0% found this document useful (0 votes)
193 views9 pages

Group Theory

This document contains exercises from Abstract Algebra by Dummit and Foote. It includes proofs of several exercises from Sections 3.1 and 3.2 relating to group theory concepts like quotient groups, normal subgroups, and Sylow theorems. The exercises cover topics such as properties of the quotient group Q/Z, conditions for a subgroup to be normal, relations between subgroups of finite index, and properties of the equivalence classes of p-tuples of group elements.

Uploaded by

너굴
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
193 views9 pages

Group Theory

This document contains exercises from Abstract Algebra by Dummit and Foote. It includes proofs of several exercises from Sections 3.1 and 3.2 relating to group theory concepts like quotient groups, normal subgroups, and Sylow theorems. The exercises cover topics such as properties of the quotient group Q/Z, conditions for a subgroup to be normal, relations between subgroups of finite index, and properties of the equivalence classes of p-tuples of group elements.

Uploaded by

너굴
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

Selected exercises from Abstract Algebra by Dummit

and Foote (3rd edition).


Bryan Félix
Abril 12, 2017

Section 3.1

Exercise 14. Consider the additive quotient group Q/Z.


a) Show that every coset of Z in Q contains exactly one representative q ∈ Q in the range
0 ≤ q < 1.

Proof. Assume there exist q1 , q2 with 0 ≤ q1 , q2 < 1 and q1 ≠ q2 such that q1 + Z = q2 + Z.


Then, for every z1 ∈ Z there exist z2 ∈ Z such that
q1 + z1 = q2 + z2 .
Without loss of generality, assume q1 < q2 . Then, we rewrite the previous equation as
q2 − q1 = z1 − z2 .
Observe that 0 < q2 − q1 < 1, in particular q2 − q1 ∉ Z. Then, since z1 − z2 is an integer, we
arrive at a contradiction.

b) Show that every element of Q/Z has finite order but that there are elements of arbitrarily
large order.

Proof. For any q + Z ∈ Q/Z let the representative element be of the form q̄ = zq with z, q ∈ Z.
Then q ⋅ (q̄ + Z) = q ⋅ (z/q) + Z = z + Z = Z. Therefore ∣q + Z∣ < ∞. Observe that the cosets
n + Z for n ∈ N have order n. Therefore there are elements of arbitrarily large order.
1

c) Show that Q/Z is the torsion subgroup of R/Z.

Proof. We show, by contradiction, that for all irrational q the coset q + Z has infinite order.
Assume that the order of q+Z is finite. Then, there exist an integer m such that m⋅(q+Z) = Z.
In other words, there exist an integer z such that m ⋅ q = z. Equivalently q = mz and q is
rational, arriving at a contradiction.

d) Prove that Q/Z is isomorphic to the multiplicative group of root of unity in C× .

Proof. It is easy to see a natural isomorphism by considering the following.


Addition over the real line works as a translation. On the other hand, the operation of
multiplication of elements in the unit circle is equivalent to a rotation . The idea here is
to map the rational numbers in [0, 1) to the angles in [0, 2π). The natural choice for the
isomorphism is q ↦ eq⋅2π⋅i .

1
Exercise 25.

a) Prove that a subgroup N of G is normal if and only if gN g −1 ⊆ N for all g ∈ G.

Proof. The first implication is trivial. If N  G then gN g −1 = N ⊆ N .


Now, assume that gN g −1 ⊆ N for all g ∈ G. We will prove that N ⊆ gN g −1 for all g ∈ G
(which implies gN g −1 = N and consequently N  G).
Let n be an element of N . Observe that

n = g(g −1 ng)g −1 .

From the assumption (gN g −1 ⊆ N for all g ∈ G) the element g −1 ng has to be an element of
N , let it be ñ. Then n = gñg −1 and therefore n is an element of gN g −1 for all g in G as
desired.

b) Let G = GL2 (Q), let N be the subgroup of upper triangular matrices with integer entries and
1’s on the diagonal, and let g be the diagonal matrix with entries 2, 1. Show that gN g −1 ⊆ N
but g does not normalize N .

Proof. Let n ∈ N have the following form

1 z
n=( )
0 1

where z is an integer. Therefore, the elements in gN g −1 look like

1 2z
gng −1 = ( ).
0 1

Clearly gng −1 is an element of N for any choice of z but some elements of N are left behind
/. Namely, all the upper triangular matrices with an odd integer in the upper right entry.
Hence, g does not normalize N .

Exercise 36. Prove that if G/Z(G) is cyclic then G is abelian. [If G/Z(G) is cyclic with
generator xZ(G), show that every element of G can be written in the form xa z for some integer
a ∈ Z and some element z ∈ Z(G).]
Proof. Assume G/Z(G) is cyclic. Then, there exist xZ(G) that generates G/Z(G). Therefore
every element of G has the form g = xa z for some a ∈ Z and z ∈ Z(G). Now, consider two
arbitrary elements of G, let them be g and h, and observe the following.

g h = xa z1 xb z2
= z1 xa xb z2 since z comutes with any element.
= z1 xb xa z2 since x commutes with itself
= x z2 x z1
b a
since z commutes with any element.
= h g.

Therefore G is abelian.

2
Exercise 42. Assume both H and K are normal subgroups of G with H ∩ K = 1. Prove that
xy = yx for all x ∈ H and y ∈ K. [Show x−1 y −1 xy ∈ H ∩ K.]
Proof. We follow the hint.
Let x ∈ H and y ∈ K. Consider the element x−1 y −1 xy. Since H is normal, the element y −1 xy is
an element of H and therefore x−1 (y −1 xy) is an element of H. In the same manner, and since
K is normal, the element x−1 y −1 x is an element of K and therefore (x−1 y −1 x)y is inside K.
Therefore x−1 y −1 xy ∈ H ∩ K. Furthermore x−1 y −1 xy = 1, and equivalently xy = yx.

Section 3.2

Exercise 9. Let G be a finite group and let p be a prime dividing ∣G∣. Let S denote the set of
p-tuples of elements of G the product of whose coordinates is 1:

S = {(x1 , x2 , . . . , xp ) ∶ xi ∈ G and x1 x2 ⋯xp = 1}.

a) Show that S has ∣G∣p−1 elements, hence has order divisible by p.

Proof. Since the p-tuples are unordered, we can arbitrarily choose the p − 1 left coordinates
in ∣G∣p−1 number of ways. Then, the coordinate xp is given by the inverse of x1 x2 ⋯xp−1

Define the relation ∼ on S by letting α ∼ β if β is a cyclic permutation of α.

b) Show that a cyclic permutation of an element of S is again an element of S.

Proof. Observe that we can translate the rightmost element to the left as follows

x1 x2 ⋯xp−1 xp = 1
x1 x2 ⋯xp−1 = x−1
p
xp x1 x2 ⋯xp−1 = 1

Since this can done indefinitely, every cyclic permutation of the p-tuple is an element of
S.

c) Prove that ∼ is an equivalence relation on S.

Proof. We proceed by proving the properties of an equivalence relation.

i) ∼ is reflexive.
Since any element α is the identity permutation of itself α ∼ α.
ii) ∼ is symmetric.
Assume α ∼ β. Then, β is a cyclic permutation of α. Furthermore, α is the inverse
cyclic permutation of β. Then β ∼ α.
iii) ∼ is transitive.
Assume α ∼ β and β ∼ γ. Then, since the composition of permutations is a closed
operation (in Sn ), γ is a permutation of α. Hence α ∼ γ.

3
d) Prove that an equivalence class contains a single element if and only if it is of the form
(x, x, . . . , x) with xp = 1.

Proof. Assume that an equivalence class contains a single element. Let it be (x1 , x2 , . . . , xp ).
Then, by part b), all of its permutations are in the same equivalence class. Therefore, it
must be true that
x1 = x2 = ⋯ = xp .
Let the element be x. then (x1 , x2 , . . . , xp ) has the desired form (x, x, . . . , x) an furthermore
xp = 1.
Now, assume that (x, x, . . . , x) is an element of S. We proceed by contradiction to show
that the equivalence class of (x, x, . . . , x) has order 1.
Assume there exist (x1 , x2 , . . . , xp ) in the equivalence class of (x, x, . . . , x) with at least one
entry xi distinct to x. Then, there must exist a cyclic permutation of (x1 , x2 , . . . , xp ) such
that (x1 , x2 , . . . , xp ) = (x, x, . . . , x). Observe that, for all permutations, the element xi = x.
Arriving at a contradiction.

e) Prove that every equivalence class has order 1 or p (this uses the fact that p is a prime).
Deduce that ∣G∣p−1 = k + pd, where k is the number of classes of size 1 and d is the number
of classes of size p.

Proof. We inspect the number of distinct cyclic permutations k of an element in the equiv-
alence class. If k = 1 or k = p we obtein the desired result. Therefore we assume that k is
of the form 1 < k < p. In this case it must be true that xi = xzk+i for all z ∈ Z. Since k
and p are relative prime zk + i generates all the indices modulo p. Therefore xi = x for all i
and we arrive at a contradiction since the permutation contains one element. Therefore, we
conclude that ∣G∣p−1 = k + p d where k is the number of equivalence classes of size 1 and d is
the number of equivalence classes of size p.

f ) Since {(1, 1, . . . , 1)} is an equivalence class of size 1, conclude from e) that there must be a
non identity element x in G with xp = 1, i.e., G contains an element of order p. [Show p∣k
and so k > 1.]

Proof. Since p ∣ ∣G∣p−1 , then p ∣ k + pd. Hence, p ∣ k and therefore k > 1. We conclude that
there must be an element x of order p.

Exercise 10. Suppose H and K are subgroups of finite index in the (possibly infinite) group
G with ∣G ∶ H∣ = m and ∣G ∶ K∣ = n. Prove that lcm(m, n) ≤ ∣G ∶ H ∩ K∣ ≤ mn. Deduce that if m
and n are relatively prime then ∣G ∶ H ∩ K∣ = ∣G ∶ H∣ ⋅ ∣G ∶ K∣.

Exercise 18. Let G be a finite group, let H be a subgroup of G and let N  G. Prove that if
∣H∣ and ∣G ∶ N ∣ are relatively prime then H ≤ N.
Proof. Since H is already a group, we only need to show that for all h in H, h is an element of
N.
Take your favourite h in H. We inspect the set hN in the quotient G/N . Note that (hN )∣G∶H∣ =
N (since ∣G ∶ N ∣ is the order of the quotient group). Furthermore, (hN )∣H∣ = (h∣H∣ )N = 1⋅N = N .
Therefore (from a previous result) (hN )gcd{∣G∶N ∣,∣H∣} = N , but gcd{∣G ∶ N ∣, ∣H∣} = 1. Hence
hN = N and, therefore, your favourite h is an element of N .

4
Exercise 21. Prove that Q has no proper subgroups of finite index. Deduce that Q/Z has no
proper subgroups of finite index.
Proof. Assume that there exist a proper subgroup of Q with finite index [Q ∶ N ] = n. Since Q
is abelian, then N  Q, and therefore Q/N is a group of order n. Therefore, for all q ∈ Q, nq is
an element of N . Then, consider the element nq ∈ Q. Observe that, by the previous assertion,
n nq = q is an element of N . Hence N = Q.
For the second part, assume there exist N ≤ Q/Z such that its index is finite. Then, consider
the two homomorphisms
ϕ ∶ Q/Z → (Q/Z)/N
with ker(ϕ) = N and
σ ∶ Q → Q/Z
with ker(σ) = Z.
Then, the homomorphism ϕ ○ σ ∶ Q → (Q/Z)/N with ker(ϕ ○ σ) = σ −1 (ϕ− 1(N )) defines a
subgroup of Q with finite index. A contradiction to the earlier problem.

100 100
Exercise 23. Determine the last two digits of 33 . [Determine 33 mod ϕ(100) and use
exercise 22.]
Proof. The goal today is to figure the value of 3(3 ) mod 100. We notice that 3 and 100 are
100

relative prime, therefore, from Euler’s formula we know that 340 ≡ 1 mod 100, (ϕ(100) = 40).
Now we compute 3100 mod 40 to further simplify our herculean task. Again, observe that 3
and 40 are relative prime, therefore 316 ≡ 1 mod 40, (ϕ(40) = 16). The latter implies that
3100 = 36(16)+4 ≡ 34 mod 40, or equivalently 3100 ≡ 81 ≡ 1 mod 40. Therefore 3100 can be written
as n ⋅ 40 + 1 for some n ∈ Z.
Now
100
33 ≡ 3n⋅40+1 mod 100
≡ (340 )n ⋅ 3 mod 100
≡ 1n ⋅ 3 mod 100
≡ 3 mod 100.

Then, the last two digits are zero and three in that order.

Section 3.3

Exercise 3. Prove that if H is a normal subgroup of G of prime index p then for all K ≤ G
either
i. K ≤ H or

ii. G = HK and ∣K ∶ K ∩ H∣ = p.
Proof. First, observe that, since H is a normal subgroup then K is a subgroup of the normalizer
of H. Therefore (by the good ol’ second isomorphism theorem) HK is a subgroup of G. We
inspect the following equality
∣G∣ ∣G∣ ∣HK∣
p= = .
∣H∣ ∣HK∣ ∣H∣
There are two cases

5
∣G∣
1. Case 1: ∣HK∣ = p and ∣HK∣
∣H∣ = 1
Then, it must be the case that H = HK since H ⊂ HK. Then, 1 ⋅ K = K is a subset of H.
∣G∣
2. Case 2: ∣HK∣ = 1 and ∣HK∣
∣H∣ = p
Here, G = HK analogous to the previous case. Furthermore, we have (from a previous
section)
∣H∣∣K∣
∣HK∣ =
∣H ∩ K∣
upon rearranging
∣K∣ ∣HK∣ ∣G∣
[K ∶ H ∩ K] = = = =p
∣H ∩ K∣ ∣H∣ ∣H∣
as desired.
Remark. The latter result is also implied from the second isomorphism theorem as
K/H ∩ K ≅ HK/H ≅ G/H .

Exercise 8. Let p be a prime and let G be the group of p-power roots of 1 in C. Prove that the
map z ↦ z p is a surjective homomorphism. Deduce that G is isomorphic to a proper quotient
of itself.
Proof. We prove that ϕ ∶ G → G given by ϕ(z) = z p is surjective. For arbitrary z ∈ G let
ϕ−1 (z) = z 1/p . Note that this number exist since z ∈ C. It is left to show that that z 1/p ∈ G.
Take n such that z (p ) = 1 and observe that
n

= z (p
n+1 n)
(z 1/p )p = 1.

Therefore z 1/p ∈ G and hence, ϕ is a surjection. Now observe that

ϕ(z1 z2 ) = (z1 z2 )p = z1p z2p = ϕ(z1 )ϕ(z2 ).

This shows that ϕ is an homomorphism. Therefore, ϕ is a surjective homomorphism. We


proceed to inspect the kernel of ϕ. By definition, ker(ϕ) = {z ∈ G ∶ z p = 1}. This is the set
of all p-complex roots of 1. Clearly, ker(ϕ) is non trivial and therefore (by means of the first
isomorphism theorm)
G/ ker(ϕ) ≅ G
where ker(ϕ) is a proper normal subgroup of G.

Exercise 9. Let p be a prime and let G be a group of order pa m, where p does not divide m.
Assume P is a subgroup of G of order pa and N is a normal subgroup of G of order pb n, where
p does not divide n. Prove that ∣P ∩ N ∣ = pb and ∣P N /N ∣ = pa−b .
Proof. Observe that P ≤ N (N ) since N is normal. Therefore by the good ol’ second isomor-
phism theorem P N ≤ G. Furthermore ∣P N ∣ divides ∣G∣ = pa m.
We have that P and N are subgroups of P N , therefore

∣P ∣ ∣ ∣P N ∣ and ∣N ∣ ∣ ∣P N ∣

or, equivalently
pa ∣ ∣P N ∣ and pb n ∣ ∣P N ∣

6
The former implies that ∣P N ∣ is of the form pa m̃ where m̃ divides m. The latter further implies
that ∣P N ∣ has the form pa nm′ where nm′ divides m.
Note that, by the second isomorphism theorem, P /P ∩ N ≅ P N /N and therefore

∣P ∣ ∣P N ∣
=
∣P ∩ N ∣ ∣N ∣

equivalently
pa p a n m′
= b
∣P ∩ N ∣ p n
and, upon rearranging
pb
∣P ∩ N ∣ = .
m′
Since p is a prime, the last assertion impies that m′ is a power of p, but m′ also has to divide
m by construction. Therefore m′ = p0 = 1.
We go back to the form of ∣P N ∣ and conclude that ∣P N ∣ = pa n. The computation of ∣P ∩ N ∣
and ∣P N /N ∣ is now trivial

∣P ∣∣N ∣
∣P ∩ N ∣ = = pb
∣P N ∣
∣P ∣
∣P N /N ∣ = = pa−b
∣P ∩ N ∣

as desired.

Exercise 10. Generalize the preceding exercise as follows. A subgroup H of a finite group G is
called a Hall subgroup of G if its index in G is relatively prime to its order: gcd(∣G ∶ H∣, ∣H∣) = 1.
Prove that if H is a Hall subgroup of G and N  G, then H ∩ N is a Hall subgroup of N and
HN /N is a Hall subgroup of G/N .
Proof. We will use the following property. If the order of one element divides the order of ∣H∣
and the other one divides the order of the index [G ∶ H] then, it must be true that their greatest
common divisor is also one.

1. To show: gcd(∣H ∩ N ∣, [N ∶ H ∩ N ]) = 1.
Observe that from the second isomorphism theorem

H/H ∩ N ≅ HN /N.

It is clear that ∣H ∩ N ∣ must divide the order of H .


Now, observe the following

∣N ∣ ∣G∣ ∣HN ∣ ∣G∣


=
∣H ∩ N ∣ ∣HN ∣ ∣H∣ ∣HN ∣
∣G∣
=
∣H∣

Therefore, the index [N ∶ H ∩ N ] divides the index [G ∶ H] as desired.

7
2. To show: gcd(∣HN /N ∣, [G/N ∶ HN /N ]) = 1.
Again, from the second isomorphism theorem we have that
∣HN ∣ ∣H∣
∣H ∩ N ∣ = ∣H ∩ N ∣
∣N ∣ ∣H ∩ N ∣
= ∣H∣

Therefore the order of HN /N divides the order of H.


Now observe that
∣G∣ ∣HN ∣ ∣HN ∣ ∣G∣
( / ) =
∣N ∣ ∣N ∣ ∣H∣ ∣H∣

Therefore the index [G/N ∶ HN /N ] divides the index [G ∶ H] as desired.

Section 3.4

Exercise 5. Prove that subgroups and quotient groups of a solvable group are solvable.
Proof. We show each statement individually.

1. Subgroups of a solvable group are solvable.


Let H be any subgroup of G and assume that G has decomposition

1 = G0  G1  ⋯  Gl = G

where Gi+1 /Gi is abelian. Consider the sets H ∩ Gi . We will show that the composition

1 = H ∩ 1 = H ∩ G0 ≤ H ∩ G1 ≤ ⋯ ≤ H ∩ Gl = H ∩ G = H

satisfies H ∩ Gi  H ∩ Gi+1 and H ∩ Gi+1 /H ∩ Gi is abelian for all i.


(a) H ∩ Gi+1  H ∩ Gi+1 .
We proceed by showing that for all k ∈ H ∩ Gi+1 , k(H ∩ Gi+1 ) = (H ∩ Gi+1 )k.
Take any x ∈ k(H ∩ Gi+1 ), then x has the form

x = kh

for some h ∈ (H ∩ Gi+1 ). Furthermore

x = (khk −1 )k.

Since k ∈ H ∩ Gi+1 , then khk −1 ∈ H. Also, since Gi  Gi+1 , k ∈ Gi+1 and h ∈ Gi the
conjugation khk −1 is an element of Gi . Hence x has the form

x = h̃k

where h̃ is an element of H ∩ Gi . In particular h̃ ∈ H ∩ Gi+1 . and therefore

x ∈ (H ∩ Gi+1 )k.

Hence k(H ∩ Gi+1 ) ⊆ (H ∩ Gi+1 )k


The reverse containment is analogous.

8
(b) The quotient H ∩ Gi+1 /H ∩ Gi is abelian for all i.
Observe that the quotient
H ∩ Gi+1 /H ∩ Gi
is a subgroup of Gi+1 /Gi which, by assumption, is abelian. Therefore H ∩Gi+1 /H ∩Gi
is abelian as well.

2. Quotient groups of a solvable group are solvable.


Let N be any normal subgroup of G and assume that G has decomposition

1 = G0  G1  ⋯  Gl = G

where Gi+1 /Gi is abelian.


Consider the quotient group G/N and the quotients Gi N /N . We will show that the
composition

1 = N = G0 N /N ≤ G1 N /N ≤ ⋯ ≤ Gl N /N = GN /N = G/N

satisfies Gi N /N  Gi+1 N /N and (Gi+1 N /N )/(Gi N /N ) is abelian for all i.

(a) Gi N /N  Gi+1 N /N
(b) (Gi+1 N /N )/(Gi N /N ) is abelian
Using the third isomorphism theorem

(Gi+1 N /N )/(Gi N /N ) ≅ Gi+1 N /Gi N

which is a subgroup of Gi+1 /Gi , and therefore abelian.

You might also like