Permutation groups are not just one example among many: every group is one, after an isomorphism. This is Cayley's theorem, and it is the reason permutations are a universal model for symmetry.
Why it works
Proof. For $g \in G$ define $\lambda_{g}: G \to G$ by $\lambda_{g}(x) = gx$.
(1) $\lambda_{g}$ is a permutation: it has the inverse $\lambda_{g^{-1}}$, since $g^{-1}(gx) = x$.
(2) $g \mapsto \lambda_{g}$ is a homomorphism: $\lambda_{gh}(x) = ghx = \lambda_{g}(\lambda_{h}(x))$, so $\lambda_{gh} = \lambda_{g} \circ \lambda_{h}$.
(3) It is one-to-one: if $\lambda_{g}$ is the identity, then $g = ge = \lambda_{g}(e) = e$, so the kernel is trivial.
So $G \cong \{\lambda_{g} : g \in G\}$, a subgroup of the permutations of $G$. $\square$
You have already seen these permutations: $\lambda_{g}$ is row $g$ of the Cayley table, read as a rearrangement of the column headings. That every row is a rearrangement was the cancellation law.
Worked example: A Cayley permutation in $U(10)$
- $\lambda_{3}(1) = 3$, $\lambda_{3}(3) = 9$, $\lambda_{3}(9) = 27 \equiv 7$, $\lambda_{3}(7) = 21 \equiv 1$.
- One cycle: $(1\;3\;9\;7)$, of length $4 = |3|$.
Answer: $(1\;3\;9\;7)$.
Worked example: The shape of every $\lambda_{g}$
- Following $x \to gx \to g^{2}x \to g^{3}x = x$ closes up after exactly $|g| = 3$ steps, and never sooner (if $g^{k}x = x$ then $g^{k} = e$).
- So every cycle has length $3$, and there are $12/3 = 4$ of them (they are the cosets $\langle g \rangle x$).
Answer: Four 3-cycles.
Worked example: Relabeling $\mathbb{Z}_{6}$
- $\lambda_{2}(x) = x + 2$: $0 \to 2 \to 4 \to 0$ and $1 \to 3 \to 5 \to 1$.
- With labels $k + 1$: $(1\;3\;5)(2\;4\;6)$.
Answer: $(1\;3\;5)(2\;4\;6)$.
Cayley's theorem is about existence, not economy: it puts a group of order $6$ inside $S_{6}$, which has $720$ elements, although $S_{3}$ is already a group of order $6$. Its value is conceptual. Any statement proved for all subgroups of symmetric groups holds for all finite groups, and any finite group can be explored by computer as a set of permutations.
Common mistake
Key idea