Core Universal Algebra
Homomorphisms and the Isomorphism Theorems
The three isomorphism theorems and the correspondence theorem are proved once here, in arbitrary type, and never need proving again for any specific structure.
- Define homomorphism, embedding, epimorphism and isomorphism for arbitrary type.
- Prove the first isomorphism theorem from the kernel correspondence.
- State the second and third isomorphism theorems in congruence language.
- Apply the correspondence theorem to intervals in Con A.
- Recognise the congruence extension property and where it fails.
01Homomorphisms in arbitrary type
A map α : A → B between algebras of the same type is a homomorphism when it commutes with every operation.
| Name | Condition | Note |
|---|---|---|
| Homomorphism | commutes with operations | the base notion |
| Embedding | injective homomorphism | image is a subalgebra isomorphic to source |
| Epimorphism | surjective homomorphism | image is all of B |
| Isomorphism | bijective homomorphism | inverse is automatically a homomorphism |
| Endomorphism | homomorphism A → A | forms a monoid under composition |
| Automorphism | isomorphism A → A | forms a group |
Unlike lattices under monotone maps, a bijective homomorphism of algebras is always an isomorphism: the inverse map automatically commutes with the operations. The lattice pathology arose because monotone maps are not homomorphisms.
02The first isomorphism theorem
- input: homomorphism α : A → B
- compute θ := ker(α) = { ⟨a,b⟩ : α(a) = α(b) } — a congruence on A
- form the quotient A/θ and the natural map ν : A → A/θ
- define β : A/θ → B by β(a/θ) := α(a)
- well defined: a/θ = b/θ means α(a) = α(b)
- β is an injective homomorphism with image α(A)
- therefore A/ker(α) ≅ α(A), and α = β ∘ ν
The factorisation is unique up to isomorphism and is the reason the congruence lattice controls the homomorphic images. Homomorphic images of A correspond exactly to congruences on A, which is the H in HSP made concrete.
03The second and third theorems
The second theorem is the one that makes interval reasoning in Con A possible: it identifies the congruence lattice of A/θ with the interval [θ, ∇] in Con A. That identification is used constantly and is the content of the correspondence theorem.
04The correspondence theorem
- Congruences on the quotient correspond to congruences above θA congruence on A/θ pulls back to a congruence on A containing θ, and pushes forward again to the original.
- The correspondence is a lattice isomorphismMeets and joins are preserved in both directions, so the interval is a faithful copy of Con(A/θ).
- Consequence: simplicity of quotientsA/θ is simple exactly when θ is a coatom of Con A — a maximal congruence below ∇. This is how maximal congruences are recognised.
05The congruence extension property
An algebra has the congruence extension property (CEP) when every congruence on every subalgebra extends to a congruence on the whole algebra. A variety has CEP when all its members do.
Groups have it; semigroups do not. Where CEP fails, a congruence on a subalgebra can be an obstruction with no global counterpart, and arguments that pass congruences between a subalgebra and its parent break down. Check CEP before using such an argument.
CEP interacts with the classification: every congruence-distributive variety generated by a finite algebra has good behaviour here, and discriminator varieties have CEP outright. It is one of the standard hypotheses in the structure theory of the Boolean Constructions stream.
Frequently asked
Why is a bijective homomorphism automatically an isomorphism?
Because the inverse of a bijective homomorphism commutes with the operations: apply α to both sides of the required identity and use that α is injective and a homomorphism. The lattice-theoretic failure discussed elsewhere concerned monotone bijections, which are not homomorphisms.
Do the isomorphism theorems require any hypotheses on the type?
None at all. They hold for algebras of arbitrary type, which is precisely the point — proving them once here removes the need to prove them separately for every structure. This is the clearest single dividend of the universal-algebraic approach.
Does the correspondence theorem hold for subalgebras too?
Not in the same clean form. There is no general order isomorphism between Sub(A/θ) and an interval in Sub(A); the third isomorphism theorem gives an embedding rather than an isomorphism. The asymmetry between congruences and subalgebras runs throughout the subject.
- S. Burris and H. P. Sankappanavar, A Course in Universal Algebra, Millennium Edition (a corrected re-typesetting of Springer GTM 78, 1981).
- G. Grätzer, Universal Algebra, 2nd edition, Springer.
- R. McKenzie, G. McNulty and W. Taylor, Algebras, Lattices, Varieties, Volume I.
Original KEVOS® explanatory article. Written from the topic map of the cited works; no text is reproduced from them.
