Core Structure Theory
Directly Indecomposable Algebras
Algebras that admit no non-trivial direct decomposition, the congruence-lattice condition characterising them, and the limits of unique factorisation.
Learning objectives
- Define directly indecomposable and characterise it via factor congruences
- Identify indecomposable algebras in familiar varieties
- State what does and does not hold about uniqueness of decomposition
Definition and criterion
An algebra A with more than one element is directly indecomposable if whenever A ≅ B × C, one of B, C is trivial.
So indecomposability is a statement purely about Con(A) — specifically about which of its elements are complemented and permuting.
Simple implies indecomposable
A simple algebra has only Δ and ∇ as congruences at all, so certainly only those as factor congruences. Every simple algebra is therefore directly indecomposable.
Directly indecomposable algebras are far more common than simple ones. The cyclic group of order 4 has three congruences — so it is not simple — but it does not decompose as a product, because its congruence lattice is a chain and a chain has no non-trivial complemented elements.
Examples
| Algebra | Indecomposable? | Reason |
|---|---|---|
| Cyclic group of prime order | Yes | Simple |
| Cyclic group of order pn | Yes | Congruence lattice is a chain |
| Cyclic group of order 6 | No | ≅ C2 × C3 |
| The two-element Boolean algebra | Yes | Simple |
| A finite Boolean algebra of size 2n | No | ≅ 2n for n > 1 |
| Any chain as a lattice | Yes | Congruence lattice has no non-trivial complements |
| A field | Yes | Simple as a ring |
The pattern for cyclic groups is the Chinese remainder theorem in disguise: Cn decomposes exactly according to the prime factorisation of n, and the prime power factors are indecomposable.
Unique factorisation
One would like every algebra to be a product of indecomposables in an essentially unique way. The situation is more delicate.
What holds
For finite algebras in congruence-modular varieties, decomposition into directly indecomposable factors exists and is unique up to isomorphism and reordering — a version of the Krull–Schmidt theorem.
What fails without modularity
Uniqueness can fail. Examples exist of finite algebras with two genuinely different decompositions into indecomposables.
What fails in the infinite case
Existence can fail: an infinite algebra need not be a product of indecomposables at all, since the decomposition process need not terminate.
Because direct decomposition is both rare and badly behaved, the subject relies instead on subdirect decomposition. Birkhoff's theorem guarantees that every algebra is a subdirect product of subdirectly irreducible algebras — with no hypotheses at all. That universality is why the next stream begins there.
Boolean products as the repair
Boolean products, developed in Chapter IV §8, sit between direct and subdirect products. They retain enough of the direct product's structure to support a representation theory, while being general enough to exist widely.
Frequently asked questions
Is every finite algebra a product of indecomposables?
Yes — the decomposition process terminates by finiteness. What can fail without congruence-modularity is uniqueness of the resulting factors.
How does this relate to the Krull–Schmidt theorem?
Krull–Schmidt for groups and modules is the congruence-modular case. The universal-algebraic version identifies modularity as the hypothesis that makes the classical argument work, which explains why it holds for groups and modules but not in general.
Source. S. Burris and H. P. Sankappanavar, A Course in Universal Algebra, The Millennium Edition — a corrected re-typesetting of Springer-Verlag Graduate Texts in Mathematics 78 (1981). Section II.7, book pages 58-61.
This page is an original exposition prepared for the KEVOS® knowledge library. It restates and reorganises mathematical results; it is not a reproduction of the source text.