1

If $G$ is a finite group of order (size) $n$ then, for any $g \in G$, the order (period) of $g$ is a divisor of $n$.

Proof: $g$ must have finite order since $G$ is finite. If the order (period) of $g$ is $m$ then the order (size) of the cyclic subgroup $\left<g\right>$ it generates is also $m$. Since $G$ has a subgroup of size $m$, Lagrange's Theorem tells us that $m$ is a divisor of $n$.

What does this proof mean by "$g$ must have finite order since $G$ is finite". Firstly, are we talking about the order (period) or order (size) and either way why does $G$ being finite mean $g$ has finite order?

user2850514
  • 3,689
  • $g$ must have finite order since $G$ is finite means there must exist $n \in \mathbb N$ such that $g^n = e$ where $e$ is the identity of $G$. If $g$ did not have finite order, that would mean that there does not exist $k \in \mathbb N$ such that $g^k = e$, we write $\vert g \vert = \infty$ in that case. – Robert Cardona Dec 23 '13 at 20:20

4 Answers4

3

If $G$ is finite, and $g \in G$, look at the sequence $g^i$ for integer $i \ge 1$, i.e. $g^i = \{ g, g^2, . . . , g^k, g^{k + 1}, . . . , \text{etc.} \}$ Since $G$ is finite, this sequence of elements of $G$ must repeat itself at some point; thus we must have $g^l = g^k$ for $l > k \ge 1$. If we then choose $k$ to be the least integer for which such repetition occurs for some $l$, and $l$ to be the first integer after $k$ for which it does occur, we must have $g^{l - k} = e$, the identity of $G$, and $g^m \ne e$ for any $m < l - k$. The order of $g$ is thus $l - k$. Thus $G$ finite implies the order of any $g \in G$ finite, and now the argument based on Lagrange's theorem shows that the order of $g$ must divide that of $G$.

Hope this helps. Happy Holidays,

and as always,

Fiat Lux!!!

Robert Lewis
  • 71,180
2

For your first question: The period of $g$.

For your second question: Suppose for a contradiction that g does not have a finite period. Then it will generate an infinite number of elements, resulting in an group with infinite number of elements which is a contradiction to the supposition that $G$ is finite.

ireallydonknow
  • 1,415
  • 2
  • 14
  • 33
  • I refer to the size of the group as the order of the group $G$ and the period/order of the element in the group $G$ as the smallest number $n$ such that $g^n=1$ where 1 is the identity. An element of the group will not have a 'size' as it is not a set and does not contain anything. – ireallydonknow Dec 23 '13 at 20:37
  • 1
    @user2850514 $g$ is not assumed cyclic, but the group $\langle g\rangle$ it generates clearly is cyclic (i.e. it is generated by a single element, namely $g$) – Hagen von Eitzen Dec 23 '13 at 20:38
  • Ok, but then surely we are talking about a subgroup of $G$, namely $\left< g \right>$ for $g\in G$, rather than the actual group $G$? – user2850514 Dec 23 '13 at 20:41
  • Yes, since the key idea in the proof is the group $\left< g \right>$ generated by the element $g$. – ireallydonknow Dec 23 '13 at 20:43
  • Ok, but if we just consider the line "$g$ must have finite order since $G$ is finite", how does $G$ being finite make $g$ finite, at this point we have not introduced the cyclic subgroup yet. If $g$ has finite order then there is an $m \geq 1$ such that $g^{m} = e$, does this mean that each $g^{i}$ for $1 \leq i < m$ must be inside $G$? – user2850514 Dec 23 '13 at 20:51
  • I've proved in my answer that if the order of the group $G$ is finite and $g\in G$ then the order of the element $g$ must be finite. For your second question, yes, $\forall 1 \leq i < m$, $g^i \in G$, since a group is closed under its binary operation. – ireallydonknow Dec 23 '13 at 21:10