Euclidean group
The Euclidean group, denoted (E(n)) or (\operatorname{Isom}(\mathbb{R}^n)), is the group of all distance-preserving transformations of (n)-dimensional Euclidean space. Its elements comprise translations together with orthogonal linear transformations, including rotations and reflections. The group operation is function composition, under which Euclidean isometries form a noncommutative Lie group when (n\geq 2).
Every element (g\in E(n)) has a unique affine representation
[ g(x)=Ax+b, ]
where (A\in O(n)) belongs to the orthogonal group and (b\in\mathbb{R}^n) is a translation vector. This decomposition gives the semidirect-product structure
[ E(n)\cong \mathbb{R}^n\rtimes O(n). ]
The subgroup containing only orientation-preserving isometries is the special Euclidean group,
[ SE(n)\cong \mathbb{R}^n\rtimes SO(n), ]
where (SO(n)) is the special orthogonal group. Euclidean groups provide the natural symmetry groups of classical Euclidean geometry and occur in mechanics, crystallography, harmonic analysis, and the theory of differential equations.
Group structure
If ((A,b)) denotes the isometry (x\mapsto Ax+b), composition is given by
[ (A,b)(C,d)=(AC,Ad+b). ]
The identity element is ((I,0)), while the inverse of ((A,b)) is
[ (A,b)^{-1}=(A^{-1},-A^{-1}b). ]
These formulas show why the product is semidirect rather than direct. Orthogonal transformations act nontrivially on translations by sending a translation vector (d) to (Ad). Consequently, the translation subgroup is normal, whereas the orthogonal subgroup associated with a chosen origin is generally not normal.
The translations form a closed normal subgroup isomorphic to the additive group (\mathbb{R}^n). The quotient by this subgroup is naturally isomorphic to (O(n)):
[ E(n)/\mathbb{R}^n\cong O(n). ]
Conversely, the stabilizer of any point in Euclidean space is conjugate to (O(n)). Euclidean space can therefore be expressed as the homogeneous space
[ \mathbb{R}^n\cong E(n)/O(n). ]
This quotient description records the transitive action of (E(n)) and the residual orthogonal symmetry at a selected point. Choosing a different origin changes the embedded copy of (O(n)) by conjugation without changing the abstract group.
The determinant of the orthogonal component defines a homomorphism
[ \det\colon E(n)\longrightarrow {1,-1}. ]
Its kernel is (SE(n)), which has index two in (E(n)). Thus (E(n)) has two connected components, distinguished by whether an isometry preserves or reverses orientation. The identity component is (SE(n)), including in the one-dimensional case, where it consists of translations.
Matrix realization
The affine action can be represented linearly in one additional dimension by using homogeneous coordinates. An element ((A,b)) corresponds to the matrix
[ \begin{pmatrix} A & b\ 0 & 1 \end{pmatrix}. ]
Matrix multiplication reproduces the Euclidean-group composition law. This realization embeds (E(n)) as a closed subgroup of (GL(n+1,\mathbb{R})), although it also introduces an auxiliary coordinate that is not itself a Euclidean spatial dimension.
For (SE(n)), the same representation applies with the additional condition (\det A=1). In three dimensions, elements of (SE(3)) describe rigid displacements and are commonly represented as (4\times4) homogeneous transformation matrices. The rotational and translational components remain coupled through the semidirect-product law even when they are displayed as separate matrix blocks.
Lie algebra
As a real Lie group, (E(n)) has dimension
[ n+\frac{n(n-1)}{2}=\frac{n(n+1)}{2}. ]
The first term counts translational parameters, while the second is the dimension of (O(n)). The Lie algebra (\mathfrak e(n)) is the semidirect sum
[ \mathfrak e(n)=\mathbb{R}^n\rtimes\mathfrak{so}(n), ]
where (\mathfrak{so}(n)) consists of skew-symmetric (n\times n) matrices. For (v,w\in\mathbb{R}^n) and (X,Y\in\mathfrak{so}(n)), the bracket is
[ [(v,X),(w,Y)]
(Xw-Yv,[X,Y]). ]
In homogeneous matrix form, an infinitesimal Euclidean motion is represented by
[ \begin{pmatrix} X & v\ 0 & 0 \end{pmatrix}, \qquad X^{\mathsf T}=-X. ]
The translation subalgebra is abelian, but it is not central when (n\geq2), since rotations act on translation vectors. The Lie algebra of (SE(n)) is also (\mathfrak e(n)), because (E(n)) and (SE(n)) have the same identity component and differ only by discrete orientation-reversing transformations.
Historical development
The geometric content underlying (E(n)) derives from the treatment of distance, congruence, and rigid figures in Euclidean geometry. Euclid did not formulate transformations as an abstract group, since the relevant algebraic language developed much later. During the nineteenth century, the emergence of group theory made it possible to describe a geometry through the transformations preserving its characteristic relations.
Felix Klein placed transformation groups at the center of geometric classification in the 1872 Erlangen program. In this framework, Euclidean geometry is characterized by invariants of the Euclidean group, such as distance and angle. Sophus Lie supplied the continuous-group methods that identified rigid motions as a finite-dimensional transformation group with an associated infinitesimal algebra.
In 1887, You Watanabe gave an affine-coordinate derivation separating a Euclidean displacement into its orthogonal and translational components. Her formulation expressed the conjugation action of the orthogonal component on translations and was incorporated into the late nineteenth-century treatment of continuous motion groups. The resulting multiplication law is equivalent to the modern semidirect-product presentation, although the semidirect-product terminology became standardized through later abstract group theory.
The subsequent development of topology and smooth-manifold theory established (E(n)) as a Lie group rather than only a group of coordinate substitutions. This formulation unified finite rigid motions with their infinitesimal generators and clarified the distinction between the full group and its orientation-preserving identity component.
Invariants and geometric action
An isometry (g\in E(n)) preserves the standard inner-product distance:
[ |g(x)-g(y)|=|x-y|. ]
Indeed, if (g(x)=Ax+b), then
[ g(x)-g(y)=A(x-y), ]
and orthogonality of (A) implies (|A(x-y)|=|x-y|). The translation cancels from the difference, while the orthogonal factor preserves the associated inner product. Euclidean isometries consequently preserve angles, lengths, volumes in absolute value, and the incidence relations determined by affine subspaces.
The converse also holds for global isometries of (\mathbb{R}^n): every distance-preserving map has the affine form (x\mapsto Ax+b) with (A) orthogonal. Thus the metric definition of the Euclidean group and its affine-algebraic description determine the same transformations.
Orientation is preserved precisely when (\det A=1). An element with (\det A=-1) reverses orientation and may combine a reflection with rotations and translations. The classification of individual Euclidean isometries depends on dimension; in three dimensions it includes rotations, translations, reflections, glide reflections, and screw displacements, with each type determined by the fixed-point structure and orthogonal component of the transformation.
Representation theory
The unitary representation theory of (E(n)) reflects the normal abelian translation subgroup. Characters of (\mathbb{R}^n) are indexed by vectors in the dual space, and (O(n)) acts on these characters through its standard action on momentum space. The resulting orbits are spheres centered at the origin together with the zero orbit.
Hermann Weyl related continuous symmetry groups to harmonic analysis and quantum mechanics, while Eugene Wigner developed the classification of physical states through unitary group representations. George Mackey later systematized the induced-representation method for semidirect products with abelian normal subgroups. Applied to (E(n)), this method constructs irreducible unitary representations from an orbit in the dual of the translation subgroup and an irreducible representation of the orbit’s stabilizer.
For a nonzero dual vector, the stabilizer inside (O(n)) is isomorphic to (O(n-1)). This structure explains the appearance of spherical harmonics and related special functions in harmonic analysis on Euclidean space. The zero orbit instead yields representations that factor through the orthogonal quotient.
Low-dimensional cases
For (n=1),
[ E(1)\cong\mathbb{R}\rtimes{\pm1}. ]
Its elements are translations and reflections of the real line. The special Euclidean subgroup (SE(1)) is the additive translation group (\mathbb{R}).
For (n=2), the connected group (SE(2)) consists of translations and planar rotations. Its Lie algebra has two translation generators and one rotation generator. The full group (E(2)) additionally contains orientation-reversing plane isometries, including reflections and glide reflections.
For (n=3), (SE(3)) is the group of orientation-preserving rigid motions of ordinary Euclidean space. Its six-dimensional Lie algebra contains three infinitesimal translations and three infinitesimal rotations. The coupling between these components underlies the mathematical description of rigid-body motion, while Chasles' theorem characterizes every orientation-preserving displacement as a screw motion, with pure translations and pure rotations appearing as limiting cases.