∈ X | P (x)} is a proper subset of
X and A 6= ∅.
Then, identify a property Q(x) such that B := {x ∈ X | Q(x)} is a proper
subset of A (i.e. B ⊂ A) and B 6= ∅.
Problem 3.11.7. Let A, B, and C be sets with A ⊆ C and B ⊆ C.
(a) Draw a Venn diagram for the sets A ∩ B and (A ∩ B).
(b) Prove that A ∩ B ⊆ (A ∩ B).
(c) Define specific sets A, B, C such that the containment is strict, i.e. A ∩ B ⊂
(A ∩ B)
(d) Define specific sets A, B, C such that A ∩ B = (A ∩ B).
Problem 3.11.8. Let S = (m, n) ∈ Z × Z | m = n2 . How does S compare to


the set T = {(m, n) ∈ Z × Z | n = m2 }? If one is a subset of the other, prove

it. If not, provide examples to show this.
Problem 3.11.9. Let (a, b) be a point on the Cartesian plane, i.e. (a, b) ∈ R × R.
Let ε (the Greek letter epsilon) be a nonnegative real number, i.e. ε ∈ R and
ε ≥ 0.
Let C(a,b),ε be the set of real numbers that are “close” to (a, b), defined as
follows: n p o
C(a,b),ε = (x, y) ∈ R × R | (x − a)2 + (y − b)2 < ε

1. Come up with a geometric description of the set C(a,b),ε .

What happens to the set as we change a and b?
What happens as we change ε?
2. What is C(0,0),1 ∩ C(0,0),2 ?
3. What is C(0,0),1 ∪ C(0,0),2 ?
4. What is C(0,0),1 ∩ C(2,2),1 ?
Problem 3.11.10. Consider the (false!) claim that
[
P([n]) = P(N)
n∈N

(a) What is wrong with the following “proof” of the claim? Point out any
error(s) and explain why it/they ruin the “proof”.
First, we will show that
[
P([n]) ⊆ P(N)
n∈N
3.11. CHAPTER EXERCISES 211

Consider an arbitrary element X of the union on the left.

By the definition of an indexed union, we know there exists some
k ∈ N such that X ⊆ [k].
Since [k] ⊆ N, and X ⊆ [k], we deduce that X ⊆ N.
Thus, X ∈ P(N).
Second, we will prove the “⊆” relationship holds in the other
direction, as well.
Consider an arbitrary Y ⊆ N.
By the definition of subset, and the fact that Y is a set of natural
numbers, we know there exists some ` ∈ N such that Y ⊆ [`].
[
By the definition of indexed union, we know Y ∈ P([n]).
n∈N

Since we have shown ⊆ and ⊇, we know the two sets are equal.
(b) Disprove the claim by defining an explicit example of a set S such that
[
S ∈ P(N) and S∈ / P([n])
n∈N

Problem 3.11.11. Let A = [3] × [4]. (Remember that [n] = {1, 2, 3, . . . , n}.)
Let B = {(x, y) ∈ Z × Z | 0 ≤ 3x − y + 1 ≤ 9}.
(a) Prove that A ⊆ B.
(b) Is it true that A = B? Why or why not? Prove your claim.
Problem 3.11.12. Let n ∈ N be a fixed natural number. Let S = [n] × [n]. Let
T be the set
T = (x, y) ∈ Z × Z | 0 ≤ nx + y − (n + 1) ≤ n2 − 1


Prove that S ⊆ T but S 6= T .

Problem 3.11.13. Suppose A and B are sets.
(a) Prove that
P(A) ∪ P(B) ⊆ P(A ∪ B)

(b) Provide an explicit example of A and B where the containment in (a) is

strict.
Problem 3.11.14. Let S and T be sets whose elements are sets, themselves.
Suppose that S ⊆ T .
Prove that [ [
X⊆ Y
X∈S Y ∈T
212 CHAPTER 3. SETS

Problem 3.11.15. Let A, B, C, D be sets.

(a) Prove that
(A × B) ∪ (C × D) ⊆ (A ∪ C) × (B ∪ D)

(b) Provide an explicit example of A, B, C, D where the containment in (a) is

strict.
Problem 3.11.16. Let A, B, C be sets. Prove that

A × (B ∩ C) = (A × B) ∩ (A × C)

and
A × (B − C) = (A × B) − (A × C)
Problem 3.11.17. Let X, Y, Z be sets. Prove that (X ∪ Y ) − Z ⊆ X ∪ (Y − Z)
but equality need not hold.
Problem 3.11.18. Find an example of a set S such that S ∈ P(N) and S contains
exactly 4 elements.
Then, find an example of a set T such that T ⊆ P(N) and T contains exactly 4
elements.
Problem 3.11.19. Find examples of sets R, S, T such that R ∈ S and S ∈ T and
R ⊆ T but R ∈/ T.
Problem 3.11.20. Identify what each of the following sets are, and prove your
claims. \ [
[n] and [n]
n∈N n∈N

Problem 3.11.21. Let I = {−1, 0, 1}. For each i ∈ I, define Ai = {i − 2, i −

1, i, i + 1, i + 2} and Bi = {−2i, −i, i, 2i}.
[
(a) Write out the elements of Ai .
i∈I
\
(b) Write out the elements of Ai .
i∈I
[
(c) Write out the elements of Bi .
i∈I
\
(d) Write out the elements of Bi .
i∈I
! !
[ [
(e) Use your answers above to write out the elements of Ai − Bi .
i∈I i∈I
! !
\ \
(f) Use your answers above to write out the elements of Ai − Bi .
i∈I i∈I
3.12. LOOKAHEAD 213
[
(g) Write out the elements of (Ai − Bi ). How does this compare to your
i∈I
answer in (e)?
\
(h) Write out the elements of (Ai − Bi ). How does this compare to your
i∈I
answer in (f)?

Problem 3.11.22. In this problem, we are going to “prove” the existence of the
negative integers! We say “prove” because we won’t really understand what
we’ve done until later but, trust us, it’s what we’re doing.
Because of this goal, you cannot assume any integers strictly less than 0 exist,
so your algebraic steps, especially in part (d), should not involve any terms that
might be negative.
That is, if you consider an equation like

x+y =x+z

we can deduce that y = z, by subtracting x from both sides, since x − x = 0.

However, if we consider an equation like

x+y =z+w

we cannot deduce that x − z = w − y. Perhaps y > w, so w − y does not exist

in our context . . .

Let P = N × N. Define the set R by

R = {((a, b), (c, d)) ∈ P × P | a + d = b + c}

(a) Find three different pairs (c, d) such that ((1, 4), (c, d)) ∈ R.

(b) Let (a, b) ∈ P . Prove that ((a, b), (a, b)) ∈ R.

(c) Let ((a, b), (c, d)) ∈ R. Prove that ((c, d), (a, b)) ∈ R, as well.

(d) Assume ((a, b), (c, d)) ∈ R and ((c, d), (e, f )) ∈ R.
Prove that ((a, b), (e, f )) ∈ R, as well.

3.12 Lookahead
Now that we’ve introduced sets, defined them, seen many examples, and talked
about operations and how to manipulate sets, it’s time to move on to logic.
We’ve already previewed some important logical ideas, specifically in Section
3.9 on how to write proofs about sets. In the next Chapter, we will make all
of these logical ideas more formal, explicit and rigorous. We will develop some
214 CHAPTER 3. SETS

notation and grammar that will help us express logical ideas more precisely and
concisely. We will use these to express our mathematical thoughts in a common
language and communicate our ideas with others. In short, we will be able to
confidently talk and write about mathematics!
Chapter 4

Logic: The Mathematical

Language

4.1 Introduction
We are moving on to learn about the language of mathematics! We will learn
how to express our ideas formally and precisely and concisely. This will require
learning some new terminology and notation, all the while thinking and writing
in a more formal way. Ultimately, this will allow us to solve problems and write
good, clear, and correct mathematical proofs.

4.1.1 Objectives
The following short sections in this introduction will show you how this chapter
fits into the scheme of the book. They will describe how our previous work
will be helpful, they will motivate why we would care to investigate the topics
that appear in this chapter, and they will tell you our goals and what you
should keep in mind while reading along to achieve those goals. Right now,
we will summarize the main objectives of this chapter for you via a series of
statements. These describe the skills and knowledge you should have gained by
the conclusion of this chapter. The following sections will reiterate these ideas
in more detail, but this will provide you with a brief list for future reference.
When you finish working through this chapter, return to this list and see if you
understand all of these objectives. Do you see why we outlined them here as
being important? Can you define all the terminology we use? Can you apply
the techniques we describe?

By the end of this chapter, you should be able to . . .

• Define variable propositions using proper notation.

215
216 CHAPTER 4. LOGIC

• Define mathematical statements using quantifiers and other proper nota-

tion, and characterize sentences that are not proper mathematical state-
ments.

• Understand and explain the difference between two types of quantifiers,

as well as how they are used.

• Define and understand the meanings of several logical connectives, as well

as use them to define more complicated mathematical statements.

• Apply proof techniques to create formal arguments that demonstrate the

truth of a mathematical statement.

• Compare and contrast different types of proof techniques, as well as un-

derstand when and how to use them, depending on the situation.

4.1.2 Segue from previous chapter

We introduced sets to have some standard, fundamental mathematical objects
to work with. You probably noticed, though, that we used a lot of phrases like
“If . . . , then . . . ” and “for every” and “there exists” and “for at least one”
and “and” and “or” and True and False and so on and so forth . . . We relied on
your intuitive understanding, and previous knowledge, of these concepts. As a
living, breathing human being who converses with others, you have some kind of
understanding of what logic is. Our goal now is to build upon those intuitions,
and help you learn to read and write and speak mathematics.

4.1.3 Motivation
In mathematics, we are interested in identifying True claims and subsequently
explaining to others how and why we know those claims are True. Thus far, we
have already presumed some familiarity with logical terminology and truth. For
instance, look back at the assumptions of the PMI (Principle of Mathematical
Induction, Theorem 3.8.5). We needed to know that if P (k) is true then P (k+1)
is true. What does this mean? What does this say about how the statements
P (k) and P (k + 1) are connected? What does it even mean for something to be
True?!?!
Our goals for this section are many, but the major emphasis is on defining
and identifying what types of statements in mathematics are meaningful and
interesting. Once we do that, we can figure out how to express those statements
in concise and precise terms. Ultimately, we will learn how to apply general
techniques to prove that those statements are True (or False, as the case may
be).

4.1.4 Goals and Warnings for the Reader

Keep this in mind throughout this chapter:
4.2. MATHEMATICAL STATEMENTS 217

You are learning a language.

Some of this material will seem difficult, some will seem boring, and some might
seem both! But this is all essential.
Have you ever learned another language? Think back to a foreign language
class you took in school, perhaps. How did you start? We bet that you didn’t
jump right in and try to write beautiful poetry. You learned the basic grammar
and syntax and vocabulary. You learned important articles, like “the” and
“an”. You learned basic verbs like “to be” and “to have” and how to conjugate
them. You learned some common nouns, like “apple” and “dog” and “friend”.
From there, you started to piece phrases together and, over time, you learned
to create more complex sentences from all of the tools you developed. All along,
you probably had some great ideas for wonderful sentences in mind, but you
just didn’t know how to express them in your new language until you learned
the necessary words and grammar.
We will be doing exactly this in the current chapter, but here our language
is mathematics. You might have some great ideas for wonderful mathematical
sentences in mind, but you just aren’t sure how to express them. Interestingly
enough, as well, we’ve already been “speaking” a lot of mathematics with each
other! We’ve solved some puzzles, we introduced a proof technique (induction),
and we worked with some building blocks of mathematical objects (sets); all
along, we’ve made sure that we’ve understood each other, being verbose with
our writing and explaining lots of details. In a way, we’ve been communicating
without setting down a common language. This is a lot like how you’d survive
in a foreign country without knowing the language: you’d use a lot of hand
gestures and charades, you’d try to listen to others speak and pick up some key
words, you’d draw pictures and make noises, and so on. This is all well and
good if you’re just on a week-long vacation, say, but if you’re going to live there,
you’ll have a lot more work ahead of you.
This is precisely the situation we face now. We’d like to inhabit this world
of mathematics, so we need to settle down and really learn the language of its
people. Once we get through that, we’ll feel more at home, like native speakers.
Then, we can maybe be a little less careful with our words and grammar, use
some slang or abbreviations or common idioms. (Think of some English exam-
ples of phrases and sentences that technically make no sense, grammatically or
in terms of vocabulary, but are still understood by your fellow speakers.) But
only then can we do so.
In the meantime, we will be much more formal and pedantic with our lan-
guage. If we don’t force ourselves to do this now, we won’t truly all speak the
same mathematical language.

4.2 Mathematical Statements

Our first step is to discuss what types of sentences are even reasonable to con-
sider as mathematical truths that need to be proven or disproven. Completing
this step is actually quite difficult! Many authors tend to gloss over this subject
218 CHAPTER 4. LOGIC

or offer a simple definition that ignores the many subtleties of mathematical

language and logic. We feel tied, as well, because the time and space provided
in this book/course are not sufficient to properly study the field of abstract
logical theory. We encourage you to investigate some books or websites that
contain relevant information. For the current context, we will have to sweep
many details under the rug, so to speak. Suffice it to say, though, there is a
very deep, rich, and fruitful field of mathematical research concerning exactly
what we will be discussing here in a more heuristic way.
Remember that we mentioned we will have to assume the existence of the
real numbers R and their usual arithmetic properties. Likewise, we will assume
many of the results and concepts of mathematical logic, often without even
realizing it (until we point it out for you). These details can be studied more
in-depth later on in your mathematical careers.

4.2.1 Definition
For now, let us discuss what we mean by a mathematical statement. We want
this term to encapsulate the kinds of “things” that we can prove or disprove.
Mathematics is unique among the sciences in that the results of this field are
proven rigorously, and not hypothesized and then “confirmed” via laboratory
experiments or real-world observations. In mathematics, we assume a set of
common axioms and then follow rigorous logical inferences to deduce truths
from these axioms (and from other truths we have proven thus far). If we
encounter a falsity, we would have to show or demonstrate that it is, indeed,
False.
With these ideas in mind, we can now consider and agree on several examples
of what such a mathematical statement or proposition could be. (We have
even proven a few already!) For instance, the sentence
For any real numbers x, y ∈ R, the inequality 2xy ≤ x2 + y 2 holds.
is a valid mathematical statement. In fact, it is True, and we will prove it later
on in Section 4.9.2. (It is sometimes referred to as the AGM Inequality, short
for the Arithmetic-Geometric Mean Inequality.) We should point out that the
word “holds” is often used in mathematics to mean “holds true” or “is a true
statement”.
Here’s another example of a mathematical statement:
For any sets S, T, U , if S ∩ T ⊆ U then S ⊆ U or T ⊆ U .
This statement, however, is False, as the following counterexample demon-
strates:
Let S = {1, 2, 3} and T = {2, 3, 4} and U = {2, 3, 5}.
Observe that S ∩ T = {2, 3} ⊆ U but S 6⊆ U and T 6⊆ U .
Why does this example disprove the statement? Do you understand? Can you
explain it? We will discuss that in more detail later on in this chapter, but we
4.2. MATHEMATICAL STATEMENTS 219

hope that, for now, we all somehow recognize that this example accomplishes
exactly that.
We can also agree that a sentence like
Why do we have class at 9:00 am?!
is definitely not a mathematical statement. It’s a perfectly valid English sen-
tence, but it is not meaningful, mathematically speaking: we can’t prove or
disprove it.
Likewise, the sentence
x2 − 1 = 0
is not a mathematical statement, despite being composed entirely of mathemat-
ical symbols. The problem is that we cannot verify whether it is True or False
purely from axioms and logical inferences. This statement depends on x, what-
ever that value is (i.e. x is a variable) and without imposing extra assumptions
about it, we cannot declare whether this sentence is True or False. This type of
sentence will be referred to later as a variable proposition: its truth depends
on a variable inside the sentence.
All of these observations and examples/non-examples motivate the following
definition:
Definition 4.2.1. A mathematical statement (or proposition or logical
statement) is a grammatically correct sentence (or string of sentences), com-
posed of English words/symbols and mathematical symbols that has exactly one
truth value, either True or False.

4.2.2 Examples and Non-examples

By “grammatically-correct” we mean that the words and symbols contained in
the sentence are used and combined correctly and make sense. This eliminates
strings of symbols/words that are nonsensical when placed together, like
1+ = 2 , Brendan2 = 1 , {{∅}} − 7 > 5π , You am smart
For instance, the third one above is not a mathematical statement because
{{∅}} is not a number, so we don’t know how to interpret “subtracting 7” from
that set.
By “has exactly one truth value”, we mean that the statement should be
either True or False, but certainly cannot be both or neither or something else
in between. This eliminates the “x2 − 1 = 0” example above, because it has no
truth value. (Without a declaration of what x is, we cannot decide, either way.)

Not Knowing the Truth Value

One strange/interesting/complicated aspect of our definition is that we might
not know the truth value for a given statement, even though we can be sure
that there is only one such value. By way of illustration, consider the following
statement:
220 CHAPTER 4. LOGIC

Any even natural number greater than or equal to 4 can be written

as the sum of two prime numbers.
Is this statement True or False? If you have a proof or disproof, then the
world of mathematics would love to see it! The statement above is known as the
Goldbach Conjecture and it is a very famous unsolved (for now, we hope!)
problem in mathematics. Nobody knows yet whether the claim is True or False,
but it is certainly the case that only one of those truth values applies. That is,
this statement cannot be both True and False, nor can it somehow be somewhere
in between. Either all even natural numbers greater than or equal to 4 do have
the stated property, or there is at least one that does not have the property. We
can state this “either/or” property even without yet knowing which of the two
possibilities is the correct one. As such, this sentence does actually satisfy our
definition of mathematical statement.
(Terminology note: In general, a conjecture is a claim that someone believes
to be true but has not yet been proven/disproven.)

Paradoxical Sentences
One way to have a sentence that does not have a truth value is to create a
paradox. Consider this sentence:
This sentence is False.
Pretty weird, right? The sentence itself is asserting something about its own
truth value. Let’s try to analyze what truth value it has:
• Let’s say the sentence is True. Then, the sentence itself tells us that it is,
in fact, False.
• Let’s say the sentence is False. Similarly, then, the sentence tells us that
it is, in fact, True.
This cannot work! This sentence is somehow both True and False at the same
time, or somehow neither, or . . . Whatever it is, it’s a bad idea. We do not want
to deal with strangeness like this in mathematics, so our definition disallows this
sentence as a mathematical statement.
(Question: What happens if you do allow sentences like this to be proper math-
ematical statements? What if you don’t adhere to the principle that every
sentence we care about must be either True or False? Think about it! Is this
somehow wrong, or is it just a different mathematical universe? . . . )
In general, self-referential sentences like the one above (that is, sentences
that make reference to themselves) are quite bizarre and can produce some
paradoxes that we want to disallow.
A variant of the above paradoxical claim is given in a cartoon drawing,
wherein Pinocchio says, “My nose will grow now!” Does it then grow? If he’s
telling the truth, then it will grow, but that only happens when he’s lying! If
4.2. MATHEMATICAL STATEMENTS 221

he’s lying, then his nose will grow (by definition), but then his statement is
actually true! Yikes!

Source: http://www.the-drone.com/magazine/wp-content/uploads/2010/04/BLA6.jpg

An even stranger example of this phenomenon is Quine’s Paradox:

“Yields falsehood when preceded by its quotation” yields falsehood

when preceded by its quotation.

We’ll let you think about that one on your own. Suffice it to say that paradoxical
claims like this are too ill-behaved for us to worry about. This is why our
definition outlaws them.

4.2.3 Variable Propositions

Other examples of sentences that are not mathematical statements are sentences
that involve unquantified variables. For instance, take the sentence

“ x2 − 1 = 0 ”

This is certainly grammatically correct and we can make sense of it, but what
is its truth value? We don’t know! If x = 1, then the sentence is True, but if
x = 8, it is False, and if x = N or x = Brendan, then the sentence doesn’t even
make sense! As such, we want to disallow sentences like this, as well. These
types of sentences are useful and common, though; we will call them variable
propositions because they make a claim that depends on some variable.
In the case of the above sentence, we might define P (x) to be the variable
proposition “x2 − 1 = 0”. We would usually write this declaration as

Let P (x) be the statement “x2 − 1 = 0”.

It is common to use capital letters to denote variable propositions and math-

ematical statements, and lowercase letters to denote the variables contained
therein. (This is not a requirement, though, merely a common convention.)
222 CHAPTER 4. LOGIC

With this variable proposition now defined, we can create proper mathemat-
ical statements by assigning particular values to the variable x in the expression.
We can say that P (1) is True and P (0) is False. We can also make quantified
claims about P (x). For instance, we claim that the following sentence is a True
mathematical statement:
There exists an x ∈ R such that the proposition P (x) is True.
whereas the following sentence is a False mathematical statement:
For every x ∈ R, the proposition P (x) is True.
Think about why these statements have the truth values we have claimed. Can
you see why they are mathematical statements to begin with? How would you
prove these claims?

Defining Variable Propositions

Notice the format we used to define variable propositions, like the one above:
(1) We give the proposition a letter name (like P ); (2) we indicate its dependence
on some number of variables, each of which has a letter (like x and y); (3) we
put quotes around the actual proposition itself; and (4) we don’t include any
new letters that have no meaning within the context of the proposition.
This format has been chosen carefully because it is precise and unambiguous.
It assigns a meaning to every letter in the proposition and clearly distinguishes
what is and isn’t part of the proposition.
For example, the following are BAD “definitions” of variable propositions.
We will give you some reasons as to why they are bad and provide some proper
amendments of the propositions.
• Let Q(y) be the proposition “x < 0”.
Reason: What is x? Where is y? We have no idea what x is, inside the
context of the proposition, so this is a poor definition.
If we had said
Let Q(x) be the statement “x < 0”.
that would have been perfect. The variable inside the parentheses is the
one that appears in the statement in quotes later. Great.
• Let P (x) be the proposition x2 ≥ 0, for every x ∈ R.
Reason: Does the writer of this sentence want to assert that x2 ≥ 0, no
matter what x ∈ R is? Is that phrase “for every x ∈ R” meant to be part
of the proposition, or not?
If we interpret this to mean that P (x) is defined as “x2 ≥ 0”, and this
definition is made for every x ∈ R, then . . . okay, that might be reasonable.
However, if we interpret this to mean that P (x) is defined as “x2 ≥ 0 for
every x ∈ R” then . . . well, this is certainly different. In fact, it’s not even
4.2. MATHEMATICAL STATEMENTS 223

a properly-defined proposition! The proposition P (x) should depend on

the input value x, but it shouldn’t be allowed to change or further quantify
that variable inside the proposition!
The way this proposition was originally written, there are two possible
interpretations and they are very different. Accordingly, this is a poor
definition.
If we had said,

Let P (x) be the statement “x2 ≥ 0”, defined for every x ∈ R.

that would have been fine. As we mention below, as well, we don’t techni-
cally have to tell the reader which values of x we want the proposition de-
fined for. Perhaps this is just some helpful information to include, though,
so it doesn’t hurt.

• Let T (x, y) = “x2 − 7 = y”.

Reason: Ugh! What does “=” mean in this context? That symbol applies
when we wish to compare two numbers and say they are equal in value (or
two sets and say they are equal in terms of their elements). The object
T (x, y) is meant to be a mathematical statement, something that is either
True or False. Thus, it does not have a numerical value to compare with
anything else.
Likewise, given a value for x and y, the statement “x2 − 7 = y” is either
True or False, so it makes no sense to say that equation “equals” something
else. It has a truth value, not a numerical value.
If we had said,

Let T (x, y) be “x2 − 7 = y”.

that would have been perfect.

Okay, that’s enough non-examples for now. We don’t want to put any bad
ideas in your head, really! However, from past experience, we know that these
are common ways for students to write propositions (either accidentally, or
without realizing why they’re wrong), so we felt compelled to share.
One final note on variable propositions. It is not essential to say where the
variables come from when defining a proposition. That can be filled in later
when the proposition is invoked, or when a specific value of a variable is used
or quantified. That is, we can make the definition

Let T (x, y) be “x2 − 7 = y”.

without needing to specify whether x and y are natural numbers, or integers,

or complex numbers, or anything like that. Later on, we can say that T (3, 2) is
True and T (π, −1) is False, and that T (∅, N) has no meaning, but we wouldn’t
need to somehow anticipate any of those interpretations when defining T (x, y).
224 CHAPTER 4. LOGIC

4.2.4 Word Order Matters!

The notion of quantifying variables is something we will discuss in detail in
the very next section. For now, we want to consider one more striking example
of a mathematical statement that illustrates the importance of word order
in sentences. Analyzing the structure of sentences like the following will be a
major goal of the following section, as well.
There is a real number y such that y = x3 for every real number x.
What does this claim say? It says we can find a number y ∈ R such that
y = x3 is True, no matter what x ∈ R is. This is ridiculous! How can there be
a single number that is the cube of all numbers? This sentence is, indeed, a
mathematical statement, but it is decidedly False. But what about the following
claim?
For every real number x there is a real number y such that y = x3 .
This one is True! Do you see the difference between the two sentences? They
contain exactly the same words and symbols, but in a different order. Whereas
the former sentence asserts that there is some number that is the cube of every
real number (which is False), the latter asserts that every real number has a cube
root, which is True. This example emphasizes the importance of word order.

4.2.5 Questions & Exercises

Remind Yourself
Answering the following questions briefly, either out loud or in writing. These
are all based on the section you just read, so if you can’t recall a specific defi-
nition or concept or example, go back and reread that part. Making sure you
can confidently answer these before moving on will help your understanding and
memory!
(1) What are the important, defining properties of a mathematical statement?
(2) What is the difference between a mathematical statement and a variable
proposition?
(3) Why is the Goldbach Conjecture a mathematical statement?
(4) What is wrong with the following attempt at defining a variable proposition?
Let Q(x, y, z) be 7x − 5y + z

Try It
Try answering the following short-answer questions. They require you to actu-
ally write something down, or describe something out loud (to a friend/class-
mate, perhaps). The goal is to get you to practice working with new concepts,
definitions, and notation. They are meant to be easy, though; making sure you
can work through them will help you!
4.2. MATHEMATICAL STATEMENTS 225

(1) For each of the following sentences, decide whether it is a mathematical

statement or not. If it is, decide whether it is True or False. If it is not,
explain why.

(a) 142857 · 5 = 714285

X n(n + 1)
(b) For every n ∈ N, k= .
2
k∈[n]

(c) For any sets A and B, if A ⊆ B, then B ⊆ A.

(d) For any sets A and B, if A ⊆ B, then P(A) ⊆ P(B).
(e) Math is cool.
(f) 1 + 2 = 0
(g) For any x, y ∈ Z, if x · y is even, then x and y are both even.
(h) For any x, y ∈ Z, if x and y are both even, then x · y is even.
(i) 1+ = 2
(j) −5 + Z ≥ π
(k) x = 7
(l) This sentence is not True.

(2) Look back through the first three chapters and identify some examples and
non-examples of mathematical statements.
Can you also find any variable propositions? Are they written in the way
we specified in this section? Can you amend them so that they are written
properly?

(3) Write a proper definition of a variable proposition that is true when two
inputted values have a non-negative sum.
Then, find two instances each of when the proposition is True and when it
is False.

(4) Let S be the set {1, 2, 3, 6, 8, 10}.

(a) Write a proper definition of a variable proposition that inputs two vari-
ables and decides whether the absolute value of their difference is an
element of S. Then, find two instances each of when the proposition is
True and when it is False.
(b) Write a proper definition of a variable proposition that inputs two sets
and determines whether their intersection is a subet of S. Then, find
two instances each of when the proposition is True and when it is False.

(Note: Given any set X and any object x, it must be that either x ∈ X or
x∈/ X.)
226 CHAPTER 4. LOGIC

(5) Come up with another mathematical statement that is True but becomes
False when we switch the order of some words. (See the example in the last
subsection for some inspiration.)

(6) For each of the following attempts at defining a variable proposition, deter-
mine if it is correct or not. Note: This does not mean determine if it is True
or False; rather, we want to know whether the statement is well-written and
sensible.
If an attempt is incorrect for some reason, explain that reason and write a
new statement that fixes that error.

(a) Let P (x) be “x > 1”.

(b) Let Q(x) be the proposition “x2 − 1 > 0”.
(c) Let R(a, b) be “a3 = b”, for every a, b ∈ R.
(d) Let P (x) be x > 1.
(e) Let T (z) = “z is prime”.
(f) Let Q(x) be the proposition “x2 − 1 > 0”, for every x ∈ R.
(g) For every x ∈ R, let Q(x) be “x2 − 1 > 0”.
(h) Let S(a) be “b2 > 4”.
(i) Let Q(x) be x2 − 1 > 0 for every x ∈ R.

4.3 Quantifiers: Existential and Universal

We will now introduce some convenient notation that allows us to shorten some
statements we have seen so far and express wordy, language-based phrases with
mathematical symbols. Another benefit of the forthcoming notation is that
we will be able to more easily express and analyze mathematical statements.
Specifically, we will now introduce the symbols “ ∀ ” and “ ∃ ”.

Definition 4.3.1. The symbol “ ∀ ” stands for the phrase “for all”.
The symbol “ ∃ ” stands for the phrase “there exists”.
We call “ ∀ ” the universal quantifier and “ ∃ ” the existential quantifier.
A mathematical statement beginning with “ ∀ ” is said to be “universally quan-
tified”, and one beginning with “ ∃ ” is said to be “existentially quantified”.

4.3.1 Usage and notation

Other common phrases that “ ∀ ” replaces are “for every” and “for arbitrary”
or “whenever” and “given any” and even “if”.
Other common phrases that “ ∃ ” replaces are “for some” and “there is at least
one” and “there is” and even “some”.
4.3. QUANTIFIERS: EXISTENTIAL AND UNIVERSAL 227

Example 4.3.2. Let’s consider some simple examples first, to get our feet wet.
In each case, we are looking to express a mathematical thought using these
symbols, or trying to interpret a quantified statement in a more “wordy” way.

• “Every real number squared is non-negative.”

This is a straightforward statement that is, in fact, True. We would write
it as:
.
∀x ∈ R x2 ≥ 0
The “big dot” separates the quantified part of the statement from the claim
made about the variable x (which was introduced in the quantification).
Another way to write this would be:

.
Define S(x) to be “x2 ≥ 0”. Then the claim is: ∀x ∈ R S(x).

• “There is a subset of N that has 7 as an element.”

This is an existence claim. It asserts that we can find an object with a
particular property. We would write it as:

∃S ∈ P(N) 7 ∈ S.
Remember that P(N) is the power set of N, the set of all subsets of N;
thus, saying S ∈ P(N) means S ⊆ N, as desired.

• “Every integer has an additive inverse (i.e. a number that, when added
to the original number, yields 0).”
This idea of an “additive inverse” is a general concept that applies to some
mathematical objects known as rings and fields. We won’t discuss those
objects in this book, but you will touch on them in any course on abstract
algebra.
We would write this claim as

. .
∀a ∈ Z ∃b ∈ Z a + b = 0

and we would read this aloud as

For any integer a, there exists an integer b such that a + b = 0.

or perhaps

No matter what integer a we are given, we can find an integer

b with the property that a + b = 0.

Again, we could shorten the notation slightly by defining I(a, b) to be

“a + b = 0”, and then writing the claim as

.
∀a ∈ Z ∃b ∈ Z I(a, b) .
228 CHAPTER 4. LOGIC

Example 4.3.3. Here are some examples of proper usage of “∀”, and some equiv-
alent formulations of how to use this symbol.
.
• ∀x ∈ R x2 ≥ 0
• For all real numbers x, we have x2 ≥ 0.
• Every real number x satisfies x2 ≥ 0.
• Whenever x is a real number, we know x2 ≥ 0.
Likewise, here are some examples of proper usage, and equivalent formulations,
of the symbol “∃”.
.
• ∃x ∈ R x2 − 4x + 4 = 0
• There exists a real number x such that x2 − 4x + 4 = 0.
• There is a real number x that satisfies x2 − 4x + 4 = 0.
• Some real number x has the property taht x2 − 4x + 4 = 0.

Reading Quantified Statements Aloud

Example 4.3.4. Now for some harder examples. Let’s look back at the phrases
we wrote at the end of the last section and express them using this new notation.
Consider this statement:
There is a real number y such that y = x3 for every real number x.
To express thi in symbolic form, we will define P (x, y) to be the proposition
“ y = x3 ” and then write the statement as

. .
∃y ∈ R ∀x ∈ R P (x, y)

This is correct, logically-speaking, but it is rather terse. For now, we will some-
times rewrite the statement using some “helping words” to aid our reading of
the statement. In particular, we would say such words when reading the sen-
tence aloud, so by occasionally writing them here, we provide you with some
extra practice in interpreting logical notation verbally. We would read the above
statement aloud as
There exists a real number y such that, for every real number x, the
statement P (x, y) holds.
The phrase “such that” is a “helper phrase” that links an existential quantifi-
cation to the rest of the phrase. The next subsection contains some important
information about when and how to use this helper phrase!
The “big dot” between the quantified parts of the statement above just
serves to separate the pieces of the statement and make it easier to read. It
corresponds to a pause or rest in speaking, like a comma, but sometimes it has
a vocalized meaning (like the “such that” after the “∃y ∈ R” part).
4.3. QUANTIFIERS: EXISTENTIAL AND UNIVERSAL 229

We don’t want to use commas, though, because we already use them for
other meanings. For example, we write
x, y ∈ S
to mean “both x and y are elements of the set S”. The “big dot” is just a
different symbol to use.
Since our mathematical careers are still young, relatively speaking, we en-
courage you to sometimes write the helping phrases like “such that” and “holds
True” to guide your understanding, whenever possible. This reminds you what
the sentences mean and helps you practice reading and writing statements like
this in such a condensed form. Remember that you are learning a language here
and you need to practice translating sentences from one language you know (En-
glish) to another (mathematics). For instance, you might want to write out the
line above as
.
∃y ∈ R such that ∀x ∈ R P (x, y) is True.
or, at least, say it this way in your head.
(By the way, when writing on a white/chalkboard or on paper, it’s common to
write “s.t.” in place of “such that”, to save a few moments of writing. That just
goes to show how ubiquitous the phrase “such that” is in mathematical writing;
we already have an agreed-upon abbreviation for it!)

4.3.2 The phrase “such that”, and the order of quantifiers

Notice that the helping phrase “such that” always follows an existential quan-
tification, and only such a quantification. This is because a claim with “ ∃ ”
asserts something about the existence of an object with a certain property,
and the rest of the sentence is the description of that special property. Thus,
“such that” makes sense and helps us read the sentence properly. Consider this
mathematical statement:
. .
∃y ∈ R ∀x ∈ R P (x, y)
What would happen if we read it out loud but misplaced the phrase “such that”
and used it after the “ ∀ ” instead of the “ ∃ ”? That would yield this sentence:
∃y ∈ R ∀x ∈ R such that P (x, y) is True.
We claim that this can be interpreted in two ways, neither of which is really the
correct intended meaning, which is why we’ve written in red!
On the one hand, one might argue that such a sentence is not grammatical
at all and has no meaning, because “such that” does not belong after a universal
quantification. This amounts to just throwing up one’s hands and saying, “I
have no idea what you meant there!”
On the other hand, one might read into the sentence a little bit and argue
that what the writer really meant was
∃y ∈ R, ∀x ∈ R, such that P (x, y) is True.
230 CHAPTER 4. LOGIC

or, writing out the words,

There exists an x ∈ R, for each y ∈ R, such that P (x, y) is True.

Here, the commas indicate an inversion of phrase order, as is common in En-

glish language. (For instance, consider the following sentence: “I laugh, at every
episode of 30 Rock, wholeheartedly.” This is the same as saying “I laugh whole-
heartedly at every episode of 30 Rock.”) This sentence would be equivalent,
then, to writing

.
∀x ∈ R ∃y ∈ R such that P (x, y) is True.

This is not the same as the original mathematical statement we considered and,
in fact, it is actually the other statement we saw in the previous section (see
Section 4.2.4), which was False! Recall that the other statement was similar but
the phrases were reversed:

There is a real number x such that, for every real number y, we have
y = x3 .

which we can symbolize as

.
∃x ∈ R ∀y ∈ R such that P (x, y) is true

Look at that! The misplacement of the phrase “such that” led to a reasonable
linguistic interpretation of the sentence that has the exact opposite meaning as
what was originally intended. Yikes! This is why we must be careful to use
“such that” always and only after an existential quantification. Remember
that we will not always write that helper phrase, so you must remember to use
it properly when reading a sentence to yourself in your head, or out loud to
others, to make sure you have the correct, intended interpretation.
The point of this example in the previous section was to point out how
important word order is. Now that we have symbols to replace some words and
phrases, we want to emphasize how important the order of those symbols is, as
well. The two mathematical statements we see above contain the exact same
words and symbols, but in a different order, and one is False whereas the other
is True. Clearly, order is extremely important!

4.3.3 “Fixed” Variables and Dependence

While we are on the topic of order of quantifiers, we will also mention the
following example to emphasize that the order of quantifiers dictates when to
consider variables as fixed in an expression.
Consider the statement “Any even natural number greater than or equal to
4 can be written as the sum of two primes.” (Recall that this is the famous
Goldbach Conjecture we discussed in the previous section.) To express this
4.3. QUANTIFIERS: EXISTENTIAL AND UNIVERSAL 231

statement logically and symbolically, we would write

Let X be the set of even natural numbers, except 2.
Let P be the set of prime numbers.
Define Q(n, a, b) to be “n = a + b”.
Then the claim is:
. .
∀n ∈ X ∃a, b ∈ P Q(n, a, b)
Notice that we used some shorthand here. A phrase like “∃a, b ∈ P ” really
means “there exists some a ∈ P and there exists some b ∈ P ”, and it would be
perfectly acceptable to express the above statement instead as
. . .
∀n ∈ X ∃a ∈ P ∃b ∈ P Q(n, a, b)
When two variables are quantified as elements from the same set, though, and
the quantifications follow one another immediately in a sentence, it is very com-
mon to combine them into one quantification. We might even see mathematical
statements like,
. .
∀x, y ∈ Z ∃a, b, c, d ∈ Z a + b + c + d = x + y and a + b 6= x and c + d 6= y
(What is this statement asserting, by the way? Is it True or False? Does it
depend on the context of Z? What if we used N or R instead in both places?)

Quantification “Fixes” a Variable

Look back at the example above, where we defined Q(n, a, b). The reason we
brough up that example was to mention that the initial quantification “∀n ∈
X” serves to fix a particular value of n that will be used for the rest of the
statement. After that, the assertion that “∃a, b ∈ P ” with the subsequent
property Q(n, a, b) depends on that fixed, but arbitrary, value of n.
The statement, as a whole, is saying that no matter what n is chosen, we
can find values a, b that satisfy the property Q. (Notice that those values of a, b
might depend on n, of course.) However, the order of quantification is telling
us that those values a, b might depend on the chosen n. This is what we want
to emphasize.
As an example, consider a particular value of the variable n in the statement.
We know 8 ∈ X because 8 is even and 8 ≥ 4. What happens when n = 8? Can
you find a, b ∈ P such that a + b = 8? Sure, we can use a = 3 and b = 5. Okay,
what about when n = 14? Can you find a, b ∈ P that satisfy a + b = 14? Surely,
your choices now have to be different than before. This is what we mean when
we say a and b depend on n. (By the way, can you find a and b in this case,
with n = 14? We can think of a couple of choices that would work!)
To make sure you’re understanding this discussion, think about the following
question and answer it: What is the difference between the statement above and
the following one?
. .
∃n ∈ X ∃a, b ∈ P Q(n, a, b)
Is this statement True or False? Why?
232 CHAPTER 4. LOGIC

4.3.4 Specifying a quantification set

Another aspect of quantifiers we want to emphasize is that we must specify a
set whenever we use quantifiers. The sentence

.
∀x x2 ≥ 0

may “look true” but it is, in fact, meaningless. What is x? Where does it
come from? “For every x . . . ” from where? What if x is not a number?
We need to specify where the object x “comes from” so that we know whether
x2 ≥ 0 is even a well-defined, grammatical phrase, let alone whether it is True.
If we amend the sentence to say

.
∀x ∈ R x2 ≥ 0

then this is a well-defined, grammatical (and True!) mathematical statement.

However, if we amend the sentence to say

.
∀x ∈ C x2 ≥ 0

then this is a well-defined yet False mathematical statement! This is because

i ∈ C but i2 = −1 < 0. (Remember, we will not make significant use of the set
of complex numbers C in this book, but it makes for interesting and enlightening
examples, like this.)
The main lesson here is that context really matters. It can change the
meaning of a statement, as well as its truth value. For this reason, we must
always be sure to specify a set from which we draw variable values.

One Exception
We sheepishly admit that there is one exception to this “always specify a quan-
tification set” rule, but there’s a good reason for the exception. Consider the
following claim:
For any sets A, B, C, the equality (A ∪ B) ∩ C = (A ∩ C) ∪ (A ∩ B)
holds true.
This is a True mathematical statement. (Can you prove it? Try using a double-
containment argument!)
How might we try to write this statement in symbolic form? This is a
universal quantification (“for any . . . ”) so we need to use a “ ∀ ” symbol. The
variables here (denoted by A and B and C) are sets. Where do they come from?
What is the set of objects we would draw them from?
We’re pretty sure you’re tempted to say “the set of all sets”. Right? That’s
a big problem, though! Remember our discussion in the previous chapter about
Russell’s Paradox? (See Section 3.3.5 to remind yourself.) That object—the
collection of all possible sets—is not, itself, a set! Thus, we cannot write this
statement symbolically as

∀A, B, C ∈ . (A ∪ B) ∩ C = (A ∩ C) ∪ (A ∩ B)
4.3. QUANTIFIERS: EXISTENTIAL AND UNIVERSAL 233

because we don’t know how to fill in the blank with a set.

Because of this issue, we will continue to use phrases like “For any sets
A, B, C . . . ”, instead of a symbolic form. When taking notes on your own, or
working out a problem on scratch paper, feel free to write “∀A, B, C” and know
that it really represents a quantification of sets. However, when writing more
formally (say, on written homework), you should stick to the phrasing used
above.

4.3.5 Questions & Exercises

Remind Yourself
Answering the following questions briefly, either out loud or in writing. These
are all based on the section you just read, so if you can’t recall a specific defi-
nition or concept or example, go back and reread that part. Making sure you
can confidently answer these before moving on will help your understanding and
memory!
(1) What is the difference between ∀ and ∃ ?
(2) How would you read the following statement out loud?

. .
∀x ∈ R ∃y ∈ R x = y 3

(3) Why is the following sentence not a proper mathematical statement?

.
∃y y + 3 > 10

What is the difference, if anything, between the following two statements?

. .
∃x ∈ N ∃y ∈ N x + y = 5

.
∃x, y ∈ N x + y = 5
Are they True or False?
(4) What is the difference, if anything, between the following two statements?

.
∃a, b ∈ Z a · b = −3

∃♥, ♦ ∈ Z. ♥ · ♦ = −3
Are they True or False?
(5) Why are the following sentences not properly quantified statements?
.
• ∃x x > 7
• ∀y ∈ Z
.
• ∀z > 2 z 2 > 4
. .
• ∀w ∈ Z w2 = t ∃t ∈ N
234 CHAPTER 4. LOGIC

Try It
Try answering the following short-answer questions. They require you to actu-
ally write something down, or describe something out loud (to a friend/class-
mate, perhaps). The goal is to get you to practice working with new concepts,
definitions, and notation. They are meant to be easy, though; making sure you
can work through them will help you!
(1) Look back at Section 4.3.3 where we expressed the Goldbach Conjecture in
symbolic notation. We defined X to be the set of all even natural numbers
except 2.
Write a definition of X using symbols, with quantifiers and set builder
notation (and perhaps set operations, depending on how you do it).
(2) Write an example of a mathematical statement that starts with a quanti-
fier, and such that the statement is True if that quantifier is “ ∃ ” but the
statement is False if that quantifier is “ ∀ ”.
(3) Write an example of a variable proposition P (x) such that
.
∀x ∈ Z P (x)
is True but
∀x ∈ N. P (x)
is False.
(4) For each of the following mathematical statements, write it in symbolic form
using quantifiers. (Be sure to properly define any variable propositions you
might need, first!) Then, determine whether the statement is True or False.
(a) There is a real number that is strictly bigger than every integer.
(b) Each integer has the property that its square is less than or equal to its
cube.
(c) Every natural number’s square root is a real number.
(d) Every subset of N has the number 3 as an element.
(5) For each of the following quantified statements, say it out loud by reading
the symbolic notation. Then, determine wheter the statement is True or
False.
. .
(a) ∀x ∈ N ∃y ∈ Z x + y < 0
. .
(b) ∃x ∈ N ∀y ∈ Z x + y < 0
.
(c) ∃A ∈ P(Z) N ⊂ A ⊂ Z
(Recall that ⊂ means “is a proper subset of”.)
(d) Let P be the set of prime numbers.
. .
∀x ∈ P ∃t ∈ Z x = 2t + 1
. . .
(e) ∀a ∈ N ∃b ∈ Z ∀c ∈ N a + b < c
. .
(f) ∃b ∈ Z ∀a, c ∈ N a + b < c
4.4. LOGICAL NEGATION OF QUANTIFIED STATEMENTS 235

4.4 Logical Negation of Quantified Statements

Let’s return to those example statements we have used before. Define P (x, y)
to be “y = x3 ” and then define Q1 to be the statement

. .
“∃y ∈ R ∀x ∈ R P (x, y)”

and Q2 to be
. .
“∀x ∈ R ∃y ∈ R P (x, y)”
Remember that Q1 is False and Q2 is True.
How is it that we know Q1 is False? It says that there is some real number
with a certain property. To declare the entire statement to be False, we might
have to verify that the property does not hold for every real number y, but
that would take a long time! The set R is infinitely large! A far more efficient
approach is to show that the negation of this statement is True.
By “negation”, we mean the logical negation, the statement that is the
“opposite” of the original statement, in the logical sense. The logical negation
of a mathematical statement has the exact opposite truth value as the original,
so if we examine Q1 ’s negation and show it is True, then we have proven that
Q1 itself is False.
How do we negate this statement, though? We already had the right idea in
mind when we noticed that we would somehow have to prove something about
every real number y, since the original statement makes a claim of existence. In
this section, we will explore how to properly negate statements like these.
We should note that there are some subtle, yet deep, mathematical concepts
underlying what we have discussed thus far. Why is it that a mathematical
statement is either True or False? Well, a cheeky (and completely correct, mind
you) response would be, “Because you defined ‘mathematical statement’ to
be that way, silly!” Yes, indeed, we did, but why did we do so? What is it about
the duality of True/False that is somehow helpful to mathematics, or essential ?
These are meaningful and difficult questions, and are definitely worth thinking
about. Discussions of these topics will necessarily delve into the philoshophy of
mathematics and human thought which are interesting and worthwhile pursuits,
certainly, but beyond the scope and goals of this book/course. We will rely on
our common, intuitive understandings of truth.

4.4.1 Negation of a universal quantification

In general, the negation of a universal (i.e. “ ∀ ”) claim is one of existence (i.e.
“ ∃ ”), and vice versa. Before we tackle the larger problem of negating any
quantified statement, let’s look at a simple case.
Assume S is a set and R(x) is a mathematical statement, defined for every
x ∈ S. The statement
.
∀x ∈ S R(x)
asserts the truth value of a variable proposition for every possible value of the
variable x from the set S. It says that no matter what element x of the set S we
236 CHAPTER 4. LOGIC

are referring to, we can necessarily conclude that the proposition R(x) is true.
Now, how could this statement be False, and how could we prove that?
If it’s False that every element x ∈ S satisfies a certain property, it must
be that at least one element does not satisfy that property. To prove this,
we would be expected to produce such a value; we would have to define (i.e.
identify) an element x and explain why R(x) does not hold for that particular
element. (Think about how we understand this negation linguistically. We do
this all the time in everyday language without even thinking about it.) The
conclusion, then, is that the negation of the original statement is

∃x ∈ S such that R(x) is False

We now introduce the notational symbol ¬ to mean “logical negation” or

“not”. With this in hand we can rewrite the negated statement

¬ ∀x ∈ S R(x) .

as
.
∃x ∈ S ¬R(x)
The concluding phrase of that statement, ¬R(x), could be simplified, de-
pending on what R(x) is.
For instance, if S = R and R(x) is “x2 ≥ 0”, then the negated statement would
read
“ ∃x ∈ R such that x2 < 0 ”
since “x2 < 0” is logically equivalent to “¬(x2 ≥ 0)”.
In general, though, we must leave it as “¬R(x)” without knowing anything
further about the proposition R. We will also point out that, in general, the
phrases “R(x) is False” and “¬R(x) is True” are logically equivalent; they both
assert that the proposition R(x) is not true.
This notion we are developing right now is what is meant by a counterexam-
ple, a term you have likely heard before. To disprove a universally quantified
statement, we must prove an existentially quantified statement; that proof in-
volves explicitly defining an element of a set that does not satisfy the specified
property, whence the word counterexample.

4.4.2 Negation of an existential quantification

A statement like
.
∃x ∈ S R(x)
makes an existence claim. It says that there must be some element x that
satisfies the property R(x). To disprove a claim like this, we would seek to show
that any value of x actually fails to satisfy the property R. Accordingly, we can
say that the statement
.
¬ ∃x ∈ S R(x)


4.4. LOGICAL NEGATION OF QUANTIFIED STATEMENTS 237

is logically equivalent to the statement

.
∀x ∈ S ¬R(x)

This makes sense if we think about how to disprove such an existential claim.
Pretend you are having a debate with a friend who told you that some kwyjibo
has the property that it is a zooqa. How would you disprove him/her? You
might say something like, “Nuh uh! Show me any kwyjibo you want to. I know
it can’t possibly be a zooqa because of the following reasons . . . ” and then you
would explain why the property fails, no matter what.
Now, when you say “show me any” you are really performing a universal
quantification! You are saying that no matter which kwyjibo you consider,
something is true; that is, for every kwyjibo, or ∀x ∈ K (where K is the set of
all kwyjibos), something is True.
Think about this and consider why the logical negations we have discov-
ered/defined make sense to you. Later on in the chapter, when we consider proof
techniques, we will explain the strategy of considering an arbitrary kwyjibo and
why this actually proves the logical negation we just wrote above. For now, we
hope it is clear that
.
∀x ∈ S ¬R(x)
and
∃x ∈ S . R(x)
have opposite truth values.

4.4.3 Negation of general quantified statements

The observations we have made so far motivate a general procedure for negating
quantified statements. The statement A we defined above is of the form

.
∃y ∈ R C(y)

where C(y) is the rest of the statement (which depends on the value of y,
of course). We think of C(y) as some property of the quantified variable y;
that property might have other quantifiers and variables inside it, but at a
fundamental level, it is merely asserting some truth about y.
To negate this statement, we follow the method discussed above and write

.
∀y ∈ R ¬C(y)

Now, we know that C(y) is a universally quantified statement itself:

.
∀x ∈ R y = x3

We know how to negate that type of statement, too! That negation, ¬C(y), is

.
∃x ∈ R y 6= x3
238 CHAPTER 4. LOGIC

This step just uses the other negation procedure that we saw above. Then,
putting it all together, we can say that ¬A is the statement

. .
∀y ∈ R ∃x ∈ R y 6= x3

This claim we can prove to be true, thus showing that the original statement
must be False.
(We leave this proof as an exercise. Hint: Given any y ∈ R, define a value of x
that will force y 6= x3 to be true. Notice that your choice of x will depend the
value of y; how does it?)
Look at how this negation came about: we recognized that the original
statement was a sequence of nested quantifiers (i.e. a sequence of several
quantified variables in a row) with a variable proposition at the end, and we
saw that we could treat part of the sequence of quantifiers as its own statement.
We then “passed the negation” from the outside quantifier to the inside one,
and pieced those negations together.
Following this same idea, we can figure out how to identify a statement with
a longer sequence of quantifiers. For instance, look at a statement like

. . . .
∀a ∈ A ∃b ∈ B ∃c ∈ C ∀d, e ∈ D Q(a, b, c, d, e)

To start negating it, we would break off the first quantification, and treat the
rest as its own proposition, R(a), that depends only on a:

. . . .
∀a ∈ A ∃b ∈ B ∃c ∈ C ∀d, e ∈ D Q(a, b, c, d, e)
| {z }

R(a)

The negation can therefore be written as

.
∃a ∈ A ¬R(a)

but we would then have to figure out another way to write ¬R(a). But hey,
we would just do the same thing! We would just separate “∃b ∈ B” from the
rest and . . . you see where this is going. Try working out the steps on your own,
and make sure that you end up with the following as the logical negation of the
original statement:

. . . .
∃a ∈ A ∀b ∈ B ∀c ∈ C ∃d, e ∈ D ¬Q(a, b, c, d, e)

In general, we can say this: To negate a statement composed only of quan-

tifiers and variable propositions, just switch every “ ∀ ” to “ ∃ ”, and vice-versa,
and negate the propositions. Don’t alter any of the sets over which we quan-
tify, merely the quantifiers themselves and the ensuing propositions; it wouldn’t
make sense to change the universe of discourse. Later on, we will look at how
to negate other types of statements, more complicated ones built from other
connectives. Before we do that, we need to move on and define and discuss
those other connectives.
4.4. LOGICAL NEGATION OF QUANTIFIED STATEMENTS 239

4.4.4 Method Summary

Let’s summarize what we have discovered in this section.
• Negating a universal quantification:
Let X be a set and let P (x) be a proposition. Then the negation of a
universal quantification, like this,

.
¬ ∀x ∈ X P (x)

is written as
.
∃x ∈ X ¬P (x)

In words, we have shown that saying

It is not the case that, for every x ∈ X, P (x) holds.

is equivalent to saying

There exists an element x ∈ X such that P (x) fails.

• Negating an existential quantification:

Let X be a set and let Q(x) be a proposition. Then the negation of an
existential quantification, like this,

.
¬ ∃x ∈ X Q(x)

is written as
.
∀x ∈ X ¬Q(x)

In words, we have shown that saying

It is not the case that there exists an x ∈ X such that Q(x) holds.

is equivalent to saying

For every element x ∈ X, Q(x) fails.

Don’t Change the Quantification Set!

We mentioned above that it wouldn’t make sense to change the universe of
discourse when negating a statement. To think about why this makes sense,
take a real-life example.
Suppose we said “Every book on this bookshelf is written in English.” How
would you prove to us that we are lying, that our statement is actually False?
You would have to produce a book on this shelf that is written in a different
language. You couldn’t bring in a French novel from the room down the hall
and say, “See, you were wrong!” That wouldn’t prove anything about the claim
we made; the realms of discourse are different, and we didn’t make any claim
240 CHAPTER 4. LOGIC

about what’s going on in any bookshelves in other rooms. We only asserted

something about this particular shelf.
For the same reason, when negate a statement like

.
∀b ∈ T P (b)

we obtain
.
∃b ∈ T ¬P (b)
without changing that realm of discourse, the set T . The original claim only
asserts something about elements of T , so its negation does only that, as well.

4.4.5 The Law of the Excluded Middle

You know what? Let’s actually discuss why we can talk about a statement
and its logical negation. Built into our definition of mathematical/logical
statement is the idea that such a sentence must have exactly one truth value,
either True or False. Why can we do this? Well, we’re in charge of the defini-
tions here! Mathematicians have to set the ground rules—the axioms—of their
systems and we want our logical system to ensure that every claim we make is
decidedly True or False, and not both, and not neither.
This dichotomy is truly an axiom of our system. It is widely adopted in
most of mathematics, and is famously known as The Law of the Excluded
Middle. The name comes from this very idea, that every claim is True or False,
so there is no middle ground between those two sides; that middle is excluded.
In essence, this makes what we do in mathematics fruitful: every claim has
a truth value, and our goal is to find that truth value. Sometimes, though, we
have to fall back on this axiom, this law we agreed upon, and just guarantee that
some claim is either True or False, without knowing which truth value actually
applies. We present an interesting and striking example of this idea here.

Proposition 4.4.1. There exist real numbers a and b that are both irrational
such that ab is rational.

(Remember that a rational number is one that can be expressed as a fraction

of integers, and an irrational number is a real number that is not rational.
Can you think of some examples of both rational and irrational numbers?)
√
Proof. We know 2 is irrational. (Question: Why? Can you prove that? Try
it now. We will prove this very soon, as well!)
√ √2
The number 2 is either rational or irrational. (This is where the Excluded
Middle is used.) Let’s consider these two cases separately.
√ √2 √
• Suppose
√ that the number 2 is, indeed, rational. Then a = 2 and
b = 2 is the example we seek, because a and b are both irrational and
ab is rational.
4.4. LOGICAL NEGATION OF QUANTIFIED STATEMENTS 241

√ √2
• Now, suppose that the number 2 is irrational. In this case, we can
√ √2 √
use a = 2 and b = 2 as the example we seek, because a and b are
both irrational and
√ √2 √ √ √2·√2 √
ab = ( 2 ) 2 = 2 = ( 2)2 = 2
which is rational.
In either case, we have found an example of real numbers a, b ∈ R such that
both a and b are irrational and yet ab is rational. This proves the claim.
This is an example of a non-constructive proof. It tells us something
exists (and narrows it down to two possibilities, even) without actually telling
us exactly which possibility is the one we sought all long. It is a direct use of
the Law of the Excluded Middle that causes this.
√ √2
(Question: Can you prove somehow that 2 is irrational? It is True, but
there is no known “simple” proof of this fact. Maybe you can find one!)
Most of the proofs we do here will be of the constructive variety (but not
all). These might be more satisfying to you, and we’re inclined to agree. If we
claim something exists, we should be able to show it to you, right? If we just
talked for a while about why some such object exists somewhere else, without
being able to point to it, you would have to believe us, but you might not feel
great about it. Constructive proofs are subjectively better because of this, and
we will always strive for one when we can. Sometimes, though, a constructive
proof is not immediately clear, and we have to make a non-constructive one,
like we did here.

4.4.6 Looking Back: Indexed Set Operations and Quanti-

fiers
Look back at Section 3.6.2, where we defined set operations—union and inter-
section, in Definitions 3.6.3 and 3.6.4, respectively—performed over index sets.
The main idea was that we could express the union/intersection of an entire
class of sets all at once using a shorthand notation.
Look carefully at those definitions. What characterized whether an object
actually is an element of an indexed union, for example? That object needed
to be an element of at least one of the constituent sets of the union. That is,
there needed to exist some set of which that obejct is an element. This sounds
like an existential quantification, doesn’t it?
Likewise, what characterized whether an object is an element of an indexed
intersection? That object needed to be an element of all the constituent sets.
That is, for all of those sets, that object must be an element thereof. This is a
universal quantification.
With those observations now made, we can rewrite those definitions of in-
dexed set operations using our new quantifier notation:
242 CHAPTER 4. LOGIC

Definition 4.4.2. Suppose I is an index set and ∀i ∈ I .A i ⊆ U , for some

universal set U . Then
[
.
Ai = {x ∈ U | ∃k ∈ I x ∈ Ak }
i∈I

A = {x ∈ U | ∀i ∈ I . x ∈ A }
\
i i
i∈I

Try working with some of the examples and exercises in that Section 3.6.2
again. Do the definitions make more sense now?

4.4.7 Questions & Exercises

Remind Yourself
Answering the following questions briefly, either out loud or in writing. These
are all based on the section you just read, so if you can’t recall a specific defi-
nition or concept or example, go back and reread that part. Making sure you
can confidently answer these before moving on will help your understanding and
memory!

(1) What is the negation of a mathematical statement? How are a statement

and its negation related?

(2) Why is the negation of a ∀ claim an ∃ one?

Why is the negation of an ∃ claim a ∀ one?

(3) What is a non-constructive proof? To what type of claim—∃ or ∀—does

this term apply?

(4) Consider the claim

.
∀x ∈ S P (x)
Why is its negation neither of the following?

∀x ∈ .
/ S P (x)

/ S . ¬P (x)
∃x ∈

Try It
Try answering the following short-answer questions. They require you to actu-
ally write something down, or describe something out loud (to a friend/class-
mate, perhaps). The goal is to get you to practice working with new concepts,
definitions, and notation. They are meant to be easy, though; making sure you
can work through them will help you!

(1) For each of the following statements, write its negation. Which one—the
original or the negation—is True?
4.4. LOGICAL NEGATION OF QUANTIFIED STATEMENTS 243

. .
(a) ∀x ∈ R ∃n ∈ N n > x
(b) ∃n ∈ N. ∀x ∈ R. n > x
(c) ∀x ∈ R. ∃y ∈ R. y = x 3

(d) ∃y ∈ R. ∀x ∈ R. y = x 3

(2) For each of the following statements, write its negation. Which one—the
original or the negation—is True?
. .
(a) ∃S ∈ P(N) ∀x ∈ N x ∈ S
(b) ∀S ∈ P(N). ∃x ∈ N. x ∈ S
(c) ∀x ∈ N. ∃S ∈ P(N). x ∈ S
(d) ∃x ∈ N. ∀S ∈ P(N). x ∈ S
(3) Let I = {x ∈ R | 0 < x < 1}.
For each of the following defined sets, write out the defining condition that
determines whether a number y ∈ R is an element of the set, using quanti-
fiers.
Then, determine what the set is, and write your answer using set-builder
notation.
(Try to prove your claim, as well, using a double-containment argument!)
(a) [
S= {y ∈ R | x < y < 2}
x∈I

(b) \
T = {y ∈ R | −x < y < x}
x∈I

(c) [
V = {y ∈ R | −3x < y < 4x}
x∈I

(4) Let P = {y ∈ R | y > 0}. Consider this statement:

. .
∀ε ∈ P ∃δ ∈ P ∀x ∈ {y ∈ R | −δ < y < δ} |x3 | < ε .
Write out the logical negation of this statement.
What does this statement say? What does its negation say?
Which one is True? Can you prove it?
(5) Let A, B, C, D be arbitrary sets.
Let P (x), Q(x, y), R(x, y, z) be arbitrary variable propositions.
Write the negation of each of the following statements.
244 CHAPTER 4. LOGIC

(a) . .
∀a ∈ A ∃b ∈ B Q(a, b)
(b) .
∀a ∈ A ¬P (a)
(c) . .
∀c ∈ C ∀d ∈ D ¬Q(c, d)
(d) . .
∃a1 , a2 ∈ A ∀d ∈ D R(a1 , a2 , d)
(e) .
∀b1 , b2 , b3 ∈ B ¬R(b1 , b2 , b3 )
(f) . . .
∃b ∈ B ∀c ∈ C ∀d ∈ D R(d, b, c)

4.5 Logical Connectives

To build mathematical statements from simpler ones (meaning ones composed of
just quantifiers and propositions) we can connect several statements with certain
words and phrases—such as “and”, “or”, and “implies”—to create more compli-
cated statements and assert further claims and truths. We call these words and
phrases logical connectives, and each of them has their own corresponding
mathematical symbol and meaning. These meanings will make sense to you,
based on our intuitive grasp of the English language and rational thought, but
we emphasize that one of the major goals of introducing mathematical logic
and its corresponding notation is to build these intuitions into rigorous and
unambiguous concepts.
Throughout this section, let us assume that P and Q are arbitrary mathe-
matical statements. These statements themselves can be composed of compli-
cated combinations of quantifiers and other connectives and all sorts of mathe-
matical notions. The point is that the way we combine P and Q into a larger
statement is independent of their individual compositions. Before, we saw that
. .
“¬(∀x ∈ X R(x))” is equivalent to “∃x ∈ X ¬R(x)”, regardless of what the
statement R(x) was and how complicated it might have been. This idea contin-
ues here. We can talk about how to combine two statements without knowing
what they are, individually.
We should also point out that these constituent statements, P and Q, may
actually be variable propositions. For instance, we will consider how to connect
two variable propositions, P (x) and Q(x), that each depend on some variable
x. The definitions and methods we develop in this section apply to these vari-
able propositions even though these propositions, themselves, do not have truth
values without being told what the value of the variable x is.
When we want to talk about those propositions meaningfully and mathe-
matically, we will have to quantify the variable x. Thus, if we have variable
propositions P (x) and Q(x), we can still meaningfully define P (x)∧Q(x) (where
∧ means “and” as you’ll see in the next section). We could then, in an example
or a problem, talk about a claim of the form
.
∃x ∈ X P (x) ∧ Q(x)
This is a mathematical statement.
Essentially, the point we want to make is that these connectives still apply
to variable propositions, but the relevant variables will have to be quantified
4.5. LOGICAL CONNECTIVES 245

somewhere in an overall statement to make the variable proposition into a proper

mathematical statement.

4.5.1 And
To say
“P and Q” is True
means that both statements have the truth value: True. If either one of the
statements P or Q were False, then the statement “P and Q” would be False,
as well. This definition encapsulates this idea:
Definition 4.5.1. We use the symbol “ ∧ ” between two mathematical state-
ments to mean “and”. For instance, we read “P ∧ Q” as “P and Q”.
This is referred to as the conjunction of P and Q.
The truth value of “P ∧ Q” is True when both P and Q are true, and the truth
value is False otherwise.
Here are some examples to demonstrate this definition:
Example 4.5.2.

.
(1 + 3 = 4) ∧ (∀x ∈ R x2 ≥ 0) True
(1 + 3 = 5) ∧ (∀x ∈ R. x 2
≥ 0) False
(1 + 3 = 5) ∧ (∃x ∈ Q. x 2
= 2) False

Notation: Parentheses
It’s sometimes common to drop the parentheses that we used in the examples
above. For example, the first line in the above example can be written equiva-
lently as
1 + 3 = 4 ∧ ∀x ∈ R x2 ≥ 0 .
Using the parentheses tends to make the statement more readable. Without
them, we have to think for a few extra moments about where one part of the
statement ends and the next one begins, but we can still eventually make sense
of it. We will try to use parentheses whenever they make a statement more
easily understandable.

Notation: Sets and Logic

You might notice the similarity between the logical connective “ ∧ ” and the set
operator “ ∩ ”. This is not a coincidence!
As we will discuss below in Section 4.5.4, we can write the definition of “ ∩ ”
using the connective “ ∧ ” because of the underlying logic of that set operator.
Try it now, and then glance ahead to that section briefly, if you’d like. In
general, though, be careful to keep these two notations separate! If A and B
are sets, the phrase “A ∧ B” is not well-defined; what was meant is “A ∩ B”.
246 CHAPTER 4. LOGIC

4.5.2 Or
To say
“P or Q” is True
means that “P is True, or Q is True”. We need to know that one of the state-
ments is True to declare that the entire statement has the truth value True. We
don’t care whether both P and Q are true or not, merely that at least one of
them is true.
This is in contrast with the so-called “exclusive OR” of computer science,
also known as XOR, which declares “P XOR Q” to be False when both P and Q
are True. In mathematics, we use the inclusive “or”. We only care whether
at least one of the statements holds.
Definition 4.5.3. We use the symbol ∨ between two mathematical statements
to mean “or”. For instance, we read “P ∨ Q” as “P or Q”.
This is referred to as the disjunction of P and Q.
The truth value of “P ∨ Q” is True when at least one of P and Q is True (even
when both are True), and the truth value is False otherwise.
Example 4.5.4.

.
(1 + 3 = 4) ∨ (∀x ∈ R x2 ≥ 0) True
(1 + 3 = 5) ∨ (∀x ∈ R. x 2
≥ 0) True
(1 + 3 = 5) ∨ (∃x ∈ R. x 2
< 0) False

Notation
The same notes about notation that we mentioned in the previous subsection
apply here, as well. First, the use of parentheses—like in the above examples—is
helpful but not technically required. We will try to use them whenever it helps,
though.
Second, you might notice the similarity between the logical connective “ ∨ ”
and the set operator “ ∪ ”. Again, this is not a coincidence! Try rewriting the
definition of “ ∪ ” using “ ∨ ”, and glance ahead briefly at Section 4.5.4. In
general, though, be careful to keep these two notations separate! If A and B
are sets, the phrase “A ∨ B” is not well-defined; what was meant is “A ∪ B”.

4.5.3 Conditional Statements

This is the hardest logical connective to work with and continually gives students
some problems, so we want to be extra careful and clear about this one. We
want the statement “If P , then Q” (sometimes written as “P implies Q”) to
have the truth value True when the truth of Q necessarily follows from the truth
of P . That is, we want this statement to be True if the following holds:
Whenever P is True, Q is also True.
4.5. LOGICAL CONNECTIVES 247

Truth Table and Definition

Since this is the hardest connective to suss out, semantically, let’s introduce the
idea of a truth table to make the notation easier:
P Q ¬P P ∧Q P ∨Q P =⇒ Q Q =⇒ P
T T F T T T T
T F F F T F T
F T T F T T F
F F T F F T T

You may have seen a truth table before, in other mathematics courses, but even
if you haven’t, don’t worry! Here’s the main idea: Each column corresponds to
a particular mathematical statement and its corresponding truth values. Each
row corresponds to a particular assigment of truth values to the constituent
statements, P and Q.
Notice that there are 4 rows because P and Q can each have one of two
different truth values, so there are 4 possible combinations of those choices.
Reading across a particular row, we find the corresponding truth values for
other statements, based on what the T and F assignments for P and Q are in
the first two columns.
Notice that the columns for P ∧ Q and P ∨ Q follow the definitions given
above. The column for P ∧ Q has only one T, and it corresponds to the case
where both P and Q are True. All other cases make P ∧ Q a False statement.
Likewise, the column for P ∨ Q has only one F, and it corresponds to the case
where both P and Q are False. All other cases make P ∨ Q a True statement.
Now, why are the last two columns the way they are? Let’s say that I make
the claim “If you work hard, then you will get an A in this course”. Here, P is
“You work hard” and Q is “You will get an A”. When can you call me a liar?
When can you declare I told the truth? Certainly, if you work hard and get an
A, I told the truth. Also, if you work hard and don’t get an A, then I lied to
you. However, if you don’t work hard, then no matter what happens, you don’t
get to call me a liar! My claim didn’t cover your situation; I was assuming all
along you would just work hard! Thus, I didn’t speak an untruth and so, by
the Law of the Excluded Middle, I did speak the truth. The negation of a lie is
a truth.
This situation—where the third and fourth rows of the P =⇒ Q column
are True—is known as a false hypothesis. When the statement on the left of
the “ =⇒ ” does not hold, then we are not in a situation that is meant to be
addressed by the claim, so we cannot assert that the claim is False. Therefore,
the claim must be True (again, by the Law of the Excluded Middle).
Let’s make the proper definition of this symbol and then consider more
examples to illustrate the definition.
Definition 4.5.5. We use the symbol “ =⇒ ” between to mathematical state-
ments to mean “If . . . then” or “implies”. For instance, we read “P =⇒ Q” as
“If P , then Q” or “P implies Q”.
This is referred to as a conditional statement.
248 CHAPTER 4. LOGIC

The truth value of “P =⇒ Q” is True assuming that Q holds whenver P holds.

The truth value is False only in the case where P is True and yet Q is False.
We refer to P as the hypothesis of the conditional statement and Q as the
conclusion.
That key word “whenever” in the definition should indicate to you why the
false hypothesis cases make sense. When we know P is true and can deduce
that Q is also true, then we get to declare P =⇒ Q as True. If P wasn’t true
to begin with, we cannot declare P =⇒ Q to be False. We only get to say
P =⇒ Q is false when Q did not necessarily follow from P , i.e. when there is
an instance where the hypothesis P is True but the conclusion Q is False.

Examples
Here are several examples to help you get the idea:

.
(1 + 3 = 4) =⇒ (∀x ∈ R x2 ≥ 0) True
(1 + 3 = 5) =⇒ (∀x ∈ R. x 2
≥ 0) True
(1 + 3 = 5) =⇒ (Abraham Lincoln is alive) True
(1 + 1 = 2) =⇒ (0 = 1) False
.
(0 = 0) =⇒ (∃x ∈ R x2 < 0) False
(Pythagorean Theorem) =⇒ (1 = 1) True
(0 = 1) =⇒ (1 = 1) True

Notice that the second and third examples are True because they have the false
hypothesis “1 + 3 = 5”. Regardless of the conclusion, the entire conditional
.
statement must be True because of this. It doesn’t matter that “∀x ∈ R x2 ≥ 0”
happens to be True or that “Abraham Lincoln is alive” happens to be False; that
false hypothesis determines the statement’s truth value.
Also, notice that the second-to-last example is True, but it doesn’t help us
determine whether or not the Pythagorean Theorem itself is True! This is what
we did in the False “spoof” of the theorem back in Chapter 1. Look back to
Section 1.1.1, specficially “Proof 2”. We assumed the Pythagorean Theorem was
True and then logically derived a True statement from that assumption. That
does not make the hypothesis valid, just because we derived a valid conclusion!
This idea is so important, that we will even show you again this striking
example. Notice that it’s logical form is exactly the same as that other spoof:
“Proof ”. Assume 1 = 0. Then, by the symmetric property of =, it is also true
that 0 = 1. Adding these two equations tells us 1 = 1, which is True. Therefore,
0 = 1.
This is the main point here:
Knowing a conditional statement, overall, is True, doesn’t tell us
anything about the truth values of the constituent propositions.
4.5. LOGICAL CONNECTIVES 249

This is also strikingly illustrated in the third and seventh statements above;
both conditional claims are True, but we certainly don’t get to conclude that
Abraham Lincoln is alive, or that 0 = 1.

“Implies” is not the same as “Can be deduced from”

There is often some confusion with using the word “implies” to mean an “If
. . . then . . . ” statement, a conditional statement. We belive this arises because
of some connotations surrounding the word “implies”; specifically, it seems to
convey some sort of causality. For instance, consider this statement:

1 + 3 = 4 =⇒ 2 + 3 = 5

This is a True conditional statement, and our minds probably recognize this
because we can just take the hypothesis, namely 1 + 3 = 4, and add 1 to both
sides, yielding the equation in the conclusion. In this sense, it seems that the
truth of the hypothesis has some influence on the truth of the conclusion: we
can deduce one directly from the other. This does not have to be the case, in
general!
Look back at the first example given above:

.
(1 + 3 = 4) =⇒ (∀x ∈ R x2 ≥ 0)

What does the fact that 1 + 3 = 4 have to do with the fact that any real number
squared is non-negative? Does it even have any connection? We don’t actually
care! We can still identify this conditional statement as True, whether or not
we can find a way to deduce the conclusion directly from the hypothesis (and
whether or not such a deduction even exists). Only the truth values of the
constituent statements matter.
Granted, when we work on proving conditional statements, we will likely try
to deduce one statement directly from another. It’s important to keep in mind,
though, that this is a consequence of our proof strategy, and not an underlying
part of how conditional statements are defined. For these reasons, we tend to
write conditional statements using the “If . . . then . . . ” form, instead of using
“implies”. We might sometimes use it, and we’re sure you’ll see it in other math-
ematical writings. For now, though, we’ll try to avoid it as much as possible,
while we’re still learning about these logical statements and connectives.

Quantifying Variables: Still Important!

In mathematics, we often want to prove conditional statements that involve
variables. For instance, we might want to show that, in the context of the real
numbers R, the following conditional claim holds:

x > 1 =⇒ x2 − 1 > 0

That sentence, written in the line above, is itself a variable proposition, and
the definition of the symbol “ =⇒ ” applies to it.
250 CHAPTER 4. LOGIC

If we knew that x > 1 and also x2 − 1 > 0, then we could declare the
conditional to be True. If we knew that x ≤ 1, then we wouldn’t even care
whether x2 − 1 > 0 or not; we could declare the conditional to be True. This is
how the definition of “ =⇒ ” applies here.
Remember, though, that the conditional claim, as written above, is not
technically a mathematical statement. We were making that claim in the context
of the real numbers, so it would really make sense to write

.
∀x ∈ R x > 1 =⇒ x2 − 1 > 0

This is, ultimately, what the writer was trying to claim, so they should just say
so! We make the same recommendation to you. These logical connectives—
∧ and ∨ and =⇒ —make sense and can be applied to variable propositions.
Outside of that scope, somewhere else in the statement you’re putting together,
there must be some kind of quantification on those variables. Only then can we
be assured that the sentence is a mathematical statement with one truth value.

Writing “ =⇒ ” using “∨”

There is a useful and important idea worth mentioning. This is partly because
we will use it later, but also partly because it might help you understand con-
ditional statements and learn how to use them.
This idea hinges on the notion of a false hypothesis. Consider a conditional
statement, P =⇒ Q. If P fails, then the entire statement is True, regardless
of Q’s truth value. However, if P holds, then we definitely need Q to hold, as
well, to declare the entire statement to be True.
These observations allow us to identify two ways in which a conditional
statement can hold, and write these two ways in an “or” statement. Either ¬P
holds (i.e. a false hypothesis), or else Q holds. In either of these situations, the
conditional statement P =⇒ Q must hold! Let’s state this claim outright for
you to consider:

The conditional statement “P =⇒ Q” and the statement “¬P ∨ Q”

have the same truth value.

This is a good example of a logical equivalence, which is a topic we will

discuss in the next section. For now, we will present the truth table for the
two statements mentioned above. Notice that they have the same truth value,
regardless of the truth values of the constituent statements, P and Q. This
serves as further verification that the statements are equivalent, in addition to
the description we provided above.

P Q ¬P ¬P ∨ Q P =⇒ Q
T T F T T
T F F F F
F T T T T
F F T T T
4.5. LOGICAL CONNECTIVES 251

Investigating More Examples

Let’s consider some more examples of conditional claims and decide whether
they are True or False. In so doing, we are helping you to better understand
how =⇒ works.
Then, we’ll move on to proof strategies and discuss how to formally and
rigorously prove claims like these, with logical connectives and quantifiers.
Example 4.5.6. We will start here with a “real world” example, to get used to
the logic involved. Throughout this example, pretend we are in a class that only
has lectures officially scheduled on Mondays, Wednesdays, and Fridays.
You’ll notice that we will take two statements, P and Q, and consider all four
possible combinations of these statements and their negations to make condi-
tional statements.
• “If we have lecture today, then it is a weekday.”
(Note: There is some implicit quantification in this sentence. We are really
saying that “For all days d in the weekly calendar, if we have lecture on
day d, then d is a weekday.” We think that the main idea is conveyed
more succinctly by the sentence above, so we will use that version. Keep
in mind that this is the meaning of the sentence and, in our discussion
below, we will have to consider the different cases of that quantification.)
This can be written logically by defining P to be “We have lecture today”
and Q to be “Today is a weekday”; then the claim is P =⇒ Q.
Is this True? Notice that the statements P and Q don’t specify what day
it is, so if we were to assert this claim to be True, that truth should be
independent of the current day. That is, whatever day it is, we would have
to show that P =⇒ Q holds. Let’s do that by considering a few cases:
– Pretend today is Saturday or Sunday. Since we never have lecture
on these days, the conditional claim is not a lie, so it is True.
– Pretend today is Monday or Wednesday or Friday. If we do, indeed,
have lecture today, then it is definitely a weekday, so the claim is
True.
– Pretend today is Tuesday or Thursday. We don’t typically have
lecture today, but even in the special case that we did (for some
rescheduling reason), it would still be a weekday, so the claim is
True.
In any of the possible cases, the claim holds. Thus, P =⇒ Q is True.
You might interject to say, “Why bother with the cases at all? Couldn’t
we just say that no matter what day it is, supposing we have lecture,
then we conclude it is definitely a weekday?” Well, yes, we could! That’s
actually a better strategy, a more direct route, you might say.
This hints at how we will prove conditional claims in the future. Since we
don’t, in fact, care about the situations where we don’t have lecture (the
252 CHAPTER 4. LOGIC

false hypothesis) we only need to suppose we have lecture on day X and

deduce that X is a weekday. This is the method we will use to directly
prove a conditional claim.
• “If it is a weekday, then we have lecture today.”
This is logically written as Q =⇒ P , using the same definitions of P and
Q as the previous example.
Is this claim True? Definitely not! We didn’t have lecture on the first
Tuesday of the semester, yet that day was a weekday. Thus, the claimer
lied in that instance! On that Tuesday, Q was True but P was False. Thus,
Q =⇒ P is False.
• “If it is not a weekday, then we don’t have lecture today.”
This is logically written as ¬Q =⇒ ¬P , using the same definitions of P
and Q as the example above.
Is this claim True? Yes, and we can prove it directly. Suppose that it is not
a weekday; that is, pretend today is Saturday or Sunday. Obviously, the
university would not be so sadistic as to schedule a lecture on a weekend,
so we can necessarily declare that we don’t have lecture, i.e. ¬P holds.
This shows that ¬Q =⇒ ¬P is a True statement.
(Question: Why did we not need to consider the case where today is a
weekday?)
• “If we don’t have lecture today, it is not a weekday.”
This is logically written as ¬P =⇒ ¬Q, using the same definitions of P
and Q as the example above.
Is this claim True? Well, let’s think about it. What if we pretend we don’t
have lecture today. What can we necessarily say? Is it definitely not a
weekday? I don’t think so! Maybe it’s a Tuesday, and we just don’t have a
scheduled lecture. This shows that the claim is False; we have an instance
where the hypothesis (¬P ) holds, and yet the conclusion (¬Q) does not
hold.
Notice that there are situations where ¬P holds and ¬Q does, as well. For
instance, if today were Saturday, then certainly we don’t have lecture and
it is not a weekday. This specific instance does not meant that the claim
is True, though! We need to verify its truth for all instances.
Example 4.5.7. Let’s do the same kind of analysis with a more “mathy” example.
Throughout this example, let A and B be arbitrary sets. Also, let P be “A ⊆ B”,
and let Q be “A − B = ∅”.
We will do what we did in the previous example and consider all four possible
ways of combining P and Q and their negations to make conditional statements.
• Is P =⇒ Q True?
4.5. LOGICAL CONNECTIVES 253

Yes! Let’s pretend A and B satisfy the relationship A ⊆ B. This means

every element of A is also an element of B. Therefore, we cannot possibly
have an element of A that does not belong to B, Since A − B is the set of
elements that belong to A and not B, we conclude that there are no such
elements, so A − B = ∅.

• Is Q =⇒ P True?
Yes! Let’s pretend A − B = ∅. This means there are no elements of A
that are also not elements of B. (Think about this.) Put another way, any
element x ∈ A cannot have the property that x ∈ / B (or else x ∈ A − B
and so A − B 6= ∅); thus, x ∈ B, necessarily. Hey, this is exactly what
A ⊆ B means! Whenever x ∈ A, we conclude x ∈ B, too. This shows
Q =⇒ P is true.

• Is ¬Q =⇒ ¬P True?
Hmm, this is harder to figure out. Let’s pretend ¬Q holds; this means
A − B 6= ∅. That is, there exists some element x that satisfies x ∈ A and
x∈/ B. Certainly, then A 6⊆ B, because we have identified an element of
A that is not an element of B (whereas the ⊆ relationship would dictate
that every element of A is also an element of B). Thus, ¬Q =⇒ ¬P is
True.

• Is ¬P =⇒ ¬Q True?
Again, we need to think about this. Let’s just write down what ¬P means.
To say A 6⊆ B means there exists some element x ∈ A that also satisfies
x∈/ B. (This is what we used in the previous case, too.) Okay, what does
that tell us? Consider the set A − B. Does it have any elements? Yes, it
has at least x as an element! Since x ∈ A ∧ x ∈/ B, we can say x ∈ A − B.
Thus, A − B 6= ∅, so we conclude that ¬P =⇒ ¬Q is True.

Observations and Facts About “ =⇒ ”

Okay, now we have some practice working with conditional statements and
determining their truth values. What you should notice from the examples we
discussed is that knowing P =⇒ Q holds does not tell us anything about
Q =⇒ P . In both of these examples above, P =⇒ Q was True; however,
Q =⇒ P was True in one example and False in the other. There is nothing we
can necessarily say, with certainty, about Q =⇒ P , even if we know the truth
value of P =⇒ Q. This idea is so important, that we will touch on it in the
next subsection.
For now, let’s make a few more remarks about the “ =⇒ ” connective.

• Remember that, given mathematical statements P and Q, the sentence

“P =⇒ Q” is, itself, another mathematical statement. It has a truth
value. That truth value depends on P and Q (in the way we defined it
254 CHAPTER 4. LOGIC

above), but it does not tell us anything about the truth values of P and
Q. So, if you just write down the claim

Blah blah =⇒ Yada yada

on your paper, we have no idea if you’re trying to assert whether “Blah

blah” or “Yada yada” are True or False! To a mathematician, this just
says:
The conditional statement “ ‘Blah blah’ implies ‘Yada yada’ ”
is True.
If you wish to make some kind of inference or deduction, then use some
helping words and sentences to indicate that. Write something like this:
P =⇒ Q because . . .
Also, P holds because . . .
Therefore, Q holds.
If you have studied formal logic before, or have seen this type of argument
in a philosophy course, then you might recognize this as Modus Ponens.
• The idea of a false hypothesis yielding a True conditional is kind of
weird; we realize this. It’s a direct consequence of the Law of the Excluded
Middle. Under a false hypothesis, we don’t get to say the overall statement
is False, so it must be True, since it must be one or the other.
• Remember that we can always write a conditional statement without the
“ =⇒ ” symbol by converting it to an “or” statement.
The statements “P =⇒ Q” and “¬P ∨ Q” always have they
always have the same truth value.

Converse and Contrapositive

Let’s give some names to the different types of conditional statements related
to a given conditional statement. We will refer to these frequently later on.
Definition 4.5.8. Let P and Q be mathematical statements. Consider the
“original” claim, P =⇒ Q.
We refer to Q =⇒ P as the converse of the original claim.
We refer to ¬Q =⇒ ¬P as the contrapositive of the original claim.
By our observations in the previous subsection, we know that the converse
does not necessarily have the same truth value as the original. What we will see
(and prove) in the next section is that the contrapositive always has the same
truth value as the original claim. (This is the notion of logical equivalence,
which we will discuss in full detail in the next section.)
4.5. LOGICAL CONNECTIVES 255

You might wonder why we need this terminology at all. Well, since the
contrapositive can be shown to be logically equivalent to the original claim, it
gives rise to a valid proof method when we try to prove conditional statements.
We will develop that very soon. That’s why we use the contrapositive.
The converse is interesting because its truth value is not necessarily tied to
that of the original statement: even knowing the original is True, the converse
might be True and it might be False. Thus, whenever we prove a claim P =⇒ Q
to be true, a mathematician (probably) immediately wonders, “Hmm, does the
converse also hold?” It’s a natural question to ask, and worth thinking about
whenever you face a conditional statement. (In fact, if you ever find yourself
at a party with mathematicians, and you hear someone talking about an “If
. . . then . . . ” statement, you should ask, “Does the converse hold, as well?”
You might impress your fellow guests.)
The converse is also the subject of a common logical fallacy you might en-
counter in everyday life. Perhaps you are trying to argue that A =⇒ B in a
debate with a friend. What if they retort, “Well, B doesn’t necessarily imply A!
Your argument is wrong!” Have you ever been frustrated by that situation? You
might have been tempted to shout, “So what? I wasn’t trying to say anything
about whether B =⇒ A. I was talking about A =⇒ B, you . . . ” (We’ll cut
ourselves off before we get mean.) Whether or not your friend is right, knowing
the truth value of the converse doesn’t tell you anything about the truth value
of your original claim. You should tell them that! Next time, just say, “You’re
talking about the converse, which is not necessarily logically connected to my
claim.”
Okay, now that we have defined all of the requisite logical symbols and seen
some examples, it’s time to move on and start proving things about them! But
first, a short aside about set operations, and then a few practice problems.

4.5.4 Looking Back: Set Operations and Logical Connec-

tives
Look back at Sections 3.4 and 3.5, where we defined subsets and set operations.
All of those definitions made use of some logical ideas, but we wrote them in
English back then, relying on our collective intuition and working knowledge of
logic. We can rewrite them now using quantifiers and connectives!
First, recall the definition of subset. We write A ⊆ B if the following holds:
whenever x ∈ A, we can also say x ∈ B. Notice that key word, “whenever”,
which signals both a universal quantification and a conditional statement. Think
about how you would rewrite the definition of A ⊆ B, using this notion, then
read on to see our version . . .

Definition 4.5.9. Let A, B, U be sets, where A, B ⊆ U (i.e. U is a universal

set). We say A is a subset of B, and write A ⊆ B, if and only if

∀x ∈ U . x ∈ A =⇒ x ∈ B
256 CHAPTER 4. LOGIC

This makes sense because it asserts that “whenever” statement we wrote in

the above paragraph: whenever x ∈ A, we must also be able to conclude x ∈ B;
“if x ∈ A, then x ∈ B” must hold.
Look back again at the definitions of the set operations we gave. Try to write
your own versions of those definitions using logical symbols and then read ours
here. Think about how they make sense, how they express the same underlying
ideas.

Definition 4.5.10. Let A, B, U be sets, where A, B ⊆ U (i.e. U is a universal

set). Then,

A ∩ B = {x ∈ U | x ∈ A ∧ x ∈ B}
A ∪ B = {x ∈ U | x ∈ A ∨ x ∈ B}
A − B = {x ∈ U | x ∈ A ∧ ¬(x ∈ B)} = {x ∈ U | x ∈ A ∧ x ∈
/ B}
A = {x ∈ U | ¬(x ∈ A)} = {x ∈ U | x ∈
/ A}

While we’re here, we can also redefine a partition of a set. This will make
use of logical connectives, but this also hearkens back to indexed sets and how
they are defined in terms of quantifiers. Everything we’ve learned is coming
together here!

Definition 4.5.11. Let A be a set. A partition of A is a collection of sets

that are pairwise disjoint and whose union is A.
That is, a partition is formed by an index set I and non-empty sets Si (defined
for every i ∈ I) that satisfy the following conditions:

.
(1) ∀i ∈ I Si ⊆ A.

.
(2) ∀i, j ∈ I i 6= j =⇒ Si ∩ Sj = ∅.
[
(3) Si = A
i∈I

Look back to Definition 3.6.9 to see how we originally defined a partition.

Do you see how we are saying the same thing here, just using logical symbols?

4.5.5 Questions & Exercises

Remind Yourself
Answering the following questions briefly, either out loud or in writing. These
are all based on the section you just read, so if you can’t recall a specific defi-
nition or concept or example, go back and reread that part. Making sure you
can confidently answer these before moving on will help your understanding and
memory!

(1) What is the difference between ∧ and ∨?

4.5. LOGICAL CONNECTIVES 257

(2) What is the difference between ∧ and ∩?

What is the difference between ∨ and ∪?

(3) Write out a truth table for the statement P =⇒ Q.

(4) Why are P =⇒ Q and ¬P ∨ Q logically equivalent statements?

(5) What is the converse of a conditional statement?

What is the contrapositive of a conditional statement?

(6) Are the truth values of a conditional statement and its converse related?

Try It
Try answering the following short-answer questions. They require you to actu-
ally write something down, or describe something out loud (to a friend/class-
mate, perhaps). The goal is to get you to practice working with new concepts,
definitions, and notation. They are meant to be easy, though; making sure you
can work through them will help you!

(1) For each of the following sentences, rewrite it using logical notation and
determine whether it is True or False.

(a) Every integer is either strictly positive or strictly negative.

(b) For any given real number, there exists a natural number that is strictly
bigger.
(c) For every real number, if it is negative, then its cube is also negative.
(d) There is a subset of Z that has the property that whenever a number
is an element of that set, so is its square.
(e) There is a natural number that is both even and odd.

(2) Rewrite each of the following ∀ claims as a conditional statement, and de-
termine whether it is True or False.

. . .
(a) ∀x ∈ {y ∈ N | ∃k ∈ Z y = 2k} x2 ∈ {y ∈ N | ∃k ∈ Z y = 2k}
(b) ∀x ∈ {y ∈ N | ∃k ∈ Z. y = 2k + 1}. x 2
.
∈ {y ∈ N | ∃k ∈ Z y = 2k + 1}
(c) ∀t ∈ {z ∈ R | z > 4}. t > 2
2

(3) Rewrite each of the following conditional statements as a ∀ claim, using

set-builder notation, and determine whether it is True or False.

.
(a) ∀x ∈ R x > 3 =⇒ x2 < 9
(b) ∀x ∈ R. x < 3 =⇒ x < 9 2

(c) ∀t ∈ R. t − 6t + 9 ≥ 0 =⇒ t ≥ 3
2
258 CHAPTER 4. LOGIC

(4) Let’s define the following variable propositions:

1
P (x) is “ <x”
2
3
Q(x) is “ x < ”
2
R(x) is “ x2 = 4 ”
S(x) is “ x + 1 ∈ N ”

For each of the following statements, determine whether it is True or False.

.
(a) ∀x ∈ N P (x)
(b) ∀x ∈ N. Q(x) =⇒ P (x)
(c) ∀x ∈ Z. Q(x) =⇒ P (x)
(d) ∃x ∈ N. ¬S(x) ∨ R(x)
(e) ∃x ∈ Z. R(x) ∧ ¬S(x)
(f) ∀x ∈ R. R(x) =⇒ S(x)
(g) ∃x ∈ R. P (x) ∧ S(x)
(h) ∀x ∈ Z. R(x) =⇒ P (x) ∨ Q(x)

(5) For each of the following conditional statements, write it in logical notation,
and use this write its converse and its contrapositive; then, determine the
truth values of all three: the original statement, the converse, and the
contrapositive.

(a) If a real number is strictly between 0 and 1, then so is its square.

(b) If a natural number is even, then so is its cube.
(c) Whenever an integer is a multiple of 10, it is a multiple of 5.

4.6 Logical Equivalence

In this section, the major goal is to introduce the idea of logical equivalence
and prove a few fundamental claims. Essentially, we want to decide when some
complicated logical statements are actually “the same”, in the sense of truth
values. Since mathematical statements may depend on some propositional vari-
ables, we might not be able to conclude anything specific about their truth
values. However, we can sometimes prove that two mathematical statements
will have the same truth value, for all possible values of the variables they con-
tain. That’s a really nice conclusion to make! We get to say that they have the
same truth value, regardless of what it is. In that sense, we are proving the two
statements to be equivalent, in a logical sense.
4.6. LOGICAL EQUIVALENCE 259

4.6.1 Definition and Uses

The following definition introduces a convenient symbol for the notion of logical
equivalence described in the above paragraph:
Definition 4.6.1. Let P and Q be mathematical statements. We use the symbol
“⇐⇒” to mean “is logically equivalent to”, or “has the same truth value as”.
That is, we write “P ⇐⇒ Q” when P and Q always have the same truth value,
regardless of whether it is T or F.
We read “P ⇐⇒ Q” aloud as “P is logically equivalent to Q” or “P if and
only if Q”.
This type of statement is known as a biconditional (or a bi-implication).
Let’s take the truth tables we saw in the last section and add a new column
for this symbol:
P Q ¬P ¬P ∨ Q P =⇒ Q Q =⇒ P P ⇐⇒ Q
T T F T T T T
T F F F F T F
F T T T T F F
F F T T T T T

In the column for P ⇐⇒ Q, an entry has the truth value T when (and only
when) P and Q have the same truth value. This happens in Row 1, where both
are T, and in Row 4, where both are F. Notice, then, that P ⇐⇒ Q has the
truth value T if and only if
(P =⇒ Q) ∧ (Q =⇒ P )
is a True statement. This is the notion of logical equivalence: P ⇐⇒ Q means
that both P =⇒ Q and Q =⇒ P hold. Whatever truth value P has, Q is
guaranteed to have the same truth value, and vice-versa:
• Supposing that P is True, then P =⇒ Q tells us that Q must also be
True.
• Supposing that P is False, then Q =⇒ P tells us that Q cannot be True
(since Q =⇒ P would be False, in that case), so Q must also be False.
Either way, P and Q have the same truth value.

Examples
Example 4.6.2. Look again at the third and fourth columns in the truth table
above. They prove the following logical equivalence:
(P =⇒ Q) ⇐⇒ (¬P ∨ Q)
Whatever the truth value of P =⇒ Q (which, of course, depends on P and
Q), it must be the same as the truth value of ¬P ∨ Q. We’ve mentioned this
equivalence before, and we will make use of it fairly often in the future.
260 CHAPTER 4. LOGIC

Example 4.6.3. Look at this truth table:

P Q ¬P ¬Q P =⇒ Q ¬Q =⇒ ¬P
T T F F T T
T F F T F F
F T T F T T
F F T T T T
Regardless of the truth values of P and Q, we find that P =⇒ Q and ¬Q =⇒
¬P have the same truth values. Thus, they are logically equivalent, and we can
write:
(P =⇒ Q) ⇐⇒ (¬Q =⇒ ¬P )
This is the fact that we stated (without proof) in the previous section:
The contrapositive of a conditional statement is logically equivalent
to the original statement.
A different proof of this fact makes use of the way to express a conditional
statement as a disjunction. Recall the logical equivalence



P =⇒ Q ⇐⇒ ¬P ∨ Q
that we mentioned in the previous example. Now, think about the contrapositive
of that original conditional statement:
¬Q =⇒ ¬P
Applying the same disjunctive form to that statement yields the following equiv-
alence:


¬Q =⇒ ¬P ⇐⇒ ¬(¬Q) ∨ ¬P
Now, noticing that ¬(¬Q) is equivalent to just Q, and remembering that the
order of a disjunction is irrelevant (i.e. P ∨ Q and Q ∨ P have the same truth
value) we find that




¬Q =⇒ ¬P ⇐⇒ ¬P ∨ Q ⇐⇒ P =⇒ Q
This shows, in another way, that a conditional statement and its contrapositive
have the same truth value!
Example 4.6.4. Later in this section, we will prove the following logical equiva-
lences, which hold no matter what the propositions P and Q and R are:
¬(P ∧ Q) ⇐⇒ ¬P ∨ ¬Q
(P ∧ Q) ∧ R ⇐⇒ P ∧ (Q ∧ R)
P ∨ (Q ∧ R) ⇐⇒ (P ∨ Q) ∧ (P ∨ R)
¬(P =⇒ Q) ⇐⇒ P ∧ ¬Q
Each of these is an assertion that the expressions on both sides of the ⇐⇒
symbol have the same truth value. Can you see why these claims are True? Can
you think of how to prove them?
4.6. LOGICAL EQUIVALENCE 261

If and Only If
Logical equivalence has a nice relationship with the phrase “if and only if”. To
say “P if and only if Q” means we are asserting that both “P if Q” and “P only
if Q” hold. One of these corresponds to P =⇒ Q and the other corresponds
to Q =⇒ P , so asserting both are true means exactly what we have described:
P ⇐⇒ Q is the same as saying (P =⇒ Q) ∧ (Q =⇒ P )
Now, which one is which, though? When we say “P if Q”, this means “If Q,
then P .” That is,
“P if Q” is the same as saying Q =⇒ P
Sussing out the other direction is a little harder! What does “P only if Q”
really mean? This sentence is asserting that, in a situation where P holds, it
must also be the case that Q holds. That is, knowing P holds means we also
immediately know Q holds. Put even another way, whenever P is true, we
necessarily know that Q is true. This is the same as saying P =⇒ Q holds!
Another way of thinking about it is as follows. Saying “P only if Q” is the
same as saying it cannot be the case that P holds and Q does not. Written
logically, we have

¬ P ∧ ¬Q
Later on in this section, we will state and prove DeMorgan’s Laws for Logic.
One of those laws tells us how to negate that statement inside the parentheses.
(You might already know these logical laws, in fact. If not, you can glance ahead
at Sections 4.6.5 and 4.6.6 for a preview.) The conclusion is:
¬P ∨ Q
Hey look, that’s logically equivalent to P =⇒ Q, as we observed already! Cool.
Just further confirmation that “P only if Q” means P =⇒ Q.

Using “⇐⇒” in Definitions

We will also often use the “⇐⇒” symbol in a definition to indicate that the
term defined is an equivalent term for the property that is used in the definition.
For example:
We say x ∈ Z is even ⇐⇒ ∃k ∈ Z x = 2k .
That is, the notion of an integer being even is equivalent to knowing that the
number is twice an integer. Similarly, we can define odd:
.
We say x ∈ Z is odd ⇐⇒ ∃k ∈ Z x = 2k + 1
These are formal definitions, mind you, and are the only way of guaranteeing
an integer is even (or odd). We will use these definitions soon to rigorously
prove some facts about integers and arithmetic. Every time we want to assert
a particular integer (call it x) is even, we need to show there exists an integer k
that satisfies x = 2k. That is, we have to satisfy the definition by appealing to
the logical equivalence given in the definition.
262 CHAPTER 4. LOGIC

Biconditional Statements: A Technical Distinction

We can also use the symbol “⇐⇒” to express two conditional statements at
once. Technically speaking, this is not exactly the same as asserting a logical
equivalence, but it conveys a similar idea, so we allow the symbol to be used in
both ways.
A logical equivalence involves some undefined propositions, and it asserts
that the two statements will have the same truth value, regardless of the truth
values of those propositions. For instance,

(P =⇒ Q) ⇐⇒ (¬P ∨ Q)

is a perfect example of a logical equivalence. Without telling you what P and

Q are, we can’t be sure exactly what P =⇒ Q and ¬P ∨ Q mean. However,
we don’t need to know what P and Q are to know that those two statements
definitely have the same truth value.
The situation is slightly different when the two statements on either side
of the “⇐⇒” are actually proper mathematical statements, with no undefined
propositions. For instance, consider this statement:

.
 
1
∀x ∈ R (x > 0) ⇐⇒ >0
x
This is a logical claim, and it asserts that, whenever x is a real number, knowing
one of those two facts—x > 0 or x1 > 0—necessarily guarantees the other. That
is, if I told you I have a real number in mind and it is positive, you get to
conclude that its reciprocal is positive, too. Also, if I told you I have a real
number in mind whose reciprocal is positive, you would get to conclude that
the number itself is positive, too. It goes both ways. (Question: What if I told
you I had a negative real number in mind? Could you conclude anything about
its reciprocal? Why or why not?)
Do you see how this is different? Given an arbitrary x ∈ R, the statement
“x > 0” is decidedly either True or False. Its truth value isn’t left undetermined.
This distinguishes it from the example given above, where the truth value of
the individual statements is unknown, yet we can still declare those truth values
must be identical.
For lack of a better, widespread term for these kinds of statements, we will
refer to them as biconditionals. This is because they are really meant to
represent two conditional statements that go “in opposite directions”:

.
     
1 1
∀x ∈ R (x > 0) =⇒ >0 ∧ > 0 =⇒ (x > 0)
x x
This is what the statement above says: each part of the statement implies the
other one.
This term is not necessarily standard in other mathematical writing, but we
wanted to point out this technical distinction so you are aware of it. You might
approach a mathematicial logician or set theorist and use the phrase “logical
4.6. LOGICAL EQUIVALENCE 263

equivalence”, and they might be confused or take offense to the way in which
you use it. This is not a big worry, mind you! Since we are learning about these
fundamental ideas now for the first time, we don’t necessarily have to keep in
mind all of the technical details that lie underneath these concepts. Also, in the
remainder of this book, we might use “logical equivalence” and “biconditional”
interchangeably. This is fine and acceptable for now.
The main point behind using the “⇐⇒” symbol is to assert that two state-
ments have the same truth value. The only difference between a “logical equiv-
alence” and a “biconditional” is whether or not those statements contained
therein have any arbitrary, undefined propositions. This is a minor distinction
to be made, in the grand scheme of things, so we will consider it only briefly
here.

4.6.2 Necessary and Sufficient Conditions

There are two terms occasionally used in mathematics that convey the two direc-
tions of a biconditional statement: necessary and sufficient. They correspond
exactly to the “only if” and “if” parts of the biconditional. These terms are
motivated by the natural types of questions mathematicians ask.

Sufficient: P , if Q
If we identify some interesting fact or property—call it P —of a mathematical
object, we might wonder, “When can we guarantee such a property holds? Is
there some condition we can check that will give us a ‘Yes’ answer right away?”
This is what a sufficient condition is, a property that guarantees P will also
hold. It is “sufficient” in the sense that it is “enough” to conclude P ; we don’t
need any other outside information.
Let’s say we have identified a proposition Q as a sufficent condition for P .
How can we express this logically? Well, knowing Q is sufficient to conclude P ,
so we can easily write this as a conditional statement:

Q =⇒ P means Q is a sufficient condition for P

Said another way, this conditional statement expresses: “P , if Q”.

Necessary: P only if Q
We also might wonder, “How can we guarantee that P fails? Is there some con-
dition we can check that will tell us this right away?” This is what a necessary
condition is, a property that is necessary or essential for the property P to hold.
This condition is not necessarily enough to conclude that P holds, but for P to
even have a chance of holding, this condition better hold, too.
Think about the logical connections here. Say we have established a prop-
erty Q that is a necessary condition for P . How can we express the relationship
between P and Q, symbolically? That’s right, we can use a conditional state-
ment. Knowing P holds tells us that Q definitely holds; it was necessary for P
264 CHAPTER 4. LOGIC

to be true. This is expressed as

P =⇒ Q means Q is a necessary condition for P

Said another way, this conditional statement expresses: “P only if Q”.

We could also think of this in terms of the contrapositive. If Q does not
hold, then P cannot hold, either. That is,

¬Q =⇒ ¬P

which is the contrapositive of the conditional statement above, P =⇒ Q. We

know these are logically equivalent forms of the same statement.

Examples
Example 4.6.5. Let P (x) the proposition “x is an integer that is divisible by 6”.
For each of the following conditions, let’s identify whether it is a necessary or
sufficient condition (or possibly both!) for P (x) to hold.

(1) Let Q(x) be “x is an integer that is divisible by 3”.

To determine whether Q(x) is a necessary condition, let’s assume P (x)
holds. Can we deduce Q(x) holds, too? Well, yes! To say an integer x is
divisible by 6 means that it is divisible by both 2 and 3. Thus, it is certainly
divisible by 3, so Q(x) holds.
To determine whether Q(x) is a sufficient condition, let’s assume Q(x) holds.
Can we deduce P (x) holds, too? Hmm . . . knowing that x is an integer
divisible by 3, is it also definitely divisible by 2, allowing us to conclude it is
divisible by 6? We think not! Consider x = 3; notice Q(3) holds but P (3)
does not.
This shows Q(x) is only a necessary condition, and not a sufficient one.

(2) Let R(x) be “x is an integer that is divisible by 12.”

Following similar reasoning to the above example, we can conclude that
R(x) is a sufficient condition for P (x), but not a necessary one (because we
can have x = 6, where P (6) holds but R(6) does not hold).

(3) Let S(x) be “x is an integer such that x2 is divisible by 6”.

We’ll let you work with this one on your own . . . Is S(x) a necessary condition
for P (x)? Is it a sufficient one?
Be careful, and notice that we specified x, itself, is an integer . . .

4.6.3 Proving Logical Equivalences: Associative Laws

Now, let’s actually prove some logical equivalences! In doing so, we will be
working on our ability to read and understand and write logical statements
using quantifiers and connectives. We will also be developing some fundamental
4.6. LOGICAL EQUIVALENCE 265

logical results that we can apply in the near future to develop proof techniques.
These techniques will be the foundation of the rest of our work, and everything
else we do will involve implementing some combination of these proof strategies
and logical concepts.
Let’s start with some of the simpler symbolic logical laws. Showing some-
thing is associative essentially means we can “move around the parentheses”
willy-nilly and end up with the same thing. You probably use the fact that
addition is associative all the time! To add x to y + z, we can just add z to x + y
instead and know we get the same answer. That is, we can rest assured that

x + (y + z) = (x + y) + z

We can move the parentheses around wherever we want to and so, ultimately,
we can just pretend as if they aren’t even there and just write

x+y+z

because the order in which we interpret the two additions is irrelevant. The same
kind of result applies to conjuctions and disjunctions of logical statements, and
that’s what we will prove now.

Theorem 4.6.6. Let P, Q, R be logical statements. Then

P ∧ (Q ∧ R) ⇐⇒ (P ∧ Q) ∧ R

and
P ∨ (Q ∨ R) ⇐⇒ (P ∨ Q) ∨ R

We will actually prove these claims in two separate ways: (1) via truth
tables, and (2) via semantics (i.e. words). They are both perfectly valid, but
we want to show you both of them to let you decide which style you like better.

Proof 1. First, we will prove these claims via truth tables. Observe the table
for conjunctions:

P Q R P ∧Q Q∧R P ∧ (Q ∧ R) (P ∧ Q) ∧ R
T T T T T T T
T T F T F F F
T F T F F F F
T F F F F F F
F T T F T F F
F T F F F F F
F F T F F F F
F F F F F F F

Thus, P ∧ (Q ∧ R) ⇐⇒ (P ∧ Q) ∧ R because their corresponding columns are

identical, in every case.
266 CHAPTER 4. LOGIC

Next, observe the table for disjunctions:

P Q R P ∨Q Q∨R P ∨ (Q ∨ R) (P ∨ Q) ∨ R
T T T T T T T
T T F T T T T
T F T T T T T
T F F T F T T
F T T T T T T
F T F T T T T
F F T F T T T
F F F F F F F
Thus, P ∨ (Q ∨ R) ⇐⇒ (P ∨ Q) ∨ R because their corresponding columns are
identical, in every case.
Proof 2. Second, let’s prove these claims by analyzing them, semantically. Con-
sider the first claim,
P ∧ (Q ∧ R) ⇐⇒ (P ∧ Q) ∧ R
To show that the two sides are logically equivalent, we need to show both of the
following conditional statements are True:
P ∧ (Q ∧ R) =⇒ (P ∧ Q) ∧ R
and
(P ∧ Q) ∧ R =⇒ P ∧ (Q ∧ R)
( =⇒ ) Let’s prove the first conditional statement. Suppose P ∧ (Q ∧ R) is True.
This means P is True and Q ∧ R is True. By definition, this means P is True
and Q is True and R is True. Certainly, then P ∧ Q is True and R is True, by
definition. Thus, (P ∧ Q) ∧ R is True, as well.
(⇐=) Now, let’s prove the second conditional statement. Suppose (P ∧Q)∧R
is True. This means P ∧ Q is True and R is True. By definition, this means P is
True and Q is True and R is True. Certainly, then P is True and Q ∧ R is True,
by definition. Thus, P ∧ (Q ∧ R) is True, as well.
Since we have shown both conditional statements, we conclude the two sides
are, indeed, logically equivalent.
Next, consider the second claim of the theorem,
P ∨ (Q ∨ R) ⇐⇒ (P ∨ Q) ∨ R
To show that the two sides are logically equivalent, we need to show both of the
following conditional statements are True:
P ∨ (Q ∨ R) =⇒ (P ∨ Q) ∨ R
and
(P ∨ Q) ∨ R =⇒ P ∨ (Q ∨ R)
( =⇒ ) Let’s prove the first conditional statement. Suppose P ∨ (Q ∨ R) is True.
This means either P is True or Q ∨ R is True. This gives us two cases.
4.6. LOGICAL EQUIVALENCE 267

1. Suppose P is True. This means P ∨ Q is True, by definition. Thus,

(P ∨ Q) ∨ R is True, also by definition.

2. Suppose Q ∨ R is True. This means either Q is True or R is True. Again,

this gives us two cases.

(a) Suppose Q is True. This means P ∨ Q is True, by definition. Thus,

(P ∨ Q) ∨ R is True, also by definition.
(b) Suppose R is True. This means (P ∨ Q) ∨ R is True, by definition.

In any case, we find that (P ∨ Q) ∨ R is True. Thus, this conditional statement

is True.
(⇐=) Let’s prove the second conditional statement. Suppose (P ∨ Q) ∨ R is
True. This means either P ∨ Q is True or R is True. This gives us two cases.

1. Suppose P ∨ Q is True. This means either P is True or Q is True. This

gives us two cases.

(a) Suppose P is True. This means P ∨ (Q ∨ R) is true, by definition.

(b) Suppose Q is True. This means Q ∨ R is True, by definition. Thus,
P ∨ (Q ∨ R) is true, also by definition.

2. Suppose R is True. This means Q ∨ R is True, by definition. Thus, P ∨

(Q ∨ R) is True, also by definition.

In any case, we conclude that P ∨ (Q ∨ R) is True. Thus, this conditional

statement is True.
Since we have shown both conditional statements hold, we conclude the two
sides are, indeed, logically equivalent.

Okay, what have we accomplished with these proofs? What have we proven,
and how? Why did it work?
Let’s mention a consequence of these proofs, before going on to discuss and
compare the proffs, themselves. We proved that the “∧” and “∨” connectives are
associative, so the order in which we evaluate parenthetical statements involving
only one such connective does not matter. For example, we now know that
“P ∧ (Q ∧ R)” has the same meaning as “(P ∧ Q) ∧ R”. Accordingly, in the
future, we will just write these statements without the parentheses: “P ∧Q∧R”.

Reflecting: Truth Tables vs. Semantics

Let’s talk about the truth tables first. Since P , Q, and R are logical statements,
they are each, individually, True or False. The eight rows of the truth tables
consider all possible assignments of truth values to those three constituent state-
ments. The first three columns tell us whether P, Q, R are True or False. The
next two columns correspond to the more complicated constituent parts of the
logical statements in the claim, and the last two columns correspond to the two
parts of the actual claim in the theorem. By comparing those last two columns,
268 CHAPTER 4. LOGIC

we can decide whether or not those two statements are logically equivalent.
(Remember that “logically equivalent” means “has the same truth value as, no
matter the assignment of truth values to P and Q and R”. Thus, observing
that the two columns have identicals entries, row by row, is sufficient to show
that the two statements are logically equivalent.)
Next, let’s talk about the semantic proofs. How do you feel about them?
They were certainly longer, right? Disregarding that, though, did they feel
like good proofs? Were they clear? Correct, even? Reread the proofs above
and think about these questions. We will emphasize that they are fully correct
proofs. The use of cases is essential when trying to prove a disjunction (an
“or” statement) holds. When we suppose something is True and deduce that
something else is True, that’s how we prove a conditional statement is True. We
will further analyze these techniques very soon, but we hope that giving you a
first example like this will help you later on.
For the remainder of this section, we will use a truth table to verify simple
claims like these. The proofs are much shorter that way! We are sure that you
can go through the details of a semantic proof, like the ones we gave above, if
you need further convincing or extra practice with interpreting symbolic logical
claims as English sentences.

4.6.4 Proving Logical Equivalences: Distributive Laws

In arithmetic, you’ve used the fact that multiplication distributes across ad-
dition. That is, we know that

.
∀x, y, z ∈ R x · (y + z) = x · y + x · z

Hey, look, we wrote this in symbolic notation! Do you see why it says what you
already know about the distributive property?
Here we will state and prove two similar laws. They will tell us that the
logical connectives “∧” and “∨” also distribute across each other.

Theorem 4.6.7. Let P , Q, and R be logical statements. Then

P ∧ (Q ∨ R) ⇐⇒ (P ∧ Q) ∨ (P ∧ R)

and

P ∨ (Q ∧ R) ⇐⇒ (P ∨ Q) ∧ (P ∨ R)

Proof. We will use truth tables to verify these two claims. Observe, for the first
4.6. LOGICAL EQUIVALENCE 269

claim, that

P Q R Q∨R P ∧Q P ∧R P ∧ (Q ∨ R) (P ∧ Q) ∨ (P ∧ R)
T T T T T T T T
T T F T T F T T
T F T T F T T T
T F F F F F F F
F T T T F F F F
F T F T F F F F
F F T T F F F F
F F F F F F F F

Notice that the last two columns are identical, thus proving the desired logical
equivalence.
Observe, for the second claim, that

P Q R Q∧R P ∨Q P ∨R P ∨ (Q ∧ R) (P ∨ Q) ∧ (P ∨ R)
T T T T T T T T
T T F F T T T T
T F T F T T T T
T F F F T T T T
F T T T T T T T
F T F F T F F F
F F T F F T F F
F F F F F F F F

Again, notice that the last two columns are identical, thus proving the desired
logical equivalence.

4.6.5 Proving Logical Equivalences: De Morgan’s Laws

(Logic)
Let’s prove some logical equivalences involving negations. The following two
laws are named for the British mathematician Augustus De Morgan. He
is credited with establishing these logical laws, as well as introducing the term
mathematical induction! We are grateful and indebted to his work in math-
ematics.
De Morgan’s Laws for Logic state some logical equivalences about negating
conjunctions and disjunctions.

Theorem 4.6.8. Let P and Q be logical statements. Then

¬ P ∧ Q ⇐⇒ ¬P ∨ ¬Q

and
¬ P ∨ Q) ⇐⇒ ¬P ∧ ¬Q
270 CHAPTER 4. LOGIC

Proof. We prove the first claim by a truth table:

P ¬P Q ¬Q P ∧Q ¬(P ∧ Q) ¬P ∨ ¬Q
T F T F T F F
T F F T F T T
F T T F F T T
F T F T F T T
And then we prove the second claim by a truth table:
P ¬P Q ¬Q P ∨Q ¬(P ∨ Q) ¬P ∧ ¬Q
T F T F T F F
T F F T T F F
F T T F T F F
F T F T F T T

These two laws are extremely useful! In fact, we can use them to prove
similar statements about sets.

4.6.6 Using Logical Equivalences: DeMorgan’s Laws (Sets)

The following statements “look a lot like” the statements in DeMorgan’s Laws
for Logic we saw above. We will see exactly why they look so similar when we
see the proof!
Theorem 4.6.9. Let A and B be any sets, and suppose that A, B ⊆ U , so the
complement operation is defined in the context of U . Then,
A∩B =A∪B
and
A∪B =A∩B
We will prove these claims using logical equivalences and DeMorgan’s Laws
for Logic. Our method will be to show that, in either case, the property of being
an element of the set on the left-hand side of the equation is logically equivalent
to being an element of the set on the right-hand side. This proves both parts of
a double-containment argument in one fell swoop.
Proof. Let’s prove the first set equality. Let x ∈ U be arbitrary and fixed. Then,
x ∈ A ∪ B ⇐⇒ x ∈
/ A∪B Definition of complement
⇐⇒ ¬ (x ∈ A ∪ B) Definition of ∈
/
⇐⇒ ¬ [(x ∈ A) ∨ (x ∈ B)] Definition of ∪ and ∨
⇐⇒ ¬(x ∈ A) ∧ ¬(x ∈ B) DeMorgan’s Law for Logic
⇐⇒ (x ∈
/ A) ∧ (x ∈
/ B) Definition of ∈
/
⇐⇒ (x ∈ A) ∧ (x ∈ B) Definition of complement
⇐⇒ x ∈ A ∩ B Definition of ∧ and ∩
4.6. LOGICAL EQUIVALENCE 271

Remember that“ ∧” is a logical operation, while “∩” is a set operation. We had

to be careful about which one made sense in every sentence we wrote. Also,
notice that we used DeMorgan’s Law for Logic in the middle of the proof, to
convert a negation of a disjunction into the conjunction of two negations.
This chain of logical equivalences shows that

x ∈ A ∪ B ⇐⇒ x ∈ A ∩ B

so, in the context of U , the property of being an element of A ∪ B is logically

equivalent to the property of being an element of A ∩ B. Therefore,

A∪B =A∩B

Let’s prove the second equality now, with a similar method. Let x ∈ U be
arbitrary and fixed. Then,

x ∈ A ∩ B ⇐⇒ x ∈
/ A∩B Definition of complement
⇐⇒ ¬ (x ∈ A ∩ B) Definition of ∈
/
⇐⇒ ¬ [(x ∈ A) ∧ (x ∈ B)] Definition of ∩ and ∧
⇐⇒ ¬(x ∈ A) ∨ ¬(x ∈ B) DeMorgan’s Law for Logic
⇐⇒ (x ∈
/ A) ∨ (x ∈
/ B) Definition of ∈
/
⇐⇒ (x ∈ A) ∨ (x ∈ B) Definition of complement
⇐⇒ x ∈ A ∪ B Definition of ∨ and ∪

This chain of logical equivalences shows that

x ∈ A ∩ B ⇐⇒ x ∈ A ∪ B

so, in the context of U , the property of being an element of A ∩ B is logically

equivalent to the property of being an element of A ∪ B. Therefore,

A∩B =A∪B

Voilà! We have proven both equalities stated in the theorem.

Notice the striking similarities between the two proofs. They used exactly
the same method, and the only real difference is flipping a “∩” to a “∪”, and vice-
versa. Because we proved something about how to do this already (DeMorgan’s
Laws for Logic), we can cite that result and make this proof short and sweet.
Wouldn’t you agree this is far easier and more concise than writing out a full
double-containment proof for these two claims? (Try it!)

4.6.7 Proving Set Containments via Conditional State-

ments
Whenever you can, go ahead and use the method we used in the previous sec-
tion, with DeMorgan’s Laws for Logic and Sets; that is, feel free to prove set
272 CHAPTER 4. LOGIC

relationships via conditional statements and logical equivalences. In general,

when you’re proving an equality, you need to be sure that all of your claims
really are “⇐⇒” claims. In the previous section, we only applied definitions
and a theorem about logical equivalences, so we were positive about all of the
directions of the “⇐⇒” arrows in the proof. Whenever you write a proof like
this, read over it again once you’re done and ask yourself at every line, “Does
this actually work? Does the implication work both ways here?”
Let’s see another example of this technique in action. It will be slightly more
complicated, in that we have to define some variable propositions because the
claim given is not fundamentally identical to DeMorgan’s Laws for Logic. We
will, though, invoke a logical law that we just proved, and use it to establish a
sets law.
Proposition 4.6.10. Let A, B, C be any sets, with A, B, C ⊆ U , where U is a
universal set. Then,
A ∩ (B − C) = (A ∩ B) − C
Much like the previous example, DeMorgan’s Laws for Sets, we will establish
a logical equivalence between being an element of the left-hand side and being
an element of the right-hand side. (Again, this is like proving both sides of a
double-containment proof all at once.) To do this, we will just establish some
variable propositions that refer to the properties of being an element of A, B,
and C, respectively. From there, the result will follow from a logical law.
Proof. Let A, B, C be any sets, with A, B, C ⊆ U , where U is a universal set.
We define the following variable propositions:

Let P (x) be “x ∈ A”
Let Q(x) be “x ∈ B”
Let R(x) be “x ∈ C”

Let x ∈ U be arbitrary and fixed. With these definitions, we can write the
following chain of logical equivalences (where “Defn” is just space-saving short-
hand for “Definition”):

x ∈ A ∩ (B − C) ⇐⇒ x ∈ A ∧ (x ∈ B − C) Defn of ∩
⇐⇒ x ∈ A ∧ (x ∈ B ∧ x ∈
/ C) Defn of −
⇐⇒ P (x) ∧ (Q(x) ∧ ¬R(x)) Defn of P (x), Q(x), R(x), ∈
/
⇐⇒ (P (x) ∧ Q(x)) ∧ ¬R(x) Associative Law for ∧
⇐⇒ (x ∈ A ∧ x ∈ B) ∧ x ∈
/C Defn of P (x), Q(x), R(x)
⇐⇒ x ∈ A ∩ B ∧ x ∈
/C Defn of ∩
⇐⇒ x ∈ (A ∩ B) − C Defn of −

This shows that

x ∈ A ∩ (B − C) ⇐⇒ x ∈ (A ∩ B) − C
4.6. LOGICAL EQUIVALENCE 273

holds True for any element x in the universe U . Therefore,

A ∩ (B − C) = (A ∩ B) − C

Think about why we needed to make sure all of these claims are truly if and
only if statements. We are ensuring that any element x that is an element of a
set on one side of the equality is also necessarily an element of the set on the
other side; but, furthermore, we are ensuring that any element x that is not an
element of one set is also not an element of the other set. The biconditional
statements “go both ways”, so we prove both the “is an element of” and “is not
an element of” parts of the claim all at once.
To illustrate our previous warnings, consider the following claim as an ex-
ample of a proof where one “direction” of a ⇐⇒ claim fails.

Proposition 4.6.11. Let X, Y, Z be any sets, with X, Y, Z ⊆ U , for some

universal set U . Then, the following containment holds:

(X ∪ Y ) − Z ⊆ X ∪ (Y − Z)

You might recognize this claim as Problem 3.11.17! In that problem, we

asked you to prove this claim using a containment argument, taking an ar-
bitrary x ∈ U and supposing it is an element of the left-hand side set, then
deducing it must also be an element of the right-hand side set. We will do (es-
sentially) the same thing here, but the argument will be recast in logical terms
and symbols. We will do this to (1) give us more practice with making these
types of arguments, but also (2) to recognize precisely where in the argument
the “reverse” direction fails. Remember that, in Problem 3.11.17, we also asked
you to find an example that shows that the ⊇ direction is not necessarily True.
This means that the logical argument working in that direction would break
down somewhere. We will see precisely where that is, and we can use it to help
us come up with that required counterexample.

Proof. Let X, Y, Z be any sets, with X, Y, Z ⊆ U , for some universal set U .

Let x ∈ U be arbitrary and fixed. We can write the following chain of logical
equivalences:

x ∈ (X ∪ Y ) − Z ⇐⇒ x ∈ X ∪ Y ∧ x ∈
/Z Defn of −
⇐⇒ (x ∈ X ∨ x ∈ Y ) ∧ x ∈
/Z Defn of ∪
⇐⇒ (x ∈ X ∧ x ∈
/ Z) ∨ (x ∈ Y ∧ x ∈
/ Z) Distr. Law

Scratch work:
274 CHAPTER 4. LOGIC

From here, what further logical equivalences could we assert? We could

simplify the right-hand side and express it as

x∈X −Z ∨ x∈X −Z

and, therefore,
x ∈ (X − Z) ∪ (Y − Z)
This is not what the claim was, but this procedure so far would be a valid
proof of a different claim, namely that

(X ∪ Y ) − Z = (X − Z) ∪ (Y − Z)

However, our right-hand side is

X ∪ (Y − Z)

but we are not trying to prove an equality, merely a containment. Thus,

the goal of the rest of our proof is to prove this conditional claim:
 
(x ∈ X ∧ x ∈ / Z) ∨ (x ∈ Y ∧ x ∈ / Z) =⇒ x ∈ X ∪ (Y − Z)

To help us figure out how to get there, let’s do some scratch work here to
rewrite the statement on the right-hand side; then, we can see how to get
there from where we are already:

x ∈ X ∪ (Y − Z) ⇐⇒ x ∈ X ∨ x ∈ Y − Z Defn of ∪
⇐⇒ x ∈ X ∨ (x ∈ Y ∧ x ∈
/ Z) Defn of −

This is similar to the last logical equivalence we derived up above, but this
one differs in the term on the left. Can you see how the one up above
implies this one? Think about it, and then read on for the rest of our
proof, resumed.

Now, we want to show that

 
(x ∈ X ∧ x ∈/ Z) ∨ (x ∈ Y ∧ x ∈
/ Z) =⇒ x ∈ X ∪ (Y − Z)

To do this, let’s suppose the expression on the left-hand side is True. This means
either
x∈X ∧ x∈ /Z
or
x∈Y ∧ x∈
/Z
(or possibly both). Thus, we have two cases:
4.6. LOGICAL EQUIVALENCE 275

1. Suppose the first expression is True, so that x ∈ X ∧ x ∈

/ Z. This certainly
means that x ∈ X, and thus x ∈ X ∨ x ∈ Y − Z holds.
2. Suppose the second expression is True, so that x ∈ Y ∧ x ∈
/ Z. This means
that x ∈ Y − Z, and thus x ∈ X ∨ x ∈ Y − Z holds.
In either case, we find that x ∈ X ∨ x ∈ Y − Z holds, and therefore,

x ∈ X ∪ (Y − Z)

holds, in either case, by the definition of ∪.

Overall, this shows that

x ∈ (X ∪ Y ) − Z =⇒ x ∈ X ∪ (Y − Z)

holds for every element x ∈ U . Therefore, by the definition of ⊆, we have

(X ∪ Y ) − Z ⊆ X ∪ (Y − Z)

Recognizing where we are, and where we wanted to go, helped us finish this
proof. We had no hope of completing it using logical equivalences alone because,
in fact, the sets given in the claim are not always equal! Looking back at the
proof, can we identify the step whose logical equivalence was invalid, and can
we use it to help construct a counterexample to the (False) claim that those sets
are always equal?
We had reached as far as this valid statement

(x ∈ X ∧ x ∈
/ Z) ∨ (x ∈ Y ∧ x ∈
/ Z)

and we used it to deduce this statement

x ∈ X ∨ (x ∈ Y ∧ x ∈
/ Z)

It seems clear, from our argument in the proof, that the first statement does
imply the second one; that is, if we suppose the first statement holds, we can
figure out that the second statement one holds, as well. The only difference
between them is in the first term, and knowing two parts of an “∧” statement
hold certainly lets us conclude a particular one of them holds.
This deduction does not work in the other direction. Suppose that second
statement holds. If it’s the right term that is valid—that x ∈ Y ∧ x ∈ / Z—
then this also makes the first statement hold. However, since we have an “∨”
statement, we have to consider the case where the left term is the one that holds.
In that case, knowing only x ∈ X does not let us deduce that x ∈ X ∧ x ∈ /Z
holds. Supposing an “∧” holds lets us deduce either one of its parts holds, but
just knowing only one part cannot tell us that both hold!
We can use this to construct a counterexample. We see that we just need to
ensure that there is some particular element x ∈ U that satisfies the left term
276 CHAPTER 4. LOGIC

in the second statement, namely x ∈ X, but does not satisfy the left term in
the first statement, namely x ∈ X ∧ x ∈ / Z. Said another way, we just need to
ensure that there is an element x ∈ X ∩ Z. The following example accomplishes
exactly that.
Example 4.6.12. We claim that

(X ∪ Y ) − Z ⊆ X ∪ (Y − Z)

holds for any sets X, Y, Z, but equaltiy need not hold. See the proof of Propo-
sition 4.6.11 to see why the containment claimed above does hold.
Now, consider the following example. Let’s define

X = {1}
Y = {2}
Z = {1, 2}

Notice that

(X ∪ Y ) − Z = ({1} ∪ {2}) − {1, 2} = {1, 2} − {1, 2} = ∅

and
X ∪ (Y − Z) = {1} ∪ ({2} − {1, 2}) = {1} ∪ {∅} = {1}
Since ∅ ⊂ {1} (a proper) subset, we conclude that

(X ∪ Y ) − Z 6= X ∪ (Y − Z)

in this case. This shows that equality need not hold in the claim above.
This strategy now lets us go back and complete many proofs involving sets
in a more efficient and rigorous manner! Rather than fumbling through the
linguistics of “ands” and “ors”, we can use our logical notation and laws that
we have proven. Many of the exercises in this section deal with sets, specifically
because of this. If you need to flip back to Chapter 3 and remind yourself of
any relevant definitions, go right ahead!

4.6.8 Questions & Exercises

Remind Yourself
Answering the following questions briefly, either out loud or in writing. These
are all based on the section you just read, so if you can’t recall a specific defi-
nition or concept or example, go back and reread that part. Making sure you
can confidently answer these before moving on will help your understanding and
memory!
(1) What are the Associative Laws for Logic?
(2) What are the Distributive Laws for Logic?
4.6. LOGICAL EQUIVALENCE 277

(3) What are DeMorgan’s Laws for Logic? What are DeMorgan’s Laws for
Sets? How are they related?

(4) What is the difference between a necessary and a sufficient condition?

(5) What happens when a condition is both necessary and sufficient?

Try It
Try answering the following short-answer questions. They require you to actu-
ally write something down, or describe something out loud (to a friend/class-
mate, perhaps). The goal is to get you to practice working with new concepts,
definitions, and notation. They are meant to be easy, though; making sure you
can work through them will help you!

(1) We used Truth Tables to prove DeMorgan’s Laws for Logic. Can you come
up with a semantic proof? Can you explain DeMorgan’s Laws to a non-
mathematician friend and convince them they are valid?

(2) Let P (x) be the variable proposition “x is an integer that is divisible by

10”. Come up with two necessary conditions and two sufficient conditions
for this statement.

(3) Let A, B, C be any sets, where A, B, C ⊆ U , for some universal set U .

Use logical equivalences and logical laws to prove the following claims.

(a) A ∩ (B ∪ C) = (A ∩ B) ∪ (A ∩ C)
(b) (A ∪ B) ∩ A = B − A
(c) A ∪ B = A ∩ B
(d) (A − B) ∩ C = A − (B ∪ C)

(4) Use conditional statements and logical equivalences to prove that the con-
tainment

A − (B ∪ C) ⊆ A ∩ B ∩ C
holds for any sets A, B, C.
Then, find an example that shows that equality need not hold.
(Hint: In general, a helpful idea in constructing a strict set containment is
to see if you can make one side the empty set.)

(5) Let D, E, F be any sets. Consider the sets

D − (E − F )

and
(D − E) − F
278 CHAPTER 4. LOGIC

How do they compare? Are they always equal? Is one always a subset of
the other, or vice-versa?
Clearly state your claims, then either prove them or provide relevant coun-
terexamples.

4.7 Negation of Any Mathematical Statement

We saw already how to negate quantified statements. With DeMorgan’s Laws
in hand, we now know how to negate ∧ and ∨ statements. What’s left? Aha,
conditional statements!

4.7.1 Negating Conditional Statements

Consider a claim of the form P =⇒ Q. It says that whenever P is true, Q is
also true. How do we negate such a statement? What would the logical negation
even mean? Think back to how we defined “ =⇒ ” as a logical connective. In
which cases did we get to call the speaker of the conditional statement a liar.
Those are precisely the cases where we would say the logical negation is True.
The only such case was when the hypothesis P was True but the conclusion Q
was False.
To prove this equivalence, we need to remember the way to write P =⇒ Q
as an “∨” statement:

(P =⇒ Q) ⇐⇒ (¬P ∨ Q)

This will help us in the proof of the following claim.

Lemma 4.7.1. The logical negation of a conditional statement is given by

¬(P =⇒ Q) ⇐⇒ P ∧ ¬Q

Proof. Observe that

¬(P =⇒ Q) ⇐⇒ ¬(¬P ∨ Q) Logical equivalence proven already

⇐⇒ ¬(¬P ) ∧ ¬Q DeMorgan’s Law for Logic
⇐⇒ P ∧ ¬Q since ¬(¬P ) ⇐⇒ P

This makes intuitive sense: to show that a conditional claim is False, we

need to find a case where the hypothesis holds but the conclusion fails.
Despite the risk of putting bad ideas into your head that weren’t already
there, we are going to point out some statements that are NOT logically equiv-
alent to ¬(P =⇒ Q). These are common mistakes that we see students use
fairly often. Let’s check them out and see why they don’t actually work. For
each of them, keep in mind that we want the logical negation ¬(P =⇒ Q) to
4.7. NEGATION OF ANY MATHEMATICAL STATEMENT 279

be guaranteed to have the exact opposite truth value of the original statement
P =⇒ Q. In each of these cases, we can see that this relationship would not
hold.
• ¬P =⇒ Q
This conditional statement has no logical connection to the original claim,
P =⇒ Q. Remember that the statement P =⇒ Q makes no claim
about whether Q is true or not in the cases where P is False. (Think
about the “If it is raining, then I carry my umbrella” example. If it’s not
raining, who knows what I’m carrying!) Thus, why would we expect Q to
be necessarily true in those cases, like this statement says?
• P =⇒ ¬Q
Again, this conditional statement has no logical connection to the original
claim. Think about the umbrella example again. This statement would
say “If it is raining, then I will not be carrying an umbrella.” Is that what
it means for the original claim to be False? Definitely not!
• P =⇒
6 Q
This one is more subtle. A mathematician would read “P =⇒ 6 Q” as
“P does not necessarily imply Q”. That is, it would say that there are
assignments of truth values where P =⇒ Q is valid and there are assign-
ments where P =⇒ Q is invalid; those cases would depend on what the
individual statements P and Q are. This is a somewhat meaningful claim
to make, depending on the situation, but it is not, strictly speaking, the
logical negation of the original claim.
In particular, we run into an issue when we try to take the logical negation
of this statement. What does it mean to say that “It is not the case that P
does not necessarily imply Q”? Does that mean there exists cases where
P does not imply Q but there are also cases where P does imply Q? That
sounds an awful lot like the actual claim P =⇒ 6 Q, itself . . .
6 . It does have
For these reasons, we want to avoid using this notation: =⇒
some kind of meaning in mathematics, but it is not really well-defined in
a symbolic logical sense. And in any event, it is definitely not the logical
negation of =⇒ .
Now that we have these common errors out of the way, let’s stress the correct
negation of P =⇒ Q. We find that it’s quite helpful to remember the “∨”
version of a conditional statement; from there, it’s easy to apply DeMorgan’s
Law and negate the statement:

¬(P =⇒ Q) ⇐⇒ ¬(¬P ∨ Q) ⇐⇒ P ∧ ¬Q
280 CHAPTER 4. LOGIC

Negating “⇐⇒”
To negate a biconditional statement, we just write it as a conjuction of two
conditional statements:

¬(P ⇐⇒ Q) ⇐⇒ [¬(P =⇒ Q) ∨ ¬(Q =⇒ P )] ⇐⇒ (P ∧ ¬Q) ∨ (Q ∧ ¬P )

If you do any kind of comptuer programming, you might recognize the statement
on the right as the XOR operator! It says that exactly one statement is True,
either P or Q, but not both.

4.7.2 Negating Any Statement

That’s it, right? We have now discussed how to negate any fundamental math-
ematical claim: ∃, ∀, ∧, ∨, and =⇒ . Everything else we write will be a
combination of these basic claims, so we should be able to just apply these
techniques over and over and negate any statement we come across. Essentially,
we just read the statement left to right and negate every piece in turn. Come
across a “∃”? Just switch it to a “∀” and negate the property that comes after!
Come across an “∨”? Negate both sides and switch it to and “∧”! Come across
a conditional? Apply the technique we just showed above!
Let’s see several examples to get the idea.
Example 4.7.2. Find the logical negation of

.
∀x ∈ R x < 0 ∨ x > 0

This statement says “Every real number x satisfies either x < 0 or x > 0.”
The logical negation is

. .
¬ (∀x ∈ R x < 0 ∨ x > 0) ⇐⇒ ∃x ∈ R x ≥ 0 ∧ x ≤ 0

Notice that we applied DeMorgan’s Law for Logic to negate the ∨ claim on the
right-hand side, and we used the fact that x 6> 0 is logically equivalent to x ≤ 0.
We see that this negation is True because 0 ∈ R and 0 ≤ 0 and 0 ≥ 0. Thus,
the original statement was False.
Example 4.7.3. Find the logical negation of

.
∃n ∈ N n ≥ 10 ∧
√
n≤3

statement says “There exists a natural number n that satisfies both n ≥ 10

This√
and n ≤ 3.
The logical negation is

. √
∀n ∈ N n < 10 ∨ n > 3

That
√ is, the logical negation says “Every natural number n satisfies either n > 10
or n < 3.
4.7. NEGATION OF ANY MATHEMATICAL STATEMENT 281

Example 4.7.4. Find the logical negation of

. .
∃x ∈ R ∀y ∈ R x ≥ y =⇒ x2 ≥ y 2

This statement says “There exists a real number x such that whenever we have
a real number y that satisfies x ≥ y, we may conclude that x2 ≥ y 2 ”.
The logical negation is

. .
∀x ∈ R ∃y ∈ R x ≥ y ∧ x2 < y 2

Can you prove that this logical negation is, in fact, the True statement? Try it!

Example 4.7.5. Find the logical negation of

. .
∀X ∈ P(Z) (∀x ∈ X x ≥ 1) =⇒ X ⊆ N

This statement says that “For every subset X of the integers Z, if every element
x of the set X satisfies x ≥ 1, then X is a subset of the natural numbers N.”
The logical negation is

. .
∃X ∈ P(Z) (∀x ∈ X x ≥ 1) ∧ X 6⊆ N

This statement says that “There is a subset X ⊆ Z such that every element
x ∈ X satisfies x ≥ 1 and yet X 6⊆ N.” We could even rewrite the last part
further by noting that

X 6⊆ N ⇐⇒ ∃y ∈ X y ∈
/N .
although this wouldn’t be totally essential.
Which statement (the original or the negation) is True? Can you prove it?

Compare the statement used in the example above with the following one:

. .
∀X ∈ P(Z) ∀x ∈ X x ≥ 1 =⇒ X ⊆ N

The only difference between them is the location of the parentheses, but
this completely changes the statement’s meaning!
The statement used in the example asserts something about every subset
of the integers. That is, no matter what subset X ⊆ Z is introduced, the
statement says that if that set has the property that all of its elements are
at least 1, then that set X is actually a subset of N, as well.
The new statement written in this box says something else: no matter
what subset X ⊆ Z is introduced and, furthermore, no matter what element
282 CHAPTER 4. LOGIC

x of that set X is introduced, the statement says that if that element x is

at least 1, then that set X is a subset of N, as well.
Do you see why this is different? The issue is where the “if” happens:
where does the quantification end and the conditional statement begin?
The first statement, from the above example, puts the quantification over
elements of X inside the “if” part of the conditional statement. The sec-
ond statement, in this box, puts that quantification before the conditional
statement entirely.
We claim that this second version, in this box, is False, and we encourage
you to figure out why (if you haven’t already).

Example 4.7.6. Let O(x) be the proposition “x is odd”, and let E(x) be the
proposition “x is even”. Find the logical negation of the statement

.
∀x, y ∈ Z O(x · y) ⇐⇒ (O(x) ∧ O(y))

This statement says that “For every two integers x and y, their product is odd
if and only if they are both odd, themselves”.
Before we find the logical negation, remember the ⇐⇒ means “ =⇒ and
⇐=”. Let’s rewrite the claim that way first, so that we can negate it properly:

.
∀x, y ∈ Z [O(x · y) =⇒ (O(x) ∧ O(y))] ∧ [(O(x) ∧ O(y)) =⇒ O(x · y)]

The logical negation is

.
 
¬ ∀x, y ∈ Z [O(x · y) =⇒ (O(x) ∧ O(y))] ∧ [(O(x) ∧ O(y)) =⇒ O(x · y)]

.
⇐⇒ ∃x, y ∈ Z ¬ [O(x · y) =⇒ (O(x) ∧ O(y))]
∨ ¬ [(O(x) ∧ O(y)) =⇒ O(x · y)]
⇐⇒ ∃x, y ∈ Z. [O(x · y) ∧ ¬(O(x) ∧ O(y))] ∨ [(O(x) ∧ O(y)) ∧ ¬O(x · y)]
⇐⇒ ∃x, y ∈ Z. [O(x · y) ∧ (E(x) ∨ E(y))] ∨ [(O(x) ∧ O(y)) ∧ E(x · y)]

That is, the logical negation says “There exist integers x and y such that either
their product is odd and yet (at least) one of them is even, or they are both odd
and yet their product is even.
Can you prove which one of these claims is True?

4.7.3 Questions & Exercises

Remind Yourself
Answering the following questions briefly, either out loud or in writing. These
are all based on the section you just read, so if you can’t recall a specific defi-
nition or concept or example, go back and reread that part. Making sure you
can confidently answer these before moving on will help your understanding and
memory!
4.7. NEGATION OF ANY MATHEMATICAL STATEMENT 283

(1) How is a mathematical statement related to its logical negation?

(2) What is the logical negation of a conditional statement?
(3) Consider the statement P =⇒ Q. What is its contrapositive? What is the
logical negation of that contrapositive? Can you see that it must have the
same truth value as the logical negation of the original statement?
(4) What is the logical negation of an if and only if statement, P ⇐⇒ Q? Why
does this make sense, considering what such a statement says about the
truth values of P and Q?

Try It
Try answering the following short-answer questions. They require you to actu-
ally write something down, or describe something out loud (to a friend/class-
mate, perhaps). The goal is to get you to practice working with new concepts,
definitions, and notation. They are meant to be easy, though; making sure you
can work through them will help you!
(1) Write out the logical negation of each of the following mathematical state-
ments.
Then, determine the truth value of each statement.
(If you’re feeling ambitious, formally prove/disprove each statement!)
. .
(a) ∃x ∈ N ∀y ∈ N y − x2 ≥ 0
(b) ∃x ∈ Z. ∀y ∈ R. xy = 0
(c) ∃x ∈ Z. ∀y ∈ Z. (y 6= 0 =⇒ xy > 0)
(d) ∀a, b ∈ Q. ab ∈ Z =⇒ (a ∈ Z ∨ b ∈ Z)
(e) ∀x ∈ R. x > 0 =⇒ ∃y ∈ R. y < 0 ∧ xy > 0

(f) ∀x ∈ R. x +  = 2 ⇐⇒ x = 1
1
 
x

(2) Let A = {1, 2, 3, 4} and B = {2, 3}.

What is the difference between the following two statements? Determine
the truth value of each one.
Then, negate each one, and explain how those negations also differ. What
are their truth values?
. .
(a) ∀x ∈ A ∀y ∈ B (x ≥ y =⇒ x2 ≥ 4)
(b) ∀x ∈ A. (∀y ∈ B . x ≥ y) =⇒ x 2
≥4
(3) Let P = {x ∈ R | x > 0}. Write the logical negation of each of the following
statements, and determine their truth values.

. . . .
   
(a) ∀ε ∈ P ∀x ∈ P ∃δ ∈ P ∀y ∈ R |x − y| < δ =⇒  x1 − y1  < ε
 
284 CHAPTER 4. LOGIC

. . . .
   
(b) ∀ε ∈ P ∃δ ∈ P ∀x ∈ P ∀y ∈ R |x − y| < δ =⇒  x1 − y1  < ε
 

Hint/suggestion: A statement like |a| < b can be written as −b < a < b.

Also, a statement like a < b < c can be written as a < b ∧ b < c. This
might help you rewrite the statements when determining their truth values.

(4) Let P (n) be the proposition “n is odd” and let Q(n) be the proposition
“n2 − 1 is divisible by 8”.
Write the logical negation of the statement

.
∀n ∈ N P (n) ⇐⇒ Q(n)

and determine its truth value.

(5) Let P = {x ∈ R | x > 0}. Write the logical negation of each of the following
statements, and determine their truth values.

. . .
(a) ∀ε ∈ P ∃δ ∈ P ∀x ∈ R 0 < x < δ =⇒ 1
x > 10
ε

(b) ∀ε ∈ P . ∃x ∈ R. ∀n ∈ N. (−1)n

n > x =⇒ n <ε
(c) ∀ε ∈ P . ∃x ∈ R. ∀n ∈ N. (−1)n

n > x ⇐⇒ n <ε

4.8 [Optional Reading] Truth Values and Sets

There is a convenient and interesting relationship between sets (and their corre-
sponding relationships and operations) and logical truth values (and their cor-
responding relationships and connectives). We will mention it here and demon-
strate some examples, and leave it to you to investigate it further, if you’d like.
We won’t need these kinds of ideas in the rest of our work, but we believe that
thinking about these ideas and sorting them out in your head will really help
your understanding of the fundamentals of logic, as well as sets.
Suppose we have two variable propositions, P (x) and Q(x). Further, suppose
these propositions make sense for any input x that comes from some universal
set U . (This set U depends on the specific statements inside P (x) and Q(x), of
course, but we don’t really care what they are for this general discussion.) For
each of these propositions, we can define a truth set; that is, we can consider
the set of all instances x from the universe U that make those propositions
evaluate as True. We define

TP = {x ∈ U | P (x) is True}
TQ = {x ∈ U | Q(x) is True}

Perhaps the propositions P (x) and Q(x) are related somehow. Let’s suppose
that, in fact,
.
∀x ∈ U P (x) =⇒ Q(x)
4.8. TRUTH VALUES AND SETS 285

holds. What does this say about those truth sets? This conditional statement
says that any x that satisfies P (x) (i.e. makes P (x) True) must also satisfy
Q(x). Written another way, using those truth sets, we have
.
∀x ∈ U x ∈ TP =⇒ x ∈ TQ
This is precisely the definition of “subset”! What we have just discovered is
that
TP ⊆ TQ
when that conditional statement above holds.
Let’s suppose even further, now, that
.
∀x ∈ U P (x) ⇐⇒ Q(x)
holds. We just discovered that TP ⊆ TQ , and applying the exact same reasoning
to the “other direction” of that ⇐⇒ statement (that is, the Q(x) =⇒ P (x)
part of it) will show us that TQ ⊆ TP , as well. By the definition of set equality,
this means that
TP = TQ
when that biconditional statement above holds.
How else might we combine the propositions P (x) and Q(x)? Let’s consider
the proposition P (x) ∧ Q(x). What are the instances x that make this conjunc-
tion True? How can we characterize those instances in terms of the truth sets we
defined? Think about it for a minute, and you’ll see that all of those instances
are characterized by the intersection of those sets; we need both P (x) and Q(x)
to hold, so we need an instance that comes from both of the truth sets.
Similarly, we can consider the conjunction P (x)∨Q(x). An instance x makes
this statement True when that x makes at least one of the propositions True.
Thus, that x must come from at least one of the sets, so it must come from their
union.
Let’s summarize these relationships we have discovered:
∀x ∈ U .

P (x) =⇒ Q(x) ⇐⇒ TP ⊆ TQ

∀x ∈ U .

P (x) ⇐⇒ Q(x) ⇐⇒ TP = TQ

∀x ∈ U .

P (x) ∧ Q(x) ⇐⇒ x ∈ TP ∩ TQ

∀x ∈ U .

P (x) ∨ Q(x) ⇐⇒ x ∈ TP ∪ TQ

Can you come up with some characterizations, using truth sets, for the
following statements? Fill in the blanks!
∀x ∈ U .

P (x) ∧ ¬Q(x) ⇐⇒
.
∃x ∈ U P (x) ⇐⇒

∀x ∈ U . ¬P (x) ⇐⇒

(Careful: how are the previous statement and the next one different?)
.
∀x ∈ U ¬P (x) ⇐⇒


286 CHAPTER 4. LOGIC

4.9 Writing Proofs: Strategies and Examples

We are now prepared to fully tackle the goal we have been building towards all
along: writing PROOFS!
In this section, we will apply all of these fundamental logical principles we
have developed in this chapter. Specifically, we will learn how to use them to
write formal arguments that demonstrate the truth (or falsity) of mathematical
statements. In general, it’s hard to describe how to figure out which mathemat-
ical statements are True and which are False. In a way, though, the strategies we
develop here will help us discover truths. More importantly, though, they will
provide us with templates and guidelines for how to actually present a truth to
someone else and describe why it is, indeed, a truth.
As we have discussed, it is not enough to figure out some interesting fact
and just hope that others will trust us about it. We need to be able to explain
that fact; we need to present an argument that will convince someone else of
its truth. We don’t necessarily have to explain where it came from, or why we
cared to investigate it in the first place (although sometimes you might want to
answer these implicit questions, if you think it would help a potential reader).
In general, we just need to make sure that someone else—a peer, a classmate,
a fellow mathematician—can pick up our written proof, read it, and afterwards
be fully convinced that what we claimed to be True is, indeed, True.

Outline of this section

Mostly, we hope you will see how the ensuing strategies come directly from the
underlying logical principles associated with propositions and quantifiers and
connectives and negations. We have split the section into several subsections,
each one corresponding to a particular quantifier or connective.
When you face a mathematical statement and have to prove it, just start
reading the statement left to right. What do you encounter first? If it’s a “∃”
quantifier, look at Section 4.9.1. If it’s a “∀” quantifier, look at Section 4.9.2.
After that, what type of claim do you face? What form does the ensuing vari-
able proposition take? Is it an “∨” statement? Look at Section 4.9.3. Is it
a conditional statement? Look at section 4.9.5. Is it a conditional statement
where the hypothesis is an “∨” statement and the conclusion is an “∧” state-
ment? Look at all three Sections—4.9.3 and 4.9.4 and 4.9.5—and combine them
appropriately! In general, every proof we write from now on (except for induc-
tion proofs, which we will return to in the next chapter) will be a combination
of these strategies. Which ones you use and how you combine them depends on
the statement you’re trying to prove and how you’ve decided to approach it.
Within each subsection, we have provided some templates and some ex-
amples. You might find the templates too restrictive, perhaps too formal; we
understand, but we think that following our formats as closely as possible for
now will help you in the long run. These templates—as well as how we’ve used
them in the examples provided—are meant to emphasize the logical principles
behind these proof strategies. Working with them closely will give you extra
4.9. WRITING PROOFS: STRATEGIES AND EXAMPLES 287

practice with these logical concepts and, we strongly believe, help you adapt
them for your own uses in the future.
For each example provided, we have boxed the proof strategy in blue and
the example implementation in green and any necessary scratch work in red .
Any other discussion of the strategy or the implementation appears outside of
those boxes.
Also, several of the examples we consider in this section (and the next one)
are interesting and useful results, in their own right. You’ll notice that some of
them have a name or a descriptive title, which is meant to indicate this fact.
While the main emphasis of this section is on the proof strategies—developing
them and seeing how to use them—we encourage you to also keep in mind these
examples as interesting facts, themselves. We’ll bring up this idea again when
it’s warranted, but we’ll keep those discussions brief, so as not to distract from
the overall structure of this section.

Direct vs. Indirect methods

You will also notice that each subsection includes strategies for both direct and
indirect methods. These terms might not be familiar to you yet. All they refer
to is whether or not we try to prove a statement (1) directly by demonstrating
that it is True, or (2) indirectly by invoking the Law of the Excluded Middle,
by demonstrating that its logical negation is False.
Both forms of strategy are, in general, equally valid, but direct proofs
are typically considered subjectively better by many readers. (Sometimes, you
might write an indirect proof that is actually hiding a direct proof inside it!)
These subjective ideas will be assessed and discussed as we work through ex-
amples and ask you to write proofs on your own, in the exercises.
You’ll notice that all of our indirect proofs begin with the phrase “Assume for
sake of contradiction”, usually abbreviated as “AFSOC”. This is an important
and helpful phrase. It signals to the reader of our proof that we are going to make
an assumption but we don’t really think that the assumption is valid. Rather,
we are going to use this assumption to derive something False, a contradiction.
This will allow us to conclude that our original assumption was invalid, so its
logical negation (i.e. our original statement that we hoped to prove) is actually
True. You’ll see that we use the symbol “ × ×
× ” to indicate a contradiction, but
×
we also make sure to point out why we have found a contradiction. We don’t
leave it to the reader to figure it out!
Alright, that’s enough preamble. Let’s dive right in and WRITE SOME
PROOFS!

4.9.1 Proving ∃ Claims

An “∃” claim is one of existence. It asserts that some particular object exists
as an element of some set and that it has a certain property. To prove such
a claim, we need to exhibit such an object and verify, for our reader, that (1)
288 CHAPTER 4. LOGIC

that object is an element of the correct set and (2) that object has the correct
property.

Direct Method

Strategy:
.
Claim: ∃x ∈ S P (x)
Direct proof strategy:
Define a specific example, y = .
Prove that y ∈ S.
Prove that P (y) holds true.

Example 4.9.1. Solving a system of linear equations:

Statement: Fix a, b, c, d, e, f ∈ R with the property that ad − bc 6= 0.
We claim that one can simultaneously solve

ax + by = e (4.1)
cx + dy = f (4.2)

for some x, y ∈ R.
Define S(x, y) to be the statement “x and y simultaenously satisfy both equa-
tions, (4.1) and (4.2), above”. Then we claim

.
∃x, y ∈ R S(x, y)

First, we must do some scratch work to construct the solution. Then, we

can write a proof that defines the objects x and y and shows why they work.

Scratch work:
We need ax + by = e and cx + dy = f , and we want to know which x and
y make this work.
Let’s multiply the first and second equations by the right coefficients
(namely, d and -b, respectively) so we can cancel the y terms by adding
4.9. WRITING PROOFS: STRATEGIES AND EXAMPLES 289

the two lines:

adx + bdy = de


+ −bcx − bdy = −bf

(ad − bc)x = de − bf
de−bf
Dividing tells us x = ad−bc , which is okay because ad − bc 6= 0.
Doing something similar, canceling the x terms, tells us how to get y:

acx + bcy = ce


+ −acx − ady = −af

(bc − ad)y = ce − af
af −ce
Dividing tells us y = ad−bc .

The main lesson here is that we do not need to show this scratch work in our
proof below! We don’t assume that a reader would care to wade through our
messy notes about how we came up with the solution to the system of linear
equations. Rather, we assume the the reader only cares about what the solution
is and why it’s a solution. Also, this makes the proof much more concise, so it
can be read more easily and quickly.

Implementation:
Proof. Since ad − bc 6= 0 (by assumption), we may define

de − bf af − ce
x= and y=
ad − bc ad − bc
and know that x, y ∈ R. Then,

(ade − abf ) + (abf − bce) ade − bce e(ad − bc)

ax + by = = = =e
ad − bc ad − bc ad − bc
(cde − bcf ) + (adf − cde) adf − bcd f (ad − bc)
cx + dy = = = =f
ad − bc ad − bc ad − bc
so S(x, y) holds.

If you’ve studied some linear algebra

 before,
 you’ll recognize the term ad−bc
a b
as the determinant of the matrix . The stipulation that ad − bc 6= 0
c d
means that we require this matrix of coefficients to have an inverse, to be “non-
290 CHAPTER 4. LOGIC

singular”. In that situation, we have a solution to the system for any e, f ∈ R.

Indirect Method (Proof by Contradiction)

This strategy relies on the logical negation of an ∃ claim:

.

¬ ∃x ∈ S P (x) ⇐⇒ ∀x ∈ S ¬P (x) .
We will assume this negation and deduce something contradictory from it, mean-
ing the negation is False so the original is True.

.
Claim: ∃x ∈ S P (x)
Indirect proof strategy:

AFSOC that for every y ∈ S, ¬P (y) holds.

Find a contradiction.

Example 4.9.2. A version of the Pigeonhole Principle:

Statement: Suppose n ∈ N and we have n real numbers, a1 , a2 , . . . , an ∈ R.
We claim that one of the numbers is at least as large as their average. That is,
n
.
∃B ∈ [n] aB ≥
1X
n i=1
ai

Proof. AFSOC none of the numbers are at least as large as the average,
i.e.
n
∀i ∈ [n] ai <. 1X
n i=1
ai

Pn
Define the constant S = i=1 ai , so that ai < Sn .
Then we can sum all of the ai s and observe that
n n
X X S S
S= ai < =n· =S
i=1 i=1
n n

This shows that the real number S is strictly less than itself: S < S. This
×
is a contradiction. ×
××
Therefore, the original assumption was false, and the claim follows.

As stated, this is a version of the Pigeonhole Principle. We will investigate

and use this principle again in Section 8.6, when we study combinatorics.
4.9. WRITING PROOFS: STRATEGIES AND EXAMPLES 291

4.9.2 Proving ∀ Claims

A “∀” claim is one of universality. It asserts that all elements of a set have some
common property. To prove such a claim, we need to show that every element
of the set has that property. To accomplish this “all at once”, we will consider
an arbitrary and fixed element of the set, and prove that it has the desired
property. Because this object is arbitary, our argument applies to every element
of the set. Because this object is fixed, we are allowed to refer to it by name
throughout the proof.

Direct Method

Strategy:
.
Claim: ∀x ∈ S P (x)
Direct proof strategy:
Let y ∈ S be arbitrary and fixed.
Prove that P (y) holds true.

Example 4.9.3. A version of the AGM Inequality:

.
Statement: ∀x, y ∈ R xy ≤ x+y 2

2

Implementation:
Proof. Let x, y ∈ R be arbitrary and fixed.
We know 0 ≤ (x − y)2 .
Multiplying out and rearranging, we get 2xy ≤ x2 + y 2 .
Adding 2xy to both sides, we get 4xy ≤ x2 + 2xy + y 2 .
Factoring, we get 4xy ≤ (x + y)2 .
Dividing by 4 and putting that into the square, we get
 2
x+y
xy ≤
2

This result is known as the AGM Inequality because it deals with the Arith-
metic Mean (AM) and the Geometric Mean (GM) of two real numbers.
x+y
The Arithmetic Mean of x and y is 2 .
√
The Geometric Mean of x and y is xy. (Notice that this only applies when
292 CHAPTER 4. LOGIC

xy ≥ 0, i.e. when x and y have the same sign be it positive or negative, or

zero.)
The AGM Inequality asserts that the AM is always at least as large as the GM.
A helpful mnemonic is to read “AGM” as “Arithmetic Mean Greater than
Geometric Mean”.
What we proved above is a slightly more general version, because it applies to
all real numbers x and y, and not just those with the same sign. Supposing that
xy ≥ 0, though, one can simply take the square root of both sides and obtain
√
the “usual” statement of the AGM Inequality: xy ≤ x+y 2 .

Indirect Method (Proof by Contradiction)

.
Claim: ∀x ∈ S P (x)
Indirect proof strategy:

AFSOC that ∃y ∈ S such that ¬P (y) holds.

Find a contradiction.

√
Example 4.9.4. 2 is irrational:
√
.
Statement: ∀a, b ∈ Z ab 6= 2
(Note: This claim is appealing
√ directly to the definition of the rational numbers
Q. It is saying that 2 ∈ / Q because that number has no representation as a
ratio of integers.)

Implementation:
√
Proof. AFSOC ∃a, b ∈ Z . a
b = 2.
a
We may assume that is given in reduced form, so that a and b have
b
no common factors. (If this were not the case, we could just reduce the
fraction and obtain such a form.)
Let such a, b ∈ Z be given.
(Note: We discuss this phrase, “let such be given”, below in Section
4.9.8. It is meant to not only assert that such an a, b ∈ Z exist, but also
that we want some particular instances of those variables so that we can
work with them for the rest of the proof.)
√ 2
This means ab = 2, so ab2 = 2.
Thus, 2b2 = a2 , so a2 is even, by definition.
Since a2 is even, this tells us a is even.
4.9. WRITING PROOFS: STRATEGIES AND EXAMPLES 293

(Note: You should prove this. We will prove it in Section 4.9.6, but you
should try to do it on your own now.)
.
Thus, ∃k ∈ Z a = 2k. Let such a k be given, so a2 = 4k 2 .
Then 2b2 = 4k 2 , so b2 = 2k 2 .
Thus, b2 is even, by definition. By the same reasoning as above with a2
and a, we deduce here that b is even.
This shows both a and b are even so, in fact, they have a common factor
of 2.
This contradicts our assumption that a and b have no common factors.
××
×
×
Therefore, the original assumption must be False, so the claim is True.

4.9.3 Proving ∨ Claims

An “∨” claim asserts that at least one of two statements must be True. If it
just so happens that one of the two statements is clearly False, then just try
to prove the other one is True. This is what the direct method is here; it is
straightforward, so we will not provide an example of implementation.

Direct Method

Strategy:
Claim: P ∨ Q
Direct proof:
Prove that P is True, or else prove that Q is True.

This relies on you being able to decide ahead of time which one of the statements
(P or Q) is True, of course. If you can do so, then this isn’t even really a “strat-
egy”. Just implement whatever strategy applies to P (or Q, as the case may be).

Indirect Method (Proof by “Otherwise”)

This method is far more interesting than the direct one. In general, it is helpful
when the statements P and Q are actually variable propositions, and for some
instances P is True whereas for other instances Q is the True one. In that case,
rather than characterize exactly which instances satisfy P and which satisfy Q,
we can just say, “Well, if P is True, then our proof is already complete. Thus,
all we need to worry about are the cases where P is False; for those cases, we
need to show that Q is still True.”
294 CHAPTER 4. LOGIC

Strategy:
Claim: P ∨ Q
Indirect proof strategy 1:
Suppose that ¬P holds. Prove that Q holds.

To reiterate the strategy: If P holds, then the claim holds, and in fact we don’t
care whether Q holds or not. Otherwise, in the case where P fails, we need to
guarantee that Q holds.
Notice that P ∨Q ⇐⇒ Q∨P (i.e. the order is irrelevant in a logical disjunction)
so we can rewrite our claim as Q ∨ P and rewrite the above strategy as:
Suppose ¬Q holds. Prove that P holds.
This is the exact same strategy on the equivalent statement Q ∨ P , i.e. with the
roles of P and Q switched.
Example 4.9.5. When a real number is less than its square:
Statement: Suppose that x ∈ R and x2 ≥ x.
We claim that x ≥ 1 or x ≤ 0.

Implementation:
Proof. Let x ∈ R be arbitrary and fixed, and suppose that x2 ≥ x.
If it were the case that x ≤ 0, we would be done; so, suppose otherwise.
That is, suppose x > 0.
By assumption, x2 ≥ x. Since x > 0, we can divide both sides by x.
This yields x ≥ 1.

This proves a necessary condition for a real number to be less than (or equal
to) its square. Is this condition—namely, x ≥ 1 ∨ x ≤ 0—also a sufficient
condition? Prove it! It’s easy, and once you’ve done so, we will have together
proven this biconditional statement:

.
∀x ∈ R x2 ≥ x ⇐⇒ (x ≥ 1 ∨ x ≤ 0)

Indirect Method (Proof by Contradiction)

This method is more like the indirect methods considered above, in that we
suppose the logical negation is valid and then deduce something absurd. We
will illustrate this strategy by applying it to the same claim in the directly
previous example.
4.9. WRITING PROOFS: STRATEGIES AND EXAMPLES 295

Strategy:
Claim: P ∨ Q
Indirect proof strategy 2:
AFSOC that ¬P ∧ ¬Q holds. Find a contradiction.

Implementation:
Proof. Let x ∈ R be arbitrary and fixed, and suppose that x2 ≥ x.
AFSOC that 0 < x and x < 1.
Since x > 0, we can multiply both sides of these inequalities by x and
preserve the sign.
Multiplying x < 1 by x, then, we find that x2 < x.
This contradicts our assumption that x2 ≥ x ×
×
××
Thus, our assumption was invalid, so the claim is True.

How does this compare to the previous implementation? We are proving the
exact same claim, but the proofs are slightly different. Which do you think is
better, in your opinion? Which do you think was easier to write? Furthermore,
could you go back and rewrite the original claim using quantifiers and “ =⇒ ”?
After doing that, do you see what these two proofs accomplish? Try it!

4.9.4 Proving ∧ Claims

An “∧” claim asserts that both of two statements are True. There’s one obvious
and direct way to do this: just prove one statement and then prove the other!
We will show you an example implementation of this method because the ∧
statement of our example comes after an ∃ claim. Thus, there’s actually some
scratch work to be done to figure out how to define an object that will, indeed,
satisfy both of the desired properties. This will be our first illustrative example
of how these proof strategies can be combined to prove statements that use
both quantifiers and connectives.

Direct Method

Strategy:
Claim: P ∧ Q
Direct proof strategy:
296 CHAPTER 4. LOGIC

Prove that P holds. Prove that Q holds.

Example 4.9.6. A smaller number whose square is bigger:

. .
Statement: ∀x ∈ R ∃y ∈ R x ≥ y ∧ x2 < y 2

Scratch work:
How does this work? Let’s take a specific x, like x = 4. We need to find a
smaller real number whose square is bigger than x2 = 16.
The key is that we want y ∈ R, so we can use negative numbers. In this
case, picking a negative number with larger magnitude, like y = −5, will
work.

Let’s take a different x, like x = −2. This number is already negative, so

just picking any smaller number, like y = −3, will work.
The proof we follow below splits into two cases, based on whether x is
positive or non-positive.

Now we are ready to prove our claim.

Implementation:
Proof 1. Let x ∈ R be arbitrary and fixed. We consider two cases.
(1) Suppose x ≤ 0.
Define y = x − 1. Notice y ∈ R.
Notice y ≤ x. Also, notice that

y 2 = (x − 1)2 = x2 − 2x + 1 = x2 − (2x − 1)

Since x ≤ 0, we know 2x ≤ 0 and so 2x − 1 ≤ −1. Thus,

x2 − (2x − 1) ≥ x2 − 1 > x2

and therefore, y 2 > x2 .

(2) Now, suppose x > 0.

Define y = −x − 1. Notice y ∈ R.
Notice y < 0 and x > 0, so y ≤ x. (In fact, y < x, even.)
Also, notice that

y 2 = (−x − 1)2 = x2 + 2x + 1 = x2 + (2x + 1)

4.9. WRITING PROOFS: STRATEGIES AND EXAMPLES 297

Since x > 0, we know 2x + 1 > 0. Thus,

x2 + (2x + 1) > x2

and therefore, y 2 > x2 .

In either case, we found a y with the desired properties, namely y ∈ R and

y ≤ x and x2 < y 2 . Therefore, the claim is True.

Why did we call this “Proof 1”? We split the proof into two cases based on
our observations in the scratch work. Specifically, we recognized that we will
define y (in terms of x) differently, depending on the sign of x. We claim that
it is possible to rewrite this proof in a way that avoids these cases. This is what
“Proof 2” will be, and we want you to write it! To reiterate the goal, we want
you to rewrite the above proof so that y is defined in terms of x in one general
way that works regardless of the sign of x.
(Hint: What is −x when x < 0? Is this a function we’ve seen before?)

Indirect Method (Proof by Contradiction)

This method is just like the other indirect methods. We just take the logical
negation of a claim, assume it holds, and deduce something absurd. This means
that the assumption was invalid, so the original statement is the True one.
We will leave it to you to try to apply this method to the claim used in the previ-
ous example. (Note: You might want to do this after finding the “second proof”
we hinted at just above this.) Then, you can compare how the two methods
played out and decide which one you prefer, in this case.

Strategy:
Claim: P ∧ Q
Indirect proof:
AFSOC that ¬P ∨ ¬Q holds.
Consider the first case, where ¬P holds. Find a contradiction.
Consider the second case, where ¬Q holds. Find a contradic-
tion.

4.9.5 Proving =⇒ Claims

It might help you to look back at Section 4.5.3, where we introduced the con-
nective “ =⇒ ”. Specifically, we want you to recall that P =⇒ Q means that
whenever P holds, Q also necessarily holds. This conditional statement is True
298 CHAPTER 4. LOGIC

in the cases where P itself (the hypothesis) is False. Thus, our proof strategy
does not need to consider such cases. All we need to do is suppose that P holds,
and deduce that Q also holds. This takes care of the “whenever P holds, so
does Q” consideration.

Direct Method

Strategy:
Claim: P =⇒ Q
Direct proof strategy:

Suppose P holds. Prove that Q holds.

Example 4.9.7. Monotonicity of squares:

.
Statement: ∀y ∈ R y > 1 =⇒ y 2 − 1 > 0
Implementation:
Proof. Let y ∈ R be arbitrary and fixed.
Suppose y > 1.
Multiplying both sides by y (since y > 0), we obtain y 2 > y.
Since y > 1, this tells us y 2 > y > 1, and so y 2 > 1.
Subtracting yields the desired conclusion: y 2 − 1 > 0.

We called this “monotonicity of squares” because it states a particular property

of real numbers that is monotone. This is a term that is used to indicate a
certain inequality is preserved under an operation. In this case, the fact that
some number being greater than 1 is preserved by the “squaring operation”.
That is, we proved that if y > 1, then y 2 > 12 , as well.
Now, this was a pretty easy example to prove, but we wanted to include it to
emphase the proof strategy for conditional statements. Let’s work with a more
difficult example now.
(You’ll also notice that Exercise 4.11.22 has a similar-looking problem statement.
Perhaps you want to work on that one after following this example.)
Example 4.9.8. Working with inequalities:
Statement: We define the following variable propositions:
x−3 1
P (x) is “ >1− ”
x+2 x
x+3 1
Q(x) is “ <1+ ”
x+2 x
4.9. WRITING PROOFS: STRATEGIES AND EXAMPLES 299

Define S = {x ∈ R | x > 0}.

We claim that

.
∀x ∈ S P (x) =⇒ Q(x)

Scratch work:
We’re guessing that a direct method will work here, so let’s try to manipu-
late the inequality stated inside P (x) and make it “look like” the inequality
inside Q(x).
So we start with that inequality
x−3 1
>1−
x+2 x
and we’ll try multiplying both sides by x + 2. Can we do this? Oh right,
x > 0 and so x + 2 > 0, as well. Phew! This gives us
x+2 2 2
x − 3 > (x + 2) − =x+2−1− =x+1−
x x x
We want to see an x + 3 somewhere, so we’ll add/subtract on both sides:
2
x−1+ >x+3
x
Can we divide by x + 2 and make the right fraction? Hmm . . . Oh wait!
We already simplified the fraction x+2 x and moved it to one side. Maybe
we shouldn’t have simplified it first, so let’s try undoing that:
 
2 x+2 1
x + 3 < x − 1 + = (x + 2) + − 4 = (x + 2) 1 + −4
x x x

Aha, that looks better! We even have some “wiggle room” in the form of
the negative 4 there. We know the right-hand side is less than what we
wanted it to be, so the result holds.

Let’s take these algebraic steps we worked on here, reorder them a bit and
explain them, and wrap it all together in a formal proof.

Implementation:
Proof. Let x ∈ S be arbitrary and fixed.
300 CHAPTER 4. LOGIC

Suppose that P (x) holds; that is, suppose

x−3 1
>1−
x+2 x
We will show that the inequality
x+3 1
<1+
x+2 x
also holds, necessarily.
Since x ∈ S, we know x > 0 and so, certainly, x + 2 > 0, as well. Thus, we
can multiply both sides of the known inequality by x + 2, yielding
 
1 x+2
x − 3 > (x + 2) 1 − =x+2−
x x

Adding 3 + x+2
x to both sides, subtracting 2 from both sides, and rewriting
in the reverse direction (for ease of reading), we obtain
x+2
x+3<x−2+
x
Since x − 2 < x + 2, we deduce that
x+2
x+3<x+2+
x
and factoring tells us
 
1
x + 3 < (x + 2) 1 +
x

Again, since x + 2 > 0, we can divide both sides by x + 2, obtaining

x+3 1
<1+
x+2 x
which was the desired inequality. This shows P (x) =⇒ Q(x), and since
x was arbitrary, we have proven the claim.

A key lesson here lies in how we took our scratch work and presented it in a
different way in our proof. We cut out the unnecessary simplification and re-
factoring steps, but we also noted why each step was valid as we performed it. A
more seasoned mathematician would likely skip several of these steps and leave
it to the reader to verify the algebraic work, but since we are early on in our
mathematical careers, we thought it would be prudent to show as many details
as possible.