Klein’s Erlangen Program: How Groups Define Geometry

Felix Klein’s Erlangen program defines a geometry as a set paired with a group of transformations that preserve “sameness” of figures on that set. Different choices of transformation group applied to the same underlying set, such as the Euclidean plane, produce different geometries (Euclidean, similarity, affine, topological) with different invariants. Any homogeneous geometry can also be described abstractly as a quotient space G/H, where G is a group and H is the stabilizer subgroup of a point.

Key Takeaways

• A geometry is defined by a set X together with a group of transformations acting on X, not by the set alone.
• Congruent triangles are related by Euclidean motions (rotations, reflections, translations); similar triangles additionally allow uniform scaling.
• The symmetry group of a cube acting on its vertices has 48 elements and is isomorphic to S_4 × Z_2.
• Moving from Euclidean geometry to similarity geometry to affine geometry to topology progressively discards invariants: length, then angle, then straightness, leaving only properties like dimension.
• Any transitive group action of G on a geometry X can be recovered as the coset space G/H, where H is the stabilizer of a chosen basepoint.
• For two-dimensional Euclidean geometry, the plane R² is isomorphic to the coset space E(2)/O(2), where E(2) is the Euclidean motion group and O(2) is the orthogonal group.

What Is Klein’s Erlangen Program?

Klein’s Erlangen program establishes a link between group theory and geometry that is often left out of introductory group-theory courses despite being conceptually accessible. The program’s central idea is that a geometry is not just a set of points but a set equipped with a group of transformations that determine which figures count as “the same.”

How Can One Set Support Multiple Geometries?

Two-dimensional Euclidean geometry, the geometry taught in high school, actually contains several distinct geometries within the same ambient set R². Which geometry is in play depends entirely on which transformations are declared to preserve sameness between figures.

Triangles illustrate the distinction clearly. Congruent triangles can be cut from paper and placed exactly on top of one another: they share equal angles and equal side lengths. Similar triangles share the same shape but not necessarily the same size: angles still agree, but side lengths are proportional rather than equal.

How Is a Geometry Formally Defined by a Transformation Group?

A geometry specifies a relation on subsets of an underlying set indicating whether two figures are “the same,” and this relation is encoded by a set of allowable transformations. Two triangles are congruent, for example, if there exists a transformation T: R² → R² carrying one triangle onto the other, where T is built from reflections, rotations, translations, or combinations of these. Similarity geometry additionally allows transformations that uniformly shrink or expand the plane.

Formally, given a set X equipped with a collection of allowable transformations, a subset A of X is called isomorphic to a subset B if some allowable transformation T satisfies T(A) = B. Three natural requirements constrain which transformations may be allowed:

• Every set A must be isomorphic to itself, which holds if the identity map on X is included among the allowable transformations.
• If A is isomorphic to B, then B must be isomorphic to A, which requires every allowable transformation to be invertible with its inverse also allowable.
• If A is isomorphic to B and B is isomorphic to C, then A must be isomorphic to C, which requires the composition of two allowable transformations to remain allowable.

These three requirements are exactly the axioms of a group, so a geometry can be defined as a set X equipped with a group of transformations X → X. Geometries are often required to act transitively on X, meaning that for any two points x and y in X there exists some transformation T in the group with T(x) = y.

What Are the Common Examples of Geometries?

Common geometries defined by their underlying set and transformation group

Geometry	Underlying set	Transformation group / form	Notes
Euclidean geometry	R²	T(x) = Ax + b, with A in O(2,R)	The Mazur–Ulam theorem shows distance-preserving bijections of R² must be affine of this form; the full motion group is the semidirect product of O(2,R) with the translation group R².
Similarity geometry	R² (or C)	T(x) = αAx + b, with A in O(2,R), α ≠ 0; equivalently T(z) = az + b over the complex numbers	Adds uniform scaling to the Euclidean motions.
Affine geometry	R^n	T(x) = Ax + b, with A in GL(n,R)	Allows any bijection sending straight lines to straight lines; the fundamental theorem of affine geometry shows such bijections must take this form.
Other geometries	Varies (hyperbolic/spherical plane, vertices of a solid)	Möbius transformations; finite symmetry groups	The symmetry group of a cube acting on its vertices is S_4 × Z_2, with 48 elements.

What Invariants Survive Under Each Geometry?

Different geometries on the same underlying set preserve different quantities, called invariants, and comparing these invariants clarifies how the geometries relate to one another. Euclidean geometry preserves length, angles, and perpendicularity. Similarity geometry drops length as an invariant but keeps angles. Affine geometry drops angles but keeps parallel lines and ratios of lengths measured along a single line.

At the extreme end of this progression, equipping R² with the full group of homeomorphisms (continuous bijections with continuous inverses) produces topology, in which very few invariants survive but dimension remains meaningful. Topology is sometimes described as the largest interesting geometry because its transformation group is so permissive that almost no finer structure is preserved.

How Do Groups and Subgroups Encode a Geometry as G/H?

A geometry can also be constructed starting only from a group G and a subgroup H, without first specifying a set of points. Given such a pair, the set of left cosets X = G/H becomes the geometry’s underlying set, and each element g in G defines a transformation T_g on X by T_g(aH) = gaH, so that G acts transitively on X. The subgroup H is the isotropy group, meaning the stabilizer, of the coset H itself.

The construction also runs in reverse. If a group G acts transitively on a geometry X, choosing any point p in X and letting H be the set of transformations in G that fix p shows that X is naturally isomorphic to G/H. This means all of the geometric information is encoded in the pair (G, H) alone. For two-dimensional Euclidean geometry, G is the Euclidean motion group E(2) and H is the stabilizer O(2), giving the identification R² ≅ E(2)/O(2). A triangle viewed as a geometry on three points can likewise be described by a corresponding pair of a group and subgroup.

Where Does the Erlangen Program Lead Next?

Viewing geometries through their transformation groups makes it possible to compare geometries directly by comparing their groups, and to extend the framework by allowing curvature and additional structure, producing what are known as Cartan–Cayley–Klein geometries. The unifying claim of the Erlangen program is that a geometry is determined by its group of allowable transformations, and that many classical geometries can be represented as homogeneous spaces G/H under this single framework.

Frequently Asked Questions

What is the Erlangen program in simple terms?

The Erlangen program, introduced by Felix Klein, defines a geometry as a set of points together with a group of transformations that preserve which figures count as “the same.” Different transformation groups applied to the same set of points yield different geometries, such as Euclidean, similarity, or affine geometry.

What is the difference between congruent and similar triangles?

Congruent triangles have equal angles and equal side lengths and can be placed exactly on top of one another after applying a Euclidean motion (rotation, reflection, or translation). Similar triangles share the same angles and proportional side lengths but are related by transformations that also allow uniform scaling.

What is a homogeneous space G/H?

A homogeneous space G/H is the set of left cosets of a subgroup H inside a group G, where G acts transitively on the coset set. Any geometry on which a group G acts transitively can be identified with G/H, where H is the stabilizer subgroup of a chosen point.

Why is topology called the largest interesting geometry?

Topology arises when the plane R² is equipped with the full group of homeomorphisms, which is a far larger transformation group than the ones used for Euclidean, similarity, or affine geometry. Because so many transformations are allowed, very few invariants survive, though dimension remains meaningful, making topology the coarsest geometry still considered informative.

How many elements are in the symmetry group of a cube?

The symmetry group of a cube acting on its vertices has 48 elements and is isomorphic to S_4 × Z_2, the product of the symmetric group on four elements and the cyclic group of order two.

What is the isotropy group in the G/H construction?

The isotropy group, also called the stabilizer, is the subgroup H consisting of transformations in G that fix a chosen basepoint. For two-dimensional Euclidean geometry, the isotropy group of the origin under the Euclidean motion group E(2) is the orthogonal group O(2).