What Stays the Same

Rotate a square by ninety degrees and you have the same square. Reflect it in a diagonal and you have the same square. Leave it alone and you have the same square, which counts too.

That last one sounds like a joke and is the key to the whole subject. Once you stop thinking of symmetry as a visual property (balanced, pleasing, the same on both sides) and start thinking of it as the *set of operations that leave a thing unchanged*, the concept becomes something you can compute with, and it applies to objects with no shape at all.

The operations have structure. Do one, then another, and the result is again a symmetry of the object. One of them does nothing. Every one can be undone. Those conditions, plus the fact that bracketing does not matter when you compose three, define a group.

That is the whole definition, and its emptiness is the point. Nothing in it mentions squares, or space, or geometry. So anything whose symmetries you can identify (a crystal, a molecule, an equation, a physical law, a wallpaper pattern) hands you a group, and facts about groups become facts about all of them at once.

The first great application was a negative one. There are formulas for solving quadratics, cubics and quartics by radicals: write the roots in terms of the coefficients using arithmetic and root extraction. For three hundred years people looked for the quintic formula.

Abel proved in 1824 that none exists. Évariste Galois, a few years later and before dying in a duel at twenty, explained why. To each polynomial he attached a group describing the symmetries among its roots: the ways they can be permuted without disturbing any algebraic relation between them. Solvability by radicals corresponds exactly to a structural property of that group, and the group of the general quintic does not have it.

This is worth restating, because it is easy to miss what kind of result it is. It does not say quintics have no solutions; they have five. It says that no expression built from the coefficients using arithmetic and roots can name them. The impossibility is proved by translating a question about equations into a question about symmetry, and answering it there.

The classification results came next. There are exactly seventeen ways to arrange a pattern repeating in two directions across a plane: seventeen wallpaper groups, enumerated by Fedorov in 1891. Every wallpaper, every tiled floor, every Islamic geometric pattern is one of them. It is often claimed that the Alhambra contains all seventeen; people who have gone and counted dispute it. The seventeen is not in dispute, because it is a theorem: there is no eighteenth to be discovered by a sufficiently inventive designer.

In three dimensions the same programme gives the 230 crystallographic space groups, which is why crystallography is a classification science, and why a diffraction pattern can be read as a statement about symmetry.

And then physics, where the idea turned out to be foundational rather than decorative.

In 1918 Emmy Noether proved that every continuous symmetry of a physical system corresponds to a conserved quantity. If the laws do not change over time, energy is conserved. If they do not change from place to place, momentum is conserved. If they do not change with orientation, angular momentum is conserved.

Before this, conservation laws were empirical regularities: things that had always been observed to hold. Afterwards they were consequences of structure, derivable from the statement that the laws have no preferred moment, position or direction. Noether proved it at Hilbert's request, to sort out a problem about energy conservation in general relativity, at a university that would not give her a paid post because she was a woman; Hilbert had to lecture under his own name with her doing the teaching.

Modern particle physics is built on this. The standard model is specified by naming its symmetry groups, particles are classified by how they transform under them, and forces arise from requiring symmetries to hold locally rather than globally. In 1962 Murray Gell-Mann arranged the known particles into a symmetry pattern, noticed a gap, and predicted a particle with specified properties to fill it. The omega-minus was found two years later, with those properties.

A definition about rotating squares, pushed hard enough, ends up predicting particles.