Contents

Part I Algebra Homeworks

0 Math 103B Homework Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

0.1 Homework #1 (Due Thursday, April 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
0.2 Homework #2 (Due Thursday, April 9) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
0.3 Homework #3 (Due Thursday, April 16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
0.4 Homework #4 (Due Thursday, April 23) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
0.5 Homework #5 (Due Thursday, April 30) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
0.6 Homework #6 (Due Thursday, May 7) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
0.7 Homework #7 (Due Thursday, May 14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
0.8 Homework #8 (Due Thursday, May 21) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
0.9 Homework #9 (Due Thursday, May 28) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
0.10 Homework #10 (Due Thursday, June 4) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Part II Math 103B Lecture Notes

1 Lecture 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.1 Definition of Rings and Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2 Appendix: Facts about finite sums . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2 Lecture 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.1 Polynomial Ring Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 Subrings and Ideals I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

3 Lecture 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.1 Some simple ring facts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2 The R [S] subrings I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.3 Appendix: R [S] rings II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4 Contents

4 Lecture 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.1 Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.2 (Zero) Divisors and Integral Domains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.3 Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

5 Lecture 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5.1 Characteristic of a Ring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

6 Lecture 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
6.1 Square root field extensions of Q . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
6.2 Homomorphisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

7 Lecture 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

8 Lecture 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

9 Lecture 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
9.1 Factor Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

10 Lecture 10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
10.1 First Isomorphism Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

11 Lecture 11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

12 Lecture 12 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
12.1 Higher Order Zeros (Not done in class) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
12.2 More Example of Factor Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
12.3 II. More on the characteristic of a ring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
12.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

13 Lecture 13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
13.1 Ideals and homomorphisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
13.2 Maximal and Prime Ideals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

14 Lecture 14 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

15 Lecture 15 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
15.1 The rest this section was not covered in class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

16 Lecture 16 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
16.1 The Degree of a Polynomial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
16.2 The evaluation homomorphism (review) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
16.3 The Division Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
16.4 Appendix: Proof of the division algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

Page: 4 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

 Contents 5

17 Lecture 17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
17.1 Roots of polynomials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
17.2 Roots with multiplicities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

18 Lecture 18 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
18.1 Irreducibles and Maximal Ideals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

19 Lecture 19 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
19.1 Irreducibles Polynomials I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

20 Lecture 20 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
20.1 Two more homomorphisms involving polynomials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
20.2 Gauss’ Lemma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

21 Lecture 21 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
21.1 mod p Irreducibility Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

22 Lecture 22 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
22.1 Eisenstein’s Criterion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
22.2 Summary of irreducibility tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

23 Lecture 23 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
23.1 Irreducibles and Primes II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

24 Lecture 24 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
24.1 Unique Factorization Domains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
24.2 Extra Topics (not covered in class) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
24.2.1 Greatest Common Divisors (not covered in class) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
24.2.2 Partial Fractions (not covered in class) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
24.2.3 Factorizing Polynomials in finite time (not covered in class) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
24.2.4 Fields of fractions (not covered in class) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

25 Lecture 25 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
25.1 Vector Spaces & Review of Linear Algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

26 Lecture 26 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
26.1 Field Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

27 Lecture 27 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
27.1 Ruler and Compass Constructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
27.2 Geometric Consequences for the Greeks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

Page: 5 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

 Part I

Algebra Homeworks
0
Math 103B Homework Problems

The problems are from Contemporary Abstract Algebra by Gallian, 6th Edi- Exercise 0.1. Let R be an integral domain with characteristic zero and ϕ, ψ :
tion [2] unless otherwise noted. We will label the problems via Chapter.Number, Q → R be two ring homomorphisms such that ϕ (1) = ψ (1) . Show ϕ (a) = ψ (a)
so that 4.12 will refer to exercise 12 in Chapter 4. for all a ∈ Q. Thus ϕ is uniquely determined by its value on 1 ∈ Q.

Hints:
0.1 Homework #1 (Due Thursday, April 2) 14.18: In other words, find all possible choices of a pair of ideals I and J of
Z which satisfy, h35i ( J ( I.
Hand in Problems: 15.28: Consider the relationship mod 3.
Extra problems for practice (do not hand in):
• Chapter 12: #2, 6, 7, 18, 22, 26, 42, 44*, 46 (∗ Hint: consider (−a)n ).
• Chapter 14, #25, 27, 59
Extra problems for practice (do not hand in): Chapter 15, #11, 13
• Chapter 12: # 14, 19, 29, 31, 37, 39, 41, 47

0.2 Homework #2 (Due Thursday, April 9) 0.4 Homework #4 (Due Thursday, April 23)

Hand in Problems: Hand in Problems:

• Chapter 13: #4, 6, 13, 14, 16, 22, 44, 54 • Chapter 14, #24, 34, 37, 54
Chapter 14: #2, 4, 10, 12 • Chapter 15, #20, 26, 36, 40.
• Chapter 16, #2, 4, 6, 14
Hints:
Extra problems for practice (do not hand in):
15.26: The 1st - isomorphism theorem for rings may help.
• Chapter 13, #5, 15, 21, 25, 29, 33, 39, 45, 53. 15.36. Use Exercise 0.1 to see that any homomorphism ϕ : Q → Q is
Chapter 14, #1, 5, 7, 11, 17, 19. determined uniquely by the value, a = ϕ (1) . Now use the multiplicativity
Hints: property of ϕ to determine the allowed values for a.
13.54: Use Lagrange’s theorem from last quarter. 15.40. Think about the first isomorphism theorem and make use of Exercise
25 of Chapter 14.
Extra problems for practice (do not hand in):
0.3 Homework #3 (Due Thursday, April 16) • Chapter 14, #29, 33
• Chapter 15, #39, 45, 51.
Hand in Problems:
• Chapter 14, #18, 40
Chapter 15, #6, 12, 14, 16, 28, 30
4 0 Math 103B Homework Problems

0.5 Homework #5 (Due Thursday, April 30) 0.8 Homework #8 (Due Thursday, May 21)
Hand in Problems: Hand in all problems below:
Exercise 0.2 (This problem is to be handed in!). Let R be a commutative • Chapter 17, #10a, c, e, l0 b & d, 32
ring with identity. Then R is a field iff R has no non-trivial proper ideals. (Recall
that I ⊂ R is the trivial ideal if I = {0} and is a proper ideal if I R.) Exercise 0.3. Prove Proposition 20.4.
• Chapter 14, #28, 32, 36, 52 Exercise 0.4. Let ψ : R → T be an onto ring homomorphism of commutative
• Chapter 15, #58, 60 rings, R and T and I := ker (ψ) .
Hints: 1. Explain why
ψ̄ : R [x] → T [x] is onto.
14.36: Let S := {a + bi : a ∈ Z4 and b ∈ Z2 } . To count the number of ele- 2. Show ker ψ̄ = I [x] .
ment in Z [i] / h2 + 2ii you might show 3. Use the first isomorphism theorem to conclude R [x] /I [x] is isomorphic
ψ T [x] .
S 3 (a + bi) −→ [a + bi] ∈ Z [i] / h2 + 2ii
Exercise 0.5. Let R be a commutative ring and I ⊂ R be an ideal. Use the
is a bijection. results of Exercise 0.4 to show R [x] /I [x] is isomorphic to (R/I) [x] .
14.52: One way is to prove that Z [i] / h1 − ii is isomorphic to Z2 .
Extra problems for practice (do not hand in): Exercise 0.6. Use Exercise 0.5 to give another proof Exercise 16.38 on page
300 of the book. This proof should be very short and similar in spirit to the
• Chapter 14, #31
proof of Gauss’ Lemma on p. 305 of the book.
• Chapter 15, #27, 29, 42, 62
Exercise 0.7. Prove Proposition 18.7.

0.6 Homework #6 (Due Thursday, May 7) Exercise 0.8. Prove Proposition 23.3

Hand in Problems: Exercise

p 0.9. Show x8 − 20/9 ∈ Q [x] is irreducible over Q. Conclude that
8
20/9 ∈
/ Q.
• Chapter 16, #12, 18, 20, 24, 30, 36, 38, 48.
Hint: Hints:
17.10b & d. Try the mod p irreducibility test, using some small prime p.
16.20. Think about the roots of the polynomial h = f − g.
Follow the method of examples 7 and 8 in the book.
Extra problems for practice (do not hand in):
17.32. One (but not the only) possibility is to show Z [x] / x2 + 1 is iso-
• Chapter 16, #1, 11, 13, 15, 19, 41 morphic to a ring we have studied frequently. Then use Theorems 14.3 and 14.4
of the book.

0.7 Homework #7 (Due Thursday, May 14)

0.9 Homework #9 (Due Thursday, May 28)
Hand in Problems:
• Chapter 17, #2, 4, 6, 8, 12, 14, 25 (17.4 is a special case of 17.25!) Hand in Problems:

Extra problems for practice (do not hand in): • Chapter 18, #4, 12, 28, 30.
• Chapter 19, #8, 22, 26.
• Chapter 17, #1, 3, 5, 7, 21

Page: 4 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

Exercise 0.10. Let f (x) := 6x3 − 14x2 − 2x + 2 ∈ Z [x] . Factor f (x) into
irreducible polynomials; 1) over R, 2) over Q, and 3) over Z. (The answers are
different in each case.) Hint: first find the rational roots of f (x) .

Extra problems for practice (do not hand in):

• Chapter 18, #1, 5, 27
• Chapter19, #7, 13, 14.

0.10 Homework #10 (Due Thursday, June 4)

Hand in Problems:
• Chapter 20, #1, 3, 5
• Chapter 21, #10, 14

• Chapter 23, #14

Exercise 0.11. Let g(x) = x3 + 2x + 1 ∈ Q [x] . Let α ∈ C be a root of g over

the complex numbers (do not try to find α explicitly; you don’t need to in order
to do this problem.)
1. Show g (x) is irreducible.
2. What is the minimal polynomial of α over Q.
3. Explain how you know that

K = Q [α] = Q + Qα + Qα2 = {a + bα + cα2 : a, b, c ∈ Q}

is a subfield of C.
4. Find explicit a, b, c ∈ Q such that α−1 = a + bα + cα2 . (Note that some
such expression must exist since K is a field.)
5. Find the degree, [Q (α) : Q] of the field extension Q ⊂ Q (α) . Recall that
this is defined to be the dimension of Q (α) as a vector space over Q.

Hints:
20.1: The hint in the back is useless. One way to “describe the elements”
is to find a basis for the given field as a Q-vector space. (Hint: it has dimension
3.)
21.14: See Examples 28.11 and 28.13.
23.14: Use Exercise 17.25 to show the relevant polynomial is irreducible.
 Part II

Math 103B Lecture Notes

1
Lecture 1

1.1 Definition of Rings and Examples Proof. Use the same proof that we used for groups! I.e. 1 = 1 · 10 = 10 and
if b, b0 are both inverses to a, then b = b (ab0 ) = (ba) b0 = b0 .
A ring will be a set of elements, R, with both an addition and multiplication
operation satisfying a number of “natural” axioms. Notation 1.6 (Subtraction) In any ring R, for a ∈ R we write the additive
inverse of a as (−a). So at a + (−a) = (−a) + a = 0 by definition. For any
Axiom 1.1 (Axioms for a ring) Let R be a set with 2 binary operations a, b ∈ R we abbreviate a + (−b) as a − b.
called addition (written a + b) and multiplication (written ab). R is called a
ring if for all a, b, c ∈ R we have Let us now give a number of examples of rings.

1. (a + b) + c = a + (b + c) Example 1.7. Here are some examples of commutative rings that we are already
2. There exists an element 0 ∈ R which is an identity for +. familiar with.
3. There exists an element −a ∈ R such that a + (−a) = 0. 1. Z = all integers with usual + and ·.
4. a + b = b + a. 2. Q = all mn such that m, n ∈ Z with n 6= 0, usual + and ·. (We will generalize
5. (ab)c = a(bc). this later when we talk about “fields of fractions.”)
6. a(b + c) = ab + ac and (b + c)a = ba + bc.
3. R = reals, usual + and ·.
Items 1. – 4. are the axioms for an abelian group, (R, +) . Item 5. says mul- 4. C = all complex numbers, i.e. {a + ib : a, b ∈ R} , usual + and · operations.
tiplication is associative, and item 6. says that is both left and right distributive (We will explicitly verify this in Proposition 3.7 below.)
over addition. Thus we could have stated the definition of a ring more succinctly
Example 1.8. 2Z = {. . . , −4, −2, 0, 2, 4, . . . } is a ring without identity.
as follows.
Example 1.9 (Integers modulo m). For m ≥ 2, Zm = {0, 1, 2, . . . , m − 1} with
Definition 1.2. A ring R is a set with two binary operations “+” = addition
and “·”= multiplication, such that (R, +) is an abelian group (with identity + = addition mod m
element we call 0), “·” is an associative multiplication on R which is both left
and right distributive over addition. · = multiplication mod n.

Remark 1.3. The multiplication operation might not be commutative, i.e., ab 6= Recall from last quarter that (Zm , +) is an abelian group and we showed,
ba for some a, b ∈ R. If we have ab = ba for all a, b ∈ R, we say R is a
[(ab) mod m · c] mod m = [abc] = [a (bc) mod m] mod m (associativity)
commutative ring. Otherwise R is noncommutative.
and
Definition 1.4. If there exists and element 1 ∈ R such that a1 = 1a = a for all
a ∈ R, then we call 1 the identity element of R [the book calls it the unity.] [a · (b + c) mod m] mod m = [a · (b + c)] mod m
Most of the rings that we study in this course will have an identity element. = [ab + ac] mod m = (ab) mod m + (ac) mod m

Lemma 1.5. If R has an identity element 1, then 1 is unique. If an element which is the distributive property of multiplication mod m. Thus Zm is a ring
a ∈ R has a multiplicative inverse b, then b is unique, and we write b = a−1 . with identity, 1.
10 1 Lecture 1

Example 1.10. M2 (F ) = 2 × 2 matrices with entries from F , where F = Z, Q, Since   

R, or C with binary operations; 0 −1 0 −1
i2 = =I
   0 0  1 0 1 0
a + a0 b + b0

ab a b
+ 0 0 = (addition) it is straight forward to check that
cd c d c + c0 d + d0
   0 0  0
ab a b aa + bc0 ab0 + bd0
 (aI+bi) (cI+di) = (ac − bd) I + (bc + ad) i and
= . (multiplication)
c d c0 d0 ca0 + dc0 cb0 + dd0 (aI+bi) + (cI+di) = (a + c) I + (b + d) i
That is multiplication is the usual matrix product. You should have checked in which are the standard rules of complex arithmetic. The fact that C is a ring
your linear algebra course that M2 (F ) is a non-commutative ring with identity, now easily follows from the fact that M2 (R) is a ring.
 
10 In this last
I=
01
.  example,
 the reader may wonder how did we come up with the
0 −1
matrix i := to represent i. The answer is as follows. If we view C as R2
For example let us check that left distributive law in M2 (Z); 1 0
in disguise, then multiplication by i on C becomes,
     
ab ef pq
+ (a, b) ∼ a + ib → i (a + ib) = −b + ai ∼ (−b, a)
cd gh rs
  
a b p+e f +q while
=       
c d g+r h+s a 0 −1 a −b
i = = .

b (g + r) + a (p + e) a (f + q) + b (h + s)
 b 1 0 b a
=
d (g + r) + c (p + e) c (f + q) + d (h + s) Thus i is the 2 × 2 real matrix which implements multiplication by i on C.
 
bg + ap + br + ae af + bh + aq + bs
= Theorem 1.12 (Matrix Rings). Suppose that R is a ring and n ∈ Z+ . Let
dg + cp + dr + ce cf + dh + cq + ds n
Mn (R) denote the n × n – matrices A = (Aij )i,j=1 with entries from R. Then
while Mn (R) is a ring using the addition and multiplication operations given by,
     
ab ef ab pq (A + B)ij = Aij + Bij and
+
cd gh cd rs X
    (AB)ij = Aik Bkj .
bg + ae af + bh ap + br aq + bs
= + k
dg + ce cf + dh cp + dr cq + ds
Moreover if 1 ∈ R, then
 
bg + ap + br + ae af + bh + aq + bs
= 
1 0 0

dg + cp + dr + ce cf + dh + cq + ds
I :=  0 . . . 0 
 
which is the same result as the previous equation.
0 0 1
Example 1.11. We may realize C as a sub-ring of M2 (R) as follows. Let
is the identity of Mn (R) .
   
10 0 −1
I= ∈ M2 (R) and i := Proof. I will only check associativity and left distributivity of multiplication
01 1 0
here. The rest of the proof is similar if not easier. In doing this we will make
and then identify z = a + ib with use of the results about sums in the Appendix 1.2 at the end of this lecture.
      Let A, B, and C be n × n – matrices with entries from R. Then
10 0 −1 a −b
aI+bi := a +b = .
01 1 0 b a

Page: 10 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

 1.2 Appendix: Facts about finite sums 11
!
X X X 1.2 Appendix: Facts about finite sums
[A (BC)]ij = Aik (BC)kj = Aik Bkl Clj
k k l Throughout this section, suppose that (R, +) is an abelian group, Λ is any set,
X
= Aik Bkl Clj and Λ 3 λ → rλ ∈ R is a given function.
k,l
Theorem 1.15. Let F := {A ⊂ Λ : |A| < ∞} . Then there is a unique function,
while S : F → R such that;

X X X
! 1. S (∅) = 0,
[(AB) C]ij = (AB)il Clj = Aik Bkl Clj 2. S ({λ}) = rλ for all λ ∈ Λ.
l l k 3. S (A ∪ B) = S (A) + S (B) for all A, B ∈ F with A ∩ B = ∅.
X
= Aik Bkl Clj . Moreover, for any A ∈ F, S (A) only depends on {rλ }λ∈A .
k,l
Proof. Suppose that n ≥ 2 and that S (A) has been defined for all A ∈ F
Similarly, with |A| < n in such a way that S satisfies items 1. – 3. provided that |A ∪ B| <
X X n. Then if |A| = n and λ ∈ A, we must define,
[A (B + C)]ij = Aik (Bkj + Ckj ) = (Aik Bkj + Aik Ckj )
k k S (A) = S (A \ {λ}) + S ({λ}) = S (A \ {λ}) + rλ .
X X
= Aik Bkj + Aik Ckj = [AB]ij + [AC]ij . We should verify that this definition is independent of the choice of λ ∈ A. To
k k see this is the case, suppose that λ0 ∈ A with λ0 6= λ, then by the induction
hypothesis we know,

Example 1.13. In Z6 , 1 is an identity for multiplication, but 2 has no multi- S (A \ {λ}) = S ([A \ {λ, λ0 }] ∪ {λ0 })
plicative inverse. While in M2 (R), a matrix A has a multiplicative inverse if = S (A \ {λ, λ0 }) + S ({λ0 }) = S (A \ {λ, λ0 }) + rλ0
and only if det(A) 6= 0.
so that
Example 1.14 (Another ring without identity). Let
S (A \ {λ}) + rλ = [S (A \ {λ, λ0 }) + rλ0 ] + rλ
= S (A \ {λ, λ0 }) + (rλ0 + rλ )
  
0a
R= :a∈R
00 = S (A \ {λ, λ0 }) + (rλ + rλ0 )
= [S (A \ {λ, λ0 }) + rλ ] + rλ0
with the usual addition and multiplication of matrices.
     = [S (A \ {λ, λ0 }) + S ({λ})] + rλ0
0a 0b 00 = S (A \ {λ0 }) + rλ0
= .
00 00 00
  as desired. Notice that the “moreover” statement follows inductively using this
10 definition.
The identity element for multiplication “wants” to be , but this is not in
01 Now suppose that A, B ∈ F with A ∩ B = ∅ and |A ∪ B| = n. Without loss
R. of generality we may assume that neither A or B is empty. Then for any λ ∈ B,
More generally if (R, +) is any abelian group, we may make it into a ring we have used the inductive hypothesis, that
in a trivial way by setting ab = 0 for all a, b ∈ R. This ring clearly has no
multiplicative identity unless R = {0} is the trivial group. S (A ∪ B) = S (A ∪ [B \ {λ}]) + rλ = (S (A) + S (B \ {λ})) + rλ
= S (A) + (S (B \ {λ}) + rλ ) = S (A) + (S (B \ {λ}) + S ({λ}))
= S (A) + S (B) .

Page: 11 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

12 1 Lecture 1
X X
Thus we have defined S inductively on the size of A ∈ F and we had no rσ(x) = rσ(x) + rσ(b) .
choice in how to define S showing S is unique. x∈B x∈B 0

NotationP1.16 Keeping the notation used in Theorem 1.15, we will denote Since σ|B 0 : B 0 → 0
P A := A \ {σP(b)} is a bijection, it follows by the induction
S (A) by λ∈A rλ . If A = {1, 2, . . . , n} we will often write, hypothesis that x∈B 0 rσ(x) = λ∈A0 rλ and therefore,
X X X
X n
X rσ(x) = rλ + rσ(b) = rλ .
rλ = ri . x∈B λ∈A0 λ∈A
λ∈A i=1

Corollary 1.17. Suppose that A = A1 ∪ · · · ∪ An with Ai ∩ Aj = ∅ for i 6= j Lemma 1.19. If {aλ }λ∈Λ and {bλ }λ∈Λ are two sequences in R, then
and |A| < ∞. Then X
(aλ + bλ ) =
X
aλ +
X
bλ .
λ∈A λ∈A λ∈A
n n
!
X X X X
S (A) = S (Ai ) i.e. rλ = rλ . Moreover, if we further assume that R is a ring, then for all r ∈ R we have the
i=1 λ∈A i=1 λ∈Ai right and left distributive laws;,
X X
Proof. As usual the proof goes by induction
P on n. For n = 2, the assertion r· aλ = r · aλ and
λ∈A rλ . For n ≥ 2, we have used
is one of the defining properties of S (A) := P λ∈A λ∈A
n
the induction hypothesis and the definition of i=1 S (Ai ) that X
!
X
aλ ·r = aλ · r.
S (A1 ∪ · · · ∪ An ) = S (A1 ∪ · · · ∪ An−1 ) + S (An ) λ∈A λ∈A
n−1 n
X X Proof. This follows by induction. Here is the key step. Suppose that α ∈ A
= S (Ai ) + S (An ) = S (Ai ) . and A0 := A \ {α} , then
i=1 i=1 X X
(aλ + bλ ) = (aλ + bλ ) + (aα + bα )
λ∈A λ∈A0
Corollary 1.18 (Order does not matter). Suppose that A is a finite subset
X X
= aλ + bλ + (aα + bα ) (by induction)
of Λ and B is another set such that |B| = n = |A| and σ : B → A is a bijective λ∈A0 λ∈A0
function. Then ! !  
X
rσ(b) =
X
ra .
X X commutativity
= aλ + aλ + bλ + bα
b∈B a∈A
and associativity
λ∈A0 λ∈A0
In particular if σ : A → A is a bijection, then
X X
= aλ + bλ .
X X λ∈A λ∈A
rσ(a) = ra .
The multiplicative assertions follows by induction as well,
a∈A a∈A ! !
X X X
Proof. We again check this by induction on n = |A| . If n = 1, then B = {b} r· aλ = r · aλ + aα = r · aλ + r · aα
and A = {a := σ (b)} , so that λ∈A λ∈A0 λ∈A0
!
X X X
rσ(x) = rσ(b) = ra = r · aλ + r · aα
x∈B a∈A λ∈A0
X
as desired. Now suppose that N ≥ 1 and the corollary holds whenever n ≤ N. = r · aλ .
If |B| = N + 1 = |A| and σ : B → A is a bijective function, then for any b ∈ B, λ∈A
we have with B 0 := B 0 \ {b} that

Page: 12 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

2
Lecture 2

Recall that a ring is a set, R, with two binary operations “+” = addition f + g = 8x2 − 2x + 6 and
and “·”= multiplication, such that (R, +) is an abelian group (with identity f g = (5x3 + 1)(3x2 − 2x + 5)
element we call 0), (·) is an associative multiplication on R which is left and right
distributive over “+.” Also recall that if there is a multiplicative identity, 1 ∈ R = 5 − 2x + 3x2 + 25x3 − 10x4 + 15x5 .
(so 1a = a1 = a for all a), we say R is a ring with identity (unity). Furthermore
One may check (see Theorem 2.4 below) that R [x] with these operations is a
we write a − b for a + (−b) . This shows the importance of distributivity. We
commutative ring with identity, 1 = 1.These rules have been chosen so that
now continue with giving more examples of rings.
(f + g) (α) = f (α) + g (α) and (f · g) (α) = f (α) g (α) for all α ∈ R where
Example 2.1. Let R denote the continuous functions, f : R → R such that ∞
limx→±∞ f (x) = 0. As usual, let f + g and f · g be pointwise addition and
X
f (α) := ai αi .
multiplication of functions, i.e. i=0

(f + g) (x) = f (x) + g (x) and (f · g) (x) = f (x) g (x) for all x ∈ R. Theorem 2.4. Let R be a ring and R [x] denote the collection of polynomials
with the usual addition and multiplication rules of polynomials. Then R [x] is
Then R is a ring without identity. (If we remove the restrictions on the functions again a ring. To be more precise,
at infinity, R would be a ring with identity, namely 1 (x) ≡ 1.)
∞
( )
X
i
Example 2.2. For any collection of rings R1 , R2 , . . . , Rm , define the direct sum R [x] = p = pi x : pi ∈ R with pi = 0 a.a. ,
to be i=0
R = R1 ⊕ · · · ⊕ Rn = {(r1 , r2 , . . . , rn ) : ri ∈ Ri all i}
where we say that pi =P 0 a.a. (read as almost always) provided that
the set of all m-tuples where the ith coordinate comes from Ri . R is a ring if ∞
|{i : pi 6= 0}| < ∞. If q := i=0 qi xi ∈ R [x] , then we set,
we define
∞
X
(r1 , r2 , . . . , rm ) + (s1 , s2 , . . . , sm ) = (r1 s1 , r2 s2 , . . . , rm sm ), p + q := (pi + qi ) xi and (2.1)
i=0
and ∞ ∞
! i
!
X X X X
i
p · q := pk ql x = pk qi−k xi . (2.2)
(r1 , r2 , . . . , rm ) + (s1 , s2 , . . . , sm ) = (r1 + s1 , r2 + s2 , . . . , rm + sm ). i=0 k+l=i i=0 k=0

The identity element 0 is (0, 0, . . . , 0). (Easy to check) Proof. The proof is similar to the matrix group examples. Let me only say
a few words about the associativity property of multiplicationPhere, since this
∞
is the most complicated property to check. Suppose that r = i=0 ri xi , then
2.1 Polynomial Ring Examples
Example 2.3 (Polynomial rings). Let R = Z, Q, R, or Z and let R [x] denote the
polynomials in x with coefficients from R. We add and multiply polynomials in
the usual way. For example if f = 3x2 − 2x + 5 and g = 5x2 + 1, then
14 2 Lecture 2
 
∞
X X Definition 2.7 (Ideals). Let R be a ring. A (two sided) ideal, I, of R is a
p (qr) =  pi (qr)j  xn subring, I ⊂ R such that RI ⊂ R and IR ⊂ R. Alternatively put, I ⊂ R is an
n=0 i+j=n ideal if (I, +) is a subgroup of (R, +) such that RI ⊂ R and IR ⊂ R. (Notice
that every ideal, I, of R is also a subring of R.)
  
∞
X X X
=  pi  qk rl  xn
Example 2.8. Suppose that R is a ring with identity 1 and I is an ideal. If 1 ∈ I,
n=0 i+j=n k+l=j
then I = R since R = R · 1 ⊂ RI ⊂ I.
∞
!
X X
= pi qk rl xn . Example 2.9. Given a ring R, R itself and {0} are always ideals of R. {0} is the
n=0 i+k+l=n trivial ideal. An ideal (subring) I ⊂ R for which I 6= R is called a proper ideal
(subring).
As similar computation shows,
Example 2.10. If R is a commutative ring and b ∈ R is any element, then the
∞
!
X X principle ideal generated by b, denoted by hbi or Rb, is
(pq) r = pi qk rl xn
n=0 i+k+l=n I = Rb = {rb : r ∈ R}.
and hence the multiplication rule in Eq. (2.2) is associative. To see that I is an ideal observer that if r, s ∈ R, then rb and sb are generic
elements of I and
rb − sb = (r − s)b ∈ Rb.
2.2 Subrings and Ideals I
Therefore I is an additive subgroup of R. Moreover, (rb) s = s (rb) = (sr) b ∈ I
We now define the concept of a subring in a way similar to the concept of so that RI = IR ⊂ I.
subgroup. Theorem 2.11. Suppose that R = Z or R = Zm for some m ∈ Z+ . Then the
subgroups of (R, +) are the same as the subrings of R which are the same as
Definition 2.5 (Subring). Let R be a ring. If S is subset of R which is itself
the ideals of R. Moreover, every ideal of R is a principle ideal.
a ring under the same operations +, · of R restricted to the set S, then S is
called a subring of R. Proof. If R = Z, then hmi = mZ inside of Z is the principle ideal generated
by m. Since every subring, S ⊂ Z is also a subgroup and all subgroups of Z are
Lemma 2.6 (Subring test). S ⊂ R is a subring if and only if S is a subgroup
of the form mZ for some m ∈ Z, it flows that all subgroups of (Z, +) are in fact
of (R, +) and S is closed under multiplication. In more detail, S is a subring
also principle ideals.
of R, iff for all a, b ∈ S, that
Suppose now that R = Zn . Then again for any m ∈ Zn ,
a + b ∈ S, − a ∈ S, and ab ∈ S. hmi = {km : k ∈ Z} = mZn (2.3)
Alternatively we may check that is the principle ideal in Zn generated by m. Conversely if S ⊂ Zn is a sub-ring,
then S is in particular a subgroup of Zn . From last quarter we know that this
a − b ∈ S, and ab ∈ S for all a, b ∈ S.
implies S = hmi = hgcd (n, m)i for some m ∈ Zn . Thus every subgroup of
Put one last way, S is a subring of R if (S, +) is a subgroup of (R, +) which is (Zn , +) is a principle ideal as in Eq. (2.3).
closed under the multiplication operation, i.e. S · S ⊂ S. Example 2.12. The set,
Proof. Either of the conditions, a + b ∈ S, −a ∈ S or a − b ∈ S for all
  
ab
a, b ∈ S implies that (S, +) is a subgroup of (R, +) . The condition that (S, ·) S= : a, b, d ∈ R ,
0d
is a closed shows that “·” is well defined on S. This multiplication on S then
inherits the associativity and distributivity laws from those on R. is a subring of M2 (R). To check this observe that;

Page: 14 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

 2.2 Subrings and Ideals I 15
   0 0 
a − a b − b0

ab a b 3. Let R := Z6 and S = h2i = {0, 2, 4} is a sub-ring of Z6 . Moreover, one sees
− = ∈S
0d 0 d0 0 d − d0 that 1S = 4 is the unit in S (42 = 4 and 4 · 2 = 2) which is not 1R = 1.
Thus again, S is not a unital sub-ring of Z6 .
and
4. The set,
   0 0  0
a a ab0 + bd0

ab a b
= ∈ S.
  
0 d 0 d0 0 dd0 a0
S= : a ∈ R ⊂ R = M2 (R),
00
S is not an ideal since,
     is a subring of M2 (R) with
00 ab 00
= ∈
/ S if a 6= 0.    
10 0d ab 10 10
1S = 6 = = 1R
00 01
Example 2.13. Consider Zm and the subset U (m) the set of units in Zm . Then
U (m) is never a subring of Zm , because 0 ∈
/ U (m). and hence is not a unital subring of M2 (R) .
Example 2.14. The collection of matrices, 5. Let v be a non-zero column vector in R2 and define,
  
0a S := {A ∈ M2 (R) : Av = 0} .
S= : a, b, c ∈ R ,
bc
Then S is a non-unital subring of M2 (R) which is not an ideal. (You should
is not a subring of M2 (R). It is an additive subgroup which is however not verify these assertions yourself!)
closed under matrix multiplication;
Remark 2.18. Let n ∈ Z+ and S := hmi be a sub-ring of Zn . It is natural to
0 a 0 a0
  0
ac0
  
ab ask, when does S have an identity element. To answer this question, we begin
= ∈
/S
b c b0 c0 cb0 ba + cc0 by looking for m ∈ Zn such that m2 = m. Given such a m, we claim that m is
Definition 2.15. Let R be a ring with identity. We say that S ⊂ R is a unital an identity for hmi since
subring of R if S is a sub-ring containing 1R . (Most of the subrings we will
(km) m = km2 = k1 m for all km ∈ hmi .
consider later will be unital.)
Example 2.16. Here are some examples of unital sub-rings. The condition that m2 = m is equivalent to m (m − 1) = 0, i.e. n|m (m − 1) .
Thus hmi = hgcd (n, m)i is a ring with identity iff n|m (m − 1) .
1. S in Example 2.12 is a unital sub-ring of M2 (R) .
2. The polynomial functions on R is a unital sub-ring of the continuous func- Example 2.19. Let us take m = 6 in the above remark so that m (m − 1) =
tions on R. 30 = 3 · 2 · 5. In this case 10, 15 and 30 all divide m (m − 1) and therefore 6
3. Z [x] is a unital sub-ring of Q [x] or R [x] or C [x] . is the identity element in h6i thought of as a subring of either, Z10 , or Z15 , or
4. The Gaussian integres, Z [i] := {a + ib : a, b ∈ Z} is a unital subring of Z30 . More explicitly 6 is the identity in
C. (For some number theoretic applications of the Gaussian integers see [1,
Sections 12.3, p. 364 – 371.].) h6i = hgcd (6, 10)i = h2i = {0, 2, 4, 6, 8} ⊂ Z10 ,
h6i = hgcd (6, 15)i = h3i = {0, 3, 6, 9, 12} ⊂ Z15 , and
Example 2.17. Here are a few examples of non-unital sub-rings.
h6i = hgcd (6, 30)i = {0, 6, 12, 18, 24} ⊂ Z30 .
1. nZ ⊂ Z is a non-unital subring of Z for all n 6= 0 since nZ does not even
contain an identity element. Example 2.20. On the other hand there is no proper non-trivial subring of Z8
2. If R = Z8 , then every non-trivial proper subring, S = hmi , of R has no which contains an identity element. Indeed, if m ∈ Z8 and 8 = 23 |m (m − 1) ,
identity. The point is if k ∈ Z8 is going to be an identity for some sub-ring then either 23 |m if m is even or 23 | (m − 1) if m is odd. In either the only
of Z8 , then k 2 = k. It is now simple to check that k 2 = k in Z8 iff k = 0 m ∈ Z8 with this property is m = 0 and m = 1. In the first case h0i = {0} is
or 1 which are not contained in any proper non-trivial sub-ring of Z8 . (See the trivial subring of Z8 and in the second case h1i = Z8 is not proper.
Remark 2.18 below.)

Page: 15 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

3
Lecture 3

3.1 Some simple ring facts Example 3.2. Consider R = h2i = {0, 2, 4, 6, 8} ⊂ Z10 . From Example 2.19 we
know that 1R = 6 which you can check directly as well. So −1R = −6 mod 10 =
The next lemma shows that the distributive laws force 0, 1, and the symbol 4. Taking a = 2 let us write out the meaning of the identity, (−1R ) · a = −a;
“−” to behave in familiar ways.
(−1R ) · a = 4 · 2 = 8 = −a.
Lemma 3.1 (Some basic properties of rings). Let R be a ring. Then;
1. a0 = 0 = 0a for all a ∈ R. Let us also work out (−2) (−4) and compare this with 2 · 4 = 8;
2. (−a)b = − (ab) = a(−b) for all a, b ∈ R
(−2) (−4) = 8 · 6 = 48 mod 10 = 8.
3. (−a)(−b) = ab for all a, b ∈ R. In particular, if R has identity 1, then

(−1)(−1) = 1 and Lastly consider,

(−1)a = −a for all a ∈ R. 4 · (8 − 2) = 4 · 6 = 24 mod 10 = 4 while
(This explains why minus times minus is a plus! It has to be true in any 4 · 8 − 4 · 2 = 2 − 8 = −6 mod 10 = 4.
structure with additive inverses and distributivity.)
4. If a, b, c ∈ R, then a (b − c) = ab − ac and (b − c) a = ba − ca.
3.2 The R [S] subrings I
Proof. For all a, b ∈ R;
1. a0 + 0 = a0 = a(0 + 0) = a0 + a0, and hence by cancellation in the abelian Here we will construct some more examples of rings which are closely related
group, (R, +) , we conclude that , so 0 = a0. Similarly one shows 0 = 0a. to polynomial rings. In these examples, we will be given a commutative ring R
2. (−a)b+ab = (−a+a)b = 0b = 0, so (−a)b = − (ab). Similarly a(−b) = −ab. (usually commutative) and a set S equipped with some sort of multiplication,
3. (−a)(−b) = − (a(−b)) = −(−(ab)) = ab, where in the last equality we have we then are going to define R [S] to be the collection of linear combinations of
used the inverting an element in a group twice gives the element back. elements from the set, ∪∞ n n
n=0 RS . Here RS consists of formal symbols of the
4. This last item is simple since, form rs1 . . . sn with r ∈ R and si ∈ S. The next proposition gives a typical
example of what we have in mind.
a (b − c) := a (b + (−c)) = ab + a (−c) = ab + (−ac) = ab − ac. A typical case will be where S = {s1 , . . . , sn } is a finite set then
Similarly one shows that (b − c) a = ba − ca. Proposition 3.3. If R ⊂ R̄ is a sub-ring of a commutative ring R̄ and S =
{s1 , . . . , sn } ⊂ R̄. Let
In proofs above the reader should not be fooled into thinking these things (
X
)
are obvious. The elements involved are not necessarily familiar things like real k
R [S] = R [s1 , . . . , sn ] = ak s : ak ∈ R with ak = 0 a.a. ,
numbers. For example, in M2 (R) item 2 states, (−I)A = −(IA) = −A, i.e. k
    
−1 0 ab −a −b where k = (k1 , . . . kn ) ∈ Nn and sk = sk11 . . . sknn with a0 s0 := a0 ∈ R. Then
= X
0 −1 c d −c −d R [s1 , . . . , sn ] is a sub-ring of R̄.
The following example should help to illustrate the significance of Lemma 3.1.
18 3 Lecture 3

ak sk and g = k bk sk , then
P P
Proof. If f = k (x + yi) (u + vi) = ux − vy + (uy + vx) i in Zm .
X The next proposition shows that this is a commutative ring with identity, 1.
f +g = (ak + bk ) sk ∈ R [S] ,
k Proposition 3.7. Let R be a commutative ring with identity and let
X
k
−g = −bk s ∈ R [S] , and R [i] := {a + bi : a, b ∈ R} ∼
= {(a, b) : a, b ∈ R} = R2 .
k
X X Define addition and multiplication of R [i] as one expects by,
f ·g = ak sk · bl sl
k l (a + bi) + (c + di) = (a + c) + (b + d) i
X X
k l k+l
= ak bl s s = ak bl s and
k,l k,l (a + bi) · (c + di) = (ac − bd) + (bc + ad) i.
!
X X Then (R [i] , +, ·) is a commutative ring with identity.
= ak bl sn ∈ R [S] .
n k+l=n Proof. This can be checked by brute force. Rather than use brute force lets
give a proof modeled on Example 1.11, i.e. we will observe that we may identify
R [i] with a unital subring of M2 (R) . To do this we take,
√
Example 3.4 (Gaussian Integers). Let i := −1 ∈ C. Then Z [i] = 
0 −1
 
10

{x + yi : x, y ∈ Z} . To see this notice that i2 = −1 ∈ Z, and therefore i := ∈ M2 (R) and 1 := I = ∈ M2 (R) .
1 0 01
∞ ∞ h
X k
X 4l 4l+1 4l+2 4l+3
i Thus we take,
ak (i) = a4l (i) + a4l+1 (i) + a4l+2 (i) + a4l+3 (i) a −b

k=0 l=0 a + ib ←→ aI + bi = ∈ M2 (R) .
b a
∞
Since
X
= [a4l + a4l+1 i − a4l+2 − a4l+3 i]    
l=0 a −b c −d
∞ ∞
! (aI + bi) + (cI + di) = +
X X b a d c
= [a4l − a4l+2 ] + [a4l+1 − a4l+3 ] i  
a + c −b − d
l=0 l=0 =
b+d a+c
= x + yi
= (a + c) I + (b + d) i
where
∞
X ∞
X and
x= [a4l − a4l+2 ] and y = [a4l+1 − a4l+3 ] . 
a −b

c −d

l=0 l=0 (aI + bi) (cI + di) =
b a d c
Example 3.5. Working as in the last example we see that
 
ac − bd −ad − bc
=
h√ i n √ o ad + bc ac − bd
Z 2 = a + b 2 : a, b ∈ Z
= (ac − bd) I + (bc + ad) i
is a sub-ring of R. we see that   
a −b
Example 3.6 (Gaussian Integers mod m). For any m ≥ 2, let S := = aI + bi : a, b ∈ R
b a
Zm [i] = {x + yi : x, y ∈ Zm } is indeed a unital sub-ring of M2 (R) . Moreover, the multiplication rules on S
and R [i] agree under the identification; a + ib ←→ aI + bi. Therefore we may
with the obvious addition rule and multiplication given by conclude that (R [i] , +, ·) satisfies the properties of a ring.

Page: 18 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

 3.3 Appendix: R [S] rings II 19
3.3 Appendix: R [S] rings II as desired. Secondly,
X X
You may skip this section on first reading. [a · (b + c)]s =
uv=s
au (b + c)v =
uv=s
au (bv + cv )

Definition 3.8. Suppose that S is a set which is equipped with an associative =

X
au bv +
X
au cv
binary operation, ·, which has a unique unit denoted by e. (We do not assume uv=s uv=s
that (S, ·) has inverses.
PAlso suppose that R is a ring, then we let R [S] consist = [a · b]s + [a · c]s = [a · b + a · c]s
of the formal sums, s∈S as s where {as }s∈S ⊂ R is a sequence with finite
support, i.e. |{s ∈ S : as 6= 0}| < ∞. We define two binary operations on R [S] from which it follows that a · (b + c) = a · b + a · c. Similarly one shows that
by, X X X (b + c) · a = b · a + c · a.
as s + bs s := (as + bs ) s Lastly if S has an identity, e, and es := 1s=e ∈ R, then
s∈S s∈S s∈S X
and [a · e]s = au ev = as
X X X X uv=s
as s · bs s = as s · bt t
s∈S s∈S s∈S t∈S from which it follows that e is the identity in R [S] .
!
X X X Example 3.10 (Polynomial rings). Let x be a formal symbol and let S :=
= as bt st = as bt u. 
xk : k = 0, 1, 2 . . . with xk xl := xk+l being the binary operation of S. No-
s,t∈S u∈S st=u
tice that x0 is the identity in S under this multiplication rule. Then for any
So really we R [S] are those sequences a := {as }s∈S with finite support with the ring R, we have
operations, ( )
Xn
k
R [S] = p (x) := pk x : pk ∈ R and n ∈ N .
X
(a + b)s = as + bs and (a · b)s = au bv for all s ∈ S.
uv=s k=0

Theorem 3.9. The set R [S] equipped with the two binary operations (+, ·) is The multiplication rule is given by
a ring.  
∞
X k
X
Proof. Because (R, +) is an abelian group it is easy to check that (R [S] , +) p (x) q (x) =  pj qk−j  xk
is an abelian group as well. Let us now check that · is associative on R [S] . To k=0 j=0
this end, let a, b, c ∈ R [S] , then
  which is the usual formula for multiplication of polynomials. In this case it is
X X X customary to write R [x] rather than R [S] .
[a (bc)] = s au (bc) = au  v bα cβ 
uv=s uv=s αβ=v This example has natural generalization to multiple indeterminants as fol-
X lows.
= au bα cβ
uαβ=s Example 3.11. Suppose that x = (x1 , . . . , xd ) are d indeterminants and k =
(k1 , . . . , kd ) are multi-indices. Then we let
while
n o
S := xk := xk11 . . . xkdd : k ∈ Nd
X X X
[(ab) c]s = (ab)α cβ = au bv cβ
αβ=s αβ=s uv=α
X X with multiplication law given by
= au bv cβ = au bα cβ = [a (bc)]s
0 0
uvβ=s uαβ=s xk xk := xk+k .

Page: 19 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

20 3 Lecture 3

Then ( ) Let us check associativity and distributivity of ·. To this end,

X
k
R [S] = p (x) := pk x : pk ∈ R with pk = 0 a.a. . [(a · b) · c] (h) =
X
(a · b) (g) · c (k)
k
gk=h
We again have the multiplication rule, X
"
X
#
  = a (u) · b (v) · c (k)
X X gk=h uv=g
p (x) q (x) =  pj qk−j  xk . X
k j≤k = a (u) · b (v) · c (k)
uvk=h
As in the previous example, it is customary to write R [x1 , . . . , xd ] for R [S] .
while on the other hand,
In the next example we wee that the multiplication operation on S need not X
be commutative. [a · (b · c)] (h) = a (u) · (b · c) (y)
uy=h
Example 3.12 (Group Rings). In this example we take S = G where G is a
 
X X
group which need not be commutative. Let R be a ring and set, = a (u) ·  b (v) · c (y)
uy=h vk=y
R [G] := {a : G → R| |{g :∈ G} : a (g) 6= 0| < ∞} . X
= a (u) · (b (v) · c (y))
We will identify a ∈ R [G] with the formal sum, uvk=h
X
X = a (u) · b (v) · c (k) .
a := a (g) g. uvk=h
g∈G
For distributivity we find,
We define (a + b) (g) := a (g) + b (g) and X X
  ! [(a + b) · c] (h) = (a + b) (g) · c (k) = (a (g) + b (g)) · c (k)
X X X gk=h gk=h
a·b= a (g) g  b (k) k = a (g) b (k) gk X
g∈G k∈G g,k∈G = (a (g) · c (k) + b (g) · c (k))
    gk=h
X X
a (g) · c (k) + b (g) · c (k)
X X X X
a (g) b g −1 h  h. =


=  a (g) b (k) h = 
h∈G gk=h h∈G g∈G gk=h gk=h

= [a · c + b · c] (h)
So formally we define,
X
X
X with a similar computation showing c · (a + b) = c · a + c · b.
a (g) b g −1 h = a (hg) b g −1 = a hg −1 b (g)


(a · b) (h) :=
g∈G g∈G g∈G
X
= a (g) b (k) .
gk=h

We now claim that R is a ring which is non – commutative when G is non-

abelian.

Page: 20 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

4
Lecture 4

4.1 Units Example 4.5. In R, the units are exactly the elements in R× := R \ {0} that is
U (R) = R× .
Definition 4.1. Suppose R is a ring with identity. A unit of a ring is an
element a ∈ R such that there exists an element b ∈ R with ab = ba = 1. We Example 4.6. Let R be the non-commutative ring of linear maps from R∞ to
let U (R) ⊂ R denote the units of R. R∞ where
R∞ = {(a1 , a2 , a3 , . . . ) : ai ∈ R for all i} ,
Notice that in fact a = b in this definition since, which is a vector space over R. Further let A, B ∈ R be defined by
a = a · 1 = a (ub) = (au) b = 1 · b = b. A (a1 , a2 , a3 , . . . ) = (0, a1 , a2 , a3 , . . . ) and
Moreover this argument shows that a satisfying au = 1 = ua is unique if it B (a1 , a2 , a3 , . . . ) = (a2 , a3 , a4 , . . . ) .
exists. For this reason we will write u−1 for a.
Then BA = 1 where
Proposition 4.2. The set U (R) equipped the multiplication law of R is a group.
1 (a1 , a2 , a3 , . . . ) = (a1 , a2 , a3 , . . . )
Proof. This is a straight forward verification – see the homework assign-
while
ment. The main point is to observe that u, v ∈ U (R) , then a := v −1 u−1
AB (a1 , a2 , a3 , . . . ) = (0, a2 , a3 , . . . ) 6= 1 (a1 , a2 , a3 , . . . ) .
satisfies, a (uv) = 1 = (uv) a, showing U (R) is closed under the multiplication
operation of R. This shows that even though BA = 1 it is not necessarily true that AB = 1.
Neither A nor B are units of R∞ .
Example 4.3. In M2 (R), the units in this ring are exactly the elements in
GL(2, R), i.e.

U (M2 (R)) = GL(2, R) = {A ∈ M2 (R) : det A 6= 0} .

4.2 (Zero) Divisors and Integral Domains

If you look back at last quarters notes you will see that we have already Definition 4.7 (Divisors). Let R be a ring. We say that for elements a, b ∈ R
proved the following theorem. I will repeat the proof here for completeness. that a divides b if there exists an element c such that ac = b.

Theorem 4.4 (U (Zm ) = U (m)). For any m ≥ 2, Note that if R = Z then this is the usual notion of whether one integer
evenly divides another, e.g., 2 divides 6 and 2 doesn’t divide 5.
U (Zm ) = U (m) = {a ∈ {1, 2, . . . , m − 1} : gcd (a, m) = 1} .
Definition 4.8 (Zero divisors). A nonzero element a ∈ R is called a zero
Proof. If a ∈ U (Zm ) , there there exists r ∈ Zm such that 1 = r · a = divisor if there exists another nonzero element b ∈ R such that ab = 0, i.e. a
ra mod m. Equivalently put, m| (ra − 1) , i.e. there exists t such that ra − 1 = divides 0 in a nontrivial way. (The trivial way for a|0 is; 0 = a · 0 as this always
tm. Since 1 = ra − tm it follows that gcd (a, m) = 1, i.e. that a ∈ U (m) . holds.)
Conversely, if a ∈ U (m) ⇐⇒ gcd (a, m) = 1 which we know implies there Definition 4.9 (Integral domain). A commutative ring R with no zero di-
exists s, t ∈ Z such that sa + tm = 1. Taking this equation mod m and letting visors is called an integral domain (or just a domain). Alternatively put,
b := s mod m ∈ Zm , we learn that b · a = 1 in Zm , i.e. a ∈ U (Zm ) . R should satisfy, ab 6= 0 for all a, b ∈ R with a 6= 0 6= b.
22 4 Lecture 4

Example 4.10. The most familiar rings to you, Z, Q, R, and C have no zero- Theorem 4.18 (Cancellation). If R is an integral domain and ab = ac with
divisors and hence are integral domains.. In these number systems, it is a fa- a 6= 0, then b = c. Conversely if R is a commutative ring with identity satisfying
miliar fact that ab = 0 implies either a = 0 or b = 0. Another integral domain this cancellation property then R has no zero divisors and hence is an integral
is the polynomial ring R [x] , see Proposition 4.13 below. domain.

Example 4.11. The ring, Z6 , is not an integral domain. For example, 2 · 3 = 0 Proof. If ab = ac, then a(b − c) = 0. Hence if a 6= 0 and R is an integral
with 2 6= 0 6= 3, so both 2 and 3 are zero divisors. domain, then b − c = 0, i.e. b = c.
Conversely, if R satisfies cancellation and ab = 0. If a 6= 0, then ab = a · 0
Lemma 4.12. The ring Zm is an integral domain iff m is prime. and so by cancellation, b = 0. This shows that R has no zero divisors.

Proof. If m is prime we know that U (Zm ) = U (m) = Zm \ {0} . Therefore Example 4.19. The ring, M2 (R) contains many zero divisors. For example
if a, b ∈ Zm with a 6= 0 and ab = 0 then b = a−1 ab = a−1 0 = 0.     
0a 0b 00
If m = a · b with a, b ∈ Zm \ {0} , then ab = 0 while both a and b are not = .
00 00 00
equal to zero in Zm .
So in M2 (R) we can not conclude that B = 0 if AB = 0 with A 6= 0, i.e.
Proposition 4.13. If R is an integral domain, then so is R [x] . Conversely if cancellation does not hold.
R is not an integral domain then neither is R [x] .

Proof. If f, g ∈ R [x] are two non-zero polynomials. Then f = an xn + l.o.ts.

4.3 Fields
(lower order terms) and g = bm xm + l.o.ts. with an 6= 0 6= bm and therefore,
If we add one more restriction to a domain we get a familiar class of objects
f g = an bm xn+m + l.o.ts. 6= 0 since an bm 6= 0. called fields.
The proof of the second assertion is left to the reader. Definition 4.20 (Fields). A ring R is a field if R is a commutative ring with
identity and U (R) = R \ {0}, that is, every non-zero element of R is a unit, in
Example 4.14. All of the following rings are integral domains; Z [x] , Q [x] , R [x] , other words has a multiplicative inverse.
and C [x] . We also know that Zm [x] is an integral domain iff m is prime.
Lemma 4.21 (Fields are domains). If R is a field then R is an integral
Example 4.15. If R is the direct product of at least 2 rings, then R has zero domain.
divisors. For example if R = Z ⊕ Z, then (0, b)(a, 0) = (0, 0) for all a, b ∈ Z.
Proof. If R is a field and xy = 0 in R for some x, y with x 6= 0, then
Example 4.16. If R is an integral domain, then any unital subring S ⊂ R is also
0 = x−1 0 = x−1 xy = y.
an integral domain. In particular, for any θ ∈ C, then Z [θ] , Q [θ] , and R [θ] are
all integral domains.

Remark 4.17. It is not true that if R is not an integral domain then every sub- Example 4.22. Z is an integral domain that is not a field. For example 2 6= 0
ring, S ⊂ R is also not an integral domain. For an example, take R := Z ⊕ Z has no multiplicative inverse. The inverse to 2 should be 21 which exists in Q
and S := {(a, a) : a ∈ Z} ⊂ R. (In the language of Section 5.1 below, S = but not in Z. On the other hand, Q and R are fields as the non-zero elements
{n · (1, 1) : n ∈ Z} which is the sub-ring generated by 1 = (1, 1) . Similar to have inverses back in Q and R respectively.
this counter example, commutative ring with identity which is not an integral Example 4.23. We have already seen that Zm is a field iff m is prime. This
domain but has characteristic being either 0 or prime would give a counter follows directly form the fact that U (Zm ) = U (m) and U (m) = Zm \ {0} iff m
example.) is prime. Recall that we also seen that Zm is an integral domain iff m is prime
so it follows Zm is a field iff it is an integral domain iff m is prime. When p is
Domains behave more nicely than arbitrary rings and for a lot of the quarter
prime, we will often denote Zp by Fp to indicate that we are viewing Zp is a
we will concentrate exclusively on domains. But in a lot of ring theory it is very
field.
important to consider rings that are not necessarily domains like matrix rings.

Page: 22 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

5
Lecture 5

In fact, there is another way we could have seen that Zp is a field, using the Example 5.3. I claim that R := Z3 [i] = Z3 + iZ3 is a field where we use the
following useful lemma. multiplication rule,

Lemma 5.1. If R be an integral domain with finitely many elements, then R (a + ib) (c + id) = (ac − bd) + i (bc + ad) .
is a field.
The main point to showing this is a field beyond showing R is a ring (see
Proof. Let a ∈ R with a 6= 0. We need to find a multiplicative inverse for −1
Proposition 3.7) is to show (a + ib) exists in R whenever a + ib 6= 0. Working
a. Consider a, a2 , a3 , . . . . Since R is finite, the elements on this list are not all formally for the moment we should have,
distinct. Suppose then that ai = aj for some i > j ≥ 1. Then aj ai−j = aj · 1.
By cancellation, since R is a domain, ai−j = 1. Then ai−j−1 is the inverse for 1 a − ib
= 2 .
a. Note that ai−j−1 ∈ R makes sense because i − j − 1 ≥ 0. a + ib a + b2
For general rings, an only makes sense
n for n ≥ 1. If 1 ∈ R and a ∈ U (R) ,
we may define a0 = 1 and a−n = a−1 for n ∈ Z+ . As for groups we then This suggest that
−1
−1
have an am = an+m for all m, n ∈ Z. makes sense for all n ∈ Z, but in generally (a + ib) = a2 + b2 (a − ib) .
negative powers don’t always make sense in a ring. Here is another very inter- In order for the latter expression to make sense we need to know that a2 +b2 6= 0
esting example of a field, different from the other examples we’ve written down in Z3 if (a, b) 6= 0 which we can check by brute force;
so far.
a 000111222
Example
√ 5.2. Lets check that C is a field. Given 0 6= a + bi ∈ C, a, b ∈ R,
−1 b 012012012
i = −1, we need to find (a + ib) ∈ C. Working formally; we expect, .
N (a + ib)
0 1 1 1 2 2 1 2 2
−1 1 1 a − bi a − bi = a2 + b2
(a + ib) = =
a + bi a + bi a − bi a2 + b2
a b Alternatively we may show Z3 [i] is an integral domain and then use Lemma
= 2 − 2 i ∈ C, 5.1. Notice that
a + b2 a + b2
which makes sense if N (a + ib) := a2 + b2 6= 0, i.e. a + ib 6= 0. A simple direct (a + ib) (c + id) = 0 =⇒ (a − ib) (a + ib) (c + id) = 0 i.e.
a2 + b2 (c + id) = 0.


check show that this formula indeed gives an inverse to a + ib;
 
a b So using the chart above, we see that a2 + b2 = 0 iff a + ib = 0 and therefore,
(a + ib) 2 − 2 i
a + b2 a + b2 if a + ib 6= 0 then c + id = 0.
1 1
a2 + b2 = 1.


= 2 2
(a + ib) (a − ib) = 2 2
a +b a +b
5.1 Characteristic of a Ring
So if a + ib 6= 0 we have shown
a b Notation 5.4 Suppose that a ∈ R where R is a ring. Then for n ∈ Z we define
(a + bi)−1 = − 2 i. n · a ∈ R by, 0Z · a = 0R and
a2 +b2 a + b2
24 5 Lecture 5
n times

 z }| { which proves Eq. (5.3).
a + ··· + a if n ≥ 1

Letting x := sgn(m)sgn(n)a, we have

n·a= |n| times .

 z
 }| { |m| times
−(a + · · · + a) = |n| · (−a) if n ≤ −1 z }| {
m · (n · a) = |m| · (|n| · x) = (|n| · x + · · · + |n| · x)
So 3 · a = a + a + a while −2 · a = −a − a. |m| times
z }| {
Lemma 5.5. Suppose that R is a ring and a, b ∈ R. Then for all m, n ∈ Z we z
|n| times
}| { z
|n| times
}| {
have = (x + · · · + x) + · · · + (x + · · · + x)
(m · a) b = m · (ab) , (5.1) = (|m| |n|) · x = mn · a.
a (m · b) = m · (ab) . (5.2)

We also have Corollary 5.6. If R is a ring, a, b ∈ R, and m, n ∈ Z, then

− (m · a) = (−m) · a = m · (−a) and (5.3) (m · a) (n · b) = mn · ab. (5.5)

m · (n · a) = mn · a. (5.4)
Proof. Using Lemma 5.5 gives;
Proof. If m = 0 both sides of Eq. (5.1) are zero. If m ∈ Z+ , then using the (m · a) (n · b) = m · (a (n · b)) = m · (n · (ab)) = mn · ab.
distributive associativity laws repeatedly gives;
m times
z }| {
(m · a) b = (a + · · · + a)b Corollary 5.7. Suppose that R is a ring and a ∈ R. Then for all m, n ∈ Z,
m times
z }| { (m · a) (n · a) = mn · a2 .
= (ab + · · · + ab) = m · (ab) .
In particular if a = 1 ∈ R we have,
If m < 0, then
(m · 1) (n · 1) = mn · 1.
(m · a) b = (|m| · (−a)) b = |m| · ((−a) b) = |m| · (−ab) = m · (ab)
Unlike the book, we will only bother to define the characteristic for rings
which completes the proof of Eq. (5.1). The proof of Eq. (5.2) is similar and which have an identity, 1 ∈ R.
will be omitted.
Definition 5.8 (Characteristic of a ring). Let R be a ring with 1 ∈ R. The
If m = 0 Eq. (5.3) holds. If m ≥ 1, then
characteristic, chr (R) , of R is is the order of the element 1 in the additive
m times m times group (R, +). Thus n is the smallest number in Z+ such that n · 1 = 0. If no
z }| { z }| { such n ∈ Z+ exists, we say that characteristic of R is 0 by convention and write
− (m · a) = − (a + · · · + a) = ((−a) + · · · + (−a)) = m · (−a) = (−m) · a.
chr (R) = 0.
If m < 0, then Lemma 5.9. If R is a ring with identity and chr (R) = n ≥ 1, then n · x = 0
for all x ∈ R.
− (m · a) = − (|m| · (−a)) = (− |m|) · (−a) = m · (−a)
Proof. For any x ∈ R, n · x = n · (1x) = (n · 1) x = 0x = 0.
and
Lemma 5.10. Let R be a domain. If n = chr (R) ≥ 1, then n is a prime
− (m · a) = − (|m| · (−a)) = (|m| · (− (−a))) = |m| · a = (−m) · a. number.

Page: 24 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

 Proof. If n is not prime, say n = pq with 1 < p < n and 1 < q < n, then

(p · 1R ) (q · 1R ) = pq · (1R 1R ) = pq · 1R = n · 1R = 0.

As p · 1R 6= 0 and q · 1R 6= 0 and we may conclude that both p · 1R and q · 1R

are zero divisors contradicting the assumption that R is an integral domain.
h√ i √
Example 5.11. The rings Q, R, C, Q d := Q + Q d, Z [x] , Q [x] , R [x] , and
Z [x] all have characteristic 0.
For each m ∈ Z+ , Zm and Zm [x] are rings with characteristic m.

Example 5.12. For each prime, p, Fp := Zp is a field with characteristic p. We

also know that Z3 [i] is a field with characteristic 3. Later, we will see other
examples of fields of characteristic p.
6
Lecture 6

6.1 Square root field extensions of Q a −b √ h√ i

ξ= + 2 ∈ Q 2 .
√ √ a2 − 2b2 a2 − 2b2
Recall that 2 is irrational. Indeed suppose that 2 = m/n ∈ Q and, with out Moreover, it is easy√to check this ξ works provided a2 −2b2 6= 0. But if a2 −2b2 =
√
loss of generality, assume that gcd (m, n) = 1. Then m2 = 2n2 from which it 0 with b 6= 0, then 2 = |a| / |b| showing 2 is irrational which
√  we know to be
follows that 2|m2 and so 2|m by Euclid’s lemma. However, it now follows that false – see Proposition 6.2 below for details. Therefore, Q 2 is a field.
22 |2n2 and so 2|n2 which again by Euclid’s lemma implies 2|n. However, we √ 
assumed that Observe that Q ( R := Q 2 ( R. Why is this? One reason is that
√ m and n were relatively prime and so we have a contradiction √ 
and
 √ hence 2 is indeed irrational. As a consequence √
of this fact, we know that R := Q 2 is countable and R is uncountable. Or it is not hard to show that
1, 2 are linearly independent over Q, i.e. if a + b 2 = 0 then a = 0 = b. an irrational number √ selected more or less
√ at random√ is not in R. For example,
you could show that 3 ∈ / R. Indeed if 3 = a + b 2 for some a, b ∈ Q then
Example 6.1. In this example we will show, √
h√ i √ 3 = a2 + 2ab 2 + 2b2
R = Q 2 = {a + b 2 : a, b ∈ Q} (6.1) √ √
and hence 2ab 2 = 3 − a2 − 2b2 . Since 2 is irrational,
√ this can only happen
√  if either a = 0 or b = 0. If b = 0 we will have 3 ∈ Q which is false and if
is a field. Using similar techniques to those in Example 3.4 we see that Q 2
a = 0 we will have 3 = 2b2 . Writing b = kl , this with gcd (k, l) = 1, we find
may be described as in Eq. (6.1) and hence is a subring of Q by Proposition
3l2 = 2k 2 and therefore 2|l by Gauss’ lemma. Hence 22 |2k 2 which implies 2|k
3.3. Alternatively one may check directly that the right side of Eq. (6.1) is a
subring of Q since;
and
√ therefore √ gcd (k, l) ≥ 2 > 1 which is a contradiction. Hence it follows that
3 6= a + b 2 for any a, b ∈ Q.
√ √ √
a + b 2 − (c + d 2) = (a − c) + (b − d) 2 ∈ R The following proposition is a natural extension of Example 6.1.
h√ i
and Proposition 6.2. For all d ∈ Z \ {0} , F := Q d is a field. (As we will see
√ √ √ √ in the proof, we need only consider those d which are “square prime” free.
(a + b 2)(c + d 2) = ac + bc 2 + ad 2 + bd(2) h√ i √
√ Proof. As F := Q d = Q + Q d is a subring of R which is an integral
= (ac + 2bd) + (bc + ad) 2 ∈ R.
domain, we know that F is again an integral domain. Let d = εpk11 . . . pknn
So by either means we see that R is a ring and in fact an integral domain by
Q ε ∈ {±1} , p1 , . .√
with √ distinct primes, and ki ≥ 1. Further let δ =
. , pn being
Example 4.16. It does not have finitely many elements so we can’t use Lemma ε i:ki is odd pi , then d = m δ for some integer m and therefore it easily
√
−1 h√ i
5.1 to show it is a field. However, we can find a + b 2 directly as follows. follows that F = Q δ . So let us now write δ = εp1 . . . pk with ε ∈ {±1} ,
√
−1
If ξ = a + b 2 , then p1 , . . . , pk being distinct primes so that δ is square prime free.
√ √
1 = (a + b 2)ξ Working as above we look for the inverse to a + b δ when (a, b) 6= 0. Thus
and therefore, we will look for u, v ∈ Q such that
√ √ √
 √  √ 
a − b 2 = (a − b 2)(a + b 2)ξ = a2 − 2b2 ξ

1= a+b δ u+v δ .
√
which implies, Multiplying this equation through by a − b δ shows,
28 6 Lecture 6
√
 √ 
a − b δ = a2 − b2 δ u + v δ Example 6.5 (Conjugation isomorphism). Let ϕ : C → C be defined by ϕ (z) =
z̄ where for z = x + iy, z̄ := x − iy is the complex conjugate of z. Then it is
so that routine to check that ϕ is a ring isomorphism. Notice that z = z̄ iff z ∈ R. There
√ a b √ is analogous conjugation isomorphism on Q [i] , Z [i] , and Zm [i] (for m ∈ Z+ )
u+v δ = − δ. (6.2)
a2 − b2 δ a2 − b2 δ with similar properties.
Thus we may define,
Here is another example in the same spirit of the last example.
 √ −1 a b √
a+b δ = − 2 δ
√  √ 
2 2
a −b δ 2
a −b δ Example 6.6 (Another conjugation isomorphism). Let ϕ : Q 2 → Q 2 be
defined by  √  √
provided a2 − b2 δ 6= 0 when (a, b) 6= (0, 0) . ϕ a + b 2 = a − b 2 for all a, b ∈ Q.
Case 1. If δ < 0 then a2 − b2 δ = a2 + |δ| b2 = 0 iff a = 0 = b.
Case 2. If δ ≥ 2 and suppose that a, b ∈ Q with a2 = b2 δ. For sake of Then ϕ is a ring isomorphism. Again this is routine to check. For example,
contradiction suppose that b 6= 0. By multiplying a2 = b2 δ though by the  √   √   √  √ 
denominators of a2 and b2 we learn there are integers, m, n ∈ Z+ such that ϕ a+b 2 ϕ u+v 2 = a−b 2 u−v 2
m n √
m2 = n2 δ. By replacing m and n by gcd(m,n) and gcd(m,n) , we may assume that
= au + 2bv − (av + bu) 2
m and n are relatively prime.


We now have p1 | n2 δ implies p1 |m2 which by Euclid’s lemma implies that while
p1 |m. Thus we learn that p21 |m2 = n2 p1 , . . . , pk and therefore that p1 |n2 . An-  √  √   √ 
other application of Euclid’s lemma shows p1 |n. Thus we have shown that p1 is ϕ a + b 2 u + v 2 = ϕ au + 2bv + (av + bu) 2
a divisor of both m and n contradicting the fact that m and n were relatively √
prime. Thus we must conclude that b = 0 = a. Therefore a2 − b2 δ = 0 only if = au + 2bv − (av + bu) 2.
a = 0 = b. √ 
Later on we will show the following; Notice that for ξ ∈ Q 2 , ϕ (ξ) = ξ iff ξ ∈ Q.

Fact 6.3 Suppose that θ ∈ C is the root of some polynomial in Q [x] , then Q [θ] Example 6.7. The only ring homomorphisms, ϕ : Z → Z are ϕ (a) = a and
is a sub-field of C. ϕ (a) = 0 for all a ∈ Z. Indeed, if ϕ : Z → Z is a ring homomorphism and
t := ϕ (1) , then t2 = ϕ (1) ϕ (1) = ϕ (1 · 1) = ϕ (1) = t. The only solutions to
Recall that we already know Q [θ] is an integral domain. To prove that Q [θ] t2 = t in Z are t = 0 and t = 1. In the first case ϕ ≡ 0 and in the second ϕ = id.
is a field we will have to show that for every nonzero z ∈ Q [θ] that the inverse,
z −1 ∈ C, is actually back in Q [θ] .

6.2 Homomorphisms
Definition 6.4. Let R and S be rings. A function ϕ : R → S is a homomor-
phism if

ϕ(r1 r2 ) = ϕ(r1 )ϕ(r2 ) and

ϕ(r1 + r2 ) = ϕ(r1 ) + ϕ(r2 )

for all r1 , r2 ∈ R. That is, ϕ preserves addition and multiplication. If we further

assume that ϕ is an invertible map (i.e. one to one and onto), then we say
ϕ : R → S is an isomorphism and that R and S are isomorphic.

Page: 28 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

7
Lecture 7

Example 7.1. Suppose that g ∈ M2 (R) is a unit, i.e. g −1 exists. Then ϕ : To each s ∈ S, there exists a ∈ R such that ϕ (a) = s. Therefore,
M2 (R) → M2 (R) defined by,
ϕ (1R ) s = ϕ (1R ) ϕ (a) = ϕ (1R a) = ϕ (a) = s
ϕ (A) := gAg −1 for all A ∈ M2 (R) , and
is a ring isomorphism. For example, sϕ (1R ) = ϕ (a) ϕ (1R ) = ϕ (a1R ) = ϕ (a) = s.

ϕ (A) ϕ (B) = gAg −1 gBg −1 = gAg −1 gBg −1 = gABg −1 = ϕ (AB) .

Since these equations hold for all s ∈ S, it follows that ϕ (1R ) is an (the) identity
in S.
Observe that ϕ−1 (A) = g −1 Ag and ϕ (I) = I.
Definition 7.5. As usual, if ϕ : R → S is a ring homomorphism we let
Proposition 7.2 (Homomorphisms from Z). Suppose that R is a ring and
a ∈ R is an element such that a2 = a. Then there exists a unique ring ho- ker (ϕ) := {r ∈ R : ϕ (r) = 0} = ϕ−1 ({0S }) ⊂ R.
momorphism, ϕ : Z → R such that ϕ (1) = a. Moreover, ϕ (k) = k · a for all
Lemma 7.6. If ϕ : R → S is a ring homomorphism, then ker (ϕ) is an ideal of
k ∈ Z.
R.
Proof. Recall from last quarter that, ϕ (n) := n · a for all n ∈ Z is a group
Proof. We know from last quarter that ker (ϕ) is a subgroup of (R, +) . If
homomorphism. This is also a ring homomorphism since,
r ∈ R and n ∈ ker (ϕ) , then
ϕ (m) ϕ (n) = (m · a) (n · a) = mn · a2 = mn · a = ϕ (mn) ,
ϕ (rn) = ϕ (r) ϕ (n) = ϕ (r) 0 = 0 and
wherein we have used Corollary 5.6 for the second equality. ϕ (nr) = ϕ (n) ϕ (r) = 0ϕ (r) = 0,

Corollary 7.3. Suppose that R is a ring with 1R ∈ R. Then there is a unique which shows that rn and nr ∈ ker (ϕ) for all r ∈ R and n ∈ ker (ϕ) .
homomorphism, ϕ : Z → R such that ϕ (1Z ) = 1R .
Example 7.7. Let us find all of the ring homomorphisms, ϕ : Z → Z10 and their
Proposition 7.4. Suppose that ϕ : R → S is a ring homomorphism. Then; kernels. To do this let t := ϕ (1) . Then t2 = ϕ (1) ϕ (1) = ϕ (1 · 1) = ϕ (1) = t.
The only solutions to t2 = t in Z10 are t = 0, t = 1, t = 5 and t = 6.
1. ϕ (0) = 0,
2. ϕ (−r) = −ϕ (r) for all r ∈ R, 1. If t = 0, then ϕ ≡ 0 and ker (ϕ) = Z.
3. ϕ(r1 − r2 ) = ϕ(r1 ) − ϕ(r2 ) for all r1 , r2 ∈ R. 2. If t = 1, then ϕ (x) = x mod 10 and ker ϕ = h10i = h0i = {0} ⊂ Z.
4. If 1R ∈ R and ϕ is surjective, then ϕ (1R ) is an identity in S. 3. If t = 5, then ϕ (x) = 5x mod 10 and x ∈ ker ϕ iff 10|5x iff 2|x so that
5. If ϕ : R → S is an isomorphism of rings, then ϕ−1 : S → R is also a ker (ϕ) = h2i = {0, 2, 4, 8} .
isomorphism. 4. If t = 6, then ϕ (x) = 6x mod 10 and x ∈ ker ϕ iff 10|6x iff 5|x so that
ker (ϕ) = h5i = {0, 5} ⊂ Z.
Proof. Noting that ϕ : (R, +) → (S, +) is a group homomorphism, it follows
that items 1. – 3. were covered last quarter when we studied groups. The proof Proposition 7.8. Suppose n ∈ Z+ , R is a ring, and a ∈ R is an element such
of item 5. is similar to the analogous statements for groups and hence will be that a2 = a and n · a = 0. Then there is a unique homomorphism, ϕ : Zn → R
omitted. So let me prove item 4. here. such that ϕ (1) = a and in fact ϕ (k) = k · a for all k ∈ Zn .
30 7 Lecture 7

Proof. This has a similar proof to the proof of Proposition 7.2. So if y, x ∈ R with y > x, then ϕ (y) − ϕ (x) = ϕ (y − x) > 0, i.e. ϕ is order
preserving.
Corollary 7.9. Suppose that R is a ring, 1R ∈ R, and chr (R) = n ∈ Z+ . Then Finally, let a ∈ R and choose rational numbers xn , yn ∈ Q such that xn <
there is a unique homomorphism, ϕ : Zn → R such that ϕ (1Zn ) = 1R which is a < yn with xn ↑ a and yn ↓ a as n → ∞. Then
given by ϕ (m) = m · 1R for all m ∈ Zn . Moreover, ker (ϕ) = h0i = {0} .
xn = ϕ (xn ) < ϕ (a) < ϕ (yn ) = yn for all n.
Example 7.10. Suppose that ϕ : Z10 → Z10 is a ring homomorphism and t :=
2
ϕ (1) . Then t2 = ϕ (1) = ϕ (1) = t, and therefore t2 = t. Moreover we must Letting n → ∞ in this last equation then shows, a ≤ ϕ (a) ≤ a, i.e. ϕ (a) = a.
have 0 = ϕ (0) = ϕ (10 · 1) = 10 · t which is not restriction on t. As we have Since a ∈ R was arbitrary, we may conclude that ϕ is the identity map on R.
seen the only solutions to t2 = t in Z10 are t = 0, t = 1, t = 5 and t = 6. Thus
ϕ must be one of the following; ϕ ≡ 0, ϕ = id, ϕ (x) = 5x, or ϕ (x) = 6x for all
x ∈ Z10 . The only ring isomorphism is the identity in this case. If ϕ (x) = 5x

Example 7.11. Suppose that ϕ : Z12 → Z10 is a ring homomorphism and let
t := ϕ (1). Then as before, t2 = t and this forces t = 0, 1, 5, or 6. In this case
we must also require 12 · t = 0, i.e. 10|12 · t, i.e. 5|t. Therefore we may now only
take t = 0 or t = 5, i.e.

ϕ (x) = 0 for all x ∈ Z12 or

ϕ (x) = 5x mod 10 for all x ∈ Z12

are the only such homomorphisms.

Theorem 7.12 (Not covered in class). If ϕ : R → R is a ring homomor-

phism, then ϕ is either the zero or the identity homomorphism.

Proof. If t = ϕ (1) , then as above, t2 = t, i.e. t (t − 1) = 0. Since R is a

field this implies that t = 0 or t = 1. If t = 0, then for all a ∈ R,

ϕ (a) = ϕ (a · 1) = ϕ (a) ϕ (1) = ϕ (a) · 0 = 0,

i.e. ϕ is the zero homomorphism. So we may now assume that t = 1.

If t = 1,
ϕ (n) = ϕ (n · 1) = n · ϕ (1) = n · 1 = n
for all n ∈ Z. Therefore for n ∈ N \ {0} and m ∈ Z,
 m m m
m = ϕ (m) = ϕ n · = ϕ (n) ϕ = nϕ
n n n
from which it follows that ϕ (m/n) = m/n. Thus we now know that ϕ|Q is the
identity.
Since ker (ϕ) 6= R, we must have ker (ϕ) = {0} so that ϕ is √ injective. In
particular ϕ (b) 6= 0 for all b 6= 0. Moreover if a > 0 in R and b := a, then
2
ϕ (a) = ϕ b2 = [ϕ (b)] > 0.

Page: 30 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

8
Lecture 8

Remark 8.1 (Comments on ideals). Let me make two comments on ideals in a As before, one shows that
commutative ring, R.
ϕ (m) = m · 6 = (6m) mod 10 and ψ (m) = m · 5 = (5m) mod 10.
1. To check that a non-empty subset, S ⊂ R, is an ideal, we should show
(S, +) is a subgroup of R and that RS ⊂ S. Since R is commutative, you Example 8.5 (Divisibility tests). Let n = ak ak−1 . . . a0 be written in decimal
do not have to also show SR ⊂ S. This is because RS = SR in when R is form, so that
commutative. Xk

2. If a ∈ R, the principle ideal generated by a is defined by; n= ai 10i . (8.1)

i=0
hai := Ra = {ra : r ∈ R} . Applying the ring homomorphism, mod 3 and mod 9 to this equation shows,

It is easy to check that this is indeed an ideal. So for example if R = R [x] k

X i
then hxi = R [x] · x which is the same as the polynomials without a constant n mod 3 = ai mod 3 · (10 mod 3)
term, i.e. p (x) = a1 x + a
2 x2 + · · · + an xn . The coefficient a0 = 0. Similarly, i=0
x2 + 1
= R [x] x2 + 1 is the collection of all polynomials which contain k
!
X
x2 + 1 as a factor. = ai mod 3
i=0
Recall from last time:
and similarly,
1. If a ∈ R satisfies a2 = a, then ϕ (k) := k · a is a ring homomorphism from !
k k
Z → R. X i
X
n mod 9 = ai mod 9 · (10 mod 9) = ai mod 9.
2. If we further assume that n · a = 0 for some n ∈ Z+ , then ϕ (k) := k · a also
i=0 i=0
defines a ring homomorphism from Zn → R.
P 
k
Example 8.2. For any m > 1, ϕ : Z → Zm given by ϕ (a) = a · 1Zm = a mod m Thus we learn that n mod 3 = 0 iff i=0 ai mod 3 = 0 i.e. 3|n iff
P 
is a ring homomorphism. This also follows directly from the properties of the 3|
k
i=0 ai . Similarly, since 10 mod 9 = 1, the same methods show 9|n iff
(·) mod m – function. In this case ker (ϕ) = hmi = Zm. P 
k
9| i=0 ai . (See the homework problems for more divisibility tests along
Example 8.3. If n ∈ Z+ and m = kn with k ∈ Z+ , then there is a unique ring these lines. Also consider what this test gives if you apply mod 2 to Eq. (8.1).)
homomorphisms, ϕ : Zm → Zn such that ϕ (1m ) = 1n . To be more explicit,
Theorem 8.6. Let R be a commutative ring with 1 ∈ R. To each a ∈ R with
ϕ (a) = ϕ (a · 1m ) = a · ϕ (1m ) = a · 1n = (a mod n) · 1n = a mod n. a2 + 1 = 0, there is a unique ring homomorphism ϕ : Z [i] → R such that
ϕ (1) = 1 and ϕ (i) = a.
Example 8.4. In Z10 , the equation, a2 = a has a solutions a = 5 and a = 6.
Notice that |5| = 2 and |6| = |gcd (10, 6)| = |2| = 5. Thus we know that for any Proof. Since Z [i] is generated by i, we see that ϕ is completely determined
k ≥ 1 there are ring homomorphisms, ϕ : Z5k → Z10 and ψ : Z2k → Z10 such by a := ϕ (i) ∈ R. Now we can not choose a arbitrarily since we must have
that
2
a2 = ϕ (i) = ϕ i2 = ϕ (−1) = −1R ,


ϕ (15k ) = 6 and ψ (12k ) = 5.
i.e. a2 + 1 = 0.
Conversely given a ∈ R such that a2 + 1 = 0, we should define
ϕ (x + iy) = x1 + ya for all x, y ∈ Z,
where ya = a + a + · · · + a – y times. The main point in checking that ϕ is a
homomorphism is to show it preserves the multiplication operation of the rings.
To check this, let x, y, u, v ∈ Z and consider;
ϕ ((x + iy) (u + iv)) = ϕ (xu − yv + i (xv + yu)) = (xu − yv) 1R + (xv + yu) a.
On the other hand
ϕ (x + iy) ϕ (u + iv) = (x1R + ya) (u1R + va)
= (x1R + ya) (u1R + va)
= xu1R + yva2 + yua + xva
= (xu − yv) 1R + (yu + xv) a
= ϕ ((x + iy) (u + iv)) .
Thus we have shown ϕ (ξη) = ϕ (ξ) ϕ (η) for all ξ, η ∈ Z [i] . The fact that
ϕ (ξ + η) = ϕ (ξ) + ϕ (η) is easy to check and is left to the reader.
Remark 8.7. This could be generalized by supposing that a, b ∈ R with b2 = b
and a2 + b = 0. Then we would have ϕ (x + yi) = x · b + y · a would be the
desired homomorphism. Indeed, let us observe that
ϕ (x + iy) ϕ (u + iv) = (xb + ya) (ub + va)
= (xb + ya) (ub + va)
= xub2 + yva2 + yua + xva
= (xu − yv) b + (yu + xv) a
= ϕ ((x + iy) (u + iv)) .
Example 8.8. Let ϕ : Z [i] → Z3 [i] be the unique homomorphism such that
ϕ (1) = 1 and ϕ (i) = i, i.e.
ϕ (a + ib) = a · 1 + b · i = a mod 3 + (b mod 3) i ∈ Z3 [i] .
Notice that
ker (ϕ) = {a + bi : a, b ∈ h3i ⊂ Z} = h3i + h3i i.
Here is a more interesting example.
Example 8.9. In Z10 we observe that 32 = 9 = −1 and also 7 = −3 has this
2
property, namely 72 = (−3) = 32 = 9 = −1. Therefore there exists a unique
homomorphism, ϕ : Z [i] → Z10 such that ϕ (1) = 1 and ϕ (i) = 7 = −3. The
explicit formula is easy to deduce,
ϕ (a + bi) = a · 1 + b · 7 = (a − 3b) mod 10.
9
Lecture 9

Lemma 9.1. If ϕ : R → S is a ring homomorphism, then ker (ϕ) is an ideal of Proof. The elements of R/I are the left cosets of I in the group (R, +).
R. There is nothing new here. R/I is itself a group with the operation + defined
by (a + I) + (b + I) = (a + b) + I. This follows from last quarter as I ⊂ R is
Proof. We know from last quarter that ker (ϕ) is a subgroup of (R, +) . If a normal subgroup of (R, +) since (R, +) is abelian. So we only need really to
r ∈ R and n ∈ ker (ϕ) , then check that the definition of product makes sense.
Problem: we are multiplying coset representatives. We have to check that
ϕ (rn) = ϕ (r) ϕ (n) = ϕ (r) 0 = 0 and the resulting coset is independent of the choice of representatives. Thus we
ϕ (nr) = ϕ (n) ϕ (r) = 0ϕ (r) = 0, need to show; if a, b, a0 , b0 ∈ R with

which shows that rn and nr ∈ ker (ϕ) for all r ∈ R and n ∈ ker (ϕ) . a + I = a0 + I and b + I = b0 + I,

Example 9.2. If ϕ : Z → Zm is the ring homomorphism defined by ϕ (a) := then ab + I = a0 b0 + I. By definition of cosets, we have i := a − a0 ∈ I and
a mod m, then j := b − b0 ∈ I. Therefore,
ab = (a0 + i) (b0 + j) = a0 b0 + ib0 + a0 j + ij ∈ a0 b0 + I
ker ϕ = {a ∈ Z : a mod m = 0} = Zm = hmi .
since ib0 + a0 j + ij ∈ I because I is an ideal. So indeed, ab + I = a0 b0 + I and we
We will see many more examples of Lemma 9.1 below. have a well defined product on R/I. Checking that product is associative and
the distributive laws is easy and will be omitted.
Example 9.5. Suppose that I = h4i = Z · 4 ⊂ Z. In this case, if a ∈ Z then
9.1 Factor Rings a − a mod 4 ∈ I and therefore,

Definition 9.3. Let R be a ring, I ⊂ R an ideal. The factor ring R/I is defined [a] = a + I = a mod 4 + I = [a mod 4] .
to be Moreover if 0 ≤ a, b ≤ 3 with a + I = b + I then a − b ∈ I, i.e. a − b is a
R/I := {r + I : r ∈ R} multiple of 4. Since |a − b| < 4, this is only possible if a = b. Thus if we let
with operations S = {0, 1, 2, 3} , then
Z/ h4i = {[m] = m + h4i : m ∈ S} = [S] .
(a + I) + (b + I) := (a + b) + I and
(a + I)(b + I) := (ab) + I. Moreover, we have
[a] [b] = [ab] = [(ab) mod 4]
We may also write [a] for a + I in which cases the above equations become, and
[a] + [b] = [a + b] = [(a + b) mod 4] .
[a] + [b] := [a + b] and [a] [b] := [ab] .
Thus the induced ring structure on S is precisely that of Z4 and so we may
Theorem 9.4. A factor ring really is a ring. conclude;
Z4 3 a → [a] = a + h4i ∈ Z/ h4i
is a ring isomorphism.
10
Lecture 10

Remark 10.1. Roughly speaking, you should think of R/I being R with the Theorem 10.3 (First Isomorphism Theorem). Let R and S be rings and
proviso that we identify two elements of R to be the same if they differ by an ϕ : R → S be a homomorphism. Let
element from I. To understand R/I in more concrete terms, it is often useful
to look for subset, S ⊂ R, such that the map, ϕ(R) = Ran ϕ = {ϕ(r) : r ∈ R} ⊂ S

S 3 a → a + I ∈ R/I and recall that I = ker ϕ := {r ∈ R : ϕ(r) = 0} is an ideal in R. Then ϕ(R) is

a subring of S and ϕ̄ : R/I → ϕ(R) defined by
is a bijection. (We will call such an S a slice.) This allows us to identify R/I
with S and this identification induces a ring structure on S. We will see how ϕ̄ ([r]) = ϕ̄ (r + I) := ϕ (r) for all r ∈ R
this goes in the examples below. Warning: the choice of a slice S is highly
non-unique although there is often a “natural” choice in a given example. The is a ring isomorphism.
point is to make S we need only choose one element from each of the cosets in Proof. We have seen last quarter that ϕ̄ : R/ ker ϕ → ϕ(R) is an (additive)
R/I. group isomorphism. So it only remains to show ϕ̄ preserves the multiplication
Example 9.5 easily generalizes to give the following theorem. We will give operations on ϕ (R) and R/I which goes as follows;
another proof shortly using the first isomorphism theorem, see 10.4 below.
ϕ̄ ([a]) ϕ̄ ([b]) = ϕ (a) ϕ (b)
Theorem 10.2 (Zm ∼
= Z/ hmi). For all m ≥ 2, the map, = ϕ (ab) = ϕ̄ ([ab]) = ϕ̄ ([a] [b]) .

Zm 3 a → [a] = a + hmi ∈ Z/ hmi (10.1)

is a ring isomorphism. Example 10.4 (Z/ (Zm) ∼ = Zm ). Let m ∈ Z+ and ϕ : Z → Zm be the ring
homomorphism, ϕ (x) = x mod m. Since ϕ (Z) = Zm and ker (ϕ) = hmi = Zm,
Proof. The distinct cosets of Z/ hmi are given by the first isomorphism theorem implies, ϕ̄ : Z/ (Zm) → Zm is a ring isomorphism
where ϕ̄ ([a]) = ϕ (a) = a mod m for all a ∈ Z.
{[k] = k + hmi : k = 0, 1, 2 . . . , m − 1}
Example 10.5. Let us consider R := Z [i] / hi − 2i . In this ring [i − 2] = 0 or
and therefore we may take S = Zm . Since [a] = [a mod m] , it is easy to see that equivalently, [i] = [2] . Squaring this equation also shows,
the map in Eq. (10.1) is a ring isomorphism.
2 2
[−1] = i2 = [i] = [2] = 22 = [4]
   

10.1 First Isomorphism Theorem from which we conclude that [5] = 0, i.e. 5 ∈ hi − 2i . This can also be seen
directly since 5 = − (i + 2) (i − 2) ∈ hi − 2i . Using these observations we learn
Recall that two rings, R and S (written R ∼
= S) are isomorphic, if there is a for a + ib ∈ Z [i] that
ring isomorphism, ϕ : R → S. That is ϕ should be a one-to-one and onto ring
homomorphism. [a + ib] = [a + 2b] = [(a + 2b) mod 5] .

Thus, if we define S = {0, 1, 2, 3, 4} , we have already shown that

 R = {[a] : a ∈ S} = [S] .

Now suppose that a, b ∈ S with [a] = [b] , i.e. 0 = [a − b] = [c] where

c = (a − b) mod 5. Since c ∈ hi − 2i we must have

c = (i − 2) (a + bi) = − (2a + b) + (a − 2b) i

from which it follows that a = 2b and

c = − (2a + b) = −5b.

Since 0 ≤ c < 5, this is only possible if c = 0 and therefore,

a = a mod 5 = b mod 5 = b.

Finally let us now observe that

[a] + [b] = [a + b] = [(a + b) mod 5] and

[a] · [b] = [ab] = [(ab) mod 5]

so that the induced ring structure on S is the same a the ring structure on Z5 .
Hence we have proved,

Z5 3 a → [a] = a + hi − 2i ∈ Z [i] / hi − 2i

is an isomorphism of rings.
11
Lecture 11

Example 11.1 (Example 10.5 revisited). In Z5 , we see that 22 = 4 = −1 and Example 11.3. Let I := hxi = R [x] x ⊂ R [x] from which it follows that [x] =
therefore there is a ring homomorphism, ϕ : Z [i] → Z5 such that ϕ (1) = 1 and 0 ∈ R [x] / hxi . Therefore if p (x) = a0 + a1 x + · · · + an xn , then
ϕ (i) = 2. More explicitly we have,
[p (x)] = [a0 + a1 x + · · · + an xn ] = [a0 ] .
ϕ (a + bi) = a · 1 + b · 2 = (a + 2b) mod 5.
Alternatively put, p (x) + I = a0 + I since a1 x + · · · + an xn ∈ I. Moreover, if
Moreover, (a + ib) ∈ ker (ϕ) iff a + 2b = 5k for some k ∈ Z and therefore, [a0 ] = [b0 ] , then a0 − b0 ∈ I which can happen iff a0 = b0 . Therefore we may
identify R [x] / hxi with S = R thought of as the constant polynomials inside of
ker (ϕ) = {−2b + 5k + ib : b, k ∈ Z} = Z (2 − i) + Z · 5. R [x] . In fact it is easy to check that

Since (2 + i) (2 − i) = 5 and 2 − i ∈ ker (ϕ) , we have, and R 3 a → a + I ∈ R [x] / hxi

h2 − ii ⊂ ker (ϕ) = Z (2 − i) + Z · 5 ⊂ h2 − ii is a ring isomorphism.

Alternatively we may use the first isomorphism theorem as follows. Let
from which it follows that ker (ϕ) = h2 − ii . Thus by the first isomorphism ϕ (p) := p (0) , then ϕ : R [x] → R is a ring homomorphism onto R with ker (ϕ) =
theorem, ϕ̄ : Z [i] / h2 − ii → Z5 defined by hxi . Therefore, ϕ̄ : R [x] / hxi → R is a ring isomorphism.

ϕ̄ ([a + ib]) = ϕ (a + bi) = (a + 2b) mod 5 Theorem 11.4 (Division Algorithm). Let F [x] be a polynomial ring where
F is a field. Given f, g ∈ F [x] both nonzero, there exists a unique q, r ∈ F [x]
is a ring isomorphism. Notice that the inverse isomorphism is given by ϕ̄−1 (a) = with f = qg + r such that either r = 0 or deg r < deg g.
[a] for all a ∈ Z5 which should be compared with Example 10.5 above.
Interpretation. We are dividing f by g and so g goes into f, q times
For what follows recall that the evaluation maps are homomorphisms. with remainder r. This is really high school polynomial division which we
will discuss in more detail a bit later. In the sequel we will sometimes denote
Theorem 11.2 (Evaluation homomorphism). Let R be a subring of a the remainder, r by f mod g.
commutative ring, R̄, and t ∈ R̄. Then there exists a ring homomorphism,
ϕt : R [x] → R̄ such that Corollary 11.5. Suppose that F is a field, p (x) = c0 + · · · + cn xn ∈ F [x] is a
6 0, and let
polynomial with cn =
n n
S := a0 + a1 x + · · · + an−1 xn−1 : ai ∈ F for i = 0, 1, . . . , n − 1 .
X X 
ϕt (p) = ak tk when p (x) = ak xk ∈ R [x] .
k=0 k=0
Then the map, ϕ : S → F [x] / hpi defined by
We will usually simply write p (t) for ϕt (p) .
ϕ (f ) = [f ] := f + hpi for all f ∈ S
The hole point of how we define polynomial multiplication is to make this
theorem true. We will give the formal proof of this theorem a bit later in the is a bijection. Moreover, S becomes a ring and ϕ a ring homomorphism provided
notes. we define
f (x) · g (x) := [f (x) g (x)] mod p
and f + g as usual polynomial addition.
38 11 Lecture 11

Proof. 1. If f ∈ F [x] , then by the division algorithm Moreover one easily shows,

f = qp + r = qp + f mod p [a + bx] + [c + dx] = [(a + c) + (b + d) x] .

and therefore, From these two facts it is now easy to check that
[f ] = [qp + r] = [q] [p] + [r] = [q] 0 + [r] = [r] . C 3 (a + ib) → [a + bx] ∈ R
Thus we have shown is an isomorphism of rings.
[f ] = [f mod p] for all f ∈ F [x] . (11.1) Second Proof. Let √ ϕ : R [x] → C be the evaluation homomorphism,
ϕ (p) = p (i) where i = −1 ∈ C. We then have ϕ (R [x]) = R [i] = C and so
2. Equation (11.1) shows ϕ : S → F [x] / hpi is onto. To see ϕ is injective, by the first isomorphism theorem, R [x] / ker (ϕ) ∼
= C. So to finish the proof we
suppose that f, g ∈ S and ϕ (f ) = ϕ (g) . Then [f − g] = 0, i.e. f − g ∈ hpi , i.e. must show,
ker (ϕ) = x2 + 1 = R [x] x2 + 1 .


f − g = q · p for some q ∈ F [x] . However this is impossible unless q = 0 and (11.2)
f = g since otherwise,
Suppose that p ∈ ker (ϕ) and use the division algorithm to write,
n − 1 ≥ deg (f − g) = deg (q) + deg (p) = deg (q) + n.
p (x) = q (x) x2 + 1 + r (x) where

Thus we have shown ϕ is injective as well, i.e. ϕ : S → F [x] / hpi is a bijection. r (x) = a + bx for some a, b ∈ R.
3. Making use of Eq. (11.1) and the fact that ϕ is a bijection shows,
As p (i) = 0 and i2 + 1 = 0, it follows that r (i) = a + bi = 0. But this happens
ϕ (f ) ϕ (g) = [f ] [g] = [f g] = [(f g) mod p] = ϕ ((f g) mod p) and iff a = 0 = b, and therefore we see that r ≡ 0 an hence that p (x) ∈ x2 + 1 .
ϕ (f ) + ϕ (g) = [f ] + [g] = [f + g] = ϕ (f + g) Thus we have shown ker (ϕ) ⊂ x2 + 1 and since x2 + 1 ∈ ker (ϕ) we must
have ker (ϕ) = x2 + 1 which completes the second proof of the theorem.
for all f, g ∈ S. Thus S equipped with the operations described in the theorem Alternative method for computing ker (ϕ) .
makes S into a ring for which ϕ is a ring isomorphism. If p ∈ ker (ϕ) , then p (i) = 0. Taking the complex conjugates of this equation
Theorem 11.6 (C as a factor ring). C ∼
= R [x] / x2 + 1 =: R. The maps, (using z + w = z̄ + w̄ and zw = z̄ · w̄ for all z, w ∈ C) we learn that p (−i) = 0
as well. As we will see in detail later, p (i) = 0 implies p (x) = (x − i) u (x) for
C3 (a + ib) → [a + bx] ∈ R [x] / x2 + 1 and some u ∈ C [x] . Moreover since,
R [x] / x2 + 1 3 [p (x)] = p (x) + x2 + 1 → p (i) ∈ C 0 = p (−i) = −2i · u (−i)
are ring isomorphisms which are inverses to one another.
we learn that u (−i) = 0 and therefore, u (x) = (x + i) q (x) with q ∈ C [x] .
Proof. We are going to give two proofs that C ∼ = R [x] / x2 + 1 . Our first Therefore,
p (x) = (x − i) (x + i) q (x) = x2 + 1 q (x) .


proof gives rise to the first map while our second gives rise to the second map.
First Proof. Let S = {a + bx : x, b ∈ R} so that It is not too hard to see (use
complex conjugation again) that in fact q ∈ R [x] .
S 3 (a + bx) → [a + bx] ∈ R [x] / x2 + 1 Conversely if p (x) = x2 + 1 q (x) with q ∈ R [x] , then p (i) = 0. Therefore we
have again proved Eq. (11.2).
   
is a bijection. Since x2 + 1 = 0, we have x2 = [−1] and therefore,

[a + bx] [c + dx] = ac + (bc + ad) x + bdx2

 

= [ac + (bc + ad) x + bd (−1)]

= [ac − bd + (bc + ad) x] .

Page: 38 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

12
Lecture 12
 
Example 12.1. Let R := Q [x] / x2 − 2 so that x2 = [2] now. Again we take Moreover [a0 + a1 x] = 0 iff a0 = a1 = 0, so we may take S =
S = {a + bx : a, b ∈ Q} and observe that {a0 + a1 x : a0 , a1 ∈ R} – the polynomials of degree less than or equal to 1.
Thus it follows that
[a + bx] [c + dx] = ac + (bc + ad) x + bdx2
 

= [ac + (bc + ad) x + bd2] R [x] / x2 = {(a0 + a1 x) + I : a0 , a1 ∈ R} ∼ R2 .

= [ac + 2bd + (bc + ad) x] . This induces a ring multiplication on R2 determined as follows;
√ 
Recalling that, in Q 2 , that [a0 + a1 x] [b0 + b1 x] = [(a0 + a1 x) (b0 + b1 x)]
= a0 b0 + (a1 b0 + a0 b1 ) x + a1 b1 x2
 
 √  √  √
a + b 2 c + d 2 = ac + 2bd + (bc + ad) 2
= [a0 b0 + (a1 b0 + a0 b1 ) x] .
it follows that Thus the multiplication rule on S should be defined by
h√ i √
Q 2 3 a + b 2 → [a + bx] ∈ Q [x] / x2 − 2 (a0 + a1 x) (b0 + b1 x) = a0 b0 + (a1 b0 + a0 b1 ) x.

is a ring isomorphism. Alternatively, if we identify S with R := R2 and equip R with the multiplication
√  and addition rules,
Example 12.2 (Example 12.1 revisited). Let ϕ : Q [x] → Q 2 be the eval-
√

uation map, ϕ (p) = p 2 . Then by the first isomorphism theorem, ϕ̄ :
√  (a0 , a1 ) · (b0 , b1 ) = (a0 b0 , a1 b0 + a0 b1 ) and (12.2)
Q [x] / ker (ϕ) → Q 2 is an isomorphism of rings. We now claim that (a0 , a1 ) + (b0 , b1 ) = (a0 + b0 , a1 + b1 ) ,

ker (ϕ) = x2 − 2 . (12.1) then

R 3 (a0 , a1 ) → a0 + a1 x + I ∈ R [x] / x2
2 2
Since x −2 ∈ ker (ϕ) we know that x − 2 ⊂ ker (ϕ) . Conversely, if p ∈ ker (ϕ)

is a ring isomorphism.
and p (x) = q (x) x2 − 2 + r (x) for some r (x) = a + bx with a, b ∈ Q, then
√  √  √  √ An important point to observe for later is that R in Example 12.3 is not a
0=p 2 =q 2 ·0+r 2 = a + b 2. field and in fact not even an integral domain.
  For example, (0, 1)·(0, 1) = (0, 0) =
√ 0. Alternatively, notice that [x] · [x] = x2 = 0, so that 0 6= [x] ∈ R [x] / x2 is
As 2 is irrational, this is only possible if a = b = 0, i.e. r (x) = 0. Thus we a zero divisor.
have shown p ∈ x2 − 2 and therefore Eq. (12.1) is valid.
Example 12.4 (Example 12.3 revisited). We let R be the ring, R2 , with usual
Example 12.3. Let I := x2 = R [x] x2 ⊂ R [x] . If p (x) = a0 + a1 x + · · · + an xn , addition and the multiplication rule in Eq. (12.2). Let ϕ : R [x] → R be the map
then p +I = a0 + a1 x + I since a2 x2 + · · · + an xn ∈ I. Alternatively, we now define by, ϕ (p) = (p (0) , p0 (0)) where p0 (x) is the derivative of p (x) computed
have x2 = 0 in R [x] / x2 , so that as usual for polynomials. Then one easily checks that ϕ is a ring homomorphism.
Moreover if p ∈ ker (ϕ) , then p (0) = 0 and therefore p (x) = xg (x) for some
[p (x)] = [a0 + a1 x + · · · + an xn ] = [a0 + a1 x] . polynomial g (x) . Since
40 12 Lecture 12

0 = p0 (0) = g (0) + 0 · g 0 (0) Since,

   
it follows that g (x) = xq (x) for some polynomial q (x) . Thus p (x) = " n
#" n # 2n k
2 2 0
X ak X bl X X b
aj
k−j  k 
x
 q (x) . Conversely  if p (x) = x q (x) , then p (0) = 0 and p 2(0) = xk xl =  x
2 0
2xq (x) + x q (x) x=0 = 0. Therefore we have shown, ker (ϕ) = x and k! l! j=0
j! (k − j)!
k=0 l=0 k=0
so by the first isomorphism theorem, it follows that R [x] / x2 3 [p (x)] → 
n

k  
 
(p (0) , p0 (0)) ∈ R is a ring isomorphism.
X 1 X k
=  aj bk−j  xk  ,
k! j=0 j
k=0

12.1 Higher Order Zeros (Not done in class) that ϕ becomes a ring homomorphisms provided we use the multiplication rule
in Eqs. (12.4) and (12.5).
Remark 12.5. Example 12.4 generalizes in the following way. Let n ∈ Z+ and Item 2. is easy since we defined the ring multiplication on R so that ϕ would
λ ∈ R and define ϕ : R [x] → R := Rn+1 by be a homomorphism.
  For item 3. let me only explain the case where n = 1 here. If p ∈ ker (ϕ) ,
ϕ (p) := p (λ) , p0 (λ) , . . . , p(n) (λ) ∈ R. (12.3) then
0 = (0, 0) = ϕ (p) = (p (λ) , p0 (λ)) .
We wish to define + and · on R so that his map is a homomorphism. Since
the derivative operation is linear we should use the ordinary vector addition Since p (λ) = 0, we know p (x) = (x − λ) u (x) for some u ∈ R [x] . Differentiating
on Rn+1 in which case ϕ will be an additive group homomorphism. For the this equation at x = λ then implies, 0 = p0 (λ) = u (λ) and therefore u (x) =
2
multiplication rule we have to use the product rule of differentiation in the (x − λ) q (x) for some q ∈ R [x] . Therefore p (x) = (x − λ) q (x) for some q ∈
2
following form, R [x] . Conversely if p (x) = (x − λ) q (x) , then
k  
(k)
X k (j)
(pq) (λ) = p (λ) q (k−j) (λ) . ϕ (p) = ϕ ((x − λ)) ϕ ((x − λ)) ϕ (q) = (0, 1) (0, 1) ϕ (q) = (0, 0) ϕ (q) = 0.
j=0
j
2
Thus if a = (a0 , . . . , an ) and b = (b0 , . . . , bn ) , we should define Thus we have shown, ker (ϕ) = R [x] (x − λ) as claimed when n = 1. (We have
also shown that (0, 1) is a zero divisor in R and hence R is not an integral
a · b := ((a · b)0 , . . . , (a · b)n ) (12.4) domain.)

where
k  
X k 12.2 More Example of Factor Rings
(a · b)k := aj · bk−j . (12.5)
j=0
j
Example 12.7. Here is another example similar to Example 11.1. In R :=
Theorem 12.6. Suppose that R = Rn+1 with the addition and multiplication Z [i] / h3 + ii , we have [i] = [−3] and therefore [−1] = [9] or equivalently
operations described in Remark 12.5. Then; [10] = 0. Therefore for a, b ∈ Z,
1. R is a ring with identity, 1 = (1, 0, . . . , 0) .
[a + ib] = [a − 3b] = [(a − 3b) mod 10] .
2. ϕ : R [x] →
D R defined in
E Eq. (12.3) is a ring homomorphism.
n+1 n+1
3. ker (ϕ) = (x − λ) = R [x] (x − λ) . Thus we should take S = {0, 1, 2, . . . , 9} . If a, b ∈ S and [a] = [b] , then [c] = 0
where c = (b − a) mod 10. Since [c] = 0, we must have
Proof. Item 1. can be proved by a straight forward but tedious verification.
However there is a better way! Consider the bijective map, c = (3 + i) (a + ib) = (3a − b) + (a + 3b) i
" n # n
ϕ X ak X ak k from which it follows that a = −3b and 3 (−3b)−b = −10b = c. Since 0 ≤ c ≤ 9,
k
R 3 (a0 , . . . , an ) −→ x := x + xn+1 ⊂ R [x] / xn+1 .
k! k! this is only possible if c = 0 and so as above if a = b. Therefore
k=0 k=0

Page: 40 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

 12.2 More Example of Factor Rings 41

S 3 a → [a] = a + h3 + ii ∈ Z [i] / h3 + ii Proof. The proof is carried out in a number of steps.

is a bijection. Moreover it is easy to see that thinking of S as Z10 , the above 1. First observe that
map is in fact a ring isomorphism.
hρ (a + ib)i = {ρ (a + ib) (s + it) : s, t ∈ Z}
Example 12.8 (Example 12.7 revisited). From Example 8.9 we have seen that
ϕ : Z[i] → Z10 defined by ϕ (a + bi) = (a − 3b) mod 10 is a ring homomorphism = {ρ [as − bt + i (bs + at)] : s, t ∈ Z} . (12.8)
2
– recall that 32 = (−3) = 9 = −1. In this case, 2. There exists s, t ∈ Z such that bs + at = 1 and so from Eq. (12.8) it follows
a + ib ∈ ker (ϕ) ⇐⇒ a − 3b = 0 in Z10 , that [ρi] = [bt − as] . Therefore every element of Z [i] / hρ (a + ib)i may be
represented in the form [x + iy] where x ∈ Z and y ∈ Zρ . Notice that
i.e. a = 3b + 10k for some k ∈ Z. Therefore,
ρ = min {β ∈ Z+ : α + iβ ∈ hρ (a + ib)i for some α ∈ Z} .
ker (ϕ) = {3b + 10k + ib : b, k ∈ Z} = Z (3 + i) + Z · 10.
3. If s, t ∈ Z such that bs + at = 0, then (s, t) = λ (a, −b) for some λ ∈ Q. In
In particular it follows that 3 + i ∈ ker (ϕ) and therefore
fact λ can not be a fraction. If it were, since both s, t ∈ Z, the denominator
h3 + ii ⊂ ker ϕ = Z (3 + i) + Z · 10. (in reduced form) of λ would have to divide both a and b and hence λ = ±1
as gcd (a, b) = 1. Thus we have λ ∈ Z.
Moreover, since (3 − i) (3 + i) = 10, we see that For such (s, t) = λ (a, −b) with λ ∈ Z we have
Z (3 + i) + Z · 10 ⊂ Z (3 + i) + Z [i] (3 + i) = Z [i] (3 + i) = h3 + ii .
ρ [as − bt + i (bs + at)] = λρ a2 + b2 .

Hence we have shown,

So the smallest positive number this expression can take is ρ a2 + b2 which
h3 + ii ⊂ ker ϕ = Z (3 + i) + Z · 10 ⊂ h3 + ii occurs when λ = 1. 

and therefore 4. From item 3. it follows that ρ a2 + b2 = 0. Combining this observation
with item 2. shows that the map, ϕ, in Eq. (12.6) is onto.
ker ϕ = Z (3 + i) + Z · 10 = h3 + ii = Z [i] (3 + i) . 5. The last main thing to prove is that the map ϕ is one to one. Suppose that
x + iy and x0 + iy 0 are in S with [x + iy] = [x0 + iy 0 ] . This happens iff
Consequently, by the first isomorphism theorem, ϕ̄ : Z[i]/ h3 + ii → Z10 , given
by [x − x0 + i (y − y 0 )] = [0] ⇐⇒ x − x0 + i (y − y 0 ) ∈ hρ (a + ib)i . (12.9)
ϕ̄ ([a + bi]) = ϕ (a + bi) = (a − 3b) mod 10
is a ring isomorphism. Again, by taking b = 0, we see that ϕ̄−1 (a) = [a] = Since |y − y 0 | < ρ, it follows form item 2. that if Eq. (12.9) holds then
y − y 0 = 0. Since |x − x0 | < ρ a2 + b2 , it now follows from item 3. that we


a + h3 + ii is the inverse isomorphism, compare with Example 12.7.
must have x − x0 = 0. Thus we have shown x + iy = x0 + iy 0 and hence ϕ
Theorem 12.9. Let ρ ∈ Z+ and a, b ∈ Z such that a+ib 6= 0 and 1 = gcd (a, b) . is one to one.
Further let 6. The assertion that when ρ = 1 the map in Eq. (12.7) is a ring isomorphism
 is left to the reader.
S := Zρ(a2 +b2 ) + iZρ = x + iy : x ∈ Zρ(a2 +b2 ) and y ∈ Zρ
where Z1 := {0} and in this case we may take S := Zρ(a2 +b2 ) . Then the map,
ϕ Example 12.10. In this example, we wish to consider, Z [x] / h2x − 1i . In this
S 3 (x + iy) −→ [x + iy] ∈ Z [i] / hρ (a + ib)i (12.6)
ring we have
is a bijection of sets. If we further assume that ρ = 1, then [1] = [2x] = [2] [x]

Z(a2 +b2 ) 3 x → [x] ∈ Z [i] / ha + ibi (12.7) which suggests that roughly speaking, “ [x] = 1/2.” Thus we might guess that

is an isomorphism of rings. Z [x] / h2x − 1i ∼

= Z [1/2] . (12.10)

Page: 41 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

42 12 Lecture 12

The general element of Z [1/2] is a rational number which has a denominator of Example 12.12. Let m ≥ 2 and
the form 2n for some n ∈ N. In order to try to prove this, let ϕ : Z [x] → Z [1/2]   
be the evaluation map, ϕ (p) = p (1/2) . Since ϕ (Z [x]) = Z [1/2] to prove Eq. ab
R = M2 (Zm ) = : a, b, c, d ∈ Zm .
(12.10) we need to show cd
ker (ϕ) = h2x − 1i . (12.11)
Then the homomorphisms above is ϕ : Z → M2 (Zm ) by
On one hand it is clear that 2x − 1 ∈ ker (ϕ) and therefore h2x − 1i ⊂ ker (ϕ) .    
For the opposite inclusion, suppose that p ∈ ker (ϕ) , i.e. p (1/2) = 0. By the 10 a0
a 7→ a · =
division algorithm, we may write p (x) = q (x) (x − 1/2) + r where r ∈ Q. Since 01 0a
p (1/2) = 0 it follows that r = 0. Let g (x) := 21 q (x) , then g (x) ∈ Q [x] satisfies,
ker ϕ = hmi, and R has the subring
p (x) = g (x) (2x − 1) .   
a0
: a, b, c, d ∈ Zm
I claim that g (x) ∈ Z [x]. To see this look at the expressions, 0a
 
Xn n−1
X which is isomorphic is Zm .
p (x) = ak xk =  bj xj  (2x − 1)
k=0 j=0
Example 12.13. If R = Zm the homomorphism ϕ : Z → Zm constructed above
is just the natural one a 7→ a mod m that we have been looking at all along and
where ak ∈ Z and bk ∈ Q. By looking at the coefficient of the xk term we learn, chr (Zm ) = m.
ak = −bk + 2bj−1 with the convention that b−1 = 0 = bn . So for k = 0 we learn
Example 12.14. If R = Z[i] = {a + bi : a, b ∈ Z}. Then ϕ : Z → Z[i], a 7→ a · 1 =
that b0 = −a0 ∈ Z, and for general k, that bk = −ak + 2bj−1 . Thus it follows
a + 0i has kernel ker ϕ = h0i. So chr (Z [i]) = 0 and Z[i] has a copy of Z/ h0i ∼
=Z
inductively that bk ∈ Z for all k.
inside it, namely {a + 0i : a ∈ Z}.
Hence we have shown if p ∈ ker (ϕ) , then p ∈ h2x − 1i , i.e. ker (ϕ) ⊂
h2x − 1i which completes the proof of Eq. (12.11).
12.4 Summary
12.3 II. More on the characteristic of a ring Let us summarize what we know about rings so far and compare this to the
group theory of last quarter.
Let R be a ring with 1. Recall: the characteristic of R is the minimum n > 1 (if
n
z }| {
any exist) such that n·1 = 1 + · + 1 = 0. If no such n exists, we call chr (R) = 0.

Theorem 12.11 (Characteristic Theorem). Let R be a ring with 1. Then

ϕ (a) := a · 1R is a homomorphism from Z → R and ker ϕ = hmi where m =
chr (R) . Moreover, R contains a copy of Z/ hmi as a subring.

Proof. Since 12R = 1R , we have already seen that ϕ (a) = a · 1R defines a

homomorphism Moreover it is clear that a · 1R = 0 iff chr (R) |a, i.e. ker (ϕ) =
hmi . The remaining statement follows by an application of the first isomorphism
theorem; i.e. Z/ hmi ∼
= ϕ(Z) = Ran ϕ. So Ran ϕ is a subring of R, and it is
isomorphic to Z/ hmi.
So the rings Z and Z/ hmi ∼= Zm are the “simplest” rings in the sense that
every ring with 1 has a copy of one of these sitting inside of it.

Page: 42 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

 Group Ring
R with (+, ·) 3 (R, +) is
an abelian group (can add
G with · ,associative,
and subtract). Associative,
Definition identity,multiplicative
distributive laws
inverse.
a(b + c) = ab + ac,
(b + c)a = ba + ca.

H ⊂ G is a subgroup
if h1 h−1
2 ∈ H for all S ⊂ R is a subring if
Sub
h1 , h2 ∈ H, a − b ∈ S, ab ∈ S
-structure
i.e. H is closed under the ∀ a, b ∈ S.
group operations

If H C G is a normal
subgroup of G, then If I ⊂ R is an ideal, then
Factor
G/H := {gH : g ∈ G} R/I = {r + I : r ∈ R} is
Structure
is the factor group of the factor ring of R by I.
G by H.

ϕ : R → S is a function
Homo- ϕ : G → H a function
between two rings R and
morphisms: between groups, G, and
S is a homomorphism if
Functions H is a homomorphism if
ϕ(r1 + r2 ) = ϕ(r1 ) + ϕ(r2 )
Preserving ϕ(g1 g2 ) = ϕ(g1 )ϕ(g2 )
and ϕ(r1 r2 ) = ϕ(r1 )ϕ(r2 )
Structure for all g1 , g2 ∈ G.
for all r1 , r2 ∈ R.
13
Lecture 13

Proof. Let us begin by showing that ϕ (J) and ϕ−1 J¯ are ideals whenever


13.1 Ideals and homomorphisms
J and J¯ are ideals. First off it is easy to verify that ϕ (J) and ϕ−1 J¯ are


Example 13.1. Let R = Z. We have already seen that the ideals of Z are exactly sub-rings if J and J¯ are subrings. Moreover, for r ∈ R, we have
h0i = {0}, h1i = Z, h2i , h3i , . . . , i.e., every ideal is a principle ideal of the form
ϕ rϕ−1 J¯ = ϕ (r) ϕ ϕ−1 J¯ = ϕ (r) J¯ ⊂ J¯





hmi for some m ≥ 0. We already know that the ideal hmi is the kernel of the
homomorphism Z → Z/ hmi ∼ = Zm .
and similarly,
Theorem 13.2 (Kernels are ideals). Let R be a ring and I ⊂ R be an ideal.
ϕ ϕ−1 J¯ r = ϕ ϕ−1 J¯ ϕ (r) = Jϕ
¯ (r) ⊂ J¯





Then the that map, π : R → R/I, defined by

π (a) := [a] = a + I for all a ∈ R, wherein we have used ϕ is surjective to conclude that ϕ ϕ−1 J¯ = J.
¯ Simi-

larly, if r̄ ∈ R̄, there exists r ∈ ϕ−1 ({r̄}) and therefore,

is a homomorphism of rings called that natural homomorphism. In particu-
lar as subset, S ⊂ R is an ideal iff S is the kernel of some ring homomorphism. r̄ϕ (J) = ϕ (r) ϕ (J) = ϕ (rJ) ⊂ ϕ (J) and
Proof. We have already seen in Lemma 7.6 that kernels of ring homo- ϕ (J) r̄ = ϕ (J) ϕ (r) = ϕ (Jr) ⊂ ϕ (J)
morphisms are ideals. We leave it to the reader to verify π : R → R/I is a
homomorphism with ker π = I. Once this is done, it follows that every ideal is which shows that ϕ (J) is an ideal as well.
also the kernel of some homomorphism – namely π. Lastly we show that the maps in Eqs. (13.1) and (13.2) are inverses to one
another. Indeed,
Lemma 13.3. Let ϕ : R → R̄ be a surjective homomorphism of rings and
J ⊂ R be an ideal. Then ϕ−1 (ϕ (J)) = J + ker (ϕ) . ϕ ϕ−1 J¯ = J¯ and

Proof. Unwinding all the definitions implies; ϕ−1 (ϕ (J)) = J + I = J.

a ∈ ϕ−1 (ϕ (J)) ⇐⇒ ϕ (a) ∈ ϕ (J) ⇐⇒ ϕ (a) = ϕ (j) for some j ∈ J In the first line we have used that fact that ϕ is surjective while in the second we
used Lemma 13.3 and the assumption that I = ker (ϕ) ⊂ J so that I + J = J.
⇐⇒ ϕ (a − j) = 0 for some j ∈ J
⇐⇒ a − j ∈ ker (ϕ) for some j ∈ J
⇐⇒ a ∈ J + ker (ϕ) . Example 13.5. Let ϕ : Z [x] → Z be the evaluation homomorphism, ϕ (f (x)) =
f (0) so that
ker (ϕ) = hxi = {f (x) ∈ Z [x] : f (0) = 0} .
Proposition 13.4. Let ϕ : R → R̄ be a surjective homomorphism of rings and For any m ∈ Z, hmi ⊂ Z is an ideal and therefore,
I := ker (ϕ) . Then the two sided ideals of R which contain I are in one to one
correspondence with the two sided ideals of R̄. The correspondence is given by, ϕ−1 (hmi) = {f (x) ∈ Z [x] : f (0) ∈ hmi}

{J : I ⊂ J ⊂ R} 3 J → ϕ (J) ⊂ R̄ and (13.1) is an ideal in Z [x] which contains hxi . In fact this is the list of all ideals of Z [x]
which contain hxi = ϕ−1 ({0}) .
{J : I ⊂ J ⊂ R} 3 ϕ−1 J¯ ←− J¯ ⊂ R̄.


(13.2)
Corollary 13.6. If I ⊂ R is an ideal than the ideals of R/I are in one to one
correspondence with the ideals of R containing I. The correspondence is given
by J ⊂ R with I ⊂ J is sent to π (J) = {[a] = a + I : a ∈ J} .
Example 13.7. Let us find all of the ideals of Z/ h10i . Since that ideals of Z
containing h10i are;
Z = {0, ±1, ±2, . . . }, h2i = 0, ±2, ±4, . . . },
h5i = {0, ±5, ±10, . . . }, and h10i = {0, ±10, ±20, . . . },
it follows by Corollary 13.6 that the ideals of Z/ h10i are given by
Z/ h10i = {0, 1, 2, . . . , 9} + h10i ,
h2i / h10i = {0, 2, 4, 6, 8} + h10i ,
h5i / h10i = {0, 5} + h10i , and
h10i / h10i = {0} + h10i .

13.2 Maximal and Prime Ideals

Let R be a ring, and I an ideal. How do we tell if R/I has properties we like?
For example, when is R/I a domain or when is R/I a field? It turns out that
this has nothing to do with R itself but rather only depends on properties of the
ideal, I. We explore these connections in the context where R is a commutative
ring with 1 ∈ R.
Definition 13.8. Let R be a commutative ring with identity 1. An ideal I is
called prime if given any a, b ∈ R with ab ∈ I, either a ∈ I or b ∈ I. An ideal
I is called maximal if given an ideal J with I ⊂ J ⊂ R either I = J or J = R.
Example 13.9. In Z, the ideal hmi is neither prime nor maximal when m is
composite. For example, h6i is not prime. We see that 2 · 3 = 6 ∈ h6i, but
2 6∈ h6i , and 3 6∈ h6i. Moreover h6i h2i Z and h6i h3i Z which shows
that h6i is not prime. On the other hand, if p is prime, then the ideal hpi is
both prime and maximal. Let us also observe that h0i = {0} is prime, but
not maximal. To see that it is not maximal simply observe, for example, that
h0i ( h2i ( Z. The fact that h0i is prime is equivalent to the statement that Z
is an integral domain, see the next lemma.
Example 13.10. Consider the ideal, I := hxi ⊂ Z [x] . This ideal is prime. Indeed,
if p, q ∈ Z [x] , then
p (x) q (x) ∈ I =⇒ p (0) q (0) = 0
=⇒ p (0) = 0 or q (0) = 0 =⇒ p (x) ∈ I or q (x) ∈ I.
On the other hand, because of Example 13.5 we know that hxi is not a maximal
ideal.
14
Lecture 14

ker (ϕ) . By what we have already

proved we know that ϕ−1 J¯ = J is prime

Lemma 14.1. Let R be a commutative ring with identity 1 = 1R . Then R is
an integral domain iff h0i = {0} is a prime ideal and R is a field iff h0i = {0} (maximal) in R iff J¯ = ϕ ϕ−1 J¯ = ϕ (J) is prime (maximal) in R̄.
is a maximal ideal. The following theorem is now an easy corollary of Lemma 14.1 and Theorem
14.2.
Proof. Suppose that a, b ∈ R. Then ab ∈ h0i iff ab = 0. If h0i is a prime
ideal then a ∈ h0i or b ∈ h0i , i.e. either a = 0 or b = 0. This shows that R is an Theorem 14.3. Let R be commutative with 1, I a prime ideal. Then
integral domain. Similarly if R is an integral domain and ab ∈ h0i , then ab = 0
and hence either a = 0 or b = 0, i.e. either a ∈ h0i or b ∈ h0i . Thus h0i is a 1. R/I is an integral domain ⇔ I is a prime.
prime ideal. 2. R/I is a field ⇔ I is maximal.
You are asked to prove on your homework that R is a field iff the only ideals Proof. Easy Proof. As usual we will write [a] for a + I and recall from
of R are {0} and R. Now {0} is maximal iff the only other ideal of R is R itself. your homework that R/I is a commutative ring with identity, [1] = 1 + I.
Let π : R → R/I be the natural homomorphism, π (a) = [a] = a + I. Then
Theorem 14.2. Suppose that R and R̄ are commutative rings with identities I = ker (π) = π −1 ({0}) . Therefore, by Theorem 14.2, I ⊂ R is prime (maximal)
and ϕ : R → R̄ is a surjective homomorphism. Then an ideal, J ⊂ R such that iff {0} ⊂ R/I is prime (maximal). But by Lemma 14.1 we know that {0} ⊂ R/I
ker (ϕ) ⊂ J is prime (maximal) iff ϕ (J) is prime (maximal) in R̄. Similarly, is prime iff R/I is an integral domain and {0} ⊂ R/I is maximal iff R/I is a
an ideal J¯ ⊂ R̄ is prime (maximal) iff ϕ−1 J¯ is prime (maximal) in R.

field.
Second Proof. To help the reader understand this theorem better, let us
Proof. We begin by proving the statements referring to J. In what follows also give a second more direct proof of the theorem.
below J will always be an ideal of R containing ker (ϕ) .
1. Suppose I is prime and [ab] = [a] [b] = [0] in R/I, then ab ∈ I. Since I is
1. Suppose that J ⊂ R is prime and let ā = ϕ (a) and b̄ = ϕ (b) are generic prime it follows that a ∈ I or b ∈ I, i.e. [a] = 0 or [b] = 0. Therefore R/I
elements of R̄. Then āb̄ ∈ ϕ (J) iff ϕ (a) ϕ (b) = ϕ (j) for some j ∈ J which has no zero divisors, i.e. R/I is an integral domain.
happens iff ab − j ∈ ker (ϕ) , i.e. ab ∈ J + ker (ϕ) = J. Since J is prime it Conversely, if I is not prime there exists a, b ∈ R \ I with ab ∈ I. Therefore
follows that either a or b ∈ J and therefore either ā or b̄ in ϕ (J) . [a] 6= 0 6= [b] while [a] [b] = [ab] = 0 which shows that R/I has zero divisors
Now suppose that ϕ (J) is prime. If a, b ∈ R such that ab ∈ J, then and hence is not an integral domain.
ϕ (a) ϕ (b) = ϕ (ab) ∈ ϕ (J) . Since ϕ (J) is prime it follows that either 2. Suppose that I is maximal and let 0 6= [a] ∈ R/I so that a ∈ R but a ∈ / I.
ϕ (a) or ϕ (b) is in ϕ (J) , i.e. either a or b ∈ ϕ−1 (ϕ (J)) = J + ker (ϕ) = J. Let J be the ideal generated by a and I, i.e.
2. If J ⊂ R is not a maximal ideal then there exists an ideal K such
that
J K R. Since ϕ−1 (ϕ (J)) = J, ϕ−1 (ϕ (K)) = K, ϕ−1 R̄ = R, J := Ra + I = {ra + b : r ∈ R, b ∈ I}.
it follows that ϕ (J) ϕ (K) R̄ which shows ϕ (J) is not maximal.
Conversely if ϕ (J) is not maximal, there (You should check that J is an ideal.) Since a ∈ J it follows that I ( J

exists an ideal, K̄ of R̄, such that and since I was maximal we may conclude that J = R. In particular 1 ∈ J
ϕ (J) K̄ R̄. Then K := ϕ−1 K̄ is an ideal of R that J ⊂ K ⊂ R.
Since ϕ (J) K̄ = ϕ (K) R̄, it follows that J K R and hence J is and hence there exists r ∈ R and b ∈ I such that 1 = ra + b. Therefore,
−1
not maximal. 1 = [1] = [ra] = [r] · [a] which shows [a] exists and is equal to [r] .
Therefore U (R/I) = (R/I) \ {0} which shows R/I is a field.
The statements referring to J¯ now follow from

what we have already proved. Conversely if I is not maximal, then there exists another ideal, J, of R such
−1
Indeed, let J¯ ⊂ R̄ be an ideal and J := ϕ−1 J¯ which is an ideal of R containing that I & J & R. Let b ∈ J \ I so that [b] = b + I 6= 0. If [b] exists, then
48 14 Lecture 14

there exist a ∈ R such that [a] [b] = 1 = [1] , i.e. i := ab − 1 ∈ I. Solving for Example 14.10. hxi ⊂ Z [x] is a prime ideal which is not maximal. Indeed from
1 gives, the previous example, hxi h2, xi Z [x] which implies hxi is not maximal. To
1 = i + ab ∈ I + J ⊂ J. see that it is prime observe that Z [x] / hxi ∼ = Z is an integral domain. Moreover
This however contradicts the fact that J is proper since R = R·1 ⊂ RJ = J. since Z is not a field it again follows that hxi is not a maximal ideal.
Thus [b] is not invertible for all b ∈ J \ I. Alternative 1. Observe that ϕ (hxi) = {0} and {0} is prime but not max-
imal in Z.
Alternative 2. If f ∈ Z [x] , we have f (x) = xq (x) + a0 and using [x] =
0 = [2] we find
Corollary 14.4. In a commutative ring R with 1, every maximal ideal is also
[f (x)] = [x] [q (x)] + [a0 ] = [a0 mod 2] .
a prime ideal.
Moreover if a, b ∈ Z2 and [a] = [b] , then a − b ∈ h2, xi which is only possible if
Proof. First Proof. This follows from the Theorem 14.3 since a field is a − b = 0. Thus it follows that we may take S = Z2 . We may now work as we
always an integral domain. have done many times before to see that Z [x] / h2, xi ∼ = Z2 .
Second Proof. Suppose I is a maximal ideal that a ∈ / I and let J :=
I + Ra = J + hai . Then J is an ideal in R which properly contains I and Definition 14.11 (Principle ideal domains). A principle ideal domain
therefore we must have J = R. Hence it follows that 1 = ra + i for some r ∈ R (PID for short) is an integral domain, R, such that every ideal I ⊂ R is a
and i ∈ I. So if ab ∈ I then principle ideal.
b = rab + ib ∈ I Example 14.12. Z and Zm for all m ∈ Z+ are principle ideal domains. We will
also see later that F [x] is a principle ideal domain for every field F.
showing I is a prime ideal.
Example 14.5 (Example 13.9 revisited). Since Z/ hmi ∼ = Zm and Zm is an inte- Example 14.13. In this example we show that Z [x] is not a principle ideal
gral domain (field) iff m is prime, we see that the following are equivalent, domain. For example, consider that ideal generated by h2, xi , i.e.

1. m is prime, h2, xi = Z [x] · 2 + Z [x] · x

2. hmi is a prime ideal of Z, and = {2a + xq (x) : a ∈ Z and q (x) ∈ Z [x]} (14.1)
3. hmi is a maximal ideal of Z.
= {f (x) ∈ Z [x] : f (0) ∈ h2i} . (14.2)
Example 14.6. In Z6 = {0, 1, 2, 3, 4, 5}, I = {0, 3} is a maximal ideal. To see this
let ϕ : Z6 → Z3 be the homomorphism, ϕ (x) = x mod 3. Then ker (ϕ) = I and If h2, xi = hp (x)i for some p ∈ Z [x] , then 2 = q (x) p (x) for some q (x) ∈ Z [x] .
since Z3 is a field it follows that h0i is a maximal ideal and hence ϕ−1 ({0}) = However, this is only possible if both q (x) and p (x) are constant polynomials
ker (ϕ) is a maximal ideal. in which case we must have p (x) = a0 ∈ {±1, ±2} . We can rule out a0 = ±1
since h2, xi is a proper ideal and hence we may assume that p (x) = 2. Noting
Example 14.7. In R [x], hxi is maximal since R [x] / hxi ∼
= R. Notice also that that x ∈ / h2i we learn that h2i $ h2, xi and therefore h2, xi is not a principle
h0i is prime since R [x] is an integral domain but not maximal since R [x] is not ideal in Z [x] .
a field.
Example 14.8. h2 − ii is maximal inside of Z[i]. This is hard to see without our
earlier result that Z[i]/ h2 − ii ' Z5 , which is a field.
Example 14.9. Let I := h2, xi = Z [x] · 2 + Z [x] · x – an ideal of Z [x] . Let ϕ :
Z [x] → Z be the evaluation homomorphism, ϕ (p (x)) = p (0) with ker ϕ = hxi .
Notice that ker ϕ ⊂ I and that ϕ (I) = 2Z = h2i . Since h2i is maximal in Z we
know that I = h2, xi is a maximal ideal of Z [x] . In fact for any prime, p ∈ N,
the same argument shows that hp, xi is a maximal ideal of Z [x] and these are
precisely all of the maximal ideals of Z [x] which contain hxi .

Page: 48 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

15
Lecture 15

• We first went over Quiz #4 in class. Suppose that J ⊂ R1 ⊕ R2 . If (a, b) ∈ J, then (a, 0) = (1, 0) (a, b) and
(0, b) = (0, 1) (a, b) are in J. It is now a simple matter to check that
Lemma 15.1. Suppose that R is an integral domain and a, b ∈ R with a 6= 0 6=
b. Then hai = hbi iff a and b are associates, i.e. a = ub for some u ∈ U (R) . I1 := {a ∈ R1 : (a, 0) ∈ J} and I1 := {b ∈ R2 : (0, b) ∈ J}
Proof. If hai = hbi then a ∈ hbi and b ∈ hai and therefore there exists
are ideals of R1 and R2 respectively. If a ∈ I1 and b ∈ I2 , then (a, b) = (a, 0) +
u1 , u2 ∈ R such that a = u1 b and b = u2 a. Thus we may conclude that b =
(0, b) ∈ J showing I1 ⊕ I2 ⊂ J. Similarly if (a, b) ∈ J, then as noted above a ∈ I1
u2 u1 b and hence by cancellation that u2 u1 = 1. This shows that a = ub with
and b ∈ I2 which implies J ⊂ I1 ⊕ I2 .
u = u1 ∈ U (R) . Conversely if a = ub with u ∈ U (R) then a ∈ hbi and b ∈ hai
since b = u−1 a. Therefore, hai ⊂ hbi and hbi ⊂ hai , i.e. hai = hbi . Corollary 15.4. Let R1 , R2 , be as in Lemma 15.3. Then the maximal ideals
Proposition 15.2 (maximal ⇐⇒ prime in PIDs). If R is a principle ideal of R1 ⊕ R2 are of the form J = I1 ⊕ R2 or J = R1 ⊕ I2 where I1 is a maximal
domain and I ⊂ R be a non-zero ideal. Then I is maximal iff I is prime. ideal of R1 and I2 is a maximal ideal of R2 .

Proof. By Corollary 14.4, we know in general that maximal ideals are prime Example 15.5 (Book problem 14.30). Find the maximal ideals, I, in R := Z8 ⊕
ideals. So we need only show that if I is a prime ideal then I is a maximal ideal. Z30 and for each maximal ideal find size of the field, R/I. The only maximal
Suppose that I = hai ⊂ R (with a 6= 0) is a prime ideal and J = hbi is another ideal of Z8 is h2i and the maximal ideals of Z30 are h2i , h3i , and h5i . Thus the
ideal such that I ⊂ J. We will finish the proof by showing either J = I or maximal ideals of R are
J = R.
As a ∈ hbi , we have a = bc for some c ∈ R. Since bc ∈ I = hai and I h2i ⊕ Z30 , Z8 ⊕ h2i , Z8 ⊕ h3i , and Z8 ⊕ h5i .
is prime, we must have either b ∈ hai or c ∈ hai . Case 1: if b ∈ hai , then
The respective fields have size, by Lagrange’s theorem or other means, 2 =
J = hbi ⊂ hai = I and hence J = I. Case 2: if c ∈ hai , then c = au for some 8 30 30
8/ gcd(2,8) , 2 = 30/ gcd(2,30) , 3 = 30/ gcd(3,30) , and 5 = 30/ gcd (5, 30) .
u ∈ R and we then have a = bc = bau. Cancelling a (here is where we use a 6= 0)
from this equation shows that 1 = bu and therefore b and 1 are associates and
so, by Lemma1 15.1, J = h1i = R.

15.1 The rest this section was not covered in class

Lemma 15.3. Suppose that R1 and R2 are two commutative rings with iden-
tities.. Then every ideal, J ⊂ R1 ⊕ R2 , is of the form J = I1 ⊕ I2 where I1 and
I2 are ideals of R1 and R2 respectively.
Proof. It is easy to check that J := I1 ⊕ I2 ⊂ R1 ⊕ R2 is an ideal whenever
I1 and I2 are ideals of R1 and R2 respectively. So let us concentrate on the
converse assertion.
1
More directly, 1 = bu implies b−1 exists and therefore 1 = b−1 b ∈ hbi = J and
hence that J = R.
16
Lecture 16

16.1 The Degree of a Polynomial In this example, deg f = 2, deg g = 1, deg h = 1, and deg (gh) = 1 (6= 2) . So
it is not always true in a polynomial ring that deg(f g) = deg g + deg h. Let us
Recall the following definition of being a polynomial. compute the “values” of g (x) ;

Definition 16.1. Let R be any commutative ring with identity. The polynomial x 012345
ring R [x] is defined to be .
g (x) 0 3 0 3 0 3

R [x] = {an xn + an−1 xn−1 + · · · + a1 x + a0 : ai ∈ R and n ≥ 0} Thus we see that deg (g) = 1 yet g (x) has three roots over Z6 .
equipped with the usual rules for addition and multiplication of polynomials. Theorem 16.5. Let R be an integral domain. Then for any f, g ∈ R [x],
Recall that the multiplicative identity in R [x] is 1, and the additive identity deg(f g) = deg(f ) + deg(g),
is 0.
and R [x] is an integral domain.
Notation 16.2 Suppose that
Proof. If f = 0 then f g = 0 and we will have deg (f g) = −∞ and deg (f ) +
f = an xn + · · · + a1 x + a0 ∈ R [x] deg (g) = −∞ + deg (g) = −∞. So we may now assume that f and g are
non-zero polynomials which we write as,
is a polynomial with an 6= 0. The we say an is the leading coefficient and
f has degree n, written deg (f ) = n. (It is convenient to use the convention f = an xn + · · · + a1 x + a0 , and
that deg (0) = −∞.) If an = 1, then f is called monic. When deg f = 0, i.e.
g = bm xn + · · · + b1 x + b0
f = a0 , we say f is a constant polynomial.

Example 16.3. In C [x], deg(4x2 + (i + 3)x + 5) = 2. The coefficient of the with ai , bi ∈ R, an 6= 0, and bm 6= 0. Then deg f = n and deg g = m. So
largest power of x is called the leading coefficient. The leading coefficient of
f g = an bm xn+m + · · · + (a1 b0 + a0 b1 )x + a0 b0 ,
4x2 + (i + 3)x + 5 is 4. The polynomial 4x2 + (i + 3)x + 5 is not monic while
g = x2 + 6 is monic. f = 5 is constant, but g = x + 5 is not. and an bm 6= 0 so that f g is not the zero polynomial and deg (f g) = m + n =
deg f + deg g.
Example 16.4. Let R = Z6 = {0, 1, . . . , 5},
Lemma 16.6. If R is an integral domain, then U (R [x]) = U (R) . In particular
f = x2 + 5, g = 2x + 1, and h = 3x.
if R = F is a field and F × := F \ {0} , then U (F [x]) = F × .
Then Proof. If p (x) ∈ U (R [x]) then there exists q (x) ∈ R [x] such that
3 2 3 2 p (x) q (x) = 1. Therefore, 0 = deg (p) + deg (q) , showing deg (p) = 0 = deg (q) .
f g = (2x + x + 10x + 5) = 2x + x + 4x + 5,
Thus p (x) = p0 ∈ R and p0 is invertible in R [x] iff it is invertible in R because
f + g = x2 + 2x + 6 = x2 + 2x, and we have seen above that deg (q) = 0 where q (x) is the inverse to p (x) = p0 .
gh = 6x2 + 3x = 3x.
52 16 Lecture 16

16.2 The evaluation homomorphism (review) f = x2 + 5 and g = 2x3 − 5x + 2.

Theorem 16.7 (Evaluation homomorphism). Let R be a subring of a Then

commutative ring, R̄, and t ∈ R̄. Then there exists a ring homomorphism, f + g = 2x3 + x2 − 5x + 7,
ϕt : R [x] → R̄ such that
f g = 2x5 + 5x3 + 2x2 − 25x + 10,
n n
ϕt (p) =
X
ak tk when p (x) =
X
ak xk ∈ R [x] . f (1) = 6, g(1) = −1,
k=0 k=0 and so
We will usually simply write p (t) for ϕt (p) . (f + g)(1) = 5 = f (1) + g(1) and
Pn
Proof. Let q (x) = l=0 bl xl , then (f g) (1) = −6 = f (1) · g(1).
Example 16.9 (Evaluation example). Suppose that R = Z6 = {0, 1, 2, 3, 4, 5},
n
! n
X X
ϕt (p + q) = ϕt (al + bl ) xl = (al + bl ) tl a = 3, and ϕ : R [x] → R is the evaluation map, f →
7 f (3). For example, if
l=0 l=0 f = 3x2 + 5x + 2, and g = x + 3, then
n
X n
X n
X
al tl + bl t l
al tl + bl tl ϕ(f ) = f (3) = 3(3)2 + 5(3) + 2 = 44 = 2 and


= =
l=0 l=0 l=0 ϕ(g) = g(3) = 3 + 3 = 6 = 0,
= ϕt (p (x)) + ϕt (q (x)) .
from which it follows that f ∈
/ ker (ϕ) while g ∈ ker (ϕ) .
Similarly,
! ! Example 16.10. Suppose that λ ∈ R and ϕ = evalλ : R [x] → R, i.e. ϕ (p) =
X X p (λ) . Then p ∈ ker ϕ iff p (λ) = 0 which happens (as we will see shortly) iff
ϕt (pq) = ϕt al bk xm p (x) = (x − λ) q (x) for some q ∈ R [x] . Therefore,
m l+k=m
X X
!
X X
! ker (ϕ) = hx − λi = R [x] (x − λ)
m m
= al bk t = al bk t
m m
for this homomorphism.
l+k=m l+k=m
!
X X X
= al tl bk tk = al tl bk tk
m l+k=m l,k
16.3 The Division Algorithm
! !
X X Definition 16.11. Let R be an integral domain and f, g ∈ R [x] . We say that
= al tl bk tk = ϕt (p) · ϕt (q) .
g divides f if f = kg for some k ∈ R [x] . We also say that g is a factor of f.
l k


The point is that the multiplication and addition rules for polynomials was Example 16.12. In Z [x] , (2x − 4) does not divide x2 − 4 . Indeed, if it did
chosen precisely so as to make this theorem true. then
x2 − 4 = (a + bx) (2x − 4) = −4a + (2a − 4b) x + 2bx2
Example 16.8. Suppose that ϕ := ϕ1 : R [x] → R is the evaluation homomor- which would imply 2b = 1 which is impossible in Z. On the other hand, working
phism, ϕ (p) = p (1) . Then in Q [x] , we have
ϕ(f g) = f g(1) = f (1)g(1) and 
1

x2 − 4 = (x − 2) (x + 2) = (2x − 4) x+1
ϕ(f + g) = [f + g](1) = f (1) + g(1). 2
For example suppose that which shows that (2x − 4) is a factor of x2 − 4 in Q [x] .

Page: 52 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

 16.3 The Division Algorithm 53

Theorem 16.13 (Division Algorithm). Let F [x] be a polynomial ring where Example 16.16 (Example 16.15). Here is alternate way to do the last example.
F is a field. Given f, g ∈ F [x] both nonzero, there exists a unique q, r ∈ F [x] First use the division algorithm over Q to find;
with f = qg + r such that either r = 0 or deg r < deg g. (We will give the proof
1
of this theorem later.) 2x + 14 ,


2x + 1 x2 + x + 2
Interpretation. We are dividing f by g and so g goes into f, q times
− x2 − 12 x
with remainder r. This is really high school polynomial division which we
1
will discuss in more detail a bit later. In the sequel we will sometimes denote 2x +2
1
the remainder, r by f mod g. − 2x − 14
7
Example 16.14. Let f := 3x3 + 5 and g = 2x + 3 in Q [x] , then 4

3 2
− 94 x + 27 that is over Q we have
2x 8

3
2x + 3 3x +5
 
1 1
− 3x3 − 92 x2 x2 + x + 2 = x+ (2x + 1) + 7/4.
2 4
− 92 x2
9 2 27 Multiplying this equation through by 4 gives,
2x + 4 x
27
4 x +5 4 x2 + x + 2 = (2x + 1) (2x + 1) + 7


27
− 4 x − 81
8
− 41 and then apply the “mod 3 homomorphism” to the coefficients implies,
8

which shows, x2 + x + 2 = (2x + 1) (2x + 1) + 1

   
3 2 9 27 41 which is the result above again.
3x3 + 5 = x − x+ (2x + 3) + −
2 4 8 8
Example 16.17. Let f (x) = 2x3 and g (x) = ix2 + 5x + 2 in C [x] , then
so that  
3 2 9 27 41
q (x) =
2
x − x+
4 8
and r (x) = −
8  −2ix + 10
ix + 5x + 2 2x3 + 0x2
2
+ 0x + 0
in this example.
2x3 − 10ix2 − 4ix
Example 16.15. Consider f (x) = x2 + x + 2 and g (x) = 2x + 1 inside of Z3 [x] . + 10ix2 + 4ix + 0
Then, using 2 · 2 = 4 mod 3 = 1, we find + 10ix2 + 50x + 20
(−50 + 4i) x − 20
 2x − 2
2x + 1 x2 + x + 2 so that
2x3 = (−2ix + 10) ix2 + 5x + 2 + (−50 + 4i) x − 20,


x2 + 2x ,
−x + 2 that is
−x − 2 q (x) = (−2ix + 10) and r (x) = (−50 + 4i) x − 20.
4=1
Corollary 16.18. Let F be a field, a ∈ F, and f (x) ∈ F [x] . Then f (a) is the
which implies
remainder in the division of f (x) by (x − a) .
x2 + x + 2 = (2x − 2) (2x + 1) + 1
Proof. By the division algorithm, there exists k (x) , r (x) ∈ F [x] such that
= (2x + 1) (2x + 1) + 1. deg (r) = 0 < 1 and

Page: 53 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

54 16 Lecture 16

f (x) = k (x) (x − a) + r (x) . (16.1) Example 16.22. Z [x] is not an principle ideal domain. For example consider the
Since deg (r) = 0, r (x) = b for some b ∈ F and hence evaluating Eq. (16.1) at ideal,
x = a implies, I := h2, xi = Z [x] · 2 + Z [x] · x.
f (a) = k (a) (a − a) + b = b. This ideal is proper since if 1 ∈ I, then 1 = 2p (x) + xq (x) which would imply
that 2p0 = 1 for some p0 ∈ Z. But this is impossible. If there exists q ∈ Z [x]
such that I = hqi , then 2 = q (x) p (x) for some p. However this would imply
Example 16.19. Let f (x) = x2 + 5 in R [x] . If we divide (x + 1) = (x − (−1)) 0 = deg (2) = deg (p) + deg (q) from which it follows that deg (q) = 0. Therefore
into f (x) the remainder will be f (−1) = 6. q (x) = q0 for some q0 ∈ Z. As I is proper we know that q0 6= ±1 and since
2 ∈ hq0 i we must have q0 = ±2. However, it should be clear that x ∈ I while
Example 16.20. If we divide x − 1 into x2 + 2 we find,
x∈ / h2i = h−2i . Thus I is not a principle ideal.
x+1


x−1 x2 +2
− x2 + x 16.4 Appendix: Proof of the division algorithm
x+2
Let us now give the formal proof of Theorem 16.13.
−x+1
Proof. Proof of Theorem 16.13. Suppose that f, g ∈ F [x] with g 6= 0.
3 We break the proof into the existence and uniqueness assertions.
2 Uniqueness. Suppose that we have two decompositions,
which gives x2 + 2 = (x + 1) (x − 1) + 3. Notice that the remainder, 3 = (1) + 2
as it should be. f = qg + r = q 0 g + r0
Theorem 16.21 (F [x] is a PID). Let F be a field, then F [x] is a principle where deg r < deg g and deg r0 < deg g. Then
ideal domain. Moreover the map,
(q − q 0 )g + (r − r0 ) = 0,
{monic polynomials} 3 p → hpi ∈ {non-zero ideals of F [x]} (16.2)

is a one to one correspondence. The inverse map is given by associating to a or equivalently,

non-zero ideal, I ⊂ F [x] , the unique monic polynomial, p ∈ I, with lowest (q − q 0 )g = (r0 − r).
degree. If r − r0 6= 0, we may take degrees of this equation to conclude,
Proof. Let us first show the map in Eq. (16.2) is one to one. So suppose deg(q − q 0 ) + deg g = deg (r − r0 ) < deg g
that p and q are monic polynomials such that hpi = hqi . Then by Lemmas 15.1
and 16.6, we know that p (x) = kq (x) for some k ∈ U (F ) = F \ {0} . Since which is a contradiction. Therefore r = r0 which then forces q = q 0 since F [x]
both p and q are monic, we must in fact have k = 1, i.e. p (x) = q (x) . is an integral domain and g 6= 0.This proves uniqueness.
Suppose that I ⊂ F [x] is an ideal. If I = {0} then I = h0i so that {0} is a Existence. Write
principle ideal (as always). So now suppose that I 6= {0} and let p (x) ∈ I be a
non-zero polynomial in I with minimal degree. By dividing p (x) by its leading f (x) = an xn + · · · + a1 x + a0 and
order coefficient, we may further assume that p (x) is monic. If f (x) ∈ I, use g (x) = bm xm + · · · + b1 x + b0 .
the division algorithm to write, f (x) = k (x) p (x) + r (x) where deg (r) <
deg (p) . Since r = f − kq ∈ I, we must have r = 0 showing that I = hpi as If m = deg g > deg f = n then we q = 0, and r = f so that f = 0 · g + f with
claimed. If q ∈ I is another monic polynomial such that deg (q) = deg (p) , then deg f < deg g.
q (x) = k (x) p (x) for some k ∈ F [x] . A simple degree argument then shows If m = deg g ≤ deg f = n, we will use induction on the degree of f . In
that deg (k) = 0 so that k (x) = k0 is a constant polynomial. Since q and p are the base case, deg f = 0, we have deg g = 0 so that f = a0 and g = b0 and
both monic, it follows that k0 = 1, i.e. q (x) = p (x) .

Page: 54 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

therefore, f = ab00 g + 0, so r = 0 in this case. Now suppose that n = deg f ≥ 1
and the existence has been established for all lower n. Let
an n−m
f 0 (x) = f (x) − x g (x)
bm
an n−m
= an xn + · · · + a1 x + a0 − x (bm xm + · · · + b1 x + b0 )
bm
= cn−1 xn−1 + · · · + c0 .

Thus deg f 0 < n and hence by the induction hypothesis, f 0 (x) = q 0 (x) g (x) +
r (x) where r = 0 or deg (r) < deg (g) . Therefore,
an n−m an n−m
f (x) = x g (x) + f 0 (x) = x g (x) + q 0 (x) g (x) + r (x)
bm bm
 
an n−m
= x + q 0 (x) g (x) + r (x)
bm
= q (x) g (x) + r (x)
an n−m
where q (x) = bm x + q 0 (x) and r = 0 or deg (r) < deg g.
17
Lecture 17

We reviewed Corollary 11.5 in preparation for the next quiz.  x2 + x + 1

x − 1 x + 0x2
3
+ 0x + 6
x3 − x2
17.1 Roots of polynomials + x2 +
+ x2 − x
Definition 17.1. Let R be an integral domain and f (x) ∈ R [x] . Then a ∈ R x + 6
is a zero or root of f (x) if f (a) = 0. x − 1
Example 17.2. Let f (x) = 2x2 + 5x − 7 in Z [x] . Then f (1) = 0 so 1 is a root 7=0
while f (2) = 11 6= 0 so 2 is not a root of f. from which it follows that
Corollary 17.3. Let F be a field, a ∈ F, and f (x) ∈ F [x] . Then
f (x) = (x − 1) x2 + x + 1 .


(x − a) |f (x) ⇐⇒ f (a) = 0, i.e. iff a is a root of f (x) .
Remark 17.4. Corollary 17.3 holds more generally in that we may replace F Let g (x) = x2 + x + 1 and notice that 0 = f (2) = g (2) so that (x − 2) must
by any commutative ring with identity. Indeed, if f (x) ∈ R [x] and f (a) = 0 divide g (x) . Indeed this is the case,
for some a ∈ R. Let g (x) := f (x + a) , so that g (x) ∈ R [x] with g (0) = 0.
x + 3
Since g (0) = 0, g (x) has no constant term which means that we may factor x 
2
out of g (x) , i.e. g (x) = xk (x) for some k (x) ∈ R [x] . This translates into the x−2 x + x + 1
statement about f (x) ; x2 − 2x
+ 3x + 1
f (x) = g (x − a) = (x − a) k (x − a) , + 3x − 6
7=0
which shows x − a is a factor of f (x) .
Example 17.5. Let f (x) = x2 + 5 has no roots over R but two roots over C. from which it follows that
Indeed,  √  √  g (x) = (x − 2) (x + 3) = (x − 2) (x − 4) .
f (x) = x − i 5 x + i 5
√
Thus as expected, we have
so that f ±i 5 = 0.
f (x) = (x − 1) (x − 2) (x − 4) .
Example 17.6 (Book Problem 17.13). Consider the polynomial, f (x) = x3 + 6
on Z7 [x] and observe that n
Corollary 17.7. Let F be a field, f (x) Qn∈ F [x] , and suppose that {ai }i=1 ⊂ F
x 0123456 is a list of n – distinct zeros of f. Then i=1 (x − ai ) divides f (x) . Alternatively
. put, there exists k (x) ∈ F [x] such that
f (x) 6 0 0 5 0 5 5

Thus we know that (x − 1) , (x − 2) , and (x − 4) are all factors of f (x) . For f (x) = k (x) (x − a1 ) (x − a2 ) . . . (x − an ) . (17.1)
example,
58 17 Lecture 17

Proof. The proof goes by induction on n. When n = 1 this is the content of 17.2 Roots with multiplicities
Corollary 17.3. Suppose the result holds for all k ≤ n for some n ≥ 1. Then if
n+1
{ai }i=1 is a list of distinct zeros of f, by corollary 17.3, there exists g (x) ∈ F [x] Definition 17.11. Let F be a field and f (x) ∈ F [x] . A root, a ∈ F, of f (x)
k k+1
such that is said to have multiplicity k ≥ 1 if (x − a) divides f (x) but (x − a) does
f (x) = g (x) (x − an+1 ) . (17.2) not.
Since
Example 17.12. In R [x] , 3 is a root of order 2 for f (x) = x2 − 6x + 9. Indeed,
0 = f (ai ) = g (ai ) (ai − an + 1) 2
n
f (x) = (x − 3) . If f (x) = x2 − 7x + 10, then f (x) = (x − 2) (x − 5) and the
and ai − an+1 6= 0 for all i ≤ n, it follows that {ai }i=1 are distinct zeros of g. only zeros of f are x = 2 and x = 5 each of which have multiplicity 1.
Therefore by the induction hypothesis, there exists, k (x) ∈ F [x] such that
Example 17.13. Since f (x) = x3 − 2x2 + x ∈ Q [x] factors as,
g (x) = k (x) (x − a1 ) (x − a2 ) . . . (x − an ) . (17.3)
2
f (x) = x x2 − 2x + 1 = x (x − 1) ,


Thus it follows from Eqs. (17.2) and (17.3) that
f (x) = k (x) (x − a1 ) (x − a2 ) . . . (x − an ) (x − an+1 ) the roots of f (x) are 0 and 1 with multiplicities 1 and 2 respectively.

which completes the induction step and the proof. Example 17.14. In R [x] the polynomial, f (x) = x4 +2x2 +1 has no roots. While
in C [x] it has two distinct roots each with multiplicity 2. To find these roots
Corollary 17.8. Let F be a field and f (x) ∈ F [x] with N := deg (f ) . Then f
observe that
has at most N distinct roots in F.

2 2 2 2
Proof. If a1 , . . . , an be distinct zeros of f (x) . Then we may write f (x) as f (x) = x2 + 1 = [(x − i) (x + i)] = (x − i) (x + i) .
in Eq. (17.1) from which it follows that
Thus the roots are ±i.
N = deg (f ) = n + deg (k) ≥ n.
Example 17.15 (17.23). Find all of the zeros and multiplicity of

f (x) = x5 + 4x4 + 4x3 − x2 − 4x + 1 ∈ Z5 [x] .

Example 17.9. Consider
f (x) = (x − 4) (x − 5) = (x + 2) (x + 1) = x2 + 3x + 2 We start by finding all of the roots;

in Z6 [x] . Clearly f (4) = f (5) = 0 but this is not all. Indeed, f (0) = 2 6= 0, x 01234
.
f (1) = 3 · 2 = 6 = 0, f (2) = 4 · 3 = 12 = 0, f (3) = (−1) (−2) = 2 6= 0. Thus f (x) 1 0 2 0 3
see that f (x) has four zeros in Z6 , namely {1, 2, 4, 5} .
Then we divide (x − 1) into f (x) to find,
Example 17.10 (Zeros of xn − 1). Suppose that z = reiθ is a zero of xn − 1 in
C. Then we must have x4 + 4x2 + 3x − 1
rn einθ = 1

x − 1 x5 + 4x4 + 4x3 − x2 − 4x + 1
which implies r = 1 and inθ = k2π for some k ∈ Z. Thus the zeros, Z, of xn − 1 x5 − x4
are, n o n o 4x3 − x2 − 4x + 1
Z = eik2π/n : k ∈ Z = eik2π/n : k ∈ Zn . 4x3 − 4x2
3x2 − 4x + 1
If we let ω := ei2π/n , then we may write
3x2 − 3x
Z = ω k : k ∈ Zn − x +1

− x +1
and ω is called a primitive nth – root of unity. 0

Page: 58 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

 to find
f (x) = (x − 1) g (x)
4 2
with g (x) = x + 4x + 3x − 1. Let us check we got this right,

(x − 1) x4 + 4x2 + 3x − 1 = x5 − x4 + 4x3 − x2 − 4x + 1

= x5 + 4x4 + 4x3 − x2 − 4x + 1.X

Notice that g (1) = 7 mod 5 = 2 6= 0 so that 1 is a root with multiplicity 1. We

now divide (x − 3) into g (x) to find;

 x3 + 3x2 + 3x + 2
x − 3 x + 0x3 + 4x2
4
+ 3x − 1
x4 − 3x3
3x3 + 4x2 + 3x − 1
3x3 − 4x2
+ 3x2 + 3x − 1
+ 3x2 − 4x
+ 2x − 1
+ 2x − 1
0

so that
g (x) = (x − 3) x3 + 3x2 + 3x + 2 .

Let h (x) := x3 + 3x2 + 3x + 2, then h (3) = 2 · 33 + 32 + 2 = 65 mod 5 = 0 so

that (x − 3) goes into h (x) . Here is the computation,

 x2 + x + 1
x − 3 x3 + 3x2 + 3x + 2
x3 − 3x2
x2 + 3x + 2 .
x2 − 3x
+ x + 2
+ x − 3
+5=0

Thus we have shown,

2
x5 + 4x4 + 4x3 − x2 − 4x + 1 = (x − 1) (x − 3) x2 + x + 1 .

so that 3 has multiplicity 2. This is rather painful way to carry this out. We will
later develop a derivative test to make finding multiplicities easier to determine.
18
Lecture 18

Corollary 17.7 has the following useful refinement. 18.1 Irreducibles and Maximal Ideals
Theorem 18.1. Suppose that a1 , . . . , an are distinct zeros of f (x) ∈ F [x] with Definition 18.3. Let R be an integral domain and a ∈ R× \U (R) . We say that
multiplicities, l1 , . . . , ln . Then there exists k (x) ∈ F [x] such that a is reducible if it admits a non-trivial factorization, i.e. a = bc for come
l l b, c ∈ R× \ U (R) . Otherwise we say that a is irreducible. So a is irreducible
f (x) = k (x) (x − a1 ) 1 . . . (x − an ) n . (18.1)
iff a 6= 0, a ∈
/ U (R) , and whenever a = bc then either b or c is in U (R) .
Pn
Proof. The proof will be by induction on N := i=1 li . If N = 1, then we
Let F be a field and recall from Lemma 16.6 that U (F [x]) = U (F ) = F × .
have n = 1, a1 = a ∈ F and l1 = 1. In this case it follows by definition of a root
Therefore the associates to f (x) ∈ F [x] are {af (x) : a ∈ F × } . So f (x) =
with multiplicity 1 that f (x) = k (x) (x − a) for some k (x) P∈n F [x] . a · h (x) with a ∈ F × and h (x) ∈ F [x] is a trivial factorization.
Now suppose N ≥ 2 and the theorem holds whenever i=1 li < N. By the
induction hypothesis there exists k1 (x) ∈ F [x] such that Example 18.4. If F is a field then f (x) ∈ F [x] is reducible iff there is a factor-
l −1 l l ization of the form f (x) = g (x) h (x) where deg g ≥ 1 and deg h ≥ 1.
f (x) = k1 (x) (x − a1 ) 1 (x − a2 ) 2 . . . (x − an ) n . (18.2)
l Lemma 18.5. If F is a field and f (x) = g ∈ F [x] is reducible, then
Moreover since (x − a1 ) 1 |f (x) it follows that (x − a1 ) divides deg (f (x)) ≥ 2. In particular if deg (f (x)) = 1 then f (x) is irreducible.
l l
f1 (x) = k1 (x) (x − a2 ) 2 . . . (x − an ) n Proof. If f (x) admits a non-trivial factorization, f (x) = g (x) h (x) , then
which implies f1 (a1 ) = 0. Since f1 (a1 ) = 0 while deg f (x) = deg h (x) + deg g (x) ≥ 1 + 1 ≥ 2.
l2 ln
(x − a2 ) . . . (x − an ) |x=a1 6= 0,

we may conclude that k1 (a1 ) = 0. Therefore that (x − a1 ) |k1 (x) , i.e. k1 (x) = Example 18.6. In Z [x] ;
k (x) (x − a1 ) for some k (x) ∈ F [x] . Using this expression for k1 (x) back in
Eq. (18.2) completes the proof. 1. x2 − 1 is reducible since x2 − 1 = (x − 1) (x + 1) .
2. 2x + 4 = 2 (x + 2) is reducible in Z [x] since both 2 and x + 2 are not units
Corollary 18.2. Let F be a field and f (x) ∈ F [x] with N := deg (f ) . Then f
in Z [x] . Similarly 5x ∈ Z [x] is reducible since f (x) = 5 · x where both 5
has at most N roots in F when counted with multiplicities. In particular, there
and x are not units.
can be at most N distinct roots of f (x) in F.
Notice that 2x + 4 = 2 (x + 2) is irreducible in Q [x] by Lemma 18.5. The
Proof. If a1 , . . . , an be zeros of f (x) with multiplicities, l1 , . . . , ln . Then
difference now is that 2 is invertible in Q [x] while it is not in Z [x] .
from Theorem 18.1 there exists k (x) ∈ F [x] such that Eq. (18.1) holds. In
particular it follows that Proposition 18.7. Suppose that D is an integral domain and 0 6= a ∈ D \
n
X n
X U (D) . Then a is irreducible iff hai is maximal among all principle ideals in
N := deg (f ) = deg (k) + li ≥ li . D. To say hai is maximal among all principle ideals in D we mean if b ∈ D
i=1 i=1 satisfies hai ⊂ hbi ⊂ R, then either hai = hbi or hbi = R.
This inequality is precisely what the Corollary states.
 Proof. You will prove this in the homework.
The following theorem is a direct consequence of Proposition 18.7 and The-
orem 14.3.

Theorem 18.8 (p irreducible ⇐⇒ hpi maximal in a PID). Suppose that

D is a PID and p ∈ D× \ U (D) . Then the following are equivalent;
1. p is irreducible,
2. hpi is a maximal ideal, and
3. D/ hpi is a field.

Example 18.9. In R [x] , x2 + 1 is irreducible. Indeed if x2 + 1 where to factor

non-trivially the factors would have to be linear and x2 + 1 would have to
have a root which it does not. Consequently we know that R [x] / x2 + 1 is a
field. We already know this since we have seen, R [x] / x2 + 1 is isomorphic
to C. Nevertheless, we have now shown that R [x] / x2 + 1 is a field without
knowledge about C and hence we may view this as a fresh construction of C.
19
Lecture 19

19.1 Irreducibles Polynomials I x 01234

f (x) 1 3 1 1 4
Let F be a field – later it will be Q. Our goal is to determine when a polynomial,
p (x) ∈ F [x] , is irreducible. We start with the following general results. showing f (x) has not roots and hence is irreducible. Therefore hf (x)i ⊂ Z5 [x]
is a maximal ideal and therefore,
Example 19.1 (Second proof of Lemma 18.5). Let F be a field and λ ∈ F. Then
ϕ : F [x] → F be the evaluation homomorphism, ϕ (p) = p (λ) . Then ϕ (F [x]) = F := Z5 [x] / hf (x)i
F, I = ker (ϕ) = hx − λi = F [x] · (x − λ) , and therefore, F [x] / hx − λi ∼
=F
and therefore x − λ is irreducible in F [x] for all λ ∈ F by Theorem 18.8. is a field. As we have,

a + bx + cx2 : a, b, c ∈ Z5 ,
 
Lemma 19.2. If F is a field, f (x) ∈ F [x] with deg (f ) ≥ 2, and f has a zero F =
in F , then f is reducible over F.
we see that the number of elements in F is 53 = 125. Thus we have constructed
Proof. Let a ∈ F be a root of f, then f (x) = (x − a) k (x) for some k (x) ∈ a field with 125 elements and characteristic 5.
F (x) with deg (k) ≥ 1. Since both (x − a) and k (x) are not units, it follows Pn
that f is reducible. Proposition 19.5. Suppose that f (x) = k=0 ak xk ∈ Z [x] . If and f (r/s) = 0
We have the following partial converse to this lemma. for some r ∈ Z and s ∈ Z+ with gcd (r, s) = 1, .then s|an and r|a0 . In particular
if f (x) is monic, i.e. an = 1, the only possible rational roots of f (x) must
Theorem 19.3. If F is a field, f (x) ∈ F [x] with deg (f ) = 2 or deg (f ) = 3, actually be in Z and in fact must be a divisor of a0 .
then f is irreducible over F iff f has no zeros in F.
Proof. See problem 17.25 in Gallian which you are assigned for homework.
Proof. Lemma 19.2 shows that f has zero implies f is reducible. Conversely
if f is reducible, then f (x) = p (x) q (x) for some polynomials p and q with √
deg (p) , deg (q) ≥ 1. Since deg (p) + deg (q) = deg (f ) ≤ 3, it follows that Example 19.6. Let k ≥ 2 and p be a prime in Z+ , then k p is irrational. Indeed
√
deg (p) = 1 or deg (q) = 1, say deg (p) = 1. Thus p (x) = ax + b for some if k p = r were rational, being a root of xk − p, r would have to be an integer
a, b ∈ F with a 6= 0 and f (x) = (ax + b) q (x) . Therefore which divides p by Proposition 19.5. Thus r would have to be ±p or ±1 none of
which will work. Hence we may conclude that x2 − p and x3 − p are irreducible
polynomials over Q (or Z) for any prime number p. (In fact by Eisenstein’s
       
−b −b −b −b
f = a +b q =0·q Criterion, Theorem 22.6 below, it is easy to see that xk − p is irreducible over
a a a a
Q for all primes, p, and k ≥ 2.) Of course
which shows that f has a zero in F.  √  √ 
x2 − 2 = x − 2 x + 2 in R [x]
Example 19.4 (Construction of finite fields). Observe that the x3 function on
Z5 is given by
and therefore x2 − 2 is reducible over R.
x 01234
.
x3 0 1 3 2 4 Example 19.7. In this example we consider the polynomial, x2 + 1 ∈ F [x] for a
Let f (x) := x3 + x + 1 ∈ Z5 [x] , then number of fields, F.
64 19 Lecture 19
2
1. If F = R, then f (x) = x2 + 1 ≥ 1 > 0 for x ∈ R and so x2 + 1 is irreducible x + x
in R [x] . Consequently, R [x] / x2 + 1 is a field. (We already have seen this x2 + x + 1 x4 + 0x3 + 0x2 + x + 1
isomorphic to C.) x4 + x3 + x2
2. If F = C, then x2 + 1 = (x − i) (x + i) is reducible over C [x] . x3 + x2 + x + 1
3. If F = Z3 , then f (0) = 1, f (1) = 2 and f (2) = 5 mod 3 = 2 and therefore x3 + x2 + x
f is irreducible over Z3 . Consequently Z3 [x] / x2 + 1 is a field which is 1
easily seen to be isomorphic to Z3 [i] .
4. If F = Z5 , then f (2) = 22 + 1 = 5 mod 5 = 0 and f (3) = 10 mod 5 = 0 and which shows that
therefore f (x) is reducible over Z5 . Notice that
x4 + x + 1 = x2 + x x2 + x + 1 + 1



f (x) = (x − 2) (x − 3) = (x + 3) (x + 2) .


and therefore x2 + x + 1 is not a factor of f (x) .
In this case Z5 [i] ∼
= Z5 [x] / x2 + 1 is not a field and in fact not even an
integral domain since [x + 2] [x + 3] = 0. Theorem 19.12 (Fundamental theorem of algebra). The complex number
field, C, is algebraically closed, i.e. every non-constant polynomial, p (x) ∈
Example 19.8. The polynomial f (x) = x3 + x + 1 is irreducible over Z2 [x] since
C [x] has a root.
f (0) = 1 = f (1) .
Example 19.9. The polynomial, f (x) = x4 + 2x2 + 1 has no roots over R yet it Proof. This is a standard result proved in a course on complex variables.
is reducible over R. Indeed we have the non-trivial factorizations, We will not give the proof in this class.
m
f (x) = x2 + 1 x2 + 1 .


Corollary 19.13. Let p (x) ∈ C [x] and {λi }i=1 be the distinct zeros of p (x)
m
and {ki }i=1 be the corresponding multiplicities. Then
This shows the hypothesis that deg (f ) ∈ {2, 3} is necessary in Theorem 19.3.
m
Y ki
3
Example 19.10. Let f (x) = 3x + 2x + 1 in Z5 [x] . Then p (x) = c (x − λi ) for some c ∈ C× . (19.1)
i=1
x 01234
f (x) 1 1 4 3 1 Proof. From Theorem 18.1 we know that exists k (x) ∈ C [x] such that
m
and it follows that f is irreducible over Z5 . Y k
p (x) = k (x) (x − λi ) i .
Example 19.11. Suppose that f (x) = x4 + x + 1 ∈ Z2 [x] . In this case f (0) = i=1

1 = f (1) so that f has no roots and hence not linear factors. Thus if f (x) is to However, k (x) must not have any roots for otherwise we would not have ac-
factor it must be of the form f (x) = p (x) q (x) where deg p (x) = 2 = deg q (x) counted for all the zeros with multiplicities of p (x) . According to Theorem
and both p (x) and q (x) have no roots. By dividing p (x) by its leading order 19.12, k (x) must be a constant polynomial c for some c ∈ C.
coefficient we may assume that p is monic. This forces q to be monic as well
since f was monic. Thus we have, p (x) = x2 + ax + b form some a, b ∈ Z2 . Since
0 6= p (0) = b we must have p (x) = x2 + ax + 1. Similarly, 0 6= p (1) = a so that
a = 1. Thus the only monic degree 2 polynomial which is itself irreducible is
x2 + x + 1. Thus if there is going to be a factorization of f (x) it must be given

2
by f (x) = x2 + x + 1 . However a simple computation shows,

2
x2 + x + 1 = 2x + 3x2 + 2x3 + x4 + 1 = x4 + x2 + 1 6= x4 + x + 1.

Alternatively we could divide x2 + x + 1 into f (x) as follows,

Page: 64 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

20
Lecture 20
n n
Corollary 20.1. The only irreducible polynomials in C [x] are those with de- X
X
ϕ ak xk = ϕ (ak ) ϕ xk


ϕ (p (x)) =
gree equal to one. The only irreducible polynomials in R [x] are the degree one
k=0 k=0
polynomials and the degree two polynomials with no roots. n n
X k
X
Proof. The first assertion should be clear from Corollary 19.13. For the = ϕ (ak ) ϕ (x) = ϕ (ak ) tk .
k=0 k=0
second assertion, suppose that p (x)
∈ R [x] is any polynomial. If λ ∈ C is a
root of p (x) , then 0 = p (λ) = p λ̄ because p (x) has all real coefficients. Let
us now suppose that p (x) is irreducible over R.
If λ ∈ R is a root of p (x) , then p (x) = (x − λ) k (x) for some k (x) ∈ R [x] . Proposition 20.4 (Changing coefficients). Let R and T be commutative
Since p (x) is irreducible we must have deg (k (x)) = 0, i.e. .k (x) is a constant rings and ψ : R → T be a ring homomorphism. Further define ψ̄ : R [x] → T [x]
and p (x) is linear. On the other hand if λ ∈ C \ R is a root of p (x) then so is by the formula,
λ̄. Therefore we know n
! n n
X X X
k

2
 ψ̄ ak x := ψ (ak ) xk for all ak xk ∈ R [x] . (20.1)
p (x) = (x − λ) x − λ̄ k (x) = x2 − 2 Re λ + |λ| k (x)


k=0 k=0 k=0

for some k (x) ∈ R [x] . Since p (x) is irreducible, we must again have deg k (x) = Then ψ̄ is a ring homomorphism. Moreover this is the unique ring homo-
2. This observation along with Theorem 19.3 completes the proof of the corol- morphism from R [x] → T [x] such that ψ̄ (x) = x and ψ̄ (a) = ψ (a) for all
lary. a ∈ R ⊂ R [x] .

Proposition 20.2. Let p (x) ∈ R [x] , then p (x) may be factored into irre- Proof. You are asked to prove this for homework.
ducibles of R [x] .
Example 20.5. Suppose that ψk : Z → Zk is the homomorphism, ψk (a) :=
Proof. If p (x) is not irreducible, there exists h (x) , k (x) ∈ R [x] with a mod k so that ψ̄k : Z [x] → Zk [x] . For example,
deg h ≥ 1, deg k ≥ 1, and p (x) = h (x) k (x) . Since deg h (x) < deg p (x) and d
ψ̄3 5x3 + 7x + 3 = 2x3 + x ∈ Z3 [x] ,


deg k (x) < deg p (x) , the result follows by a simple induction argument.
ψ̄5 5x3 + 7x + 3 = 2x + 3 ∈ Z5 [x] , and

ψ̄7 5x3 + 7x + 3 = 5x3 + 3 ∈ Z7 [x] .

20.1 Two more homomorphisms involving polynomials
Corollary 20.6. Let R and T be commutative rings and ϕ : R → T be a
Lemma 20.3. Let R and T be commutative rings, ϕ : R [x] → T be a ring ring homomorphism. Then for each t ∈ T there exists a ring homomorphism
homomorphism, and let t := ϕ (x) . Then ϕt : R [x] → T such that ϕt (a) = ϕ (a) for all a ∈ R and ϕt (x) = t. More
n
! n n specifically, if
X X X
k
ϕ ak x = ϕ (ak ) tk for all ak xk ∈ R [x] . n
! n n
X X X
k
k=0 k=0 k=0 ϕt ak x = ϕ (ak ) tk for all ak xk ∈ R [x] .
k=0 k=0 k=0
Proof. Using the ring homomorphism properties we find,
66 20 Lecture 20

Proof. Let ψ : R [x] → T [x] be the homomorphism in Proposition 20.4 and Proof. First Proof. One way to compute the gcd of a bunch of numbers
evt : T [x] → T be the evaluation homomorphism at t. Then ϕt = evt ◦ ψ : is by looking at their prime number decompositions. In terms of this decompo-
R [x] → T is the desired homomorphism. sition,
N
Y
Example 20.7. As an application of Corollary 20.6, let R be a commutative gcd (a0 , . . . , an ) = pki i
ring, ϕ : R → R [x] be the homomorphism taking a ∈ R to the constant i=1
polynomial a ∈ R [x] , and t (x) ∈ R [x] be any polynomial. Then there is a N
where {pi }i=1 are the distinct primes which appear in these decomposition and
unique homomorphism, ϕt : R [x] → R [x] such that ϕt (a) = a for all a ∈ R
and ϕt (x) = t (x) which is given formulaically as, ki ∈ {0, 1, 2, . . . } is chosen so that pki i |aj for all j but pki i +1 does not divide one
of the aj . Using this description it is easy to see the truth of the lemma.
ϕ (p (x)) = p (t (x)) for all p (x) ∈ R [x] . Second Proof. Let d := gcd (a0 , . . . , an ) and c := gcd (ma0 , . . . , man ) then
we know there exists si ∈ Z such that
In particular if t (x) = x − a for some a ∈ R the map, p (x) → p (x − a) is a ring
isomorphism from R [x] to R [x] with inverse being given by p (x) → p (x + a) . d = s0 a0 + · · · + sn an .

From this it follows that

20.2 Gauss’ Lemma md = s0 ma0 + · · · + sn man .

We now go back to finding more ways to determine when f (x) ∈ Q [x] is irre- Thus every common divisor of mai for all i is a divisor of md and in particular
ducible. Are first goal is to change the question into one involving a polynomial c| (md) . Since md is a divisor of mai for all i we also know that md|c. Thus we
g (x) ∈ Z [x] being irreducible or not – see Corollary 21.5 below. The main tool may conclude that md = c as claimed.
here will be Gauss’ Lemma 20.10 below.
Pn Corollary 20.12. If f, g ∈ Z [x] then c (f g) = c (f ) · c (g) . In particular,
Definition 20.8. Let f (x) = k=0 an xn ∈ Z [x] , then c (f ) := gcd (a0 , . . . , an ) f (x) g (x) is primitive iff both f (x) and g (x) are primitive.
is called the content of f. A polynomial f ∈ Z [x] is said to be primitive if
1 1
c (f ) = 1. Proof. Let f0 (x) := c(f ) f (x) and g0 (x) := c(g) g (x) so that both f0 (x) and
g0 (x) are primitive in Z [x] . Then by Gauss’ lemma, f0 (x) g0 (x) is primitive
Example 20.9. If f (x) = 14x2 +10x+6, then c (f ) = 2 and 21 f (x) = 7x2 +5x+3 and since f (x) g (x) = c (f ) c (g) f0 (x) g0 (x) , it follows from Lemma 20.11 that
is a primitive polynomial. c (f g) = c (f ) c (g) .
Lemma 20.10 (Gauss’ Lemma). If f (x) , g (x) ∈ Z [x] are primitive polyno- Corollary 20.13. Suppose that f (x) , g (x) ∈ Z [x] , g (x) is primitive, and
mials, then f (x) g (x) is primitive as well, i.e. c (f ) = 1 = c (g) =⇒ c (f g) = 1. k (x) ∈ Q [x] . If f (x) = k (x) g (x) in Q [x] , then k (x) ∈ Z [x] . In words if
g (x) divides f (x) in Q [x] then the quotient is in fact back in Z [x] .
Proof. Suppose P that c (f g) > 1 and let p be a prime divisor of c (f g) .
n
Further let f¯ (x) := k=0 (an mod p) xn and similarly for ḡ (x) . Observe that Proof. Let t ∈ Z+ be chosen so that h (x) := tk (x) ∈ Z [x] – for example
¯
f → f is a ring homomorphism from Z [x] → Zp [x] so that f g = f¯ḡ. Since let t be the least common multiple of all of the denominators in the coefficients
p divides all of the coefficients of f g we know f g = 0 and since Zp [x] is an of k (x) . We then have tf (x) = h (x) g (x) therefore that tc (f ) = c (tf ) =
integral domain, we may conclude that either f¯ = 0 or ḡ = 0, i.e. p|c (f ) or c (h) c (g) = c (h) . It now follows that
p|c (g) . Thus at least one of the polynomials, f or g, is not primitive.
1 1 1 1 1
Lemma 20.11. Suppose that a0 , . . . , an ∈ Z (not all zero) and m ∈ Z+ . Then f (x) = tf (x) = tf (x) = h (x) g (x) = h (x) g (x) .
c (f ) c (tf ) c (h) c (h) c (h)
gcd (ma0 , . . . , man ) = m gcd (a0 , . . . , an ) . In particular if f (x) ∈ Z [x] and
m ∈ Z× , then c (mf ) = |m| c (f ) . Therefore
1
f (x) = c (f ) h (x) g (x) = k (x) g (x)
c (h)

Page: 66 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

and hence  
1
k (x) = c (f ) h (x) ∈ Z [x] .
c (h)

Example 20.14. If g (x) is not primitive, the results in Corollary 20.13 may fail.
For example take f (x) = x2 , g (x) = 2x and k (x) = 21 x.
21
Lecture 21

Lemma 21.1. Suppose that f (x) , g (x) ∈ Z [x] are primitive polynomials and stf (x) = g1 (x) h1 (x) =⇒ stc (f ) = c (g1 ) · c (h1 )
there exists α, β ∈ Q× such that αf (x) = βg (x) as polynomials in Q [x] . Then and therefore
|α| = |β| and f (x) = ±g (x) .
1 1 1 1
f (x) = stf (x) = g1 (x) · h1 (x) .
Proof. Apply Corollary 20.13 with k (x) = β/α to learn that β/α = t ∈ Z. c (f ) stc (f ) c (g1 ) c (h1 )
Since f (x) = tg (x) it follows that 1 = c (f ) = |t| c (g) = |t| so that β/α = t =
±1. Therefore we have found a non-trivial factorization of f (x) over Z in the desired
Alternative Proof. Let α := ab and β := uv with a, b, u, v ∈ Z\ {0} . Then form, namely;
vaf (x) = ubg (x) . From Lemma 20.11 we know
   
1 1
f (x) = c (f ) g1 (x) · h1 (x)
c (g1 ) c (h1 )
|va| = c (vaf (x)) = c (ubg (x)) = |ub|    
s t
= c (f ) g (x) · h (x) =: u (x) v (x) .
from which it follows that f (x) = ±g (x) . c (g1 ) c (h1 )

Example 21.2. The polynomial, f (x) = x2 − 5x + 6 ∈ Z [x] has a factorization

over Q as   Example 21.4. Notice that x ∈ Z [x] is irreducible. Indeed if x = f (x) g (x) for
1 3 some f (x) , g (x) ∈ Z [x] , then 1 = deg (f (x)) + deg (g (x)) and we may suppose
f (x) = x− (2x − 4) .
2 2 that deg (f ) = 1 and deg (g) = 0. Moreover we know that f (x) and g (x) are
Of course if we move the factor of 2 from the second factor to the first we also primitive and therefore g (x) = ±1 ∈ U (Z [x]) . Thus the factorization is trivial
have the factorizations of f (x) over Z as, and x ∈ Z [x] is irreducible.
On the other hand, f (x) = 5x ∈ Z [x] is reducible over Z but irreducible
f (x) = (x − 3) (x − 2) . over Q. The point being that 5 ∈ U (Q) while 5 ∈ / U (Z) .

The next theorem shows this sort of rejiggering can always be done. However we do have the following corollary to Theorem 21.3.
Corollary 21.5. If f (x) ∈ Z [x] is primitive, then f (x) is irreducible over
Theorem 21.3. Let f (x) ∈ Z [x] . If f (x) is reducible over Q then f (x) is
Z [x] iff f (x) is irreducible over Q [x] .
reducible over Z. In fact f (x) is reducible over Q [x] iff f (x) factors in Z [x] in
the form, f (x) = u (x) v (x), where u (x) , v (x) ∈ Z [x] with deg u (x) ≥ 1 and Proof. 1. First observe that if g (x) ∈ / U (Z [x]) and g (x) is primitive then
deg v (x) ≥ 1. deg g (x) ≥ 1. Indeed if deg g (x) = 0 and g (x) is primitive, then g (x) = ±1,
Consequently: if f (x) is irreducible over Z then f (x) is also irreducible i.e. g (x) ∈ U (Z [x]) .
over Q. 2. Suppose that f (x) is primitive and reducible over Z, i.e. f (x) = g (x) h (x)
for some g (x) , h (x) ∈ / U (Z [x]) . Since 1 = c (f ) = c (g) c (h) it follows that
Proof. We need to show that any non-trivial factorization over Q gives rise c (g) = 1 = c (h) and therefore by the first part we must have deg g (x) ≥ 1 and
to a non-trivial factorization over Z. So suppose f (x) ∈ Z [x] and h (x) , g (x) ∈ / U (Q [x]) ∼
deg h (x) ≥ 1. Therefore g (x) , h (x) ∈ = Q× and so the factorization
Q [x] with deg (h) ≥ 1, deg (g) ≥ 1, and f (x) = g (x) h (x) in Q [x] . Choose f (x) = g (x) h (x) is nontrivial over Q as well, i.e. f (x) is reducible in Q [x] .
s, t ∈ Z+ such that g1 (x) := sg (x) and h1 (x) := th (x) and in Z [x] . Then we Since the converse was already proven in Theorem 21.3, the proof is complete.
have
21.1 mod p Irreducibility Tests

Theorem 21.6 (mod p test). Let f (x) ∈ Z [x] , p ∈ Z+ be a prime, and

f¯ (x) ∈ Zp [x] be the reduction of f (x) mod p, i.e. reduce all of the coefficients
of f (x) mod p. If deg f¯ = deg f,1 and f¯ (x) is irreducible over Zp , then f (x) is
irreducible over Q.

Proof. If f (x) is reducible over Q, then by Theorems 21.3, there f (x) =

g (x) h (x) with deg (g) , deg (h) ≥ 1, then f¯ (x) = ḡ (x) h̄ (x) . Since deg ḡ ≤
deg g and deg h̄ = deg h and deg f = deg f¯ by assumption,

deg f = deg f¯ = deg ḡ + deg h̄ ≤ deg g + deg h = deg f.

From this equation it follows that we must have deg ḡ = deg g ≥ 1 and deg h̄ =
deg h ≥ 1. Therefore f¯ (x) is reducible over Zp as well.

Remark 21.7 (mod p test failures). We will see in examples below if f¯ (x) ∈
Zp [x] is reducible or deg f¯ < deg f, then the only thing we may conclude is
that the test fails and we need to try another p.

Example 21.8. Consider f (x) = x4 + x + 1 ∈ Z [x] . We have seen in Example

19.11 that
f¯ (x) = f (x) mod 2 = x4 + x + 1 ∈ Z2 [x]
is irreducible and therefore f (x) = x4 + x + 1 is irreducible in Z [x] and hence
also on Q [x] .
Notice that if we had decided test this using the mod 3 test, we would have
seen that f¯ (1) = 1 + 1 + 1 = 3 mod 3 = 0 so that f¯ (x) is reducible in Z3 [x] .
The only thing we can conclude from this is that the test has failed.

Example 21.9. Now suppose that f (x) = x3 + x2 + x − 1 ∈ Z [x] . In this

case the mod 2 test fails but the mod 3 test wins the day. So suppose that
f¯ (x) = f (x) mod 3 and observe that

f¯ (0) mod 3 = 2, f¯ (1) mod 3 = 2, and f¯ (2) mod 3 = 1.

Since f¯ is third order and has not roots it is irreducible and therefore f (x) is
irreducible over Z and hence also Q.

1
Alternatively put, we are assuming that the leading the order coefficient of f (x) is
not divisible by p.
22
Lecture 22

Let f (x) ∈ Z [x] and for any prime p ≥ 2, let fp (x) := f (x) mod p. Then Example 22.3 (Degree hypothesis is necessary). Let
combining Theorems 21.3 and 21.6 we have the following implication;
  f (x) = 2x2 + 3x + 1 = (2x + 1) (x + 1) ∈ Q [x] . (22.1)
∃ p such that  
 deg fp (x) = deg f (x) &  =⇒ f (x) is irreducible . Clearly this polynomial is reducible. On the on the other hand if we reduce f
over Q
fp (x) ∈ Zp [x] is irreducible mod 2 we get the polynomial, f¯ (x) = x+1 which is linear and hence irreducible.
Example 22.1. Let us consider the polynomial, Notice that under the mod 2 reduction, the factorization in Eq. (22.1) becomes,
8 11 x + 1 = 1 · (x + 1)
f (x) := 5x4 − x3 + 3x + ∈ Q [x] .
3 3
Then f (x) is irreducible iff which is a trivial factorization of x + 1.

g (x) := 3f (x) = 15x4 − 8x3 + 9x + 11 Example 22.4. Let us determine if

is irreducible over Q. As g (x) ∈ Z [x] and c (g) = 1 we need only show that 3 3 5 1
g (x) is irreducible over Z [x] . To test this out, let ḡ (x) be g (x) mod 2, so that f (x) := x + x + ∈ Q [x]
7 7 7
ḡ (x) = x4 + x + 1 ∈ Z2 [x] . is irreducible or not. We need only consider 7f (x) = 3x3 + 5x + 1 which is
We have already seen in Example 19.11 that ḡ (x) is irreducible and therefore primitive in Z [x].
by the mod 2 test, g (x) is irreducible and therefore so is 3f (x) ∈ Z [x] and
hence f (x) ∈ Q [x] .
Example 22.2 (Converse of mod p test is false). Let f (x) := x2 +5x+25 ∈ Z [x]
and f¯ (x) = x2 be f (x) mod 5. Notice that f¯ is reducible over Z5 yet f (x) is
irreducible over Q. Indeed, it is easy to see that f (x) has not roots in Z and
hence not roots in Q. Alternatively, apply the mod 2 test in which case we
consider x2 + x + 1 which is irreducible over Z2 as it has no roots. Alternatively,
by the quadratic formula the roots of f (x) are given by
√ √
−5 ± 25 − 100 −5 ± i 75
= Plot of 7f (x) .
2 2
so there are not even any real roots. This can also be seen by completing the Let us consider
squares to see,
 2  2 g (x) := 7f (x) mod 2 = x3 + x + 1.
5 5
f (x) = x + + 25 − Since g (0) = g (1) = 1, it follows that g is irreducible and since deg (g) =
2 2
 2 deg (7f ) , we learn that 7f and hence f is irreducible over Q.
5 3 75 Alternatively we may use Proposition 19.5 to argue that 7f (x) and hence
= x+ + 25 · ≥ .
2 4 4 f (x) has no roots in Q. According to Proposition 19.5 we need only look for
72 22 Lecture 22

roots of the form a/3 where a ∈ {±1} that is we need only consider {0, ±1/3} . implies p| (g1 h1 ) which contradicts the assumption that p does not divide the
By direct calculation we learn leading order coefficient of f (x) .
With this as a warm-up we go to the general case. Notice that a0 = g0 h0
25
7f (−1/3) = −7/9 6= 0 and 7f (1/3) = 6= 0. and an = gr hs . Since p|a0 while p2 - a0 it follows that p divides exactly one of
9 g0 or h0 but not both. Suppose that p|g0 and p - h0 . Also notice that p - an
This again shows that f (x) is irreducible over Q [x] . implies p - gr so there is a first 0 < t ≤ r < n such that p - gt . Since,

Example 22.5. Let us show again the f (x) = x3 + 5 is irreducible over Q. at = g0 ht + g1 ht−1 + · · · + gt−1 h1 + gt h0 ,

1. Method 1. By Proposition 19.5 the only possible roots of f (x) are for x ∈ and p|at , it follows that p|gt h0 which then implies p|h0 which is a contradiction.
{±5} and it is clear that neither of these are roots of f (x) .
2. Method 2. Let us try the mod p tests. For a second proof of this theorem, see Theorem 24.8 below.
a) For p = 2 and p = 3 the tests fails since 1 is a root in each of these
cases. Example 22.7. The polynomial, x5 +9x4 +12x2 +6 is irreducible over Q [x] since
b) For p = 5, f¯ (x) = x3 is clearly reducible so the test fails again. p = 3 divides all of the coefficients except the leading order one and 32 - 6.
c) For p = 7 we find For n ∈ N, let
f (0) mod 7 = 5, f (1) mod 7 = 6, f (2) mod 7 = 6, f (3) mod 7 = 4, Φn (x) = 1 + x + x2 + · · · + xn−1 .
f (4) mod 7 = 6, f (5) mod 7 = 4, f (6) mod 7 = 4 This geometric series may be summed in the usual way,

and the test has succeeded in showing that x3 + 5 is irreducible over Q. xn − 1

xΦn (x) − Φn (x) = xn − 1 =⇒ Φn (x) = .
3. Method 3. Use Eisenstein’s criteria in Theorem 22.6 below. x−1
Corollary 22.8. For any prime, p ∈ N, Φp (x) is irreducible over Q.

22.1 Eisenstein’s Criterion Proof. Observe that

p   p  
Pn k p k p k
k=0 ak x ∈ Z [x]
p
X X
Theorem 22.6 (Eisenstein’s Criterion). Let f (x) := (x + 1) − 1 = x −1= x ,
with n ≥ 2. If there is a prime p such that p - an , p2 - a0 while p|aj for k k
k=0 k=1
j = 0, 1, 2, . . . , n − 1 then f (x) is irreducible. over Q. (x + 1) − 1 = x, and hence
p Pp p k

p  
Proof. If f (x) were reducible over Q [x] then, by Theorem 21.3, there exists (x + 1) − 1 k=1 k x
X p k−1
g (x) , h (x) ∈ Z [x] with deg (g) , deg (h) ≥ 1 such that f (x) = g (x) h (x) . Let Φp (x + 1) = = = x
(x + 1) − 1 x k
k=1
r s        
p p p p
= xp−1 + xp−2 + xp−3 + . . .
X X
g (x) = gk xk and h (x) = hk xk x+ .
p−1 p−2 2 1
k=0 k=0

where r + s = n. For 1 ≤ k < p, since there is no factor of p in k! or (p − k)!, it follows that

In order to absorb the general proof better let us first suppose that r = s = 1 
p

p (p − 1)! (p − 1)!
so that = =p·
p−k k! (p − k)! k! (p − k)!
f (x) = (g0 + g1 x) (h0 + h1 x) = g0 h0 + (g0 h1 + g1 h0 ) x + g1 h1 x2 . p


is an integer which is divisible by p. Thus p| p−k for 1 ≤ k ≤ p − 1 while
p


Since p| (g0 h0 ) and p2 - (g0 h0 ) , we may assume that p|g0 but p - h0 . Combining 2
p - 1 = p and hence the Eisenstein’s Criterion, Φp (x + 1) is irreducible. This
this with the assumption that p| (g0 h1 + g1 h0 ) implies that p|g1 . But this then then is easily seen to imply Φp (x) is irreducible.

Page: 72 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

22.2 Summary of irreducibility tests

Theorem 22.9 (Reducibility tests). Let F be a field and f (x) ∈ F [x] . If

deg f (x) = 1, then f (x) is irreducible so suppose that deg f (x) ≥ 2. Then;

   
f (x) has a root f (x) is reducible
=⇒ .
in F over F
Moreover if deg f (x) = 2 or deg f (x) = 3, then
   
f (x) has a root f (x) is reducible
⇐⇒ .
in F over F
Pn
Theorem 22.10 (Finding roots). Suppose that f (x) = k=0 ak xk ∈ Z [x]
with an 6= 0 6= a0 . Then the rational roots of f (x) are contained in
nr o
: r ∈ N, s ∈ Z+ 3 s|an and r|a0 .
s
Pn
Theorem 22.11 (Q – Irreducibility tests). Suppose f (x) = k=0 ak xk ∈
Z [x] with n := deg f (x) ≥ 2. Then;
 
deg f (x) = 1 or
 deg f (x) = 2 or 3 
 
 & f (x) has no 
roots in Q
Eisenstein’s
mod p test ⇓
   Criterion 
∃ a prime p 3   ∃ a prime p
 deg fp (x) = deg f (x)  f (x) is irreducible  such that 
 & fp (x) is irreducible  =⇒
  ⇐=  .
over Q  p|ak for k < n 
over Zp p - an & p2 - a0
 ⇑ 
f (x) is irreducible
over Z

Moreover if f (x) is primitive (i.e. c (f ) = 1) then

   
f (x) is irreducible f (x) is irreducible
⇐⇒ .
over Q over Z
23
Lecture 23

23.1 Irreducibles and Primes II equation to see that 1 = tc. This shows that c ∈ U (R) and so a only admits
trivial factorizations, i.e. a is irreducible.
Definition 23.1. Let D be an integral domain (i.e. a commutative ring with 1 The next example shows that the converse of Theorem 23.4 is false in general,
which has no zero divisors, i.e. cancellation holds in D) and a, b, c ∈ D. Then; i.e. it is not always true that irreducibles are primes.
1. a is associated to b if a = b (mod) U (R) , i.e. a ∈ bU (R) or equivalently Example 23.5 (Example 1, p. 321). In this example we will show that
b ∈ aU (R) . √ √ 
2. a ∈ D is irreducible iff a 6= 0, a ∈
/ U (R) and whenever a = bc then either α := 1 + −3 ∈ R := Z −3
b ∈ U (R) or c ∈ U (R) . (So a ∈ D is irreducible if it can not be factored in is irreducible but not prime which we will now verify. First observe√that R is
a non-trivial way.) Similarly we say that 0 6= a ∈ D is reducible if it admits a sub-ring of C√which will simplify our computations. Given x = a + b −3 ∈ R,
a factorization of the form a = bc with b, c ∈
/ U (R) . 2
let x̄ := a − b −3 and N (x) := |x| = xx̄ = a2 + 3b2 ∈ N. Since |xy| = |x| |y|
3. a ∈ D is prime if 0 6= a ∈ / U (R) and if a satisfies; whenever a|bc then for all x, y ∈ C A direct calculation shows1 that N (xy) = N (x) N (y) for all
either a|b or a|c. (So a ∈ D is prime if the conclusion of Euclid’s lemma x, y ∈ R.
holds.)
• U (R) = {±1} . If xy = 1, then 1 = N (1) = N (x) N (y) implies N (x) =
Example 23.2. Let R = Z so that U (Z) = {±1} and the associates to m are 1 = N (y) . Conversely if N (x) = 1, then x−1 = x̄ and therefore
{±m} . Now 0 6= a ∈ Z is irreducible iff whenever a = bc we must have b or c
is ±1. (That is |a| is prime in our old sense of the word.) It therefore follows U (R) = {x ∈ R : N (x) = 1} = {±1} .
by Euclid’s lemma (also see Theorem 23.7)) that if a|bc then a|b or a|c which
means that a is prime in our new sense of being prime as well. As will see in • α is irreducible. If α = xy with x, y ∈ R, then 4 = N (α) = N (x) N (y) . If
Theorem 23.4 below, prime always implies irreducible. Thus we are new notion neither x nor y is a unit, we must have N (x) = 2 = N (y) which is impossible
of prime gives {±2, ±3, ±5, ±7, . . . } for the primes in Z. since N (x) = a2 + 3b2 which is never 2. Thus no such decomposition exists
and α is irreducible.
Proposition 23.3 (hprimei is prime). Let D be an integral domain and a ∈ • α is not prime. Since αᾱ = N (α) = 4, α|4 while 4 = 2√· 2. So if α were
D with 0 6= a ∈
/ U (R) . Then the following are equivalent; prime we would have α|2, i.e. 2 = αx for some x = a + b −3. However if
2 = αx then 2ᾱ = α · ᾱx = N (α) x = 4x. Thus we must have 2x = ᾱ, i.e.
1. a is prime. √ √
2. hai = Ra is a prime ideal in D. 2a + 2b −3 = 1 − −3
3. R/ hai is an integral domain.
which is impossible to do unless we take a = 1/2 and b = −1/2 neither of
Proof. See the homework. which are in R. Thus we have shown α does not divide 2 and so α is not
prime.
Theorem 23.4 (Prime =⇒ irreducible). If D is an integral domain and • Similarly, one shows that 2 irreducible but not prime. To see that 2 is not
a ∈ D is prime then a is irreducible. prime, notice that 2|αᾱ = 4 while 2 - α. If 2 were reducible, then 2 = xy
with x, y ∈
/ U (R) . Since 4 = N (2) = N (x) N (y) and neither N (x) = 1 or
Proof. Suppose that a ∈ D is a prime. If a = bc then a|b or a|c. Say that
N (y) = 1 we must have N (x) = 2 which is impossible.
a|b, i.e. b = at. Then a = atc and since D is a domain we may cancel a form this
1
You should check this by direct calculation if this is not familiar to you.
76 23 Lecture 23

The next theorem gathers in one place all of the results we know involving Example 23.8. We have x ∈ Z [x] is prime and hence irreducible. Neverthe-
maximal and prime principle ideals and prime and irreducible elements of an less, hxi is not a maximal ideal as we have seen above. For example, hxi $
integral domain R. {p (x) : p (0) is even} $ Z [x] . Thus we can not drop the assumption that D is
a PID in Theorem 23.4. Nevertheless, we will see below in Theorem 24.1 that
Theorem 23.6. Let R be an integral domain and hai = Ra where a ∈ R with
f (x) ∈ Z [x] is prime iff it is irreducible in Z [x] .
a 6= 0 and a ∈
/ U (R) . Then;
hai is maximal ideal ⇐⇒ R/ hai is a field Example 23.9. In Z5 , x2 + 1 has two zeros, namely x = 2 and x = 3. There-
fore, x2 + 1 factors as (x − 2) (x − 3) from which it follows that x2 + 1
⇓ (1) is not a maximal ( ⇐⇒ prime) ideal in Z5 [x] . Therefore it follows that
hai is a prime ideal ⇐⇒ R/ hai is an integral domain Z5 [x] / x2 + 1 ∼ = Z5 [i] is not a field and not even an integral domain. Ex-
m (2) plicitly we have (i − 2) (i − 3) = 6 − 1 − 5i = 0 in Z5 [i] .
a is a prime element in R Remark 23.10. On the other hand, it is easy to verify that x2 + 1 has no zeros
⇓ (3) over Z3 and therefore x2 + 1 is irreducible over Z3 . From this we may conclude
a is an irreducible element in R. that Z3 [i] ∼
= Z3 [x] / x2 + 1 is a field. This was already proved with some
effort in Example 5.3.
If we further assume that R is a PID, then the converse of all of the above
implications hold.2
Proof. The horizontal equivalences are contained in Theorem 14.3. The ver-
tical implication (1) has been explained in Corollary 14.4. The vertical equiva-
lence (2) is Proposition 23.3 and (3) is Theorem 23.4.
If R is now a PID, we know from Theorem 18.8 that a is irreducible implies
hai is maximal. Therefore all statements in the theorem are now equivalent.
If time permits we will consider some other interesting integral domains
which are not PID’s and for which some of the equivalences in Theorem 23.6
no longer hold.
Remark 23.7 (Prime ⇐⇒ irreducible in a PID). Here is an alternative proof
that if D is a PID (principle ideal domain) and a ∈ D is irreducible then a is
prime. Suppose that a ∈ D is irreducible and a|bc for some b, c ∈ D. Let
I := aD + bD := {ax + by : x, y ∈ D} .
Notice that I is an ideal of D and hence I = hdi for some d ∈ D. Since a ∈ I we
have a = dr for some r ∈ D. As a is irreducible, either d ∈ U (R) or a ∈ U (R) .
If d ∈ U (R) , then I = hdi = R and in particular 1 ∈ I. Thus there exists
x, y ∈ D such that ax + by = 1. Therefore,
axc + ybc = c
from which it follows that a|c. On the other hand of r ∈ U (R) , then I = hdi =
hai . Since b ∈ I it follows that b = at for some t ∈ D, i.e. a|b. So we have shown
that a|bc implies a|b or a|c, i.e. a is prime.
2
More succinctly, if R is a PID and a ∈ R× \ U (R) is irreducible, then hai is a
maximal ideal in R.

Page: 76 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

24
Lecture 24

It is possible for all irreducibles to be prime even if the ring is not a PID as 2. The factorization into irreducibles is unique up to associates and the order
the next theorem shows. in which the factors appear. To be more precise, if a1 , . . . , an and b1 , . . . , bm
are irreducibles in R such that a1 . . . an = b1 . . . bm then m = n and there
Theorem 24.1 (Irreducible ⇐⇒ prime in Z[x]). If f (x) ∈ Z [x] is irre- exists a permutation, σ ∈ Sn , such that ai and bσ(i) are associates for each
ducible then f (x) is prime in Z [x] . i.
Proof. In the proof to follow we suppose that g (x) , h (x) ∈ Z [x] and that Example 24.3. The fundamental theorem of arithmetic asserts that Z is a unique
f (x) |g (x) h (x) , i.e. there exists k (x) ∈ Z [x] such that factorization domain.
√
g (x) h (x) = k (x) f (x) . (24.1) Example√ 24.4(Z[ −3] is not a UFD). Example 23.5 easily allows us to see that
R := Z −3 ⊂ C is not a unique factorization domain. (see Definition 24.2
Case 1. Suppose that deg (f ) = 0, i.e. f (x) is constant. Since f (x) is irre- below). We have see that 4 ∈ R may be factored as 4 = 2 · 2 and also as 4 = α · ᾱ
ducible we must have f (x) = p where p is a prime in Z which we assume to be where
positive for simplicity. Applying the content function, c (·) , to Eq. (24.1) gives √ √
α := 1 + −3 and ᾱ := 1 − −3.
c (g) c (h) = c (k) c (f ) = c (k) · p. Since the units of R are {±1} it is clear neither α nor ᾱ is associated to 2. (See
Chapterh18 iof the Gallian [2] and in particular the comments on p. 333 as to
Therefore p|c (g) c (h) which implies p|c (g) or p|c (h) , i.e. p|g (x) or p|h (x) . √
when Z d is a UFD.)
Case 2. Suppose that deg (f ) ≥ 1. In this case c (f ) = 1 for otherwise

1
 Lemma 24.5. Suppose that R is an integral domain, a ∈ R is prime, and
f (x) = c (f ) f (x) b1 , . . . , bn ∈ R are irreducibles. If a| (b1 . . . bn ) then a is associated to bi for
c (f )
some i.
would be a non-trivial factorization of f (x) over Z [x] . Since f (x) is primitive
Proof. Since a is prime, a|bi for some i. As bi is irreducible this can only
and irreducible over Z [x] it is irreducible over Q [x] by Corollary 21.5. Therefore
happen if bi = au where u ∈ U (R) , i.e. a is associated to bi .
f (x) is prime in Q [x] and it follows that f (x) |g (x) or f (x) |h (x) in Q [x] . But
by Corollary 20.13 this implies that f (x) |g (x) or f (x) |h (x) in Z [x] , i.e. f (x) m
Proposition 24.6. Suppose that R is an integral domain, 1 ≤ m ≤ n, {ai }i=1
is prime in Z [x] . n
are primes in R and {bj }j=1 are irreducibles in R. If

a1 . . . am = b1 . . . bn ,
24.1 Unique Factorization Domains n
then m = n and after a possible reordering of the {bi }i=1 , we have ai and bi
Definition 24.2 (UFD). A unique factorization domain (UFD for short) are associates for each i.
is an integral domain, R, such that
Proof. Before proving the general case let first do two special cases as a
1. Every non-zero element of R which is not a unit may be written as a product warm up.
of irreducibles in R. 1) Suppose that m = 2 and n = 3 so that a1 a2 = b1 b2 b3 . By Lemma 24.5,
bi = u1 a1 for some i and u1 ∈ U (R) . By relabeling the bi if necessary we may
78 24 Lecture 24

assume that i = 1 and we now have a1 a2 = a1 u1 b2 b3 and so by cancellation, As an application of this theorem let us give another proof of Eisenstein’s
a2 = (u1 b2 ) b3 . Another application of Lemma 24.5 shows a2 is associated to Criterion – see Theorem 22.6. For the readers convenience we repeat the state-
u1 b2 (hence b2 ) or b3 . Again we may relabel the bi if necessary and suppose ment here.
that a2 is associated to b2 , i.e. b2 = u2 a2 for some u2 ∈ U (R) . Thus we may Pn
conclude that a2 = a2 u1 u2 b3 which implies 1 = u1 u2 b3 . This shows that b3 must Theorem 24.8 (Eisenstein’s Criterion). Let f (x) := k=0 ak xk ∈ Z [x] .
be a unit which contradicts the assumption that b3 was prime and in particular If there is a prime p such that p - an , p2 - a0 while p|aj for j = 0, 1, 2, . . . , n − 1
not a unit. Thus this case can not happen. then f (x) is irreducible. over Q.
2) Suppose that m = n = 2 so that a1 a2 = b1 b2 . Working as in case 1), we
Proof. If f (x) were reducible over Q [x] then, by Theorem 21.3, it would
easily show, after relabeling the bi if necessary, that a1 is associated to b1 and
be reducible over Z [x] and in fact there would exists g (x) , h (x) ∈ Z [x] with
a2 is associated to b2 .
deg (g) , deg (h) ≥ 1 such that f (x) = g (x) h (x) . Letting f¯, ḡ and h̄ be the
The formal proof goes by induction on n. When n = m = 1 there is nothing
reductions of f, g, and h mod p we find, ān xn = ḡ (x) h̄ (x) wherein we have use
to prove. Now suppose that n ≥ 2 and the result holds for lower n. Then working
ān := an mod p 6= 0 in Zp . Since Zp is a field, Zp [x] is a UFD and because deg ḡ
as above, after relabeling the b’s is necessary we may assume that b1 = a1 u1 for
and deg h̄ are both less than n we must have x divides both ḡ (x) and h̄ (x) . (In
some u1 ∈ U (R) . If m = 1, this would imply that 1 = b2 . . . bn showing that all
fact there must exist b, c ∈ Z× p such that ḡ (x) = bx
deg g
and h̄ (x) = cxdeg h .)
the remaining bi are units which is impossible. Therefore we must have m ≥ 2 2
This then implies that p|g (0) and p|h (0) and therefore p |a0 as a0 = g (0) h (0) .
and we now have
However this contradicts the assumption that p2 - a0 .
u−1


1 a2 a3 . . . am = b2 b3 . . . bn .

Since u−1 Theorem 24.9 (Z[x] is a UFD). Z [x] is a unique factorization domain.
1 a2 is still prime, it follows by the induction hypothesis that m = n
n
and that after relabeling the {bi }i=2 if necessary that Proof. Existence. Let f (x) ∈ Z [x] and set f0 (x) := c(f 1
) f0 (x) so that
u−1 f (x) = c (f ) f0 (x) where f0 (x) is primitive. Since we may factor c (f ) into
1 a2 = u2 b2 , a3 = u3 b3 , . . . , an = un bn
prime integers which are irreducible in Z [x] it suffices to factor f0 (x) into
for some ui ∈ U (R) . This completes the proof since a2 = u1 u2 b2 and u1 u2 ∈ irreducibles. So from now on we assume that f (x) is primitive.
U (R) . If f (x) is not irreducible, the f (x) = h (x) g (x) with neither h (x) or g (x)
Because of this proposition we see that the uniqueness of factorizations into being invertible. Since c (h) c (g) = c (f ) = 1 we know that both h and g are
irreducibles will hold for an integral domain R when R has the property that primitive and therefore deg h ≥ 1 and deg g ≥ 1. Since deg f = deg h+deg g, we
irreducibles and primes are the same, i.e. irreducible implies prime. This is not must have deg h < deg f and deg g < deg f. Thus the existence of a factorization
always the case as was shown in Example 32.2. However we have seen that of f (x) when f (x) is primitive goes just as in the proof of Theorem 24.7.
irreducibles and primes are the same in principle ideal domains (Theorem 23.6) Uniqueness. The uniqueness of the factorization follows by combining
like F [x] and in R = Z [x] , see Theorem 24.1. Proposition 24.6 with Theorem 24.1.

Theorem 24.7 (F [x] is a UFD). If F is a field, then F [x] is a unique fac- Fact 24.10 (Facts about UFD’s) Here are two more facts about unique fac-
torization domain. torization domains which we will not prove in class.

Proof. Let f (x) ∈ F [x] with deg (f ) ≥ 1. If f (x) is not irreducible, then 1. Every PID is a UFD. This is proved in Gallian [2, p. 327 – 329] and is
f (x) = g (x) h (x) for some g (x) , h (x) ∈ F [x] with deg g ≥ 1 and deg h ≥ 1. fairly similar to the proof of Theorem 24.7.
Since deg g + deg h = deg f we must have deg g < deg f and deg h < deg f. 2. If R a UFD then R [x] is a UFD. Our proof that Z [x] is a UFD is rather
By further decomposing g (x) and h (x) is possible we must eventually arrive similar to the general case – see Knapp [3, pgs. 384-396] for details. (This
at a decomposition of f (x) into irreducibles which gives the desired existence. proof would make use of the field of fractions, see Subsection 24.2.4 below
(You may add the formal the induction proof if you wish.) The uniqueness of if you are interested.)
this factorization follows from Proposition 24.6 and the fact that primes and
irreducibles are the same in a PID like F [x] , see Theorem 23.7.

Page: 78 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

 24.2 Extra Topics (not covered in class) 79
24.2 Extra Topics (not covered in class) The next result has already been proved in Theorem 18.8. Nevertheless it is
instructive to give another proof.
24.2.1 Greatest Common Divisors (not covered in class)
Corollary 24.16. Suppose that F is a field and p (x) ∈ F [x] is irreducible.
Definition 24.11. Let R be an integral domain. We say that c ∈ R is a common Then F [x] / hp (x)i is a field.
divisor of a1 , . . . , an ∈ R if c|ai for all i. We say that d ∈ R is a greatest
common divisor of a1 , . . . , an ∈ R if d is a divisor and if c|d for any other Proof. If f (x) ∈ F [x] with deg f (x) < deg p (x) , then f (x) and p (x) are
divisor of a1 , . . . , an ∈ R. relatively prime since the only possible non-trivial common factor would be
p (x) which is impossible since deg f (x) < deg p (x) . Therefore by Proposition
If c and d are two greatest common divisors of a1 , . . . , an ∈ R, then c = ud 24.12, there exists s (x) , t (x) ∈ F [x] such that
and d = vc for some u, v ∈ R. Therefore, c = uvc and since R is an integral
domain, uv = 1 by cancellation. Thus c and d are associates and the gcd of s (x) f (x) + t (x) p (x) = 1.
a1 , . . . , an is well defined modulo U (R) . −1
Since [p (x)] = 0 it follows that [s (x)] [f (x)] = 1 and we have shown [f (x)] =
Proposition 24.12. If R is a PID, then every list of elements, a1 , . . . , an ∈ [s (x)] exists. Thus every non-zero element of F [x] / hp (x)i is invertible and
R× , has a greatest common divisor, d. Moreover there exists r1 , . . . , rn ∈ R hence F [x] / hp (x)i is a field.
such that −1
d = r1 a1 + · · · + ri ai + · · · + rn an . (24.2) Remark 24.17. It is interesting to see how to compute [f (x)] in the context
of Corollary 24.16 explicitly. This can be done using the same technique as
Proof. Let J = ha1 , . . . , an i = ha1 i + · · · + han i be the ideal of R generated for finding inverses in Un = U (Zn ) . That is we repeatedly use the division
by (a1 , . . . , an ) and choose c ∈ R such that J = hci . Then by definition of J, algorithm as follows;
c may be expressed as in Eq. (24.2) and ai ∈ hci for all i. The latter condition
asserts that c|ai for all i so that c is a common divisor and it follows from Eq. p (x) = k0 (x) f (x) + f1 (x)
(24.2) that in fact c is a greatest common divisor. f (x) = k1 (x) f1 (x) + f2 (x)
Definition 24.13. As in the case R = Z we will say that elements, a1 , . . . , an f1 (x) = k2 (x) f2 (x) + f3 (x)
in a PID, R, are relatively prime if the greatest common divisors are units. ..
.
Alternatively, this is equivalent to saying 1 ∈ gcd (a1 , . . . , an ) .
fl−1 (x) = kl (x) fl (x) + fl+1
Here is the analogue of Euclid’s Lemma in this more general context.
where fl+1 ∈ F × . We then work backwards from this set of equation to solve
Lemma 24.14 (Euclid’s Lemma II). Let R be a PID and a, b, c ∈ R with for fl+1 . Rather than explain this in general, let me show how this works in an
a, b being relatively prime. If a| (bc) then a|c. example.

Proof. From Proposition 24.12 we know that 1 = sa + tb for some s, t ∈ R. Example 24.18. Let I := x3 + 2x + 1 ⊂ Q [x] . Notice that p (x) := x3 +2x+1
Therefore c = sac + tbc from which it follows that a|c. −1  −1
is irreducible by the mod 3 test. We now wish to compute [x] and x2 .
The first case is rather simple since,
Remark 24.15. If F is a field and p (x) , q (x) ∈ F [x] are two non-zero polyno-
mials, one may find gcd (p, q) using the division algorithm just as we did for the p (x) = x3 + 2x + 1 = x x2 + 2 + 1


integers, Z. The point is to repeatedly make use of the fact that
and therefore
gcd (p, q) = gcd (q, q mod p) 1 = p (x) − x x2 + 2

where q mod p denotes the remainder from dividing p into q.where q mod p de- so that
1 = [p (x)] − [x] x2 + 2 = − [x] x2 + 2
   
notes the remainder from dividing p into q.

Page: 79 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

80 24 Lecture 24
−1 
which shows that [x] = − x2 + 2 . which is to say,
x4 + 4x2 + 4 = x · p (x) + 2x2 − x + 4
 2 −1
Now to compute x . By the division algorithm we have
and therefore,
p (x) = x3 + 2x + 1 = x · x2 + 2x + 1  2 −1 
= x · p (x) + 2x2 − x + 4 = 2x2 − x + 4
  
x
and
1 1
 2x − 4 as before. Alternatively, from Eq. (24.3) we may simply use
2x + 1 x2  4 
x = x · x3 = [x · (−2x − 1)] = − 2x2 + x
  
1
x2 + 2x
− 12 x to discover,
− 12 x − 1
4  2
−1
= x4 + 4x2 + 4 = − 2x2 + x + 4x2 + 4
1
  

4
[x]
= 2x2 − x + 4 .
 
so that  
2 1 1 1
x = x− (2x + 1) + .
2 4 4 Remark 24.19. If p (x) ∈ F [x] is not irreducible, then we may still compute
U (F [x] / hp (x)i) as we did for the rings Zm where m was not prime. The result
Working backwards this implies,
is,
 
1 1 1
U (F [x] / hp (x)i) = [f (x)] : deg f (x) < deg p (x) and gcd (f (x) , p (x)) ∈ F × .

= x2 − x− (2x + 1)
4 2 4
  −1
1 1 The division algorithm gives us again a way to compute [f (x)] when
= x2 − p (x) − x · x2


x− gcd (f (x) , p (x)) ∈ F × .
2 4
     
1 1 1 1
= x2 1 + x− x − x− p (x) . 24.2.2 Partial Fractions (not covered in class)
2 4 2 4

Therefore, In this section let F be a fixed field. We are going verify in this sections stan-
       dard facts about partial fraction decompositions which you were probably first
1 2 1 1 introduced to when studying integral calculus.
[1] = 4 · =4· x 1+ x− x
4 2 4
 2  2  2  Theorem 24.20 (Partial fractions). Let p (x) , q (x) , and f (x) be in F [x]
= x [4 + (2x − 1) x] = x 2x − x + 4 and assume that p and q are relatively prime and that deg f (x) < deg p (x) +
deg q (x) = deg (p (x) q (x)) . Then there exists unique polynomials, a (x) and
which shows that  2 −1  2  b (x) with deg a (x) < deg p (x) and deg b (x) < deg q (x) such that
x = 2x − x + 4 .
f (x) a (x) b (x)
As a check we should have = + . (24.4)
p (x) q (x) p (x) q (x)
 2 −1  −1 2 h 2
2 i  4
= x + 4x2 + 4 .

x = [x] = x +2 (24.3) Proof. Choose t (x) and s (x) ∈ F [x] such that 1 = t (x) p (x) + s (x) q (x) .
Then
However
f (x) f (x)
 x = (t (x) p (x) + s (x) q (x))
x + 2x + 1 x4 + 0x3 + 4x2 + 0x + 4
3 p (x) q (x) p (x) q (x)
s (x) f (x) t (x) f (x)
x4 + + 2x2 + x = + .
2x2 − x + 4 p (x) q (x)

Page: 80 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

 24.2 Extra Topics (not covered in class) 81
n n
Let a (x) := s (x) f (x) mod p (x) and b (x) := t (x) f (x) mod q (x) in which case n
X ai (x) X n−i
0 = p (x) = ai (x) p (x) =⇒ ai (x) = 0 ∀ i.
we have, p (x)
i
i=1 i=1
f (x) a (x) b (x)
= + + k (x)
p (x) q (x) p (x) q (x) To see this observe that
for some k (x) ∈ F [x] . Multiplying this last equation by p (x) q (x) then implies, n
X n−i n−1
0= ai (x) p (x) = a1 (x) p (x) + r (x)
f (x) = a (x) q (x) + b (x) p (x) + k (x) q (x) p (x) . i=1

So on one hand, where

n
X n−i n−1
deg k (x) + deg (q (x) p (x)) = deg (k (x) q (x) p (x)) r (x) = ai (x) p (x) and deg r (x) < deg p (x) .
i=2
= deg (f (x) − a (x) q (x) − b (x) p (x)) < deg (p (x) q (x))
Therefore, by the division algorithm (or directly by degree arguments) we must
from which we conclude that deg k (x) = −∞, i.e. k (x) = 0. So we have proved have that a1 (x) = 0 and r (x) = 0. Repeating this procedure (induction again)
the existence of the decomposition in Eq. (24.4). then shows that all of the ai (x) must be zero.
To prove uniqueness, it suffices to show that if f (x) = 0 then a (x) = 0 = m
b (x) in Eq. (24.4) or equivalently that Corollary 24.22 (Partial fractions). Let F be a field and {pi (x)}i=1 be a
Pmin F [x] . Further let ki ≥ 1
collection of distinct monic irreducible polynomials
a (x) q (x) + b (x) p (x) = 0 =⇒ a (x) = 0 = b (x) . for each i and f (x) ∈ F [x] with deg f (x) < i=1 ki deg pi (x) . Then there
exists aij (x) ∈ F [x] for 1 ≤ i ≤ m and 1 ≤ j ≤ ki such that deg aij (x) <
However, since a (x) q (x) = −b (x) p (x) it follows that q (x) |b (x) p (x) and as deg pi (x) for all i and j and
q (x) and p (x) are relatively prime, we must have q (x) |b (x) . but deg b (x) <
deg q (x) and therefore this is only possible if b (x) = 0 which then implies f (x) X ki
m X
aij (x)
a (x) = 0 as well. Qm ki
= j
.
i=1 pi (x) i=1 j=1 pi (x)
Proposition 24.21. Suppose that p (x) ∈ F [x] is a polynomial with deg p (x) ≥
n
1, n ∈ Z+ , and f (x) ∈ F [x] has deg f (x) < deg p (x) = n deg p (x) . Then Proof. Repeatedly make use of Theorem 24.20 and Proposition 24.21.
there exists unique polynomials, ai (x) with deg ai (x) < deg p (x) such that
n 24.2.3 Factorizing Polynomials in finite time (not covered in class)
f (x) X ai (x)
n = i
. (24.5)
p (x) i=1 p (x) The goal of this section is to show that it is possible to factor f (x) ∈ Z [x]
with a finite number of operations. The point is Mignotte’s bound in Corollary
Proof. By the division algorithm, there exists k (x) , r (x) ∈ R [x] such that 24.27 which states that if h (x) ∈ Z [x] is a factor of f (x) , then there is an a
f (x) = k (x) p (x) + an (x) where deg an (x) < deg p (x) . Moreover we must priori bound on the size of all of the coefficients of h (x) in terms of a bound
n−1
have deg k (x) < deg p (x) since otherwise we would violate the assumption on the coefficients of f (x) and the degrees of the polynomials, f (x) and h (x) .
n
that deg f (x) ≥ deg p (x) . Therefore, Thus when looking for factorizations of f (x) we may reduce the question to a
finite number of possibilities to try out. The material presented below is taken
f (x) k (x) p (x) + an (x) k (x) an (x) directly from [4, p. 162 - 164].
n = n = n−1 + n
p (x) p (x) p (x) p (x) Pn
Notation 24.23 Given f (x) = k=0 fk xk ∈ C [x] , let
and iterating this procedure (i.e. by induction) we easily prove the existence of
!1/2
the decomposition in Eq. (24.5). X X
Pn 2
For uniqueness, it suffices to show that if i=1 ap(x)i (x)
i = 0 with deg ai (x) <
kf k∞ := max |fk | ,
k
kf k1 := |fk | , and kf k2 := |fk | .
k k
deg p (x) , then ai (x) = 0 for all i. Equivalently we must show if

Page: 81 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

82 24 Lecture 24
Q Q
Further we let Then M (f ) = |fn | i∈L |zi | = fn i∈L z̄i = |gn | and therefore,
n
Y
M (f ) := |fn | max (1, |zi |) M (f ) ≤ |gn | ≤ kgk2 .
i=1
n Q
where {zi }i=1 are the roots of f counted with multiplicities, i.e. Letting h (x) := fn i∈S (x − zi ), we have, using Lemma 24.24 repeatedly, that
n
Y
f (x) = fn (x − zi ) . (24.6)
Y Y
kg (x)k2 = h (x) (z̄i x − 1) = h (x) (x − zi ) = kf (x)k2 .
i=1 i∈L i∈L
2 2
Lemma 24.24. If z ∈ C, and f (x) ∈ C [x] , then
Combining the last two equations completes the proof.
k(x − z) f (x)k2 = k(z̄x − 1) f (x)k . ×
Theorem 24.26. Suppose that h (x) , f (x) ∈ (C [x]) , deg h = m ≥ 1, deg f =
Proof. Letting f−1 = 0 and fk = 0 for k > deg f, we have n ≥ m, and h (x) |f (x) . Then
∞
hm
khk2 ≤ khk1 ≤ 2m M (h) ≤ 2m
2
X 2
k(x − z) f (x)k2 = |fk−1 − zfk | kf k2 .
fn
k=0
∞ 
X 2 2
 Proof. Write h is factored form as
= |fk−1 | + |zfk | − 2 Re fk−1 f¯k z̄
m
k=0 Y
  ∞
X h (x) = hm (x − ui )
2 2
Re fk−1 f¯k z̄ .


= 1 + |z| kf (x)k2 −2 i=1
k=0 m n
where each factor, (x − ui ) , is also a factor of f (x) , i.e. {ui }i=1 ⊂ {zi }i=1 with
On the other hand, multiplicities. The k th coefficient of h (x) is given by
∞
X X Y
2 2 hk = hm (−ui )
k(z̄x − 1) f (x)k2 = |z̄fk−1 − fk |
k=0 S⊂{1,2,...,m} i∈S
#S=m−k
∞ 
X 2 2

= |z̄fk−1 | + |fk | − 2 Re z̄fk−1 f¯k and therefore,
k=0
∞ X Y

2

2
X |hk | ≤ |hm | |ui |
Re fk−1 f¯k z̄ .


= 1 + |z| kf (x)k2 −2
S⊂{1,2,...,m} i∈S
k=0 #S=m−k
 
Comparing these last two equations completes the proof. X m
≤ M (h) = M (h) .
k
Theorem 24.25 (Landau’s inequality). For all f (x) ∈ C [x] , M (f ) ≤ S⊂{1,2,...,m}
#S=m−k
kf k2 .
Summing this equation on k (using the binomial theorem) shows,
Proof. Let f (x) be factored as in Eq. (24.6) and S := {1 ≤ i ≤ n : |zi | ≤ 1}
and L := {1 ≤ i ≤ n : |zi |Q
> 1} – so S represents the small roots and L the large khk2 ≤ khk1 ≤ 2m M (h) .
ones. Then M (f ) = |fn | i∈L |zi | . Now let
|hm |
!
Y  Finally, looking at the definition of M (·) , we see that M (h) ≤ |fm | M (f ) .
Y Y Y 1 Y
g (x) := fn (z̄i x − 1) (x − zi ) = fn z̄i · x− (x − zi ) .
z̄i
i∈L i∈S i∈L i∈L i∈S

Page: 82 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

 24.2 Extra Topics (not covered in class) 83
Corollary 24.27 (Mignotte’s bound). Suppose that f, g, h ∈ Z [x] have de- Now generally we start with a ring R, which is commutative with 1. We are
grees deg f = n ≥ 1, deg g = m, and deg h = k, and that gh|f in Z [x] . Then unhappy if we can’t divide one element of R by another. So we want to replace
R by a bigger ring S, with R ⊆ S where S is a field. That is, in S you can
kgk∞ khk∞ ≤ kgk2 khk2 ≤ kgk1 khk1 (abstract nonsense) divide and two elements, as long as you don’t divide by zero. When can we do
≤ 2m+k kf k2 ≤ (n + 1)
1/2
2m+k kf k∞ (24.7) this? How can we do this?

Example 24.28. Let R = Z. Then we know Z ⊂ Q where Q is a field. Q is also

and
the “smallest” field with this property in some sense.
khk∞ ≤ khk2 ≤ 2k kf k2 ≤ 2k kf k1 and
Lemma 24.29. Suppose S is a field and R ⊂ S is a subring. Then R is an
1/2
khk∞ ≤ khk2 ≤ (n + 1) 2k kf k∞ . (24.8) integral domain.

Proof. First off, notice that M (g) M (h) = M (gh) ≤ M (f ) since (gh)n = Proof. Since S is a field, S is an integral domain. Now if ab = 0 for some
gm hk |fn . By Theorems 24.26 and 24.25, a, b ∈ R, then ab = 0 in S. Since S is an integral domain, a = 0 or b = 0 in S,
and so also in R.
kgk1 khk1 ≤ 2m M (g) 2k M (h) = 2m+k M (gh) So if the R we start with is not a domain, we can’t hope to form a bigger
1/2 ring containing R which is a field. The next theorem shows this is the only
≤ 2m+k M (f ) ≤ 2m+k kf k2 ≤ (n + 1) 2m+k kf k∞ .
obstruction to embedding a commutative ring, R, with identity into a “field of
The second set of inequalities follow by taking g = 1 in Eq. (24.7). fractions.”

Theorem 24.30. Let R be a integral domain with 1 and R× := R \ {0}. Then

24.2.4 Fields of fractions (not covered in class) we may define an equivalence relation on R × R× by,

Suppose you met an alien from a planet where they had invented rings, but (a, b) ∼ (c, d) ⇐⇒ ad = bc.
because everything on their planet involved whole units, they had never invented
a
the idea of division. So this guy knows about Z, but not about fractions. How Let b= [(a, b)] denote the equivalence class determined by ∼ and S :=
a
R and b ∈ R× . Further define two binary operations, + and ·, on

do you explain in an algebraic way about Q? b : a ∈
Idea. An element of Q is a fraction ab where a, b are integers, that is, elements S by,
in Z and b 6= 0. But, the word fraction is undefined here, and the notation will be a c ad + cb a c ac
+ := and · := . (24.9)
difficult to justify. So instead we could do the following—we identify ab with an b d bd b d bd
ordered pair (a, b) with a, b ∈ Z and b 6= 0. Secretly thinking of fractions, now we Then with these operations, S becomes a field and the map,
can say how to multiply and add in Q. To multiply, we set (a, b)·(c, d) = (ac, bd),
and to add, we set (a, b) + (c, d) = (ad + bc, bd), for b, d 6= 0. We can also divide R 3 a → a/1 ∈ S
them: (a, b) ÷ (c, d) = (ad, bc), when b, c, d 6= 0.
is an injective rings isomorphism. In the future we identify a ∈ R with a/1 ∈ S
However, there are problems to this approach. For example
and views R as a sub-ring of S. The new ring S is called the field of fractions
(1, 2)[(1, 2) + (1, 2)] = (1, 2)(4, 4) = (4, 8) of the integral domain, R.

(1, 2)(1, 2) + (1, 2)(1, 2) = (1, 4) + (1, 4) = (8, 16). Proof. We’re almost already done. The proof is the same as for Z!

These are not the same, so distributivity fails. So to solve this problem, we 1. Define Se = {(a, b) : a, b ∈ R, b 6= 0}.
need to say when two fractions are the same. We put an equivalence relation 2. Put an equivalence relation on Se by saying (a, b) ∼ (c, d) if ad = bc in R.
on Q saying that (a, b) ∼ (c, d) when ad = bc. Then we let Q be the equivalence We start by checking that “∼” is an equivalence relation:
classes and refine our notion of ab to mean the equivalence class containing a) (a, b) ∼ (a, b) since ab = ba in R.
(a, b) . This works! b) (a, b) ∼ (c, d) ⇒ (c, d) ∼ (a, d) since ad = bc ⇒ cd = da.

Page: 83 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

84 24 Lecture 24

c) If (a, b) ∼ (c, d) and (c, d) ∼ (e, f ), then (a, b) ∼ (e, f ) since ad = and you manipulate these gadgets in the obvious way. The traditional no-
bc, cf = de ⇒ adf = bcf = bce so daf = dbe and af = be. Here we used tation for S is R(x).
that R is a domain in order to use cancellation.
3. We now must show that operations in Eq. (24.9) makes sense. That is if Example 24.32.

a a0 c c0 x x2 5x + x2 (x + 2) x3 + 2x2 + 5x
= 0 and = 0 , i.e. ab0 = ba0 and cd0 = dc0 , + = =
b b d d x+2 5 5(x + 2) 5x + 10

then we must show, Example 24.33. R = Z[i] = a + bi : a, b ∈ Z, i2 = −1 . S consists of fractions

ac a0 c0 a + bi
= 0 0 , i.e. that acb0 d0 = a0 c0 bd and { : c + di 6= 0, i2 = −1}
bd bd c + di
ad + cb a0 d0 + c0 b0
= , i.e. that (ad + cb) b0 d0 = (a0 d0 + c0 b0 ) bd. under the equivalence relation
bd b0 d0
These are easily checked; a + bi a + bi a − bi ac + bd + (bc − da)i ac + bd (bc − ad)i
∼ = 2 2
= 2 + 2
c + di c + di c − di c +d c + d2 c + d2
acb0 d0 = (ab0 ) (cd0 ) = (ba0 ) (dc0 ) = a0 c0 bd
In fact, it is not too hard to show S ∼
= Q[i] = {a + bi : a, b ∈ Q}.
and
Example 24.34. An infinite field of characteristic p. Let p be prime. Zp is a
0 0 0 0 0 0 0 0 0 0
(ad + cb) b d = ab dd + bb cd = ba dd + bb dc = (a d + b c ) bd 0 0 0 0 finite field with characteristic p. Zp [x] is an infinite ring with characteristic p.
Let S be the field of fractions of Zp , written Zp (x). Then S is an infinite field
as desired. of characteristic p.  
4. We must now show that (S, ·, +) satisfies the axioms of a ring. There is lots f
S= : f, g ∈ Zp [x] , g 6= 0
to do here and we leave the tedious details to the reader. However, observe g
that 0S = 10 and 1S = 11 .
Example 24.35. Suppose we start with a field. What is its field of fractions?
5. The new ring S is a field. This is almost immediate. Given ab ∈ S × , then
a −1
For example, what is the field of fractions of Q? S = {(a, b) : a, b ∈ Q, b 6= 0}
b × a b ab 1
= ab as was to be expected



a ∈ S and b · a = ab = 1 = 1S . Thus b under (a, b) ∼ (c, d) if ad = bc. Then (a, b) ∼ ( ab , 1). Thus the map Q → S by
and hence S is a field. q 7→ (q, 1) is an isomorphism.
6. Finally we observe that ϕ : R → S defined by ϕ (r) = 1r is a one to one
0
ring homomorphism. It is one-to-one because 1r = r1 iff r · 1 = r0 · 1, i.e. iff
r = r0 . In this way we may now identify R with ϕ (R) ⊂ S.

Once we know that we can do this, we think of elements of the field of

fractions S as fractions, and we write them that way, instead of as ordered
pairs.

Example 24.31. Let R = R [x]. Then the field of fractions S of R looks like
 
 a xn + · · · + a x + a m 
n 1 0
X
j
S= : ai , bj ∈ R, bj x 6
= 0
 bm xm + · · · + b1 x + b0 
j=0

Page: 84 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

25
Lecture 25
m
25.1 Vector Spaces & Review of Linear Algebra X
uj = vi aij for all 1 ≤ j ≤ n.
i−1
Definition 25.1. Let F be a field (called “scalars”). A set V (called ”vectors”)
is called a vector space over F if n Pn
Thus if {bj }j=1 ⊂ F we will have j=1 bj uj = 0 iff
1. V is an Abelian group under addition (written +). n
X m
X m
X n
X
2. For each a ∈ F , v ∈ V , there is an element av ∈ V (this is called “scalar 0= bj vi aij = vi aij bj .
multiplication”). j=1 i−1 i−1 j=1
3. a(v + u) = av + au ∀a ∈ F, u, v, ∈ V (scalar multiplication is group homo- Pn
morphism). Thus j=1 bj uj = 0 if (b1 , . . . , bn ) is a solution to the system of equations,
4. (a + b)v = av + bv ∀a, b ∈ F, v ∈ V . n
5. a(bv) = (ab)v ∀a, b ∈ F, v ∈ V .
X
aij bj = 0 for 1 ≤ i ≤ m.
6. 1v = v ∀v ∈ V , where 1 = 1F is the identity of F . j=1

These look a lot like ring axioms. But in general, a vector space V is not a However, from basic row reduction techniques one knows that such a homoge-
ring. The two elements participating in this multiplication come from different nous system of m – linear equations with n > m unknowns always have non-
n
sets. Let us recall a couple of basic notions from Math 20F before we go onto trivial solutions and {uj }j=1 can not be linearly independent.
giving examples.
Example 25.5. In Math 20F, one concentrates on examples like:
Definition 25.2. A set S of vectors linearly dependent if there are vectors
• Suppose F = R, and
v1 , . . . , vn ∈ S and scalars a1 , . . . , aN not all zero with a1 v1 + · · · + an vN = 0.
If S is not linearly dependent it is called linearly independent. V = R2 = {(x, y) : x, y ∈ R}.

Then V is a vector space with the usual vector addition (so V is an abelian
Definition 25.3. A set S is a basis for the vector spacePV if S is independent group, 0 = (0, 0) is an additive identity) and scalar multiplication a(x, y) =
and spans V in that sense that ever v ∈ V is of the form ai vi for some ai ∈ F (ax, ay). Here dimR R2 = 2 with basis {(1, 0) , (0, 1)} for example.
and vi ∈ S. • Suppose F = R, and V is Euclidean 3-space,
Theorem 25.4. Every vector space V over F has a basis. Moreover, the number V = R3 = {(x, y, z) : x, y, z ∈ R}
of elements in the basis is unique, and is called the dimension of V or more
precisely the dimension of V over F which we denote by dimF V . which again is a vector space in the analogous way. In this case, dimR R3 = 3
and {(1, 0, 0) , (0, 1, 0) , (0, 0, 1)} is a basis.
Proof. See the Gallian [2, Theorem 19.1] for one proof. Alternatively go • More generally we might have taken F = R and V = Rn for any n in which
back to your Math 20F course and see how it was done there. Here is a sketch case dimR V = n.
of that method. The main point is to show if β := {v1 , . . . , vm } is a basis and
γ := (u1 , . . . , un ) is a list of any n – vectors in V with n > m, then γ is linearly Now that we have some ring theory, we can note that weirder examples have
dependent. The point is, since β is a basis, there exists aij ∈ F such that already come up. There are many examples of a ring R which is also a vector
space over some field F .
86 25 Lecture 25

Example 25.6. Let R = F [x] for some field F . Then in addition to R being a Example 25.11. Think of M2 (R) as a R-vector space. Then M2 (R) is a vector
ring, R is a vector space over F . R is already an abelian group under addition space of dimension 4 over R. A basis is
of polynomials. We define scalar multiplication in the obvious way by,        
10 01 00 00
, , ,
a [bn xn + · · · + b1 x + b0 ] := abn xn + · · · + ab1 x + ab0 . 00 00 10 01

It is easy to verify the axioms of a vector space. In fact, note that what is These are called elementary matrices. Thus dimR M2 (R) = 4.
happening here is F is a subring of F [x], just the scalar polynomial functions.
Example 25.12. Here is an important example. Let F ⊂ E where F and E
So we already have a way of multiplying a ∈ F, f ∈ F [x] together, the usual
are fields and F is a subring of E. Then we say F is a subfield of E. Notice
multiplication in F [x]. This is really all that scalar multiplication is doing. So
that E is a vector space over F . The scalar multiplication is defined using the
in fact, all of the vector space axioms reduce to some special cases of the ring
multiplication in E, i.e. if a ∈ F and v ∈ E then av ∈ E. The vector space
axioms!
axioms follow from ring axioms for E.
Example 25.7. Suppose that F is any field and V = F [x]. Then V is an infinite Notation 25.13 If F is a subfield of E we let [E : F ] := dimF (E) be the
dimensional vector space over F with basis, β := {1, x, x2 , x3 , x4 , . . . }. Indeed, dimension of E as an F vector space.
everything in F [x] may be written as an xn + ·P· · + a0 , i.e. a linear combination
n
of the {xn } . Moreover, β is independent since i=0 ai xi = 0 happens iff ai = 0 Example 25.14. R ⊂ C. Then C is an R-vector space with operation for r ∈ R,
for all i by the very definition of the polynomial ring. a + bi ∈ C, r(a + bi) = ra + (rb)i and [C : R] = 2 – a basis being {1, i} .
Proof. Every element of C has the form a + bi = a(1) + b(i) for some
Definition 25.8. Finally recall that if V is a vector space over a field, F. Then a, b ∈ R. Moreover if a(1) + b(i) = 0 then a = 0 = b. [This justifies thinking of C
W ⊂ V is a subspace of V if W is a subgroup of V which is closed under as the plane which we tend to identify with R2 .] Of course only the number of
scalar multiplication. 0
√ is unique. S = {−1, −i} is also a basis for C over R. Similarly,
basis elements
00
S = {1 + i, 3 − 2i} is as well.
Example 25.9. Suppose that F is any field, n ∈ N, and
√ √
Example 25.15. Q ⊂ Q[ 2]. Then Q[ 2] (which √ is a field,√as we proved)
√ is a
V = {f ∈ F [x] : deg f (x) ≤ n} . space with operation for q ∈ a + b 2 ∈ 2], q(a + b
Q-vector √ √ Q, √  Q[  √  2) =
 qa + (qb) 2. Then {1, 2} is a Q – basis for Q 2 so that Q[ 2] : Q = 2.
The V is a subspace of F [x] , β = 1, x, x2 , . . . , xn is a basis for V and hence √
dim V = n + 1. The proof is as
√ in the previous example since every√ element of Q( 2) is of the
form a(1) + b( 2) for some a, b ∈ Q √ and a(1) + b( 2) = 0 happens iff a = 0 = b.
Example 25.10. M2 (F ) is a vector space over F with scalar multiplication The latter fact is a consequence of 2 being irrational.

wx
 
aw ax
 Example 25.16. We have that Q is a sub-field of R. However [R : Q] = ∞. One
a = . way to see this is to notice that any finite dimensional vector space over Q would
y z ay az
have to be countable. Since R is uncountable, R must be an infinite dimensional
Identify F as a subring of M2 (F ), via vector space over Q.
   
a0 b0
 Corollary 25.17 (of Theorem 25.4). If F is a finite field of characteristic
F 3a↔ ∈ b∈F . p, then # (F ) = pn for some n ∈ Z+ .
0a 0b

Then scalar multiplication is really left multiplication by Proof. Identify Zp with the image of the ring homomorphism,
  Zp 3 k → k · 1 ∈ F.
a0
.
0a Then F may be viewed as a vector space over Zp . Let n := dimZp F ∈ Z+ . Then
# (F ) = pn by a homework problem.

Page: 86 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

26
Lecture 26

Example 26.1. Let F = Z2 , and let f = x3 + x + 1, an irreducible polynomial The setup for the time being will be that E is a field and F ⊂ E is a sub-
over Z2 . Then R = Z2 / x3 + x + 1 is a field with 8 elements. Then we can field. Typically we might have E = C and F = Q or F is some other sub-field
think of F as a subfield of R by identifying a ∈ F with a + x3 + x + 1 in of C.
R. As such, R is a vector space over F and we have dimF R = 3. (Notice that
dim R Definition 26.4. We say an element α ∈ E is algebraic over F is there exists
8 = 23 = # (F ) F .)
p (x) ∈ F [x] such that p (α) = 0. Otherwise we say α ∈ E is transcendental
Theorem 26.2. Suppose that E is a field, G ⊂ E is a subfield of E, and F ⊂ G over F.
is a subfield of G. If m := [E : G] < ∞ and n := [G : F ] < ∞, then [E : F ] < ∞
and Remark 26.5. For each n ∈ Z+ there are countably many polynomials in Q [x]
[E : F ] = [E : G] [G : F ] . of degree n and each of these polynomials have at most n - distinct zeros.
m n
Therefore there are only countable many Q – algebraic elements in C. Since C
Moreover if {ui }i=1 is a basis for E over G and {vj }j=1 is a basis for G over is uncountable, most of the element in C are therefore transcendental. However,
F, then β := {ui vj : 1 ≤ i ≤ m, 1 ≤ j ≤ n} is an F – basis for E. proving a specific given complex number is transcendental is not so easy. We
Pm state here without proof that both e and π are transcendental numbers.
Proof. For any a ∈ E there exists ai ∈PG such that a = i=1 ai ui and for
n
each i there exists aij ∈ F such that ai = j=1 aij vj . Therefore it follows that Notation 26.6 If S is a subset of E we let F (S) denote the smallest subfield
m X
n
of E which contains F and S. This sub-field exists. Indeed,
X
a= aij ui vj
F (S) := ∩ {K ⊂ E : K is a sub-field 3 S ∪ F ⊂ K} .
i=1 j=1

and hence that the F – span of β is E. So to finish Lemma 26.7. If S and T are subsets of E, then F (S) (T ) = F (S ∪ T ) =
Pthe proof we need only show
m Pn F (T ) (S) .
β is linearly independent. If aij ∈ F satisfy, 0 = i=1 j=1 aij ui vj , then
  Proof. Sine S ∪ T ⊂ F (S) (T ) it follows that F (S ∪ T ) ⊂ F (S) (T ) . Con-
m n
X X versely it is clear that F (S) ⊂ F (S ∪ T ) and that T ⊂ F (S ∪ T ) , therefore
0=  aij vj  ui .
F (S) (T ) ⊂ F (S ∪ T ) .
i=1 j=1
√
 √
m Pn Example 26.8. R [i] = C and Q 2 = a + b 2 : a, b ∈ Q as we have see
Since {ui }i=1 is a basis for E over G it follows that j=1 aij vj = 0 for each i
n before.
and since {vj }j=1 is a basis for G over F it follows that aij = 0 for each j. This
completes the proof. Theorem 26.9 (Adjoining a root). Suppose that α ∈ E is algebraic over F
and ϕ : F [x] → E is the evaluation homomorphism, ϕ (f (x)) = f (α) . Then;

26.1 Field Theory 1. ker ϕ = hp (x)i where p (x) is an irreducible polynomial which may be taken
to be the unique monic polynomial f (x) ∈ F [x] of minimal degree which
Definition 26.3. Given fields E, F , with F ⊂ E where F is a subring of E, has α as a root. (This polynomial is called the minimal polynomial of
we say that F is a subfield of E. We call E an extension field of F and the α ∈ E.)
entire setup, F ⊆ E, is called a field extension.
88 26 Lecture 26
√ √
√
√

2. We have Example 26.11. Let us now consider Q 3, 5 = Q 3 5 . Notice that
√ √

5 is a root of x2 − 5. Moreover we claim that x2 − 5 is irreducible over Q 3 .
F (α) = F [α] = {f (α) : f ∈ F [x]} √ √
√

The point is that ± 5 ∈ / Q 3 and so x2 − 5 has no roots in Q 3 . Indeed,
= {f (α) : f ∈ F [x] with deg f (x) < deg p (x)} . √ √ √

if 5 = a + b 3 ∈ Q 3 then it would follow that
3. The map, ϕ̄ : F [x] / hp (x)i → F (α) defined by ϕ̄ ([f (x)]) = f (α) is an  √ 2 √
isomorphism of fields. 5 = a + b 3 = a2 + 3b2 + 2ab 3.
4. [F (α) : F ] = deg (p (x)) √
5. [F (α) : F ] ≤ deg g (x) where g (x) ∈ F [x] is any non-zero polynomial such Since 3 ∈ / Q, we would have to have ab = 0, i.e. a2 = 5 or 3b2 = 5 for some
that g (α) = 0. a, b ∈ Q which is not possible since x2 − 5 and 3x3 − 5 are both irreducible over
√

Q. Thus we may conclude that, x2 − 5 is the minimal polynomial over Q 3
Proof. 1. Since F [x] is a principle ideal domain and ker ϕ is an ideal, we √
for 5 and h √ √  √ i
know that ker ϕ = hp (x)i where p (x) may be taken to be the unique monic
3 = deg x2 − 5 = 2.


polynomial in ker ϕ = {f (x) ∈ F [x] : f (α) = 0} with minimal degree. If p (x) Q 3, 5 : Q
were not irreducible, it would have a non-trivial factorization, p (x) = h (x) k (x) Consequently, by Theorem 26.2,
where deg h (x) , deg k (x) < deg p (x) . Moreover p (α) = 0 would then imply
h √ √  i h √ √  √ i h √  i
h (α) = 0 or k (α) = 0. But then h (x) or k (x) would be in ker ϕ with degree Q 3, 5 : Q = Q 3, 5 : Q 3 Q 3 : Q = 2 · 2 = 4 (26.1)
less than p (x) which is impossible.
2. and 3. By the first isomorphism theorem we know that  √ √ √ √ √ √ √

and moreover 1, 3, 5, 3 5 = 15 is a basis for Q 3, 5 .
ϕ̄ : F [x] / hp (x)i → F [α] = {f (α) : f ∈ F [x]}
Proposition 26.12. If α ∈ C is transcendental, then ϕ : Q [x] → C defined
so defined above is an isomorphism of rings. Since p (x) is irreducible we know by ϕ (f (x)) = f (α) has trivial kernel and therefore Q [x] ∼
= Q [α] . Moreover,
hp (x)i is maximal and therefore F [x] / hp (x)i is a field. Therefore F [α] is a if Q (x) denotes the field of rational function with rational coefficients, then
field and we are justified in writing F (α) = F [α] . Lastly if f (x) ∈ F [x] , Q (α) ∼
= Q (x) .
then f (x) = k (x) p (x) + r (x) with deg r (x) < deg p (x) . Therefore, f (α) =
From now on we will be sticking with the algebraic numbers inside of C.d
k (α) p (α) + r (α) = r (α) and this shows that

F (α) = {f (α) : f ∈ F [x] with deg f (x) < deg p (x)} .

4. Since ϕ̄ is a field isomorphism it is also a vector space of F – isomorphism.

Therefore,

[F (α) : F ] = dimF (F [x] / hp (x)i)

= dim {f (x) ∈ F [x] : deg f (x) < deg p (x)} = deg p (x) .

5. This simply follows from item 4. since deg (p (x)) ≤ deg (g (x)) .
2
√ Since x − 3 is irreducible over Q. Therefore the minimal poly-
Example 26.10.
nomial for 3 is x2 − 3,
h √  i
3 : Q = deg x2 − 3 = 2,


Q
√
√
and the general element of Q 3 may be written as a + b 3 with a, b ∈ Q.

Page: 88 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

27
Lecture 27

27.1 Ruler and Compass Constructions

Let us identify the plane with the complex numbers. The type of question
we would like to address, given 0, 1 ∈ C (i.e. fix an origin in the plan and a
unit length) what complex numbers in the plane are constructible just using
a straight edge and a compass. Let us begin with some (hopefully) familiar
constructions from high school geometry.

Fig. 27.2. A perpendicular bisector.

Fig. 27.1. This figure should remind you of how to bisect a line segment ab to get
point c. Alternatively if you start with point c on the line l, this picture indicates how
to construct the line perpendicular to l which goes through c.

Using the abilities of this construction it now follows that given a line l and
a point p not in l we can construct the line l0 going through p which intersects
l at a right angle, see Figure 27.2. Fig. 27.3. Constructing parallel lines using an intermediary perpendicular line.
Making use of dropping perpendicular lines, given a line l and a point p not
in l we we can construct the line l0 which goes through p that is parallel to l as
in Figure 27.3.
Proof. The proofs of these assertions are encoded in the following pictures.
The other standard construction from high school geometry is the ability to
To construct the square root of r the reader is referred to Figure 27.7 below.
bisect an angle which is demonstrated in Figure 27.4.
Using the notation in this figure we learn by Pythagorean’s theorem that
We these constructions in hand we may now work out the numbers which
are constructible. We do this in the next few results. 2 2 2
( s + 1) = |ib − 1| + |ib − s| .
Lemma 27.1. Suppose that r, s ∈ (0, ∞) are given are representable by line Expanding our both sides of this equation gives,
segments of these lengths where we are given
√ a unit length segment as reference.
Then we can also represent; 1/r, rs, and s. s2 + 2 s + 1 = b2 + 1 + b2 + s2 = s2 + 2b2 + 1,
90 27 Lecture 27

Fig. 27.6. Given segments of length 1, r, and s here is how to construct a segment
of length rs.
Fig. 27.4. Bisecting a given angle.

Fig. 27.7. This is a semi-circle of radius (d + 1) /2.

√
which implies that b2 = s, i.e. b = s.

It should be obvious that given ways to measure r, s ∈ R we can construct

Fig. 27.5. Given segments of length 1 and r here is how to construct a segment of the point r + is ∈ C with a straight edge and compass, see Figure 27.8.
length 1/r.
Remark 27.2. Suppose E ⊂ C is a sub-field of C which is closed under complex
conjugation. Let us now look for a square root, x + iy, of a + ib ∈ E. Thus we
must have,
x2 − y 2 = a and 2xy = b.
Solving the last equation for y and substituting the result into the first gives
an equation for x, namely,

Page: 90 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

 27.1 Ruler and Compass Constructions 91

3. z̄ is representable.
4. zw is representable.
5. If z 6= 0, then 1/z is representable.
6. If z 6= 0, the two square-roots of z are representable.
Proof. Most of these are simple to realize.

Fig. 27.8. Constructing a complex number from its real and imaginary parts. 1. The real and imaginary parts are found by right angle projections of z onto
that x and y – axis which we know how to do.
2. It should be clear how to represent z + w and z − w when z and w are real.
 2 Thus we may do the complex case by considering the real and imaginary
2 b parts. (Better yet, use the usual graphical methods for vector addition which
x − = a ⇐⇒ 4x4 − 4ax2 − b2 = 0.
2x only need the ability to construct parallel lines going through given points.)
3. Similarly if z = x + iy then we can represent x and y and hence x and −y
We are looking for real roots of this equation. By the quadratic formula it
and therefore z̄ = x − iy. (Alternatively you may easily do this graphically
follows that √ √
4a ± 16a2 + 16b2 a ± a2 + b2 as well.)
2
x = = 4. Again this can be done by working with the real and imaginary parts of zw
2·4 2
and using that we know how to work with real lengths. Alternatively write
and we must choose the positive root here to learn, z = reiθ and w = ρeiα . Then the line going through ei(θ+α) as in Figure
√ 27.9 and mark the point on this line which is distance rρ from the origin.
2 a + a2 + b2 2
x = . 5. For this simply observe that z̄z = |z| from which it follows that
2
1 1
Thus we find that = z̄ 2
s √ z |z|
a+ a2 + b2
x=± which is representable by what we have already proved.
2
6. Case 1. Suppose that z = reiθ = r cos θ + ir sin θ with 0 ≤ θ ≤ π. To
and therefore that
construct the square roots of z, draw the line through 0 ∈ C which bisects
b b the angle θpbetween 0 1 and 0 z. Then mark the two points on this line
y=± q √ = ±p √ . which are |z| from the origin. These are the graphical representations of
2
a+ a +b2
2a + 2 a2 + b2
2 2 the two square roots of z.
Case 2. Suppose that z = reiθ = r cos θ + ir sin θ with 0 ≤ θ ≤ π. In this
√ q √
 
2 2
a2 + b2 , a+ a2 +b . case we may write z = re−i(2π−θ) and then work as in the previous example
√ 
Thus we see that x2 ∈ E a2 + b2 and that x, y ∈ E
to bisect the angle between 0 1 and 0 z and mark the points as before. In
3 doing this we will construct,
As an example if a + ib = 5 + 45 i then
√ √ √
q ± re−i(2π−θ)/2 = ± re−iπ+θ/2 = ∓ reiθ/2
3 3 2 4 2



5 + 5 + 5 4 2
x = 2
= and x = √ . which again represents the two square roots of z.
2 5 5
Lemma 27.3. Suppose that z, w ∈ C are given complex numbers which have Are results so far easily give the following proposition.
put on the complex plane along with 1. Then;
Proposition 27.4. If {0, 1} are plotted in the complex plane, we may construct
1. the real and imaginary parts of z are representable. √ ⊂ C using only a compass and a straight edge. Moreover we may
Q [i] = Q+iQ
2. z + w and z − w are representable. construct z for all z ∈ Q [i] .

Page: 91 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

92 27 Lecture 27

Better yet if n is even draw lines through ω k and ω̄ k to find the center as
the common intersection point while if n is odd draw lines through ω k and the
midpoint of the opposite side to find the center, see Figure 27.10.

Fig. 27.9. Multiplying complex numbers with a straight edge and a compass.
Fig. 27.10. Finding the center of a pertagon and a hexagon.

Proposition 27.5. Suppose that S ⊂ C is a set containing {0, 1} and F (S)

are the complex numbers which are constructible from S using compass and a
straight edge. Then S is a sub-field of C which contains Q. Moreover if z ∈ F (S)
then Re z, Im z, and z̄ are all back in F (S) .
Proof. From Lemma 27.3, F (S) is closed under subtraction, multiplication 27.2 Geometric Consequences for the Greeks
and complex conjugation. From this it follows that F (S) is a field which is
Theorem 27.7 (Theorem 9.24 of Knapp, p. 466). The set, C, of x coordi-
closed under complex conjugations. Since
nates that can be constructed from x = 0 and x = 1 by straightedge and compass
z + z̄ z − z̄ forms a subfield of R such that the square root of any positive element of the
Re z = and Im z =
2 2i field lies in the field. Conversely the members of C are those real numbers lying
it now follows that Re z, Im z ∈ F (S) . in some subfield Fn of R of the form,
√ √ √

Lemma 27.6. Suppose that D is a regular n – gon in C, then the center of D F1 = Q ( a0 ) , F2 = F1 ( a1 ) , . . . Fn = Fn−1 an−1 ,
may be determined by summing the coordinates of the vertices of the n – gon D
where aj ∈ Fj and a0 , . . . , an−1 ≥ 0.
and in particular the center of D may be found using only a ruler and compass.
Proof. Suppose that the vertices of D are described by xk := z + αω k for Proof. We have already seen that we can construct the elements described
k = 0, 1, 2 . . . , n − 1 where ω is a primitive nth – root of unity, α 6= 0 in C and in Fn by a ruler and straightedge so it only remains to see that we can not
z is the center we are trying to find. Then making use of the relation, construct anything else by these methods. To see this, observe that the only
points we may construct are gotten by intersection lines going through two
n−1 points already constructed with other such lines or with circles centered at
X ωn − 1
ωk = =0 (27.1) points with radii from points we have already constructed.
ω−1
k=0 So suppose that we have constructed a sub-field, E ⊂ Q using a ruler and
we see that straightedge. Then a line joining (a1 , b1 ) and (a2 , b2 ) is described by
n−1 n−1
1X 1X y − b1 b2 − b1
z + αω k = z.


xk = =
n n x − a1 a2 − a1
k=0 k=0

Page: 92 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

 27.2 Geometric Consequences for the Greeks 93

or equivalently of the form y = mx + b or x = my + b with m, b ∈ E. An allowed Example 27.9 (Doubling the cube). In this example we show that it is impossible
circle is of the form, to “double the cube” with a compass and straight edge. That is we can not use
2 2 a compass and straight edge only to construct a cube with volume
√ 2. To do this
(y − u) + (x − v) = c2 with u, v, c ∈ E. we would need to construct a cube with side length equal to 3 2 which have
1. Then the intersection of two lines; just seen is impossible.
Example 27.10 (Trisecting an angle). In this example we show that it is impos-
y = mx + b and y = m0 x + b0
sible to trisect a 60◦ angle. First off notice that
gives a linear equation for x of the form (m − m0 ) x + b − b0 = 0 which only √
cos 60 = 1/2 and sin 60 = 3/2
has solutions from E. So no new points are found this way.
2. If we intersect a line and a circle, we find an equation for x of the form, which are constructible and therefore so is 60◦ . In order to trisect this angle it
2 2
would have to be possible to construct a := cos (20◦ ) . To see this is not possible,
(mx + b − u) + (x − v) = c2 observe that
3
which is quadratic in x. If there is (real) solution to this equation it will lie cos 3θ = Re ei3θ = Re (cos θ + i sin θ) = cos3 θ − 3 cos θ sin2 θ


√

in E γ for some γ ∈ E with γ > 0.
= cos3 θ − 3 cos θ 1 − cos2 θ = 4 cos3 θ − 3 cos θ.


3. Now suppose we intersect two circles say,
2 2 2 2 Thus taking θ = 20◦ we learn that
(y − u1 ) + (x − v1 ) = c21 and (y − u2 ) + (x − v2 ) = c22
1
4α3 − 3α = cos 60 = ,
with all of the u’s, v’s and c’s coming from E. By translating the whole 2
picture by (u2 , v2 ) we may assume that u2 = v2 = 0. Thus it suffices to i.e. α is a root of f (x) = 8x3 − 6x − 1.
consider the intersection points of
2 2
(y − u) + (x − v) = c2 and y 2 + x2 = d2 with u, v, c, d ∈ E.
Since
2 2
c2 = (y − u) + (x − v) = u2 − 2vx − 2uy + v 2 + x2 + y 2
at the intersection points we must have
c2 = u2 − 2vx − 2uy + v 2 + d2
which is a linear equation for x. Thus we are back to case 2. namely the
intersection of a line and a circle and the proof is complete.
The graphs of f (x) .
But the mod 5 test shows this polynomial is irreducible since,
Corollary 27.8. Letting Fn be as in Theorem 27.7 we have [Fn : Q] = 2k for
some 0 ≤ k ≤ n. x 01234
.
√ f (x) mod 5 4 1 1 2 2
As an application of this corollary, let us
show that 3 p ∈ / Fn for any prime
√ √
p. The reason is that if 3 p ∈ Fn then Q 3 p ⊂ Fn and we must have Alternatively one may check that {±1, ±1/2, ±1/4, ±1/8} are not1 roots of
√ √ f (x) so that f (x) has no rational roots. No matter how this is done we learn
[Fn : Q] = [Fn : Q ( 3 p)] [Q ( 3 p) : Q] .
1
 √
 From the graph in Figure 27.10, we see that −1/4 and −1/8 are the most likely can-
However, x3
−p is irreducible over Q by Eisenstein and therefore Q 3 p : Q = didates to be a rational roots of f (x) – however f (−1/4) = 83 6= 0 and f (−1/8) =
√
deg x3 − p = 3. But 3 - 2k and so it is impossible for 3 p to be in Fn . − 17
64
6= 0.

Page: 93 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

94 27 Lecture 27

that [Q (α) : Q] = deg (f (x)) = 3 which does not divide 2k for any k and hence Theorem 27.13. It is possible to construct the regular n -gon with a straight-
is not constructible. edge and compass iff n = 2k p1 . . . pl , where k ≥ 0 and p1 , . . . , pl are and distinct
primes each of which is of the form, 2m + 1 for some m ≥ 2.
Example 27.11 (No 9 - gons). The angle 40◦ is not constructible. Let α :=
cos 40, then working as above For example when 5 = 22 +1 so pentagon’s are constructible as we have seen.
On the other hand 7 6= 2m + 1 and therefore the 7 - gon is not constructible.
1
4α3 − 3α = cos (3 · 40) = cos (120) = . Also 9 = 3 · 3 so the 9 -gon is not constructible since the odd prime, 3, appears
2 twice in the decomposition of 9. On the other hand, 17 = 24 + 1 is prime so the
Again this is not constructible by above. As a consequence it follows that a 17 - gon is constructible.
regular 9 – gon is not constructible. For if it were, then 360/9 = 40 would be a
constructible angle.

Example 27.12 (Yes to pentagons). In order to construct a pentagon we must

be able to construct θ := 2π/5 or equivalently α := cos (2π/5) . However, recall
from Eq. (27.1) that

1 + eiθ + e−iθ + ei2θ + e−i2θ = 0.

Taking the real part or this identity gives,

2 cos 2θ + 2 cos θ + 1 = 0.

Then using the double angle formula,

cos 2θ = cos2 θ − sin2 θ = 2 cos2 θ − 1

allows us to conclude that

2 2 cos2 θ − 1 + 2 cos θ + 1 = 0,

i.e. α is a root of f (x) := 4x2 + 2x − 1. This polynomial is irreducible since

x 012
.
f (x) mod 3 2 2 1

Alternatively use the quadratic formula to see the roots of f (x) are given by
√ √ √
−2 ± 16 + 4 −2 ± 2 5 −1 ± 5
= =
8 8 4
√

so that f (x) has no rational roots. Thus we conclude that Q (α) = Q 5 so
that α is constructible. Actually we have one the game when we found that α
was a root of some f (x) ∈ Q [x] with deg f (x) = 2.

The general result along these lines (see Gallian [2, Theorem 33.5, p. 570])
is the following theorem due to Guass, 1796.

Page: 94 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

28
Lecture 28

28.1 Splitting fields over Q Example 28.6. According to Theorem 28.4, f (x) = x3 + 1 ∈ Q [x] must split
over C – but how? It is not necessarily easy to find the factors in general!
Definition 28.1. Given a field F ⊆ E and a polynomial f ∈ F [x], we say that However in this case we see that −1 is a root of f and by the division algorithm
f splits over E if f factors into irreducibles as that x3 + 1 = (x + 1)(x2 − x + 1). The roots of x2 − x + 1 are
√ √
f (x) = c(x − a1 )(x − a2 ) . . . (x − an ) 1± 1−4 1 3i
= ±
2 2 2
for some ai ∈ E and c ∈ F × .
and thus we may conclude that (over C) f (x) factorizes as;
Example 28.2. f (x) = x2 − 2 does not split over Q but does split over R;
√ ! √ !
√ √ 1 3i 1 3i
x2 − 2 = (x − 2)(x + 2). f (x) = (x + 1) x − ( + ) x−( − ) .
2 2 2 2
Theorem 28.3. Given any polynomial f ∈ F [x] for some field F , there exists
The splitting field of f (x) over Q is then
a field extension F ⊆ E such that f splits over E.
√ √ !
We’ll mostly stick to the case F = Q. In this case to find a field over which a 1 3i 1 3i
K = Q −1, − , + .
polynomial splits we may make use Theorem 19.12 – the fundamental theorem 2 2 2 2
of algebra which we restate here.
We can simplify our expression for K and in the process get a better idea
Theorem 28.4 (Fundamental theorem of algebra). Every polynomial f ∈ what is in it. Since −1 ∈ Q already, the inclusion in the list defining K is
√ 
C[x] splits over C. 1
irrelevant. Also observe that 2 + 2 3i
∈ K and 12 ∈ Q ⊂ K allows us to
Now suppose f (x) ∈ Q[x]. We know f (x) splits over C, but there is a big conclude that √ √ !
distance between Q and C, and sometimes we want to change our fields only a √ 3 1 3i 1
−3 = 2 i=2 + − ∈K
little to a find a field in which f (x) splits. 2 2 2 2
√
Definition 28.5. Let F ⊂ E be a field extension, and f (x) ∈ F [x]. Suppose and therefore, Q( −3) ⊆ K. Conversely,
f (x) splits over E. The splitting field for f (x) is the smallest subfield K with √ √
F ⊆ K ⊆ E such that f (x) splits over K. Explicitly, if f (x) = (x − a1 ) . . . (x − 1 3i 1 3i √ √
−1, + , − ∈ Q( −3) =⇒ K ⊂ Q( −3)
an ) with a1 , . . . , an ∈ E, then K = F (a1 , . . . , an ) is the smallest subfield of E 2 2 2 2
containing F and a1 , . . . an .
and hence we may conclude that
Mostly we will use this definition in the case F = Q, E = C. Then for any √ √  √
K = Q( −3) = Q −3 = {a + b −3 : a, b ∈ Q}.
polynomial f (x) ∈ Q[x], f (x) splits over C, so f (x) has a splitting field inside
C. √
−1
Let us recall how to compute a + b −3 ;
96 28 Lecture 28
√ √ √

√
−1 1 a − b −3 Lemma 28.10. Suppose that p and q are distinct primes. Then p ∈
/Q q .
a + b −3 = √ · √
a + b −3 a − b −3 √ √
√ √
√ Proof. If p ∈ Q q then p = a + b q for some a, b ∈ Q. Squaring this
a − b −3 a b √
= 2 = 2 − 2 −3 equation would then imply that
a + 3b2 a + 3b 2 a + 3b 2
√
√ p = a2 + b2 q + 2ab q. (28.1)
which is well defined provided that a + b −3 6= 0. Let us summarize what we
have proved. √
Since q is irrational, we must have ab = 0, i.e. a = 0 or b = 0. So in order to
√
√  solve Eq. (28.1), we must solve one of the equations, qb2 = p or a2 = p with
Lemma 28.7. The splitting field for x3 + 1 ∈ Q [x] is Q −3 = Q −3 .
a, b ∈ Q. But it is easy to check that both qx2 − p and x2 − p are irreducible of
This is also the splitting field for x2 − x + 1 and for x2 + 3.
Q so that these equations have not solutions and hence there is no solution to
√ √

Example 28.8. In this example we wish to find the splitting field for x4 − 3 ∈ Eq. (28.1), i.e. p ∈ / Q q . To see that qx2 − p and x2 − p are irreducible of Q
Q [x] . Let ω := ei2π/4 = eiπ/2 = i be a primitive 4th root of unity. Then the observe that polynomials are primitive and therefore we may apply Eisenstein
roots of x4 − 3 are criterion using the prime, p.
√ √
√ √ n √ √ Example √ 28.11. √
Consider Q( 3, 5). This is the smallest subfield of C contain-
o
4
3 · 1, ω, ω 2 , ω 3 = 3 · {±1, ±i} = · ± 3, ±i 3 .
4 4 4
√ √ √ √

ing Q, 3, and 5. We claim that Q( 3, 5) = Q( 3 + 5)
√ √ √ √ √ √
Thus the splitting field for f (x) is
√ Since √3 +√ 5 is in Q( 3, 5), and Q ⊂ Q( 3, 5), it follows that
√Proof.
 √ √  √  Q( 3 + 5)√⊆ Q(√ 3, 5). The harder part√is to√prove the reverse √ √inclusion.
4 4 4
Q ± 3, ±i 3 = Q 3, i . Since Q( 3 + 5) is a field, we know ( 3 + 5)−1 ∈ Q( 3 + 5). Since
√
√ √ √ √
Since x2 + 1 is irreducible1 over Q 4 3 and x4 − 3 is irreducible (Eisenstein’s 1 1 3− 5 3− 5 1√ 1√
√ √ =√ √ √ √ = =− 3+ 5
criteria for example) over Q, it follows that 3+ 5 3+ 5 3− 5 −2 2 2
h √  i h √   √ i h  √  √ √ √ √
it follows that − 12 3 + 21 5 ∈ Q( 3 + 5). Therefore,
i
4 4 4 4
Q 3, i : Q = Q 3, i : Q 3 Q 3 :Q
√ √ 1√ 1√ √ √
 
= deg x2 + 1 · deg x4 − 3 = 2 · 4 = 8.



3 − 5 = −2 − 3+ 5 ∈ Q( 3 + 5)
√ 2 2
4


A basis for Q 3, i over Q is
and hence √ √ √ √  √ √
n o n √ √ 3 √ √ 3
o 3+ 5± 3 − 5 ∈ Q( 3 + 5)
{1, i} · 1, 31/4 , 31/2 , 33/4 = 1, 3, 3, 3 4 , i, i 3, i 3, i3 4 .
4 4

√ √ √ √
from
√ which
√ it follows
√ √that 3, 5 ∈ Q( 3 + 5). Thus we have shown
Q( 3, 5) ⊆ Q( 3 + 5) which completes the proof.
28.2 More practice on understanding field extensions of Q √
Example 28.12. −3 has minimal polynomial x2 + 3 over Q. Why? It is monic,
√
Theorem 28.9 (Primitive element theorem. Steinitz, 1910). If f (x) ∈ −3 is a root, and it is irreducible over Q, that is, has no roots in Q.
Q[x] and Q ⊆ K ⊆ C is the splitting field for f (x), then K = Q(a) for some Finding minimal polynomials can be challenging. Here is an example.
single element a ∈ C. √ √
Example 28.13. Let us find the minimal polynomial for α := 3 + 5 over Q.
Won’t prove (but see Gallian[2, Theorem 21.6 on p.375]). But in certain We first look for a monic polynomial, p (x) ∈ Q [x] such that p (α) = 0. We
examples, we can see how to find a single element by guessing. begin by observing that
√
1
It is irreducible over R and hence over Q 4
3 since x2 + 1 has not real roots.

√ √ √
α2 = ( 3 + 5)2 = 8 + 2 15

Page: 96 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

 28.2 More practice on understanding field extensions of Q 97
√ √ √
√
and therefore, Example 28.15. Let us now consider Q 3, 5, 7 . I claim that 7 ∈ /

2  √ 2 √ √
√
√
α2 − 8 = 2 15 = 4 · 15 = 60. (28.2) Q 3, 5 . To see this let x, y ∈ Q 3 so that x + y 5 is the general
√ √
√ √
element of Q 3, 5 . Thus we would have to have 7 = x + y 5 for some
Thus we see that √

x, y ∈ Q 3 . Squaring this equation then implies that

2
p (x) := x2 − 8 − 60 = x4 − 16x2 + 4 (28.3) √
7 = x2 + 5y 2 + 2xy 5.
is a polynomial having α as a root. √ √

Since we have already see that 5 ∈ / Q 3 , it will only be possible to solve
We claim this polynomial is the minimal polynomial for α. One way to see
this equation if xy = 0, i.e. either x = 0 √
or
y = 0. Thus we must either solve
this is that
x2 = 7 or 5y 2 = 7 for some x or y in Q 3 . or equivalently. But working as
h √ √  √ i h √  i √

[Q [α] : Q] = Q 3, 5 : Q 3 Q 3 : Q = 2 · 2 = 4. in the previous example, these equations have no solutions over Q 3 and we
are done. From these observations we may conclude that
√
h √ √ √ 
Notice that k (x) := x2 − 5 ∈ Q 3 [x] is an irreducible polynomial. The i
√ √
√
Q 3, 5, 7 : Q = 23 = 8.
point is that 5 ∈/ Q 3 and so x2 − 5 has no roots in Q 3 . Indeed, if
√ √ √

5 = a + b 3 ∈ Q 3 then it would follow that
Example 28.16. What is the minimal polynomial of R(i)? Answer; x2 + 1.
 √ 2 √  √ √

5 = a + b 3 = a2 + 3b2 + 2ab 3. Example 28.17. In this example we wish to compute n := Q 3 2, 4 3 : Q .
Using
√
Since 3 ∈ / Q, we would have to have ab = 0, i.e. a2 = 5 or 3b2 = 5 for some h √ √ 
3 4
i h √ √ 
3 4
 √ i h  √ 
3 3
i
a, b ∈ Q which is not possible 2 3 Q 2, 3 : Q = Q 2, 3 : Q 2 Q 2 :Q
 √since √
x − 5√and
 3x − 5 are both irreducible h √ √   √ i
over Q. This shows that Q 3, 5 : Q 3 = 2. Therefore the minimal 3 4 3
= Q 2, 3 : Q 2 ·3
polynomial for α must be fourth order and vanish on α. As p (x) in Eq. (28.3)
satisfies these criteria and is monic it must be the minimal polynomial. and
h √ √  i h √ √   √ i h  √  i
3 4 3 4 4 4
Remark 28.14. Another way to show that p (x) in Eq. (28.3) is the minimal Q 2, 3 : Q = Q 2, 3 : Q 3 Q 3 :Q
polynomial such that p (α) = 0 would be to show that p (x) is irreducible over h √ √ 
3 4
 √ i
4
Q. One way to do this is to first find all of the roots of p (x) . From Eq. (28.3), = Q 2, 3 : Q 3 ·4

2
p (a) = 0 for some a ∈ C then a2 − 8 = 60 and therefore,
it
 √ follows
√
that√

3|n and 4|n √ and therefore that 12|n.
√
Since
√ √
2
a = 8 ± 60 = 8 ± 2 15. Q 3 2, 4 3 : Q 3 2 ≤ 4 as 4 3 is a root of x4 − 4 ∈ Q 3 2 [x] we
also know that n ≤ 12. Therefore n = 12.
n p √ o
So we may conclude that a (ε, δ) = ε 8 + δ2 15 are all the roots of
ε,δ=±1
p (x) . To see that p (x) is irreducible it suffices to observe that p (x) can not
be factored into a product of quadratic polynomials since a (ε, δ) a (ε0 , δ0 ) ∈
/Q
whenever (ε, δ) 6= (ε0 , δ0 ) .
Note: We could also try to use the mod p test here. However the test fails
for p = 2 and p = 3. For example, if p̄ (x) = p (x) mod 3, then

2
p̄ (x) = x4 − x2 + 1 = x4 + 2x2 + 1 = x2 + 1

is reducible. So using a mod p test is becoming rather painful.

Page: 97 job: 103bs macro: svmonob.cls date/time: 29-May-2009/15:15

29
Lecture 29 (Review Day)

The Final Exam Questions will be similar to Quiz Questions. domains and which aren’t? Which are fields? Which are commutative and which
are noncommutative? Which have an identity element and what is it? What is
the characteristic of each ring?
29.1 Definitions you should know:
• Rings of numbers: Z, Q, R, C.
• Ring (but I will not ask you to check that something is or is not a ring • Zm , the integers modulo m, for any m ≥ 2.
directly from the definition). • Matrix rings: M2 (F ),which is 2 × 2 - matrices with entries from F . Here F
• Identity element of a ring. could be any of the rings of numbers above, or even Zm for√ some m.
• Commutative ring. • The Gaussian integers Z[i] = {a+b : a, b ∈ Z}, where i = −1. Also studied
• Unit. Zm [i] a bit.
• Subring. • The ring of polynomials F [x] , which consists of all elements of the form
• Zero-divisor and the cancellation property.
• Integral domain axn + an−1 xn−1 + · · · + a1 x + a0 ,
• Field.
where the coefficients ai all come from F . Here F could be any of the rings
• Characteristic of a ring. (I only care about characteristic for rings R with
of numbers √above, or even Zm for some m.
identity, and I define it to be the smallest positive integer n such that n · 1 =
• The ring Z[ m], where m is a positive integer which is not a square. This
0, or if no such n exists the characteristic is defined to be 0.)
ring consists of all elements
• Ideal.
• Factor ring. √
{a + b m : a, v ∈ Z}.
• Prime ideal.
• Maximal ideal. • Given any two rings R and S, the direct sum of R and S is a new ring
• Homomorphism and isomorphism.
• Kernel and image of a homomorphism. R ⊕ S = {(r, s) : r ∈ R and s ∈ S},
• Polynomial rings, R [x] and most importantly F [x] where F is a field.
• Principle ideal domain. with component-wise addition and multiplication.
• Unique factorization domain. • F [α] := {p (α) : p ∈ F [x]} where α ∈ E and E is some field extension of F.
√

• F (α) where F is a sub-field of a field E and α ∈ E, e.g. Q 4 5 .
• [E : F ] = dimF (E) .
• Algebraic and transcendental elements of an extension field. 29.3 Important theorems and techniques
• Minimal polynomial.
• Know how to check if a subset of a ring is a subring.
• Know how to check if a subset of a ring is an ideal of the ring.
29.2 Examples of Rings • Know the theorem that a finite integral domain with identity is a field.
• Know that Zm is a field precisely when m is prime, and understand why
We have only studied a few classes of rings. You should know all of these and this fails when m is not prime.
√ 
their basic properties. How do you multiply and add in each one? Which are • Understand the example Q 2 and understand the proof that it is a field.
• Know the theorem that the characteristic of a domain is a prime number
(or 0).
• Given a commutative ring R with element a, know the definition of the
principle ideal generated by a, written as hai .
• Understand the definition of a factor ring and how to do addition and mul-
tiplication in such a ring.
• Understand some important examples where factor rings can be shown to
be the same as other familiar rings. Z/ hmi ∼= Zm . R [x] / hxi ∼
= R. There are
also various problems where one looks at factor rings of Z [i]. For example,
we showed Z [i] / h2 − ii ∼
= Z5 .
• Know the theorem that a ideal I of a commutative ring R is prime if and
only if R/I is a domain, and that I is maximal if and only if R/I is a field.
As a corollary, know that a commutative ring R with identity is a field if
and only if R and {0} are the only ideals of R.
• Be able to check if a function between two rings is a homomorphism, and if
it is a isomorphism.
• Know that the kernel of a homomorphism ϕ : R → S is always an ideal of
R , and the image of a homomorphism is always a subring of S. Know the
statement of the 1st – isomorphism theorem: R/ ker (ϕ) ∼ = Im ϕ and how to
use it.
• Know that given any ring R with identity, there is a homomorphism Z → R
sending a to a · 1. The kernel of this homomorphism is exactly hmi , where
m is the characteristic of R.
• Know how to determine the homomorphisms, ϕ : Z [i] → R such that ϕ (1) =
1 – recall i must be sent to t ∈ R such that t2 = −1R .
• Understand how to manipulate polynomials and how to carry out long di-
vision in F [x] where F is a field like Q or Zp where p is prime.
• Understand how to find the roots of polynomials and their relationship to
linear factors.
• Know how to check if a polynomial in Q [x] is irreducible or not. See Section
22.2 for a summary of this point.
• Know the relationships between prime (maximal) principle ideals and prime
and maximal elements of an integral domain. See Theorem 23.6 for a sum-
mary.
• If F ⊂ E ⊂ G are fields then [G : F ] = [G : E] [E : F ] .
• F (α) ⊂ E is the smallest subfield of E containing F and α.
• Recall that if α is algebraic over F , then F (α) = F [α] and [F (α) : F ] =
deg f (x) where f (x) ∈ F [x] is the minimal polynomial over F of α. See
Theorem 26.9 for details.
References

1. Richard A. Dean, Classical abstract algebra, Harper & Row, Publishers, Inc., New
York, 1990.
2. Joseph A. Gallian, Contemporary abstract algebra, Houghton Mifflin, Boston, MA,
2002.
3. Anthony W. Knapp, Basic algebra, Cornerstones, Birkhäuser Boston Inc., Boston,
MA, 2006, Along with a companion volume ıt Advanced algebra. MR MR2257570
(2007e:00001)
4. Joachim von zur Gathen and Jürgen Gerhard, Modern computer algebra, second
ed., Cambridge University Press, Cambridge, 2003. MR MR2001757 (2004g:68202)