Introduction
Upon having gotten to know groups a little better, it seems only natural to ask how we can know whether two groups are fundamentally different. In what contexts do two groups have completely distinctive structures and in what contexts do two groups in fact have identical structures? A step in the right direction in order to answer these questions is to explore what are known as homomorphisms.

Let G and H be groups. A group homomorphism from G to H is a function \alpha:G \rightarrow H such that \alpha(g_1g_2)=\alpha(g_1)\alpha(g_2), \forall g_1,g_2 \in G. Notice that on the left hand side, the operation acting on the two elements is the operation from G while on the right hand side the operation acting on the two elements is the operation from H. Hence, if there exists a homomorphism between two groups G and H and the group operations are \ast_G and \ast_H respectively, it essentially means that whether we apply \ast_G on two elements in G or move the elements to H and then apply \ast_H on them, the result will be the same.

We define the kernel of \alpha as the set ker(\alpha)=\{g \in G: \alpha(g)=e_H\}.

Examples
In order to establish a homomorphism from the positive real numbers under multiplication to the real numbers under addition, we define \alpha(x):= log(x) and notice that the function \alpha:\mathbb{R^{+}} \rightarrow \mathbb{R} satisfies the equation \alpha(ab)=\alpha(a)+\alpha(b), \forall a,b \in \mathbb{R^{+}}. It is quite clear that ker(\alpha)=1 as log(1)=0

In order to establish a homomorphism between GL(2,\mathbb{Q}) and the set \mathbb{Q} \backslash\{0\} under multiplication, we define \alpha(x):=det(x) and notice that the function \alpha:GL(2,\mathbb{Q}) \rightarrow \mathbb{Q} \backslash\{0\} satisfies the equation \alpha(ab)=\alpha(a)\alpha(b), \forall a,b \in GL(2,\mathbb{Q}) since the determinant is multiplicative. Note that GL(2,\mathbb{Q}) denotes the group of invertible two by two matrices with rational entries under matrix multiplication. For this homomorphism, the kernel is the set of all matrices in GL(2,\mathbb{Q}) with determinant 1.

Standard Results
Having hopefully now gotten a bit of a feel for how homomorphisms work, I’ll elucidate some simple results about homomorphisms and kernels.

Proposition: Let \alpha:G \rightarrow H be a homomorphism, and g \in G. Then:

(1) \alpha(e)=e
(2) \alpha(g^{n})=(\alpha(g))^{n}, \forall n \in \mathbb{Z}
(3)|\alpha(g)| divides |g|, assuming |g|<\infty

Proof: (1) We know that \alpha(e)=\alpha(ee)=\alpha(e)\alpha(e). After cancelling, the result follows.

(2) For n>0, note that \alpha(g^{n})=\alpha(gg...g)=\alpha(g)\alpha(g)...\alpha(g)=(\alpha(g))^{n}. In the case n=0, we simply use the first thing we showed: \alpha(e)=e. In the case n=-1, we notice that \alpha(g)\alpha(g^{-1})=\alpha(gg^{-1})=\alpha(e)=e and similarly, \alpha(g^{-1})\alpha(g)=\alpha(g^{-1}g)=\alpha(e)=e. It follows that \alpha(g^{-1})=\(\(\alpha(g))^{-1}. Using what we have shown thus far, the case where n<-1 follows immediately.

(3) This follows from the second result.Letting |g|=m, we have that (\alpha(g))^{m}=\alpha(g^{m})=\alpha(e)=e and hence |\alpha(g)| divides m. \quad \blacksquare

Now for some standard results about kernels.

Proposition: Let \alpha:G \rightarrow H be a homomorphism. Then:

1. ker(\alpha) \unlhd G
2. \alpha is injective if and only if ker(\alpha)=\{e\}

Proof: (1) Let K=ker(\alpha). We’ll begin by showing that K is a subgroup of G. As shown earlier, \alpha(e)=e and so e \in K. Suppose k_1,k_2 \in K. \alpha(k_1k_2)=\alpha(k_1)\alpha(k_2)=ee=e and hence k_1k_2 \in K Also, \alpha(k_1^{-1})=\alpha(k))^{-1}=e^{-1}=e thus k_1^{-1} \in K. Hence K is a subgroup. Now we just have to show that it’s normal. If k \in K and g \in G then \alpha(g^{-1})\alpha(k)\alpha(g)=\alpha(g^{-1})\alpha(g)=(\alpha(g))^{-1}\alpha(g)=e. Thus g^{-1}kg \in K and hence K is normal.

(2) If \alpha is injective then e can only be mapped to one element and given that \alpha(e)=e, it must be that ker(\alpha)=\{e\}. Conversely, suppose that \ker(\alpha)=\{e\}. If \alpha(g_1)=\alpha(g_2) then \alpha(g_1)(\alpha(g_2))^{-1}=e. But this means that \alpha(g_1g_2^{-1})=e \implies g_1g_2^{-1} \in K=\{e\}. Thus g_1=g_2 and so \alpha is injective. \quad \blacksquare

Now for some more definitions. Let \alpha:G \rightarrow H be a homomorphism. If L \leq G then the image of L is \alpha(L)=\{\alpha(l): l \in L\}. If M \leq H then the preimage (or inverse image) of M is \alpha^{-1}(M)=\{g \in G: \alpha(g) \in M\}. Although we use the notation \alpha^{-1}(M), bear in mind that this doesn’t imply that the function \alpha is necessarily invertible.

Proposition: Let \alpha: G \rightarrow H be a homomorphism.

1. L \leq G \implies \alpha(L) \leq H
2. L \unlhd G \implies \alpha(L) \unlhd H
3. L is cyclic \implies \(\alpha(L) is cyclic
4. L is abelian \implies \alpha(L) is abelian
5. M \leq H \implies \alpha(M) \leq G
6. M \unlhd H \implies \alpha^{-1}(M) \unlhd G

Proof: (1) We know that e \in L so e=\alpha(e) \in \alpha(L). If \alpha(l_1),\alpha(l_2) \in \alpha(L), then \alpha(l_1)\alpha(l_2)=\alpha(l_1l_2) \in \alpha(L) since l_1,l_2 \in L. Lastly, (\alpha(l))^{-1}=\alpha(l_1^{-1})) \in \alpha(L) since l_1^{-1} \in L.

(2) If l \in L, g \in G, then (\alpha(g))^{-1}\alpha(l)\alpha(g)=\alpha(g^{-1}lg) \in \alpha(L) as g^{-1}lg \in L

(3) If L= \langle k \rangle then \forall \alpha(l) \in \alpha(L), l=k^{m} for some m \in \mathbb{Z}. Then \alpha(l)=\alpha(k^{m})=(\alpha(k))^{m}= \langle \alpha(k) \rangle

(4) If l_1,l_2 \in L then \alpha(l_1)\alpha(l_2)=\alpha(l_1l_2)=\alpha(l_2l_1)=\alpha(l_2)\alpha(l_1)

(5) We know that \alpha(e)=e \in M, and thus e \in \alpha^{-1}(M). If g_1,g_2 \in \alpha^{-1}(M), then \alpha(g_1g_2)=\alpha(g_1)\alpha(g_2) \in M since \alpha(g_1),\alpha(g_2) \in M. Hence g_1g_2 \in \alpha^{-1}(M). Also, \alpha(g_1^{-1})=(\alpha(g_1))^{-1} \in M since \alpha(g_1) \in M. Hence g_1^{-1} \in \alpha^{-1}(M)

(6) If a \in \alpha^{-1}(M), g \in G, then \alpha(g^{-1}ag)=(\alpha(g))^{-1}\alpha(a)\alpha(g) \in M since \alpha(a) \in M) and M is normal. Hence g^{-1}ag \in \alpha^{-1}(M) \quad \blacksquare

What’s next?
While homomorphisms are useful in elucidating whether two groups have similar structure or not, the existence of a homomorphism between two groups can’t alone quite tell us whether two groups have truly identical structure. For that, we require the notion of an isomorphism, which is a bijective homomorphism. There is a lot about isomorphisms to explore, and so they will be covered across other posts.

Trending