Group Actions

Group actions are the connective tissue between abstract algebra and almost every field that studies symmetry. In pure algebra, the Orbit-Stabilizer Theorem and the class equation (built from the conjugation action) are the direct engines of the Sylow Theorems, which classify subgroups of prime-power order in any finite group and remain the standard tool for analyzing finite group structure. In geometry, actions of isometry groups on Euclidean space or on tilings classify wallpaper patterns and crystallographic groups, directly informing crystallography and materials science, where the 230 space groups arise exactly as actions of symmetry groups on $\mathbb{R}^3$. In combinatorics, Burnside's lemma (counting orbits via fixed points) is the standard tool for counting configurations up to symmetry — necklaces, graph colorings, and molecular isomers — turning "how many essentially different objects" questions into averages of fixed-point counts. In physics, group actions on state spaces encode conserved quantities and selection rules, from the rotation group acting on angular momentum states in quantum mechanics to gauge groups acting on field configurations. In cryptography, actions of groups on sets (such as elliptic curve point groups acting via scalar multiplication) underlie discrete-log-based protocols, and isogeny-based post-quantum schemes explicitly exploit group actions on isogeny classes. In computer science, actions of permutation groups on data structures are used in symmetry-breaking algorithms for constraint satisfaction and in canonical labeling for graph isomorphism testing.

Ask the tutor: How does the Orbit-Stabilizer Theorem turn the numeric pattern $|\mathrm{Orb}(x)|\cdot|G_x| = |G|$ seen in these examples into a general proof, and how is that identity then used to derive the Sylow theorems?

Picture a set $X$ as a collection of dots and $G$ as a machine that scrambles the dots around, one permutation for every group element. The two axioms are exactly what you need for "scrambling" to behave sensibly: doing nothing ($e$) should really do nothing, and scrambling by $g$ then by $h$ should be the same as scrambling by the single combined motion $hg$ applied in the order the axiom specifies (first $h$, then $g$, matching function composition read right to left). This is why an action is really just a way of letting $G$ live inside the symmetric group of $X$: each $g$ becomes an actual bijection $X \to X$, and multiplying group elements corresponds to composing those bijections. Once you have this picture, orbits and stabilizers become geometric objects you can see. Take $X$ to be the vertices of a square and $G$ the group of rotations and reflections of the square (order 8). Every vertex can be moved to every other vertex, so there is a single orbit consisting of all 4 vertices; the stabilizer of a single vertex is the subgroup of order 2 fixing it (identity and the reflection through the diagonal at that vertex). The orbit is big, the stabilizer is small, and their sizes multiply to the size of the group — this is the phenomenon the Orbit-Stabilizer Theorem will name precisely.

The heart of the theory is that a group action is nothing more, and nothing less, than a homomorphism into a symmetric group. This translation lets us bring the entire machinery of Group Homomorphisms to bear on actions.

Example 1: A non-symmetric action of $S_3$ on the vertices of an asymmetric labeled configuration. Let $G = S_3$, the symmetric group on $\{1,2,3\}$, and let $X = \{1,2,3\}$ with $G$ acting by $\sigma \cdot i := \sigma(i)$. This is a group action: identity holds since the identity permutation fixes every point, and compatibility holds since $(\sigma\tau)(i) = \sigma(\tau(i))$ is exactly how permutation composition is defined. Now enlarge the picture: let $X' = \{1,2,3\} \times \{a,b\}$, a 6-element set, with $S_3$ acting only on the first coordinate: $\sigma\cdot(i,c) = (\sigma(i), c)$. This is still a genuine action (the axioms are inherited immediately from the action on $\{1,2,3\}$), but now let's compute orbits and stabilizers for a specific asymmetric subgroup rather than all of $S_3$, to see the theory at work on a non-trivial case. Take $H = \langle (1\,2\,3) \rangle \leq S_3$, the cyclic subgroup of order 3 generated by the 3-cycle $(1\,2\,3)$, acting on $X'$ by restriction of the above action. Compute the orbit of the point $(1,a)$:

So $\mathrm{Orb}(1,a) = \{(1,a),(2,a),(3,a)\}$, an orbit of size 3. The stabilizer $H_{(1,a)}$ consists of those $h \in H$ with $h\cdot(1,a) = (1,a)$; from the computation above only $e$ fixes $(1,a)$, so $H_{(1,a)} = \{e\}$, of size 1. Note $|\mathrm{Orb}(1,a)| \cdot |H_{(1,a)}| = 3 \cdot 1 = 3 = |H|$ — a numerical instance of the Orbit-Stabilizer phenomenon. By contrast, consider the point $(1,a)$ under the full group $S_3$ acting on $X'$. Now the orbit is still $\{(1,a),(2,a),(3,a)\}$ (the second coordinate never changes under this action), size 3, but the stabilizer in $S_3$ is

of size 2. Check: $3 \cdot 2 = 6 = |S_3|$. This example shows the same orbit arising under two different acting groups, with correspondingly different stabilizers, while the product of orbit size and stabilizer size always recovers the order of whichever group is acting — never a coincidence, always forced by the theory.

Example 2: Conjugation action and counting a specific fixed point set. Let $G$ be any group and let $G$ act on itself by conjugation: $g \cdot x := gxg^{-1}$. This is an action: $e\cdot x = exe^{-1}=x$, and $g\cdot(h\cdot x) = g(hxh^{-1})g^{-1} = (gh)x(gh)^{-1} = (gh)\cdot x$. Take the concrete case $G = D_4 = \langle r, s \mid r^4 = s^2 = e,\ srs = r^{-1}\rangle$, the dihedral group of order 8 (symmetries of a square), and compute the orbit and stabilizer of the specific element $x = r$ under conjugation. The orbit of $r$ is its conjugacy class: $\{grg^{-1} : g \in D_4\}$. We compute directly on the 8 elements $e, r, r^2, r^3, s, sr, sr^2, sr^3$:

since powers of $r$ commute with $r$. For reflections, using $srs^{-1} = srs = r^{-1} = r^3$ (from the relation $srs=r^{-1}$, noting $s^{-1}=s$):

One checks similarly $sr^k \cdot r \cdot (sr^k)^{-1} = r^3$ for each $k$, since every reflection conjugates $r$ to $r^{-1}$. So the orbit is

The stabilizer of $r$ under conjugation is $(D_4)_r = \{g \in D_4 : grg^{-1} = r\}$, i.e. the centralizer of $r$. From the computation, $e, r, r^2, r^3$ all centralize $r$ (they commute with every power of $r$), while none of the four reflections do (each sends $r \mapsto r^3 \ne r$). So

Check: $|\mathrm{Orb}(r)|\cdot|(D_4)_r| = 2\cdot 4 = 8 = |D_4|$, again matching the group order exactly. This particular computation — the size of a conjugacy class times the size of a centralizer equals $|G|$ — is precisely the identity that powers the class equation and, ultimately, the proof strategy behind the Sylow Theorems.