The Orbit-Stabilizer Theorem: Counting Symmetries

The precise statement

Let a group G act on a set X — meaning there is a map G × X → X, written (g, x) ↦ g·x, satisfying e·x = x and g·(h·x) = (gh)·x for all g, h ∈ G. Fix a point x ∈ X. Two objects attach to it:

• The orbit Orbit(x) = { g·x : g ∈ G } ⊆ X, the set of all places x can be sent.
• The stabilizer Stab(x) = { g ∈ G : g·x = x }, which is a subgroup of G.

Theorem (Orbit-Stabilizer). There is a natural bijection between the orbit of x and the set G/Stab(x) of left cosets of its stabilizer. Hence |Orbit(x)| = [G : Stab(x)]. When G is finite, Lagrange gives the division form |Orbit(x)| = |G| / |Stab(x)|. In particular orbit size always divides |G|.

The picture: trading motion for rigidity

Think of G as a bag of symmetries and x as a marked spot you can shove around. The orbit is the shadow x sweeps out; the stabilizer is the redundancy — the symmetries that move everything else but leave x exactly where it was. The theorem says these two quantities are exactly complementary: the more symmetries fix x, the fewer distinct places it can reach, and their product is a constant, |G|.

Concretely, imagine rotating a square (dihedral group, |G| = 8) and watching one corner. It can land on any of the 4 corners, so the orbit has size 4. That forces the stabilizer to have size 8/4 = 2 — indeed the identity and the reflection through that corner's diagonal fix it. You never had to hunt for the stabilizer; the orbit size told you it had 2 elements. This complementarity — freedom of motion inversely proportional to rigidity — is the whole theorem in one sentence.

The key idea: one clean bijection

The engine is a single well-defined map. Write H = Stab(x). Define φ : G/H → Orbit(x) by φ(gH) = g·x.

• Well-defined. If gH = g′H then g′ = gh for some h ∈ H, so g′·x = g·(h·x) = g·x. The output doesn't depend on the coset representative.
• Injective. If g·x = g′·x, then (g⁻¹g′)·x = x, so g⁻¹g′ ∈ H, i.e. gH = g′H. This step is exactly where the subgroup property of H is used.
• Surjective. Every orbit element is g·x = φ(gH) by definition.

A bijection between a set of cosets and the orbit gives equality of cardinalities — and it works whether or not G is finite. The finite division formula is then just Lagrange's theorem applied to H ≤ G. That is the entire proof; its elegance is that it constructs, rather than merely counts.

Worked example: the rotations of a cube

Let G be the group of rotational symmetries of a cube, acting on X = the 6 faces. Pick the top face x.

• Orbit. Any face can be rotated to the top, so Orbit(x) is all 6 faces: |Orbit(x)| = 6.
• Stabilizer. Rotations fixing the top face are the spins about the vertical axis: 0°, 90°, 180°, 270°. So |Stab(x)| = 4.

Orbit-Stabilizer then reads off |G| = 6 × 4 = 24 — no enumeration needed. Cross-check with a different action: G also acts on the 8 vertices. One vertex's orbit is all 8 vertices, and its stabilizer is the three rotations (0°, 120°, 240°) about the long diagonal through it, size 3. Again |G| = 8 × 3 = 24. The same |G| emerges from any transitive action, factored differently — a robust sanity check that also proves the cube's rotation group is isomorphic to S₄.

Why the hypotheses matter, and how it connects

The bijection form needs only that G acts on X — no finiteness required. What finiteness buys is the division statement |Orbit(x)| = |G|/|Stab(x)|. Drop it and this can fail: let G = ℤ act on X = ℝ by n·t = t + n. The orbit of 0 is ℤ (countably infinite), the stabilizer is trivial, and "∞/1" is not the useful arithmetic statement — only the coset bijection [G : Stab] survives, and it correctly gives an infinite index. The one indispensable structural fact is that Stab(x) really is a subgroup; injectivity of φ collapses without it.

Connections. Take G acting on itself by left multiplication: every stabilizer is trivial, so every orbit has size |G| — that is the regular representation. Take the action on cosets of a subgroup H and you recover Lagrange's theorem. Take conjugation and orbits become conjugacy classes, yielding the class equation — the springboard for Cauchy's and the Sylow theorems.

Applications and significance

Orbit-Stabilizer is the load-bearing beam under a surprising amount of algebra and combinatorics:

• Counting symmetric objects. Combined with the averaging step it becomes Burnside's lemma (#orbits = average number of fixed points), the workhorse for counting distinct colorings of a necklace, cube, or molecule up to symmetry, and the foundation of Pólya enumeration.
• Structure of finite groups. The class equation |G| = |Z(G)| + ∑ [G : C(xᵢ)] proves that every p-group has nontrivial center, drives Cauchy's theorem and the Sylow theorems, and underlies the classification effort.
• Geometry and beyond. Identifying a homogeneous space as G/Stab — the sphere as SO(3)/SO(2), for instance — is the orbit-stabilizer theorem for Lie group actions, the entry point to representation theory and the study of symmetric spaces.

Its significance is conceptual: it is the precise dictionary translating geometry (orbits, where things can go) into algebra (subgroups and indices, how many symmetries do what).

Orbit-Stabilizer and its close relatives in the counting toolkit

Result	What it counts	Precise statement	Relationship
Orbit-Stabilizer	Size of one orbit	|Orbit(x)| = [G : Stab(x)]	The base identity
Lagrange's theorem	Cosets of a subgroup	|G| = [G:H]·|H|	Special case: G acts on G/H by left mult., Stab = H
Class equation	Conjugacy classes	|G| = |Z(G)| + ∑ [G : C_G(xᵢ)]	Orbit-Stab applied to G acting on itself by conjugation
Burnside's lemma	Number of orbits	#orbits = (1/|G|) ∑_{g∈G} |Fix(g)|	Sums Orbit-Stab counts to average fixed points
Cauchy's theorem	Elements of order p	p | |G| ⇒ ∃ element of order p	Proved via a ℤ/p action + orbit sizes ∈ {1, p}