Symmetries of space
How are rigid transformations organized?
What does it mean for a space to have symmetry?
Take the Euclidean plane and imagine three familiar motions: translate everything two units to the right, rotate everything by a quarter turn around the origin, or reflect everything across a line. Each motion preserves distance and angle. Each is rigid.
Now perform two of them in succession.
Translate, then rotate. Reflect, then translate. Rotate, then rotate again. Every time, the result is still a rigid transformation of the same space.
That is the next structural fact geometry needs. Rigid motions form a system closed under composition. This is where symmetry stops being a loose visual intuition and becomes an organized mathematical object.
A symmetry of a geometric space is a rigid transformation of that space into itself. It preserves exactly the spatial structure the space carries. If the space is Euclidean, that means distance, angle, and therefore shape. The key point becomes organization.
If one rigid motion preserves the space and a second rigid motion preserves it as well, then doing one after the other still preserves the space. There is also an identity motion that leaves every point where it is. And every rigid motion can be undone by another rigid motion of the same space.
Those three facts are exactly the pattern we have already met in a different setting:
- composition
- identity
- inverse
So the rigid transformations of a space assemble into a group.
(space)
V
/|\
rot refl trans
\|/
V
(identity)
Read the diagram as a compressed picture of allowable motions acting on one and the same space. Rotations, reflections, and translations belong to one coherent world of rigid transformations, and that world has an identity and is closed under composition.
This reintroduces an earlier structural idea in a new role. In algebra, groups arose as reversible transformations preserving some form or operation. In geometry, the same pattern returns as the organization of rigid motions of space itself.
That return matters because besides features that remain invariant geometry is also about the full family of motions under which those invariants are preserved.
Different spaces therefore have different symmetry groups.
A square in the plane has only finitely many rigid self-motions that preserve it as the same placed figure: quarter turns, a half turn, the identity, and certain reflections. The whole Euclidean plane has far more: every translation, every rotation around every point, every reflection across every line, and compositions of these.
So symmetry is always relative to the space or figure under discussion.
The plane itself has a rich symmetry group because many rigid motions preserve the whole space. A specific figure inside the plane usually has fewer symmetries, because preserving that figure is a stricter requirement than preserving the surrounding space.
This also clarifies the difference between a rigid motion and a symmetry of a given object. A rigid motion is any distance-preserving transformation of the ambient space. It counts as a symmetry of a figure only when it sends that figure to itself.
For a square, a translation is rigid but not a symmetry of the square as a placed figure. A quarter turn about its center is both.
That distinction is worth keeping sharp. Geometry studies spaces and figures through their allowable motions, and the symmetry group records which of those motions leave the object under study unchanged.
Once the group of symmetries is visible, new invariants become available. Some spaces are highly symmetric, others far less so. Some figures admit reflections, others only rotations. Some have continuous families of symmetries, others only finitely many. The symmetry group becomes a structural fingerprint of the object.
At the same time, the picture is still conceptual rather than computational. We now know which motions belong to the geometry and how they organize, but we have not yet given ourselves a convenient numerical language for describing points, lines, curves, or the transformations acting on them. The structure is clear, while calculation is still awkward.
That is exactly the pressure that leads to analytic geometry. Coordinates do not create geometric space. They give us a way to describe it, compute inside it, and express its symmetries numerically without confusing the representation with the geometry itself.
The structural move here can be stated plainly. The objects are geometric spaces and figures within them, while the allowed morphisms are rigid transformations that preserve the full spatial structure. What composes is rigid motion, and composition remains inside the same world because isometries compose and invert. The invariants are the geometric features already identified, together with the symmetry group that records the full organization of allowable self-motions. The defining relation is still distance preservation, but now applied to endomorphisms of one space so that the resulting motions form a group. Coordinates and equations are still absent.
What becomes possible once we assign numbers to space?
And what changes under a change of coordinates while the geometry itself stays the same?