1

I do recognize that this problem from Dummit&Foote Abstract Algebra has been posted before. However the full completed proof was not shown and there was debate regarding the correctness of the definition in the hint. I am hoping that my attempt can be verified.

previous post

PROOF

Let G be a finite group of even order.

Notice that for an element $g$ of $G$, if $g=g^{-1}$, then $g^2=gg=gg^{-1}=e=1$. We have shown previously that this holds if order of $g$, $\vert g\vert$, is $1$ or $2$. Also, we know that only the identity of a group has order $1$.

Define $t(G)=\{g\in G \vert g\ne g^{-1} \}$

By definition, $G$ contains a unique identity $e=1$. Further $e=e^{-1}$, so $e\notin t(G)$.

If $t(G)$ is empty, as $G$ has even order, it follows that $\exists g\in G$, distinct from $e$. Since $g=g^{-1}$ and $g\ne e$, $\vert g\vert = 2$.

Assume $t(G)$ is nonempty. By definition a group contains the inverse of each element. Thus $\forall x\in t(G)$, $x^{-1}\in t(G)$. Hence $t(G)$ contains an even number of elements. But $e\in G-t(G)$. Therefore for $G$ to have even order, $\exists g\in G$, distinct from $e$, such that $g\notin t(G)$. As above $\vert g\vert = 2$.

Thus a finite group of even order contains an element of order 2.

RJM
  • 1,109
  • 1
    Yes, this is correct. You're pairing up elements with their inverses. This partitions $G$ into subsets of size $2$ and subsets of size $1$, the latter of which must satisfy $g^2=e$. But there is at least one element, $e$, in a singleton, so because $G$ has even order, there must be a second singleton, which gives you an element of order $2$. – Robert Shore May 11 '21 at 02:53
  • @Robert Shore thank you. Out of curiosity, could your statement in your comment be considered enough for a proof? – RJM May 11 '21 at 02:57
  • 1
    I think so. You might want to show explicitly that breaking $G$ up into subsets as described is indeed a partition. You can define a relation $gRh \iff (g=h \lor g=h^{-1})$ and then show that $R$ as defined is in fact an equivalence relation, which means its equivalence classes partition $G$. – Robert Shore May 11 '21 at 03:03
  • @Robert Shore Excellent. I appreciate your help. I know my posts are simple in relation to others, but sometimes in isolation something seems right, when it is really not. – RJM May 11 '21 at 03:07
  • Why don't you apply Lagrange's theorem directly? – GHR01 May 11 '21 at 15:23

0 Answers0