Cyclic Groups and Generators

Cyclic groups are the computational backbone of modern cryptography: the multiplicative group of nonzero residues mod a large prime $p$, and elliptic curve groups, are used precisely because they contain large cyclic subgroups in which discrete exponentiation is easy to compute but "discrete logarithm" (finding $k$ given $a^k$) is believed to be hard — the basis of Diffie–Hellman key exchange and ElGamal encryption. In number theory, the group $(\mathbb{Z}/n\mathbb{Z})^\times$ of units modulo $n$ is cyclic exactly when $n \in \{1,2,4, p^k, 2p^k\}$ for an odd prime $p$; this single fact (Gauss) underlies primitive root theory and is proved using exactly the machinery of generators and orders developed here. In physics and crystallography, cyclic groups of rotations ($C_n$ symmetry) classify molecules and crystal lattices with $n$-fold rotational symmetry; the order of the generating rotation is the number of distinct orientations before the pattern repeats, precisely $|\langle a \rangle|$ in the sense of this article. In computer science, cyclic redundancy checks, hash table probing sequences, and pseudo-random number generators (linear congruential generators) all exploit the wraparound structure of $\mathbb{Z}/n\mathbb{Z}$ as an additive cyclic group, where the "period" of the generator is exactly the order of its step size.

Ask the tutor: How does knowing $|\langle a \rangle|$ equals the order of $a$ lead directly to the classification of all cyclic groups as either $\mathbb{Z}$ or $\mathbb{Z}/n\mathbb{Z}$?

Think of an element $a$ as a single move — a step forward. Composing $a$ with itself, over and over, traces out a path: $e, a, a^2, a^3, \dots$ Since groups also allow inversion, you can retrace your steps backward: $a^{-1}, a^{-2}, \dots$ The set of every place you can possibly reach, forward or backward, from any number of applications of the move $a$, is exactly $\langle a \rangle$. Two qualitatively different pictures can emerge. If the path never returns to where it started, you get an infinite line of distinct points, indexed by all of $\mathbb{Z}$ — this is the case of infinite order, and $\langle a\rangle$ behaves exactly like the integers under addition. If instead the path eventually loops back to the identity after $n$ steps, the powers wrap around into a cycle of length $n$: $a^n = e$, so $a^{n+1} = a$, and the whole orbit collapses onto a clock with $n$ positions. This is why the order of $a$ (the smallest positive $n$ with $a^n = e$, or $\infty$ if no such $n$ exists) governs everything about $\langle a \rangle$: it is not just a number describing $a$, it is literally the size of the group $a$ generates.

Example 1: The cyclic subgroup generated by $10$ in $\mathbb{Z}/24\mathbb{Z}$. Consider $G = \mathbb{Z}/24\mathbb{Z}$ under addition, with identity $0$. Take $a = 10$. Since the operation is addition, "powers" of $a$ become integer multiples of $a$: $\langle 10 \rangle = \{k \cdot 10 \bmod 24 : k \in \mathbb{Z}\}$. First we find the order of $10$: the least positive $n$ with $10n \equiv 0 \pmod{24}$. We have

Since $\gcd(10,24) = 2$, the order is $n = 24/2 = 12$. Now list the multiples of $10$ mod $24$ for $k = 0, 1, \dots, 11$, by repeatedly adding $10$ and reducing mod $24$:

(and then $14 + 10 = 24 \equiv 0$, closing the cycle, confirming order $12$ as predicted by part (3) above). So

exactly the even residues mod $24$ — a set of size $12$, matching the theorem's formula $|\langle a \rangle| = n$ exactly. Notice this coincides with $\langle 2 \rangle$ (the subgroup generated by $2$), illustrating that different elements can generate the same cyclic subgroup; here both $10$ and $2$ generate the unique subgroup of order $12$ inside $\mathbb{Z}/24\mathbb{Z}$.

Example 2: A cyclic subgroup inside a permutation group. Let $G = S_5$, the symmetric group on $\{1,2,3,4,5\}$, and let $\sigma = (1\,3\,5)(2\,4)$, a product of a disjoint $3$-cycle and a $2$-cycle. Recall (from the theory of the order of a permutation) that the order of a permutation written as disjoint cycles is the least common multiple of the cycle lengths. Here the cycle lengths are $3$ and $2$, so

By part (3) of the theorem, $\langle \sigma \rangle$ has exactly $6$ elements: $e, \sigma, \sigma^2, \sigma^3, \sigma^4, \sigma^5$. Let's compute them explicitly, tracking where each element of $\{1,\dots,5\}$ goes, using $\sigma: 1\to3\to5\to1$ and $2\to4\to2$.

• $\sigma^1 = (1\,3\,5)(2\,4)$
• $\sigma^2$: apply $\sigma$ twice. The $3$-cycle becomes $(1\,5\,3)$ (since $1\to5\to1$... let's verify: $\sigma(\sigma(1)) = \sigma(3) = 5$, $\sigma(\sigma(3))=\sigma(5)=1$, $\sigma(\sigma(5))=\sigma(1)=3$, so on $\{1,3,5\}$, $\sigma^2$ sends $1\to5,\ 5\to3,\ 3\to1$, i.e. $\sigma^2|_{\{1,3,5\}} = (1\,5\,3)$). The $2$-cycle squared is the identity on $\{2,4\}$ since $(2\,4)^2 = e$. So $\sigma^2 = (1\,5\,3)$.
• $\sigma^3$: the $3$-cycle part cubed is identity ($(1\,3\,5)^3 = e$); the $2$-cycle part cubed is $(2\,4)^3 = (2\,4)^2(2\,4) = (2\,4)$. So $\sigma^3 = (2\,4)$.
• $\sigma^4 = \sigma^3 \sigma = (2\,4)(1\,3\,5)(2\,4) = (1\,3\,5)$ acting on top of $(2\,4)^2=e$: concretely $\sigma^4 = \sigma^3\cdot\sigma = (1\,3\,5)$ (the $3$-part cycles once more, the $2$-part returns to identity since $(2\,4)^4 = e$).
• $\sigma^5 = \sigma^4\sigma = (1\,3\,5)(1\,3\,5)(2\,4) = (1\,5\,3)(2\,4)$.
• $\sigma^6 = e$, confirming order $6$.

So explicitly,

a set of exactly $6$ permutations, matching $|\sigma| = 6$ from the theorem — and this subgroup sits inside $S_5$ (order $120$), illustrating that a cyclic subgroup can be a small, tightly structured piece of a much larger and highly non-abelian ambient group.