Lecture Notes for Abstract Algebra I James S. Cook Liberty University Department of Mathematics Fall 2016 2 preface Abstract algebra is a relatively modern topic in mathematics. In fact, when I took this course it was called Modern Algebra. I used the fourth ed. of Contemporary Abstract Algebra by Joseph Gallian. It happened that my double major in Physics kept me away from the lecture time for the course. IlearnedthissubjectfirstfromreadingGallian’stext. Inmyexperience, itwasanexcellent and efficient method to initiate the study of abstract algebra. Now, the point of this story is not that I want you to skip class and just read Gallian. I will emphasize things in a rather different way, but, certainly reading Gallian gives you a second and lucid narrative to gather your thoughts on this fascinating topic. I provide these notes to gather ideas from Gallian and to add my own. sources I should confess, I have borrowed many ideas from: 1. Contemporary Abstract Algebra by Joseph Gallian 2. the excellent lectures given by Professor Gross of Harvard based loosely on Artin’s Algebra 3. Dummit and Foote’s Abstract Algebra 4. Fraleigh 5. Rotman style guide I use a few standard conventions throughout these notes. They were prepared with LATEX which automatically numbers sections and the hyperref package provides links within the pdf copy from the Table of Contents as well as other references made within the body of the text. I use color and some boxes to set apart some points for convenient reference. In particular, 1. definitions are in green. 2. remarks are in red. 3. theorems, propositions, lemmas and corollaries are in blue. 4. proofs start with a Proof: and are concluded with a (cid:3). However, I do make some definitions within the body of the text. As a rule, I try to put what I am defining in bold. Doubtless, I have failed to live up to my legalism somewhere. If you keep a listofthesetransgressionstogivemeattheendofthecourseitwouldbeworthwhileforallinvolved. The symbol (cid:3) indicates that a proof is complete. The symbol (cid:79) indicates part of a proof is done, but it continues. As I add sections, the Table of Contents will get longer and eventually change the page numbering of the later content in terms of the pdf. When I refer to page number, it will be the document numbering, not the pdf numbering. Contents 1 Group Theory 1 1.1 Lecture 1: an origin story: groups, rings and fields . . . . . . . . . . . . . . . . . . . 1 1.2 Lecture 2: on orders and subgroups . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 1.3 Lecture 3: on the dihedral group and symmetries . . . . . . . . . . . . . . . . . . . . 13 1.4 Lecture 4: back to Z-number theory . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 1.4.1 Z-Basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 1.4.2 division algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 1.4.3 divisibility in Z . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 1.5 Lecture 5: modular arithmetic and groups . . . . . . . . . . . . . . . . . . . . . . . . 27 1.6 Lecture 6: cyclic groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 1.7 Lecture 7: classification of cyclic subgroups . . . . . . . . . . . . . . . . . . . . . . . 46 1.8 Lecture 8: permutations and cycle notation . . . . . . . . . . . . . . . . . . . . . . . 50 1.9 Lecture 9: theory of permutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 2 On the Structure of Groups 61 2.1 Lecture 10: homomorphism and isomorphism . . . . . . . . . . . . . . . . . . . . . . 62 2.2 Lecture 11: isomorphism preserves structure . . . . . . . . . . . . . . . . . . . . . . . 67 2.3 Lecture 12: cosets and Lagrange’s Theorem . . . . . . . . . . . . . . . . . . . . . . . 73 2.4 Lecture 13: on dividing and multiplying groups . . . . . . . . . . . . . . . . . . . . . 78 2.5 Lecture 14: on the first isomorphism theorem . . . . . . . . . . . . . . . . . . . . . . 83 2.5.1 classification of groups up to order 7 . . . . . . . . . . . . . . . . . . . . . . . 83 2.5.2 a discussion of normal subgroups . . . . . . . . . . . . . . . . . . . . . . . . . 84 2.5.3 first isomorphism theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 2.6 Lecture 15: first isomorphism theorem, again! . . . . . . . . . . . . . . . . . . . . . . 89 2.7 Lecture 16: direct products inside and outside . . . . . . . . . . . . . . . . . . . . . . 93 2.7.1 classification of finite abelian groups . . . . . . . . . . . . . . . . . . . . . . . 97 2.8 Lecture 17: a little number theory and encryption . . . . . . . . . . . . . . . . . . . 99 2.8.1 encryption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 2.9 Lecture 18: group actions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 2.10 Lecture 19: orbit stabilizer theorem and conjugacy . . . . . . . . . . . . . . . . . . . 110 2.11 Lecture 20: Cauchy and Sylow theorems . . . . . . . . . . . . . . . . . . . . . . . . . 113 2.12 Lecture 21: lattice theorem, finite simple groups . . . . . . . . . . . . . . . . . . . . 116 2.13 Lecture 22: Boolean group, rank nullity theorem . . . . . . . . . . . . . . . . . . . . 116 3 Introduction to Rings and Fields 117 3.1 Lecture 23: rings and integral domains . . . . . . . . . . . . . . . . . . . . . . . . . . 118 3.2 Lecture 24: ideals and factor rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 3 4 CONTENTS 3.3 Lecture 25: ring homomorphism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 3.4 Lecture 26: polynomials in an indeterminant . . . . . . . . . . . . . . . . . . . . . . 140 3.5 Lecture 27: factorization of polynomials . . . . . . . . . . . . . . . . . . . . . . . . . 146 3.6 Lecture 28: divisibility in integral domains I . . . . . . . . . . . . . . . . . . . . . . . 153 3.7 Lecture 29: divisibility in integral domains II . . . . . . . . . . . . . . . . . . . . . . 157 3.8 Lecture 30: extension fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 3.9 Lecture 31: algebraic extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170 3.10 Lecture 32: algebraically closed fields . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 Chapter 1 Group Theory The organization of these notes loosely follows Gallian. For the most part I include every theorem which Gallian includes. However, I include some extra examples and background. I lack interesting quotes12 1.1 Lecture 1: an origin story: groups, rings and fields In a different notation, but with the same essential idea, the fact that solutions to ax2+bx+c = 0 √ are given by x = −b± b2−4ac has been known for millenia. In contrast, the formula for solutions of 2a thecubicequationax3+bx2+cx+disonlyaboutahalf-milleniaold. DelFerrosolvethecubic3 circa 1500, Tartaglia solved it around 1530 then it was published by Cardano in his Ars Magna in 1545. Cardano’s student Ferrari solved quartic4 and that can also be found in the Ars Magna. Nearly the same tricks give closed form equations for the cubic and quartic. Euler, Lagrange and other 18th century mathematicians knew volumes about how to factor and solve polynomial equations. It seemed it was just a matter of time to find a formula for the solution of ax5+bx4+cx3+dx2+ex+f = 0. But, after a great effort by Lagrange there was no formula forthcoming. Moreover, it began to be clear that such a formula would be impossible due to the structure of Lagrange’s study. At the dawn of the nineteenth century Ruffini gave the first (incomplete in 1799 and again in 1813) proofs that there could not exist a general quintic formula. Abel, at the age of 19, gave a complete proof of the non-existence of the quintic formula in 1821. In 1831 a young Frenchman named Evariste Galois found a way to explain when it was possible to find the solutions to a 5-th order polynomial equation (for example, x5 − 1 = 0 is easy to solve). Galois insight was to identify the patterns in Lagrange’s work which involved permutations of the roots of the equation. In retrospect, this was the birth of Group Theory. In short, Galois said there was a nice solution to a quintic if the Galois group is solvable. If a group is simple5 then it cannot be broken down further, they’re sort of atomic6. So, in particular, if you show the Galois group of a polynomial is simple then, game-over, 1I make up for these with odd footnotes. 2for example, this or this. No Rickroll, I promise. 3forgive me if I don’t reproduce the formula here. See this for example 4this is quite a formula, it takes about a page, for example see this 5we later define simple and solvable groups, the details are not too important for our current discussion. 6moreabitlateronhowthetermatombreaksdown: Neutrons, Protons, electronsthenontoquarksandsuch... 1 2 CHAPTER 1. GROUP THEORY no solution7. This helps you understand why mathematicians were so happy we finally8 classified all finite simple groups in 20049. To give a specific example of Galois’ Theory’s power, 3x5−15x+5 = 0 is not solvable by radicals. Gallian gives the group theoretic argument on why that is on page 559 of our text. Interestingly, Galois’ contribution was not recognized until several decades af- ter his death. In 1846 Lioville understood the importance of Galois’ work and began to promote Galois’ group concept. By 1870, Jordan10 understood Galois’ well-enough to write a text on it. That said, I don’t have much more to say about Galois theor in this course. It is interesting, pow- erful, andmotivationaltothestudyofgrouptheory. But, ourfocusisonmoreelementarymaterial. Initially,groupswereallaboutpermutations,but,asthestorycontinuesmathematiciansdiscovered the structure of a group was not unique to permutations. For example, the symmetry groups promoted by Klein and Lie in the late nineteenth century. Thinking of groups abstractly came a bit later. Gallian credits this to Dyck and Weber circa 1883. Dyck, a student of Klein, emphasized the importance of invertibility in a paper about Tesselations11. Let pause our historical tour to examine the definition of a group and a few elementary examples. Definition 1.1.1. A set G with an operation12 (cid:63) : G×G → G forms a group if (i.) Associativity: (a(cid:63)b)(cid:63)c = a(cid:63)(b(cid:63)c) for all a,b,c ∈ G, (ii.) Identity: there exists e ∈ G such that a(cid:63)e = e(cid:63)a = a for each a ∈ G, (iii.) Invertibility: for each g ∈ G there exists h ∈ G such that h(cid:63)g = g(cid:63)h = e. If a(cid:63)b = b(cid:63)a for all a,b ∈ G then we say G is an abelian or commutative group. If there exist a,b ∈ G for which a(cid:63)b (cid:54)= b(cid:63)a then G is a non-abelian group. Thenotation(cid:63)isnottypicallyusedaswestudyspecificexamples. Infact,todenotea(cid:63)bwetypically use either juxtaposition (ab) or in the case of an abelian group we use additive notation (a+b). It is customary to only use + for a commutative operation. Example 1.1.2. Let G = Z is a group under addition with identity 0: In particular, we know for a,b,c ∈ Z there exists −a ∈ Z and 0 ∈ Z for which: (i.) (a+b)+c = a+(b+c), (ii.) a+0 = a = 0+a, (iii.) a+(−a) = 0 = (−a)+a. Moreover, we know whenever a,b ∈ Z the sum a+b ∈ Z. 7ok, to be precise, no closed-form solution in terms of radicals and such, a fifth order polynomial with real coefficients has a zero by the intermediate value theorem. But, the existence of such a zero is not the same as the existence of a nice formula for the zero 8 in 2004, Aschbacher and Smith published a 1221-page proof for the missing quasithin case 9we wont get to that in this course, its about 10,000 pages, including for example the paper of Feit-Thompson which alone is 250 pages, but, I will loosely cover the appropriate section later in Gallian in due time 10of the Jordan form, yes, sorry bad memories for my Math 321 class 11see Dr. Runion’s office for an example 12thisnotationindicatesthat(cid:63)isafunctionfromG×GtoG. Inotherwords,(cid:63)isabinary operation. Thisis sometimes identified as an axiom of a group known as closure. 1.1. LECTURE 1: AN ORIGIN STORY: GROUPS, RINGS AND FIELDS 3 You might wonder how we know such properties hold for Z. To be precise, we could build the integers from scratch using set-theory, but, to properly understand that construction it more or less begs an understanding of this course. Consequently, we will be content13 to use Z,C,R and Q as known objects complete with their standard properties. That said, as our understanding of abstract algebra increases we will begin to demonstrate how these standard number systems can be constructed. Example 1.1.3. Actually, this is a pair of non-examples. First, Z with subtraction is not a group. Second, Z with multiplication is not a group. why ? The next example is a bit meta. Example 1.1.4. Let V be a vector space then V,+ where + denoted vector addition forms a group where the identity element is the zero vector 0. The definition of a vector space includes the assumption (x+y)+z = x+(y+z) for all x,y,z ∈ V hence Axiom (i.) holds true. Axiom (ii.) is satisfied since x+0 = 0+x = 0 for each x ∈ V. Finally, Axiom (iii.) for each x ∈ V there exists −x ∈ V such that x+(−x) = 0. In summary, any vector space is also an abelian group where the operation is understood to be vector addition14 I should pause to note, the examples considered thus far are not the sort of interesting examples which motivated and caused mathematicians to coin the term group. These examples are just easy and make for short discussion. Let me add a few more to our list: Example 1.1.5. Let Q× = Q−{0} denote the set of nonzero rational numbers. Q× forms a group with respect to multiplication. The identity element is 1. Example 1.1.6. Let R× = R−{0} denote the set of nonzero real numbers. R× forms a group with respect to multiplication. The identity element is 1. Example 1.1.7. Let C× = C−{0} denote the set of nonzero complex numbers. C× forms a group with respect to multiplication. The identity element is 1. Example 1.1.8. Let Z× = Z−{0} denote the set of nonzero integers. Z× does not form a group since 2x = 1 has solution x = 1/2 ∈/ Z. Let me give at least one interesting explicit example in this section. This group is closely tied to invertible linear transformations on Rn: Example 1.1.9. Let GL(n,R) = {A ∈ Rn×n | det(A) (cid:54)= 0}. We call GL(n,R) the general linear group of n×n matrices over R. We can verify GL(n,R) paired with matrix multiplication forms a nonabelian group. Notice, matrix multiplication is associative; (AB)C = A(BC) for all A,B,C ∈ GL(n,R). Also, the identity matrix I defined15 by I = δ has AI = A = IA for each ij ij A ∈ GL(n,R). It remains to check closure of multiplication and inversion. Both of these questions are nicely resolved by the theory of determinants: if A,B ∈ GL(n,R) then det(AB) = det(A)det(B) (cid:54)= 0 13in an intuitive sense, numbers exist independent of their particular construction, so, not much is lost here. For example, I can construct C using vectors in the plane, particular 2×2 matrices, or via equivalence classes of polynomials. Any of these three could reasonably be called C 14 Of course, there is more structure to a vector space, but, I leave that for another time and place. 15δ is one of my favorite things. This Kronecker delta is zero when i(cid:54)=j and is one when i=j. ij 4 CHAPTER 1. GROUP THEORY thus AB ∈ GL(n,R) hence we find matrix multiplication forms a binary operation on GL(n,R). Fi- nally, we know det(A) (cid:54)= 0 implies there exists A−1 for which AA−1 = I = A−1A and det(AA−1) = det(A)det(A−1) = det(I) = 1 thus det(A−1) = 1/det(A) (cid:54)= 0. Therefore, we find A ∈ GL(n,R) implies A−1 ∈ GL(n,R) The previous example is more in line with Klein and Lie’s investigations of transformation groups. Many of those groups will appear as subgroups16 of the example above. At this point I owe you a few basic theorems about groups. Theorem 1.1.10. In a group G there can be only one identity element. Proof: let G be a group with operation (cid:63). Suppose e and e(cid:48) are identity elements in G. We have (i.) e(cid:63)a = a = a(cid:63)e and (ii.) e(cid:48)(cid:63)a = a = a(cid:63)e(cid:48) for each a ∈ G. Thus, by (i.) with a = e(cid:48) and (ii.) with a = e, e(cid:63)e(cid:48) = e(cid:48) = e(cid:48)(cid:63)e & e(cid:48)(cid:63)e = e. We observe e(cid:48)(cid:63)e = e(cid:48) = e. In summary, the identity in a group is unique. (cid:3) An examination of the proof above reveals that the axiom of associativity was not required for the uniqueness of the identity. As a point of trivia, a group without the associativity axiom is a loop. Here is a table17 with other popular terms for various weakenings of the group axioms: Relax, I only expect you to know the definition of group for the time being18. Theorem 1.1.11. Cancellation Laws: In a group G right and left cancellation laws hold. In particular, ba = ca implies b = c and ab = ac implies b = c. Proof: let G be a group with operation denoted by juxtaposition. Suppose a,b,c ∈ G and ba = ca. Since G is a group, there exists a−1 ∈ G for which aa−1 = e where e is the identity. Multiply ba = ca by a−1 to obtain baa−1 = caa−1 hence be = ce and we conclude b = c. Likewise, if ab = ac 16ask yourself about this next lecture 17I borrowed this from the fun article on groups at Wikipedia 18As my adviser would say, I include the table above for the most elusive creature, the interested reader 1.1. LECTURE 1: AN ORIGIN STORY: GROUPS, RINGS AND FIELDS 5 then a−1ab = a−1ac hence eb = ec and we find b = c. (cid:3) Cancellation is nice. Perhaps this is also a nice way to see certain operations cannot be group multiplications. For example, the cross product in R3 does not support the cancellation property. For those who have taken multivariate calculus, quick question, which group axioms fail for the cross product? Theorem 1.1.12. Let G be a group with identity e. For each g ∈ G there exists a unique element h for which gh = e = hg. Proof: let G be a group with identity e. Suppose g ∈ G and h,h(cid:48) ∈ G such that gh = e = hg & gh(cid:48) = e = h(cid:48)g In particular, we have gh = gh(cid:48) thus h = h(cid:48) by the cancellation law. (cid:3) At this point, I return to our historical overview of abstract algebra19 Returning to Lagrange and √ Euler once more, they also played some with algebraic integers which were things like a + b n in order to attack certain questions in number theory. Gauss instead used modular arithmetic in his master work Disquisitiones Arithmeticae (1801) to attack many of the same questions. Gauss √ also used numbers20 of the form a + b −1 to study the structure of primes. Gauss’ mistrust of Lagrange’s algebraic numbers was not without merit, it was known that unique factorization broke down in some cases, and this gave cause for concern since many arguments are based on √ √ factorizations into primes. For example, in Z[ −5] = {a+b −5 | a,b ∈ Z} we have: √ √ (2)(3) = (1+ −5)(1− −5). It follows the usual arguments based on comparing prime factors break down. Thus, much as with Abel and Ruffini and the quintic, we knew something was up. Kummer repaired the troubling ambiguity above by introducing so-called ideal numbers. These ideal numbers were properly con- structed by Dedekind who among other things was one of the first mathematicians to explicitly use congruence classes. For example, it was Dedekind who constructed the real numbers using so- called Dedekind-cuts in 1858. In any event, the ideals of Kummer and Dedekind and the modular arithmetic of Gauss all falls under the general concept of a ring. What is a ring? Definition 1.1.13. A set R with addition + : R×R → R and multiplication (cid:63) : R×R → R is called a ring if (i.) (R,+) forms an abelian group (ii.) (a+b)(cid:63)c = a(cid:63)c+b(cid:63)c and a(cid:63)(b+c) = a(cid:63)b+a(cid:63)c for all a,b,c ∈ R. If there exists 1 ∈ R such that a(cid:63)1 = a = 1(cid:63)a for each a ∈ R then R is called a ring with unity. If a(cid:63)b = b(cid:63)a for all a,b ∈ R then R is a commutative ring. Rings are everywhere, so many mathematical objects have both some concept of addition and mul- tiplication which gives a ring structure. Rings were studied from an abstract vantage point by Emmy Noether in the 1920’s. Jacobson, Artin, McCoy, many others, all added depth and appli- cation of ring theory in the early twentieth century. If ab = 0 and neither a nor b is zero then a 19I have betrayed Cayley in this story, but, have no fear well get back to him and many others soon enough 20if a,b∈Z then a+bi is known as a Gaussian integer 6 CHAPTER 1. GROUP THEORY and b are nontrivial zero-divisors. If ab = c then we say that b divides c. Notice, zero always is a divisor of zero. Anyway, trivial comments aside, if a ring has no zero divisors then we say the ring is an integral domain. Ok at this point, it becomes fashionable (unless youre McCoy) to assume R is commutative. A good example of an integral domain is the integers. Next, if a has b for which ab = 1 then we say a is a unit. If every nonzero element of a ring is a unit then we call such a ring a field. Our goal this semester is to understand the rudiments of groups, rings and fields. Well focus on group structure for a while, but, truth be told, some of our examples have more structure. We return to the formal study of rings after Test 2. Finally, if you stick with me until the end, Ill explain what an algebra is at the end of this course. Since we have a minute, let me show you a recent application of group representation theory to elementary particle physics. First, the picture below illustrates how a quark and an antiquark combine to make Pions, and Kaons: These were all the rage in early-to-mid-twentieth century nuclear physics. But, perhaps the next pair of examples will bring us to something you have heard of previously. Let’s look at how quarks can build Protons and Neutrons: Let me briefly explain the patterns. These are drawn in the isospin-hypercharge plane. They show how the isospin and hypercharge of individual up, down or strange quarks combine together to make a variety of hadronic particles. The N and P stand for Neutron and Proton. These patterns were discovered before quarks. Then, the mathematics of group representations suggested