Core Universal Algebra

The Congruence Lattice Con A

Con A is the central invariant of the subject. Constraining its shape across a variety is what the Mal'cev programme does, and almost every later theorem is such a constraint.

Engineering · Mathematics4 min readKV-MATH-0214

Learning objectives

Show Con A is a complete algebraic sublattice of Eq(A).
Read simplicity and subdirect irreducibility off the lattice.
Identify the monolith of a subdirectly irreducible algebra.
State the three principal congruence conditions on varieties.
Explain why Con A rather than Sub(A) carries the classification.

01Con A is complete and algebraic

Arbitrary intersections of congruences are congruences, giving completeness by the one-sided criterion. The generation operator Θ is finitary, giving algebraicity. The compact elements are the finitely generated congruences, and every congruence is the join of the principal congruences it contains.

Structure of Con A
Feature	Value	Note
Bottom	Δ	The identity relation; quotient is A itself.
Top	∇	The all relation; quotient is trivial.
Meet	intersection	Computed pointwise; cheap.
Join	Θ(union)	Alternating composites unless permutable.
Compacts	finitely generated congruences	Joins of principal congruences.
Position	complete sublattice of Eq(A)	Meets and joins agree with Eq(A).

Key resultCon A is a complete sublattice of Eq(A), not merely a subset

Both meets and joins computed in Eq(A) of families of congruences are again congruences. So no discrepancy arises between the two lattices, and one may compute in whichever is convenient. This is not automatic for sublattices in general and is worth checking rather than assuming.

02Reading structure off the lattice

Two elements

Simple

Con A = {Δ, ∇}. The only quotients are A and the trivial algebra. Equivalently every non-trivial principal congruence is ∇.

Unique atom

Subdirectly irreducible

Con A has a least non-trivial congruence, the monolith μ. Every non-trivial congruence contains μ.

Otherwise

Subdirectly reducible

Δ is the intersection of two or more incomparable non-trivial congruences, and A decomposes as a subdirect product.

Simple implies subdirectly irreducible, since ∇ is then the unique atom. The converse fails: the four-element Boolean algebra is subdirectly irreducible in the variety of Boolean algebras only in the two-element case, and there are many subdirectly irreducible algebras that are not simple in other varieties.

03The monolith

When A is subdirectly irreducible the least non-trivial congruence is called the monolith. It is the intersection of all non-trivial principal congruences and is itself principal, being Θ(a, b) for any pair generating it.

μ = ⋂ { θ ∈ Con A : θ ≠ Δ } and μ ≠ Δ

Subdirect irreducibility is exactly the assertion that this intersection is non-trivial. If it equals Δ, the algebra is a subdirect product of proper quotients.

The monolith is the handle by which subdirectly irreducible algebras are studied. Bounding the size of subdirectly irreducibles in a variety — a recurring theme in the finite basis theorems — typically proceeds by bounding what the monolith can look like.

04Classifying varieties by congruence conditions

The three principal conditions
Condition	Requirement on Con A	Mal'cev characterisation	Examples
Congruence-permutable	θ ∘ φ = φ ∘ θ for all θ, φ	ternary p with p(x,y,y) ≈ x, p(x,x,y) ≈ y	groups, rings, modules, quasigroups
Congruence-distributive	Con A distributive	Jónsson terms d₀,…,dₙ	lattices, Boolean algebras, Heyting algebras
Congruence-modular	Con A modular	Day terms	groups, rings, modules, and all of the above
Arithmetical	permutable and distributive	Pixley term	Boolean algebras, discriminator varieties

Permutable implies modular; distributive implies modular; arithmetical is the conjunction of permutable and distributive. Each condition on the lattice turns out to be equivalent to the existence of certain terms, which converts an infinitary lattice condition into a finite syntactic check — this is the whole point of the Mal'cev programme and is developed in the Equational stream.

05Why Con A carries the classification

Quotients are what vary
Two varieties may have similar subalgebra behaviour and wildly different quotient behaviour. Quotient behaviour is the discriminating invariant.
Congruence conditions are Mal'cev definable
Each is equivalent to a term condition, hence preserved by H, S and P, hence a property of the variety rather than of individual algebras.
Structure theorems take them as hypotheses
Jónsson's lemma requires distributivity. The commutator theory requires modularity. Discriminator varieties require arithmeticity. Nothing comparable is built on Sub(A).
Decidability tracks them too
The known decidability results for locally finite varieties are stated in terms of congruence conditions and the commutator, not in terms of subalgebras.

Frequently asked

Can Con A be any algebraic lattice?

For arbitrary algebras, yes — that is the Grätzer–Schmidt theorem. For finite algebras the question of which finite lattices arise as Con A of a finite algebra is much harder and was a major open problem well beyond the source's 1981 vintage.

If Con A is distributive, is A in a congruence-distributive variety?

Not necessarily. Congruence-distributivity of a variety requires every member to have a distributive congruence lattice, and a single algebra can have a distributive Con A while generating a variety containing algebras that do not. The Mal'cev characterisation is what makes the variety-level condition checkable.

What is the practical cost of non-permutability?

Joins become expensive. In a permutable setting θ ∨ φ = θ ∘ φ, a single relational composition. Without permutability the join is a transitive closure with unbounded chain length, so computing in Con A for lattices or semigroups is materially harder than for groups.

Sources and further reading

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.