Let $n \in \mathbb {N}$ and $G$ a group of order $4n+2$. Prove $G$ isn't simple.

129 Views Asked by At

I have finished writing the proof and I'm afraid I made a mistake on the way because I'm not sure I can deduce the bottom line of the proof.

Here it goes:

Let us examine the action of $G \mapsto G$ on itself by multiplication from the left.

This action induces a homomorphism $\phi: G \to S_{|G|}$. From Cayley's theorem we have $G \cong Im\phi \le S_{4n+2}$.

Let us mark $H = Im\phi$. We know $A_{4n+2} \lhd S_{4n+2}$ and from Diamond theorem we have $H \cap A_{4n+2} \lhd H$.

Let us mark $H \cap A_{4n+2} = N$. We want to show $N \lt H$ and also $N$ isn't trivial.

From Cauchy's theorem for $p=2$: exists $g\in H$ such that $|g| = 2$.

Let us assume $g$ is constructed by foreign cycles marked $c_i$ for $1 \le i \le m$ for some $m \in \mathbb N$.

Then we have $g = c_1 ... c_m$. We know $|g| = {\rm lcm}(c_i)$ and since $|g| = 2$ we have $\phi(g)$ is a product of 2-cycles.

Notice $|g|=2 \implies g \neq e$ and therefore for all $g^{'} \in G$ we have $gg^{'} \neq g$ meaning $\phi(g)$ doesn't "leave" any object in it's place.

Therefore we have $\phi(g)$ is a product of exactly $2n+1$ 2-cycles and this implies $\phi(g) \notin A_{4n+2}$ and therefore $N \lt A_{4n+2}$.

Now notice, for all $n \in \mathbb N$ exists some $p>2$ primal (and obviously odd) such that we have some $e \neq g \in H$ such that $|g| = p$.

In this case we have $\phi(g)$ is a product of oddsized-cycles and therefore $\phi(g) \in A_{4n+2}$ and therefore $N$ isn't trivial.

This implies $N \lhd H$ a proper subgroup.

Now this is the part I'm not sure I'm allowed to deduce

Since we know $G \cong H$ we have some $H^{'} \lt G$ such that $H^{'} \cong N$ therefore meaning $H^{'} \lhd G$ a proper subgroup, thus implying $G$ isn't simple.

Another problem

I feel like I'm proving a stronger theorem since I could've used all the statements about $4n+2$ for every even number? or maybe I'm just mixed up. Thanks in Advance!

1

There are 1 best solutions below

3
On BEST ANSWER

To react to your remark "Is the current proof all right now"? First of all, the Cayley embedding induces a homomorphism $\phi: G \to S_{|G|}$ (and not to $S_{|G\backslash H|}$). This is an injection so the First Isomorphism Theorem guarantees that $G \cong \phi(G)=H$. Now, as you noted $H \cap A_{4n+2} \unlhd H$. Assume that we have equality here, then $H \cap A_{4n+2} = H$, which is equivalent to $H \subseteq A_{4n+2}$.

You then observe rightly that by Cauchy $G$ has an element of order $2$, hence $\phi(g)$ is an element of order $2$. Then you describe $\phi(g)$ correctly: it is a product of $2n+1$ transpositions ($2$-cycles), hence an odd permutation, so $\phi(g) \notin A_{4n+2}$, a contradiction. We conclude that $H \cap A_{4n+2}$ is a proper normal subgroup of $H$. Your argument for being non-trivial is correct. The part you are not sure about is also correct. The point is that you again can use the First Isomorphism Theorem and look at $\phi^{-1}(H \cap A_{4n+2})$, being a non-trivial normal subgroup of $G$ (remember that the map $\phi: G \rightarrow \phi(G)$ is an isomorphism!)

So what did you learn: the whole motivation behind the proof is embedding/representing $G$ in another group where you have more tools or control to study $G$. This happens more oftenly in mathematics: represent your structure in something bigger where more tools are available: think of calculating difficult real integrals by interpreting them over $\mathbb{C}$. Later you might hit representation theory of finite groups where you represent them for example in the linear groups $GL(n,\mathbb{C})$ and suddenly you can apply linear algebra, Galois theory and (algebraic) number theory to prove nice things. Burnside used this last "trick" to prove that any group of order $p^aq^b$ ($p$, $q$ primes) is solvable. No need for the sledge hammer of Feit-Thompson for $p, q$ odd (which did not yet exist at that time!)