Kernel and Image, Cosets, Normal Subgroups, Conjugation, The Correspondence Theorem
§ Kernel and Image
Definition (Kernel and Image). Consider a homomorphism f:G→H between groups G and H. Then the kernel of f is ker(f):={g∈G:f(g)=1H}, and the image of f is im(f):={h∈H:h=f(g) for some g∈G}.
Note that ker(f) is a subgroup of G, and im(f) is a subgroup of H. (Check this!)
Example. (Application of the Kernel) Take a homomorphism f:G→H, and suppose that f(g1)=f(g2) for g1,g2∈G. What can we say about g1 and g2?
Solution: Observe the following:
f(g1)=f(g2)⟺f(g1g2−1)=1⟺g1g2−1∈ker(f)⟺g1=kg2 for some k∈ker(f).
If we multiply by the left inverse instead of the right inverse in the first step, we also find:
f(g1)=f(g2)⟺f(g2−1g1)=1⟺g2−1g1∈ker(f)⟺g1=g2k for some k∈ker(f).
Consolidating these results together, we have:
{f(g1)=f(g2)⟺g1=kg2 for some k∈ker(f)f(g1)=f(g2)⟺g1=g2k for some k∈ker(f)
Example. Take the homomorphism f:(C,+)→(C×,⋅) defined by f(z):=ez. Then the image of f is im(f)=C× (i.e. f is surjective), and the kernel of f is ker(f)=2πi⋅Z.
Example. Take the homomorphism f:GLn(R)→R× defined by f(M)=detM. Then the image of f is im(f)=R× (i.e. f is surjective), and the kernel of f is ker(f)=SLn(R).
Example. Take the homomorphism f:Sn→{−1,1} sending permutations π∈Sn to the sign of π. The kernel of f must be a subgroup of Sn; this subgroup is the alternating groupAn, consisting of all even permutations in Sn.
Example. Let G be a group, and fix an element g∈G. Then we may take the homomorphism f:(Z,+)→G defined by f(n):=gn. In this case,
The image of f is the subgroup generated by {g}. We denote this by im(f)=⟨g⟩.
The kernel of f is ker(f)=kZ, where g has order k. (If g has infinite order, then ker(f)={0}.)
§ Cosets
Example. Recall that 3Z is a subgroup of Z. This subgroup can be “copied-and-translated” twice to form the subsets 3Z+1 and 3Z+2 in Z. (Note that these subsets are not subgroups!)
Let's give a name to these “copied-and-translated” subsets associated with a subgroup: cosets.
Definition (Cosets). Let H be a subgroup of G. Then a left coset of H is a subset gH:={gh:h∈H}⊆G for some g∈G. Similarly, a right coset of H is a subset Hg:={hg:h∈H}⊆G for some g∈G.
The phrase “coset” (without further specification) should be assumed to mean “left coset”.
Theorem. (Coset Properties) Let H be a subgroup of G. The cosets of H in G satisfy the following properties:
Any two cosets of H have the same size.
Any two distinct cosets are disjoint.
Proof: To prove the first fact, simply note that the size of any coset gH is equal to ∣H∣. This is because the natural map T:H→gH defined by T(h)=gh is a bijection. ■
To prove the second fact, consider two distinct cosets C1=g1H and C2=g2H. Suppose, for the sake of contradiction, that these two cosets are not disjoint. Then there exists h1,h2∈H such that g1h1∈C1 and g2h2∈C2 are the same; that is, g1h1=g2h2. Equivalently, g1=g2h2h1−1. However, this implies
C1=g1H=(g2h2h1−1)H=g2(h2h1−1H)=g2H=C2.
So in fact, we have shown that if there is just a single pair of elements g1h1 and g2h2 in C1 and C2 that match, then we can show the entirety of C1 and C2 must match completely! ■
Here's an equivalent, more intuitive rephrasing of the above theorem:
Theorem. (Cosets form Partitions) Let H be a subgroup of G. Then the set of all left cosets of H partitions the elements of G into disjoint subsets of size ∣H∣.
Note that all of the above theorems hold equally true if every instance of “left” is replaced with “right”. However, the partitioning of G formed by left cosets usually does not equal the partitioning formed by right cosets!
Example. Consider the group G=(R2,+) and a subgroup H:={(a,2a):a∈R}. Then cosets of H look like translated copies of lines parallel to y=2x. These lines tile the plane R2.
Definition (Index). The number of cosets of H in G is the index of H in G, denoted [G:H]. Note that the theorem above implies ∣G∣=[G:H]⋅∣H∣.
Theorem. (Lagrange's Theorem) Let H be a (finite) subgroup of a (finite) group G. Then ∣H∣ divides ∣G∣.
Proof: Follows from the equation ∣G∣=[G:H]⋅∣H∣, along with the fact that [G:H] is an integer. ■
Lagrange's Theorem is very powerful! Here's an interesting corollary of the above theorem.
Theorem. (Groups of Prime Order) Any group of prime order must be cyclic.
Proof: Suppose G is a group of prime order p. Pick some g∈G that is not the identity. Then H=⟨g⟩ is a subgroup of G. Importantly, ∣H∣>1 (since g is not the identity), and ∣H∣ divides p. So it must be that ∣H∣=p, meaning that G and H are the same group. And since H was a cyclic group, G must be cyclic, too.
§ Normal Subgroups
How do cosets and homomorphisms relate?
Example. (Application of the Coset) Take a homomorphism f:G→H, and suppose that f(g1)=f(g2) for g1,g2∈G. What can we say about g1 and g2?
Solution: Recall the following result we derived earlier:
{f(g1)=f(g2)⟺g1=kg2 for some k∈ker(f)f(g1)=f(g2)⟺g1=g2k for some k∈ker(f)
We can now use cosets to rephrase the above statement as follows:
{f(g1)=f(g2)⟺g1 and g2 are in the same right coset of ker(f).f(g1)=f(g2)⟺g1 and g2 are in the same left coset of ker(f).(★)
The above motivates the following definition:
Definition (Normal). A subgroup N⊆G is normal if gN=Ng for all g∈G.
This is because the above result (★) immediately implies the following theorem:
Theorem. (Normal Kernel) For a homomorphism f:G→H, its kernel K=ker(f) is a normal subgroup of G.
Example. Consider G=S3 and a subgroup N={e,(12)}. Then:
The left cosets of N are N={e,(12)}, (13)N={(13),(123)}, and (23)N={(23),(132)}.
The right cosets of N are N={e,(12)}, N(13)={(13),(132)}, and N(23)={(23),(123)}.
Thus, N is not normal, since (e.g.) N(13)=(13)N. In other words, the partitioning of G formed by the left cosets of N is different from the partitioning of G formed by the right cosets of N.
Remark. By corollary, there is no homomorphism f:G→H to any other group H such that ker(f)=N.
Example. If G is abelian, then every subgroup N⊆G is normal. This is because:
gN={gn:n∈N}={ng:n∈N}=Ng.
Example. If N is a subgroup of G such that [G:N]=2, then N must be normal. This is because any partitioning of G into cosets—whether they be left cosets or right cosets—must look like G=N⊔(G∖N). So the (two) left cosets of N must match the (two) right cosets of N.
§ Conjugation
Definition (Conjugation). Given elements a,b∈G, the conjugation of a by b is defined to be bab−1. More generally, if H is a subgroup of G, then the conjugation of H by a is aHa−1.
It turns out (easy to check) that if H is a subgroup, then aHa−1 is also always a subgroup. Thus, we can think of conjugation by g∈G as an action that sends subgroups of G to subgroups of G.
Theorem. (Normality by Conjugation Condition) A subgroup N⊆G is normal if and only if N does not change under conjugation by any element g∈G.
Proof: Recall that N is normal if and only if gN=Ng for all g∈G. Equivalently, gNg−1=N for all g∈G, which may directly be read as: the conjugation of N by g equals N for any g∈G. ■
Theorem. (Intersecting Normal Subgroups) If N1 and N2 are normal subgroups of G, then their intersection N1∩N2 is also normal.
Proof: We argue using the formulation of normality in terms of conjugation described above. For any g∈G,
g(N1∩N2)g−1===={gng−1:n∈N1∩N2}{gng−1:n∈N1}∩{gng−1:n∈N2}(gN1g−1)∩(gN2g−1)N1∩N2 since N1 and N2 are normal.
Thus, N1∩N2 remains unchanged by conjugation by g, so N1∩N2 is normal. ■
§ Correspondence Theorem
Here's a useful theorem that we won't prove today.
Theorem. (Correspondence Theorem) Let f:G→G′ be a surjective homomorphism, and let K=ker(f) be its kernel. Consider the following collections of subgroups:
A:={subgroups of G containing K} and B:={subgroups of G′}.
Then there is a bijection from A to B that sends subgroups H∈A to subgroups f(H):={f(h):h∈H}∈B.
Example. (Correspondence Theorem Application #1) Consider the determinant homomorphism f=det:G→G′, with G=GLn(R) and G′=R×. Then K=ker(f)=SLn(R). What does the Correspondence Theorem say?
Solution: The set B:={subgroups of R×} is easier to analyze; for example, Q× is a subgroup in B.
By the Correspondence Theorem, the subgroup Q× must correspond to a new, quirky subgroup of GLn(R) containing K=SLn(R) that we didn't know of before. This subgroup is:
f−1(Q×)={M∈GLn(R):detM∈Q×}∈A.
There are many, many other subgroups of GLn(R) that live between K and G.
Example. (Correspondence Theorem Application #2) Take G=Z and G′=Cn=⟨g⟩, and define the homomorphism f:G→G′ by f(k)=gk. What does the Correspondence Theorem say?
Solution: Observe that ker(f)=nZ. Then:
A:={subgroups of G containing K}={dZ:d∣n}.
It then follows that we can classify the set of all subgroups of G′=Cn:
B:={subgroups of G′}=f(A)={f(dZ):d∣n}={⟨gd⟩:d∣n}.
So in fact, this tells us every subgroup of Cn is of the form ⟨gd⟩ for some d∣n.