CH 9
CH 9
1. Introduction: A factor group is a way of creating a group from another group. This new
group often retains some of the properties of the original group.
2. Normal Subgroups:
We see that we have ()H = H(), (12)H = H(12), (13)H = H(13), (23)H = H(23),
(123)H = H(123), (132)H = H(132).
Example: Consider the subgroup H = {(), (12)} of S3 . Observe that (23)H = {(23), (132)}
but H(23) = {(23), (123)}. Since we have a left coset not equal to a right coset the sub-
group is not normal.
(b) Theorem (Normal Subgroup Test): A subgroup H of G is normal iff gHg −1 ⊆ H
for all g ∈ G.
Proof:
⇒: Suppose H / G. We claim gHg −1 ⊆ H for any g ∈ G. Let g ∈ G and then an element
in gHg −1 looks like ghg −1 for some h ∈ H. Then observe that ghg −1 = h0 gg −1 = h0 ∈ H.
⇐: Suppose gHg −1 ⊆ H for all g ∈ G. We claim gH = Hg. Note that gH = gHg −1 g ⊆
Hg and that Hg = gg −1 Hg ⊆ gH. In the latter we have g −1 Hg ⊆ H becaues the
supposition is true for g −1 .
Example: The subgroup SL2 R of 2 × 2 matrices with determinant 1 forms a normal sub-
group of GL2 R. To see this note that if g ∈ GL2 R and s ∈ SL2 R then det(gsg −1 ) =
det(g) det(s)(1/ det(g)) = det(s) = 1 and so gsg −1 ∈ SL2 R.
3. Factor Groups:
Definition/Theorem: Let G be a group and let H / G. Then we define G/H (read ”G mod
H”) to be the set of left cosets of H in G and this set forms a group under the operation
(aH)(bH) = abH.
Proof: We have a few things to show here:
• Any given left coset will have multiple representatives because we know that aH and a0 H
can be identical for a 6= a0 . Consequently we first need to be sure that our operation is
well-defined, meaning that if we choose a0 H = aH and b0 H = bH and we do (a0 H)(b0 H) =
a0 b0 H we get the same result as if we do (aH)(bH) = abH. In other word we must verify
that abH = a0 b0 H. Since a0 H = aH and since a0 ∈ a0 H we have a0 = ah1 and likewise
b0 = bh2 for some h1 , h2 ∈ H. It follows that a0 b0 H = ah1 bh2 H = abh01 h2 H = abH.
• The identity is eH.
• The inverse of aH is a−1 H.
• Associativity follows since (aH)(bHcH) = (aH)(bcH) = abcH = (abH)(cH) = (aHbH)cH.
Example: If G = Z and h = 4Z then there are four distinct cosets:
Note: The terminology ”G mod H” arises from the analogy with modular arithmetic. When
we work in Z mod 5, for example, we say that 8 = 3 mod 5 because 8 = 3+5 = 3 mod 5 because
the 5 ”gets absorbed” into the modulus. That is, 8 mod 5 = (3 + 5) mod 5 = 3 + (5 mod 5) =
3 mod 5. Similarly if we’re looking at gH and if g = g 0 h then gH = g 0 hH = g 0 H because the
h gets absorbed by the H.
4. Applications:
(a) Theorem: If G/Z(G) is cyclic then G is Abelian.
Proof: Since G/Z(G) is cyclic we know there is some g0 ∈ G such that G/Z(G) =
hg0 Z(G)i. Thus every coset has the form g0k Z(G) for some k. Given a, b ∈ G we know
that each is in some coset so a ∈ g0j Z(G) and b ∈ g0k Z(G) for some j, k and moreover then
a = g0j z1 and b = g0k z2 for z1 , z2 ∈ Z(G). Then observe that:
ab = g0j z1 g0k z2 = g0j g0k z1 z2 = g0k g0j z2 z1 = g0k z2 g0j z1 = ba
QED
Example: Suppose G is non-Abelian and |G| = pq where p, q are distinct primes then
G has trivial center consisting only of {e}. This is because a bigger center would have to
have order p, q or pq by Lagrange’s Theorem. The first two fail by this theorem and the
third fails because G is non-Abelian.
Note: This has meaningful results. For example suppose we know that |G| = pq where
p, q are prime and suppose we find just one g0 ∈ G with g0 ∈ Z(G) and g0 6= e. Since
Z(G) ≤ G we know that by Lagrange’s Theorem we must have |Z(G)| = 1, p, q or pq.
Since |Z(G)| =6 1 we know it’s p, q or pq. If |Z(G)| = pq then G is Abelian. Without loss
of generality if |Z(G)| = p then |G/Z(G)| = pq/p = q and snce groups of prime order are
cyclic we have G/Z(G) cyclic and then G Abelian. So this goes to show that in such a
group if we find one single non-identity element in the center then the group is Abelian
and everything is in the center.