So πβπ is a bijection, and πβπ is a binary operation on Sym(π₯).
Composition is associative.
We have an identity, idπ:πβπ;π₯β¦π₯, which satisfies πβidπ₯=π=idπβπ, βπβSym(π).
We have inverses: πβπβ1=πβ1βπ=idπ.β β‘
Definition 1.13
If π={1,2,β¦,π}, ππβSym(π) is the πth symmetric group.
Remark 1.14
Sym(π) can also be interpreted as the permutations of π.
Example 1.15 (fields)
Let (πΉ,+,Γ) be a field.
Then (πΉ,+) is an abelian group, with identity element 0πΉ, whereas (πΉ\{0πΉ},Γ) (sometimes denoted πΉβ) is also an abelian group with identity element 1πΉ.
For example, βββββ all form groups under addition, or under multiplication when zero is excluded.
Also β€πββ non-zero integers mod p, for prime π, is a group under multiplication.
Example 1.16 (matrices)
(ππ(πΉ),+) is an abelian group (where ππ(πΉ) is the set of πΓπ matrices over πΉ).
Definition 1.17
Define GLπ(πΉ)β{π΄βππ(πΉ)|detπ΄β 0}; (GLπ(πΉ),β ) is the πth general linear group over πΉ.
Definition 1.18
Let SLπ(πΉ)β{π΄βGLπ(πΉ)|detπ΄=1}; this is the πth standard linear group over πΉ under matrix multiplication.
Example 1.19
Definition 1.20
Let π be a set, let πβπ.
Let Stab(π)β{πβSym(π):π(π)=π}; that is, Stab(π) is the set of permutations that map π to itself.
Proposition 1.21
(Stab(π),β) is a group.
Proof.Necessary to prove: βπ,πβStab(π),πβπβStab(π): (πβπ)(π)=π(π(π))=π(π)=π.
Associativity is inherited from the symmetric group.
idπ(π)=πβΉidπβStab(π).
Let πβStab(π), then π(π)=π. Hence πβ1(π(π))=π, and πβ1(π(π))=πβ1(π), so πβ1=π, so πβ1βStab(π).β β‘
Definition 1.22
(Stab(π),β) is the stabiliser subgroup of π.
Definition 1.23
Let πΊ be a group, with binary operation β. A subset π»βπΊ is a subgroup if β restricts to a map β:π»Γπ»βπ» which then makes (π»,β/π») into a group. We write π»β€πΊ.
Example 1.24
(β,+)β€(β,+)
Proposition 1.25 (Subgroup Test)
Let πΊ be a group. A subset π» of πΊ is a subgroup iff :
πβπ»
βπ₯,π¦βπ»,π₯π¦β1βπ»
Proof.
ββΉβ:
Assume π» is a subgroup of πΊ. Then π» is closed under multiplication and inversion. Therefore βπ₯,π¦βπ»,π₯π¦β1βπ» and also πβπ».
ββΈβ: Associativity on π» is inherited from πΊ.
(i) βΉπβπ».
Apply (ii) with π₯=π, giving βπ¦βπ»,π¦β1βπ».
So (π»,β/π») is a group.β β‘
Example 1.26
Let πΊ=GLπ(β).
Let O(π)={πβπΊ|πβ€π=π}, the orthogonal matrices.
For πββ€>0, πΆπ={π,π,π2,β¦,ππβ1} is a cyclic group with binary operation
ππππ={ππ+πifπ+πβ€πβ1ππ+πβπifπβ€π+πβ€2πβ2and inverse (ππ)β1=ππβπ for 0β€π<π. πΆπ is the cyclic group of order π.
Remark 1.34
A cyclic group can have more than one generator; for instance, (β€,+) also has β1 as a generator.
that is, it is the group of distance-preserving transformations that stabilise the regular π-sided polygon centred at the origin.
We denote the rotation by 2π/π anticlockwise by π, and the reflection in an axis (which one depends on which way we choose to draw the polygons) by π . Then the elements of π·2π are π,π,π2,β¦,ππβ1,π ,ππ ,π2π ,β¦,ππβ1π . Note that this is 2π elements.
We write permutations on the right, i.e. if πβ{1,β¦,π},πβππ, then ππβπ(π).
This means that composition is backwards from the usual order.
Definition 2.3
A cycle is a permutation πβππ such that β{π1,β¦,ππ}β{1,β¦,π} s.t. πππ=ππ+1 for πβ{1,β¦,πβ1} and πππ=π1, and for all π₯β{1,β¦,π}\{π1,β¦,ππ}, π₯π=π₯.
Every πβππ can be written as a product of disjoint cycles.
Moreover, this factorisation is unique up to cycling of elements within cycles, and order of (commuting) factors.
Proof.Fix π1β{1,β¦,π} and consider π1,π1π, π1π2,β¦.
There exists 0β€π<π such that π1ππ=π1ππ. We choose such a pair (π,π) with minimal π.
Then applying πβπ, we get that π1=π1ππβπ, and note that π<πβπ. Since π was minimal, π=0.
Let π>0 be the least such that π1ππ=π1, then π1,π1π,π1π2,β¦,π1ππβ1 must all be distinct.
The set {π1,π1π,β¦,π1ππβ1} is known as the orbit of π1 under π, orbit(π1).
Note that orbit(π1)={π1ππ|πββ}={π1ππ|πββ€}.
Now pick π2β{π,β¦,π}\orbit(π1). orbit(π2) must be disjoint from orbit(π1), otherwise π2βorbit(π1), since if π1ππ=π2ππ , then π2=π1ππβπ .
Then π=(π1π1πβ¦π1ππ1β1)(π2π2πβ¦π2ππ2β1)β¦(πππππβ¦πππππβ1), where ππ=|orbit(ππ)| for πβ{1,β¦,π}.
If π=π1β π2β β¦β ππ=π1β²β π2β²β β¦.β ππβ², then 1 must appear in one of the ππβ²s, supposing W.L.O.G. that π=1. Also 1 appears in π1. Then π1=(11πβ¦1ππ1β2), and π1β²=(11πβ¦1ππ1β²β1), hence π1=π1β² and π1=π1β². Repeat for all ππβ²s, then the factorisation is unique up to commutation and cycling.β β‘
Proposition 2.7
Let π=π1π2β¦ππβππ where the ππs are disjoint cycles. Then π(π)=lcm(π(π1),β¦,π(ππ)) and π(ππ)=length ofππ.
Proof.If π is a cycle of length π, π=(π1π2β¦ππ), then ππππ=ππ, whereas if π₯β ππ then π₯π=π₯ so π₯ππ=π₯. Therefore ππ=π.
Conversely, if ππ=π for π>0 then π1ππ=π1π=π1, so πβ₯π therefore π(π)=π.
Then let πΏβlcm(ππ),ππ=length(ππ)=π(ππ).
Then ππβ£πΏβπ, so ππΏ=(π1π2β¦ππ)πΏ=π1ππ2πβ¦πππΏ (note that this is not a valid application of distributivity in the general case!).
Lemma 2.8
If πβπΊ has finite order, then for πββ€, ππ=π iff π(π)β£π.
Proof.ββΈβ:
If π(π)β£π, then π=π(π)π, hence ππ=ππ(π)π=(ππ(π))π=ππ=π.
ββΉβ:
We write π=ππ(π)+π for πββ€ and 0β€π<π(π). Then ππ=πβΉπππ(π)+π=π hence ππ=π. But π<π(π), so π=0 by the defn of π(π).β β‘
We have that ππΏ=π; by the lemma, π(π)β£πΏ.
Now consider ππ(π)=π. Then (π1β¦ππ)π(π)=π hence π1π(π)β¦πππ(π)=π.
If we take π¦β{1,β¦,π}, saying W.L.O.G. that π¦ appears in ππ. Then π¦ does not appear in any other ππ, hence π¦πππ(π)=π¦βΉπππ(π)=π for all π.
By the lemma, π(ππ)β£π(π), so πΏ=lcm(π(ππ))β£π(π).β β‘
Definition 2.9
For πβππ, we define ππβGLπ(β) to besuch that the πth row of ππ has a 1 in column ππ, and has all other entries zero. Then ππ is a permutation matrix.
Transitivity: Suppose π₯βΌπ¦ and π¦βΌπ§, then π§=πβ1π¦π=πβ1ββ1π₯βπ=(βπ)β1π₯(βπ), so π₯βΌπ§.β β‘
Definition 3.3
The conjugacy class of π₯βπΊ is the equivalence class π₯ under βΌ. We denote this ππΊ. (This is not standard notation).
Remark 3.4
πΊ is the disjoint union of its conjugacy classes,
πΊ=β¨{ππΊ|πβπΊ}.
Proposition 3.5
If πΊ is an abelian group, then π¦βΌπ₯ iff π¦=π₯.
For π,πβππ, πβΌπ iff π and π have the same cycle type (that, is they have the same number of cycles of each length when factored into cycles).
Proof.ββΉβ: Suppose πβΌπ, then it follows from unqieueness of cycle factorisation, Lemma 3.6 and distributivity that π and π have the same cycle type.
ββΈβ:
Suppose π and π have the same cycle type, writing π=π1β¦ππ and π=π1β²β¦ππβ². By Lemma 3.6, we can choose any πβππ such that each ππ=πβ1ππβ²π, then by distributivity π=πβ1ππ so πβΌπ.β β‘
βββ: It suffices to prove (STP) that βπ»β€πΊ s.t. π»βπ, π€(πβͺπβ1)βπ».
So let π€βπ€(πβͺπβ1). Then π€=π 1π 2β¦π π for π πβπβͺπβ1. Since πβπ», each π πβπ». Since π»β€πΊ, π€βπ».
If π₯,π¦βπ€(πβͺπβ1), then π₯=π 1β¦π π, π¦=π‘1β¦π‘π for π π,π‘πβπβͺπβ1. Then π₯π¦β1=π 1β¦π ππ‘πβ1β¦π‘1β1βπ€(πβͺπβ1).
For all π βπ, π is a word in π, so πβπ€(πβͺπβ1).β β‘
Then π:πΊβπΆπ is defined by ππβ¦π₯π.
Define π:β€βπΊ;πβ¦ππ.
Then π(π+π)=ππ+π=ππππ=π(π)π(π).
π is clearly surjective as πΊ={ππ|πββ€}. If π(π)=π(π), then ππ=ππ, so ππβπ=π; since πΊ has infinite order, πβπ=0 so π=π. Hence π is also injective, so it is a bijection.
π₯β²π¦β²βπ₯π’=π₯β²(π¦β²βπ¦)+(π₯β²βπ₯)π¦, which is an integer multiple of π if πβ£π₯βπ₯β² and πβ£π¦βπ¦β². Hence π₯β²π¦β²=π₯π¦modπ.
β β‘
Proposition 5.7
There following are well-defined binary operations on β€π:
πΜ +πΜ βπ+πΜ
πΜ ΓπΜ βππΜ
Proof.It suffices to prove (STP) that these definitions do not depend on the choices made for the two πΌ,π½ββ€π:
pick πβπΌ,πβπ½
set πΌ+π½=π+πΜ .
If πβ²βπΌ and πβ²βπ½, then πβ²+πβ²β‘π+πmodπ by the Lemma (TODO ref), so πβ²+πβ²Μ =π+πΜ . Similarly πβ²πβ²Μ =ππΜ .β β‘
Proposition 5.8
(β€π,+) is an abelian group, and moreover is cyclic.
(β€π,Γ) satisifies associativity and commutativity, and admits an identity element
π₯(π¦+π§)=π₯π¦+π₯π§ for all π₯,π¦,π§ββ€π.
0Μ +π₯Μ =0+π₯Μ =π₯Μ =π₯+0Μ =π₯Μ +0Μ , so (β€π,+) admits an identity.
The inverse of π₯Μ is βπ₯Μ .
This group is cyclic because πΜ =1Μ +β¦+1Μ βπtimes for all πββ€>0.
Hence (β€π,+) is generated by 1Μ . 1Μ has order π(1Μ )=π.
Inherited from the integers in a similar manner to the above, with 1Μ as the identity.
Also inherited from the integers.
β β‘
Remark 5.9
β€π is a unital commutative ring.
Lemma 5.10
The function from {0,1,β¦,πβ1}ββ€π defined by πβ¦πΜ is a bijection.
Proof.
Suppose 0β€π,πβ€πβ1 and πΜ =π Μ . Then πβ‘π modπ, so πβ£πβπ , i.e. πβπ =ππ for some πββ€. WLOG suppose π β₯π, then if πβ 0 then |π βπ|β₯π. But |π βπ|β€πβ1, so π=0 and π=π . Hence the function is injective.
Now take πΌββ€π, and say πΌ=πΜ for some πββ€. Divide π by π with remainder to get π=ππ+π for some π,πββ€ with 0β€πβ€πβ1.
Then πβ£ππ=πβπ, so πβ‘πmodπ. Note that πΌ=πΜ =πΜ , so the function is surjective.
Hence the function is a bijection.β β‘
Proposition 5.11
Let π₯Μ in β€π. Then π₯Μ admits a multiplicative inverse iff gcd(π₯,π)=1.
Define β€πΓβ{π₯Μ ββ€π|βπ¦ββ€πs.t.π₯Μ π¦Μ =1Μ }. Then β€πΓ is an abelian group.
For prime π, β€πΓ=β€π\{0Μ }, so β€π is a field.
Proof.
exercise
If πΜ ,πΜ ββ€πΓ, then πΜ πΜ =1Μ =πΜ πΜ for some πΜ ,πΜ ββ€π. Hence (πΜ πΜ )(πΜ πΜ )=(πΜ πΜ )(πΜ πΜ )=1Μ β 1Μ =1Μ .
Let 0Μ β π₯Μ ββ€π. Then πβ€π₯. Since π is prime, gcd(π,π₯)=1. Hence by (a), π₯Μ ββ€πΓ.β β‘
6. Cosets & Lagrangeβs Theorem
Definition 6.1
Let π»β€πΊ not necessarily finite.
A left coset of π» is a subset of πΊ of the form ππ» for some πβπΊ, where
ππ»β{πβ|ββπ»}.
A right coset of π» is a subset of πΊ of the form π»π for some πβπΊ, where
π»πβ{βπ|ββπ»}.
Definition 6.2
We write πΊ/π» (βπΊ mod π»β) to mean the set of left cosets of π» in πΊ; we write π»\πΊ to mean the set of right cosets of π» in πΊ
Then π2β1π1=β2β1β1βπ» as π»β€πΊ.
Hence by the Coset Equality Lemma, π1π»=π2π». Therefore distinct left cosets are disjoint.
Say that πΊ/π»={π1π»,π2π»,β¦,πππ»} where π=|πΊ/π»|.
The function π:π»βπππ»;ββ¦ππβ is a bijection, with inverse π¦β¦ππβ1π¦. Hence |πππ»|=|π»|=|πππ»|.
Then |πΊ|=βπ=1π|πππ»|=π|π»|=[πΊ:π»]|π»|.
β β‘
Corollary 6.9.1
For πΊ a finite group and πβπΊ, then π(π)β£|πΊ|.
Suppose π is prime and πΊ is a group of order π. Then πΊβ πΆπ, and all non-identity elements of πΊ have order π.
Proof.By Corollary 6.9.1, for πβπΊ, π(π)=1 or π(π)=|πΊ|. π(π)=1 iff π=π, so for all πβπΊ\{π}, π(π)=π.
But π₯Μ =π₯Μ β1βΊπ₯Μ 2=1Μ βΊ(π₯Μ β1Μ )(π₯Μ +1Μ )=0.
Since β€π is a field, π₯Μ 2=1Μ βΉπ₯Μ =1Μ or π₯Μ =β1Μ .
Hence (πβ1)!Μ =1Μ β β1Μ =β1Μ .
β β‘
Corollary 6.12.1
Given a group πΊ with |πΊ| even, then βπβπΊ with π(π)=2.
Proof.Consider {π,πβ1} for each πβπΊ. |{π,πβ1}|=2 iff πβ πβ1 iff π2β π.
That is, |{π,πβ1}|=1 iff π2=π iff π=π or π(π)=2.
Now πΊ=βπβπΊ{π,πβ1} so |πΊ|β‘1+|{πβπΊ:π(π)=2}|mod2. Since |πΊ| is even by assumption, there must exist at least one (in fact, an odd number of) π with π(π)=2.β β‘
Theorem 6.13
Let πβ₯3 be prime, and let πΊ be a group with |πΊ|=2π. Then either πΊβ πΆ2π or πΊβ π·2π.
Suppose then βπβπΊ with π(π)=2π.
If βπβπΊ,π(π)=1 or 2, then π2=π for all πβπΊ. In this case, πΊβ πΆ2π for some πββ. Then 2π=|πΊ|=2π which is a contradiction.
Therefore, βπ₯βπΊ with π(π₯)=π.
Moreover, by Corollary TODO, βπ¦βπΊ with π(π¦)=2.
Then π¦π₯π¦β1=π₯π for some 0β€πβ€πβ1.
Now π₯=π¦2π₯π¦β2=π¦(π¦π₯π¦β1)π¦β1=π¦(π₯π)π¦β1=(π¦π₯π¦β1)π=(π₯π)π=π₯π2.
Hence π2β‘1modπ as π(π₯)=π.
Since π is prime, this forces π=1 or πβ‘β1modπ, so πβ1 or π=πβ1.
If π=1, then π¦π₯π¦β1=π₯βΉπ¦π₯=π₯π¦, so π(π₯π¦)β£2π. (π₯π¦)2=π₯2π¦2=π₯2β πβΉπ(π₯π¦)β 2. Similarly (π₯π¦)π=π₯ππ¦π=π¦β πβπ(π₯π¦)β π. So π=πβ1.
Let πΊ,π» be groups. A function π:πΊβπ» is a homomorphism if π(π1βπΊπ2)=π(π1)Γπ»π(π2) for all π2,π2βπΊ.
Remark 7.2
An isomorphism is a bijective homomorphism.
Definition 7.3
A monomorphism is an injective homomorphism. We write πΊβͺοΈπ».
Definition 7.4
A epimorphism is a surjection homomorphism. We write πΊβ π».
Example 7.5
Definition 7.6
The trivial homomorphism is the homomorphism π:πΊβπΊ;πβ¦ππ».
Example 7.7
Linear maps between vector spaces are group homomorphisms (under addition).
Theorem 7.8
Hom(β€,πΊ)β πΊ.
Proof.Let ππ:β€βπΊ be defined by πβ¦ππ for each πβπΊ.
Then STP that the function πΊβHom(β€,πΊ);πβ¦ππ is bijective.
If π,ββπΊ and ππ=πβ, then π=ππ(1)=πβ(1)=β, so πβ¦ππ is injective.
Let π:β€βπΊ be a homomorphism, then if πβπ(1)βπΊ, then ππ(π)=ππ=π(1)π=π(π) for all πββ€. Then π=ππ, so πβ¦ππ is surjective.β β‘
Definition 7.9
An automorphism of πΊ is an isomorphism π:πΊβπΊ.
Remark 7.10
Automorphisms form a group Aut(πΊ) under composition.
Definition 7.11
Fix πβπΊ. Define ππ:πΊβπΊ by π₯β¦πβ1π₯π; ππ is the conjugation by π.
Proposition 7.12
ππ is an automorphism.
Proof.ππ(π₯π¦)=πβ1π₯π¦π=πβ1π₯ππβ1π¦π=ππ(π₯)ππ(π¦). Hence each ππ is a homomorphism.
Moreover, ππ is invertible : ππβ1=ππβ1.β β‘