Artin-Wedderburn theorem

Imagine you are handed a complicated ring with the lovely property that everything splits — every module breaks cleanly into simple pieces. You might expect a vast zoo of such rings. The Artin-Wedderburn theorem delivers the opposite, beautiful surprise: there is essentially only one recipe. Every such ring is a short product of matrix algebras over division rings, and nothing else. It is the classification theorem that organizes the entire subject.

Precisely: a ring R is semisimple if and only if it is isomorphic to a finite direct product M_{n_1}(D_1) × ... × M_{n_r}(D_r), where each D_i is a division ring and each n_i ≥ 1. The data (r; n_1, ..., n_r; D_1, ..., D_r) is unique up to reordering the factors and isomorphism of the division rings. For a simple Artinian ring r = 1, so R is a single matrix ring M_n(D).

The proof is a model of structural reasoning: a semisimple R decomposes as a direct sum of its minimal left ideals, these group into isotypic components, Schur's lemma shows endomorphisms of a simple module form a division ring, and the double-centralizer pattern realizes R as matrices over that division ring. The numbers n_i are multiplicities of the simple modules and the D_i are their endomorphism rings.

The reach is enormous. Over an algebraically closed field every finite-dimensional division algebra is the field itself, so a semisimple algebra is just a product of matrix algebras M_{n_i}(k); this is the algebraic skeleton behind the character theory of finite groups, where the dimensions n_i are the degrees of the irreducible representations and the relation sum of n_i^2 equals the dimension recovers the order of the group.