11 - Setup for Group Morphisms

In section 3.1 we'll try to setup quotient groups. But before getting into the definition, let's consider an Reasoning Example for Quotient Group Definitions:

Consider the map φ:ZC× where nin. Let's verify φ is a group Homomorphism (or Group Morphism). Take any n1,n2Z. Notice:
φ(n1+n2)=in1+n2=in1in2=φ(n1)φ(n2)
If we consider the values of n that map to in:

Value n φ(n)
... ...
-1 i
0 1
1 i
2 1
3 i
4 1
... ...
Notice it's very similar to Z over nZ - The Integers Modulo n for n=4. So the image of φ is {1,i,1,i}. In general the formula is:
$
\varphi(n) = \begin
1 & n \equiv 0 \bmod 4 \
i & n \equiv 1 \bmod 4 \
-1 & n \equiv 2 \bmod 4 \
-i & n \equiv 3 \bmod 4
\end
$

Notice here that we've:

  • partitioned Z into four subsets:
    • C1=φ1(1)={,4,0,4,}, where here φ1 is the preimage (not the inverse)
    • C2=φ1(i)={,3,1,5,}
    • C3=φ1(1)={,2,2,6,}
    • C4=φ1(i)={,1,3,7,}
  • ... but only one of these partitions is a Kernel of φ. Namely:
    • C1={nZ|φ(n)=1}=Ker(φ) since 1 is the identity in C×. Thus C1Z.
    • The other 3 are not subgroups, but they look very similar to C1, only with some offset in the integers. You can say each of the C2,C3,C4 is a "translation" of C1. For example:
      • C2={,3,1,5,}=1+C1
      • Similarly C3=2+C1
      • C4=3+C1.

Here we have a clear bijection between {C1,C2,C3,C4}{1,i,1,i}. Here the right is a Group, and since we have a bijection then the left set is also a group, provided we do similar operations from C×. Namely, the operation on {C1,,C4} is to "add representatives" to get to the resulting partition, for each nCk being our representative for Ck.

Operation Result
C2C3 9+10=19C4 so C2C3=C4. Doing 1+(2) or any other example yields similar results.
C3C3 C1 since 6+6=12C1.
... ...
What's the pattern? Notice that:
Note

Iff you have a,bCk for some k{1,2,3,4} then:
φ(a)=φ(b)
which is true by the definition of Ck. This shows that φ is well-defined. But using our notation for groups then:
φ(a)φ(b)1=1C×
iff:
φ(ab)=1C×
which only happens when abKer(φ).

We've defined an equivalence relation on Z by:
abφ(a)=φ(b)abKer(φ)
So we'll adjust our notation:
C1=[0]=[4]= or use C1=0=4=
Thus:
C2=[1],C3=[2],C4=[3]
Then:
C2C3=C4[1]+[2]=[1+2]=[3]
So instead of using we can just use normal integer addition to fill our table.

[0] [1] [2] [3]
[0] [0] [1] [2] [3]
[1] [1] [2] [3] [0]
[2] [2] [3] [0] [1]
[3] [3] [0] [1] [2]
So we claim that:
Claim

  1. ({[0],[1],[2],[3]},) is a Group.
  2. The map γ:Z{[0],[1],[2],[3]} where n[n] is a Homomorphism (or Group Morphism)
  3. The map α:{[0],[1],[2],[3]}{1,i,1,i} is an Isomorphism

The General Case

For the more general scenario, say we have some φ:GH is a Homomorphism (or Group Morphism). Then:

  • Ker(φ)=φ1(1H) (see Kernel) is a normal subgroup of G
  • For each hH if himage(φ) then φ1(h)=. Otherwise then φ1(h)=gKer(φ)={gk|kKer(φ)}. Altnernatively, this set is Ker(φ)g={kg|kKer(φ)} for any gG with gφ1(h). We call these the cosets of Ker(φ).

So in our Ci example we had 4 cosets. Similar to above, there's an equivalence relation on G where:
g1g2φ(g1)=φ(g2)g1g21Ker(φ)
The set of distinct cosets (ie: the set of equivalence classes) is a group under the operation:
[g1][g2]=[g1g2]
Then this group is isomorphic to the image of φ.

We'll do a lot more of this among many different lemmas.