Terms, Free Algebras and Equational Logic

Mal'cev Conditions I: Congruence Permutability

A single ternary term decides whether congruences permute across an entire variety. Mal'cev's theorem converts an infinitary lattice condition into a finite syntactic check.

Engineering · Mathematics4 min readKV-MATH-0224

Learning objectives

State the Mal'cev term condition for congruence permutability.
Prove that a Mal'cev term implies permutability.
Exhibit Mal'cev terms for groups, rings, modules and quasigroups.
Show lattices have no Mal'cev term.
Explain what makes a condition a Mal'cev condition.
Relate permutability to modularity and to direct decomposition.

01The problem Mal'cev conditions solve

Congruence permutability is a condition on every pair of congruences on every member of a variety — a statement quantifying over a proper class. Checking it directly is impossible. Mal'cev's insight is that it is equivalent to the existence of a single term satisfying two identities.

Key resultMal'cev's theorem

A variety V is congruence-permutable if and only if there is a ternary term p in the language of V such that V satisfies p(x, y, y) ≈ x and p(x, x, y) ≈ y.

The right-hand condition is finite, syntactic and checkable. Since a variety is generated by any of its generating algebras, one can even search for the term in a single finite algebra.

02From the term to permutability

ProcedureA Mal'cev term forces congruences to permute

in: Mal'cev term p → out: θ ∘ φ = φ ∘ θ for all congruences

input: variety V with term p satisfying p(x,y,y) ≈ x and p(x,x,y) ≈ y
take A ∈ V, congruences θ, φ ∈ Con A, and ⟨a, c⟩ ∈ θ ∘ φ
so there is b with ⟨a,b⟩ ∈ θ and ⟨b,c⟩ ∈ φ
consider the element d := p(a, b, c)
⟨a, d⟩ = ⟨p(a,b,b), p(a,b,c)⟩ ∈ φ since ⟨b,c⟩ ∈ φ and p is compatible
(using p(a,b,b) ≈ a)
⟨d, c⟩ = ⟨p(a,b,c), p(c,c,c)⟩... more directly:
⟨d, c⟩ = ⟨p(a,b,c), p(b,b,c)⟩ ∈ θ since ⟨a,b⟩ ∈ θ
(using p(b,b,c) ≈ c)
so ⟨a, c⟩ ∈ φ ∘ θ; by symmetry θ ∘ φ = φ ∘ θ

The element p(a, b, c) is the 'detour' witnessing the reversed composite. For groups p(a,b,c) = a·b⁻¹·c and the construction is the familiar coset argument. Caveat: the converse direction — permutability implies a term — uses the free algebra on three generators and is the harder half.

The converse is proved by working in F_V(x, y, z): permutability of two specific congruences on the free algebra produces exactly the required element, which is a term in three variables. This pattern — deduce the term from the free algebra — is how every Mal'cev-type characterisation is proved.

03Mal'cev terms in practice

Terms for standard varieties
Variety	Mal'cev term	Check
Groups	p(x,y,z) = x·y⁻¹·z	p(x,y,y)=x·y⁻¹·y=x; p(x,x,y)=x·x⁻¹·y=y
Rings	p(x,y,z) = x − y + z	immediate
Modules	p(x,y,z) = x − y + z	same as the additive group
Quasigroups	p built from left and right division	uses both division operations
Heyting algebras	none	congruence-distributive but not permutable
Lattices	none	see below
Semigroups	none in general	some subvarieties do have one

The group case explains why quotient arguments in group theory feel effortless: permutability makes θ ∨ φ = θ ∘ φ, so joins of congruences are single compositions and the second isomorphism theorem takes its familiar concrete form.

04Why lattices have no Mal'cev term

Suppose p were a Mal'cev term for lattices. Evaluate in the two-element lattice, where every term operation is monotone in each argument. Set the arguments to make p(x, y, y) ≈ x and p(x, x, y) ≈ y both hold and derive a contradiction from monotonicity.

CautionMonotonicity obstructs permutability

Every lattice term operation is monotone in each variable. A Mal'cev term must decrease in its middle argument in one identity and depend on it critically in the other, which monotone functions on the two-element lattice cannot do. So no Mal'cev term exists, and lattice congruence joins genuinely require unbounded alternating composites.

Lattices are nonetheless congruence-distributive, which is a different and in some ways stronger structural property. The two conditions are independent: neither implies the other, and their conjunction is arithmeticity.

05What makes a condition a Mal'cev condition

A Mal'cev condition is a property of varieties expressible as: there exist terms satisfying a prescribed finite set of identities. A weak Mal'cev condition allows a countable disjunction of such statements.

Strong Mal'cev condition

One finite term system

Permutability is of this kind: a single ternary term and two identities. Checking reduces to a finite search.

Weak Mal'cev condition

A chain of conditions

Congruence-distributivity is of this kind: for some n, there exist Jónsson terms d₀,…,dₙ. The n is not bounded in advance.

Mal'cev conditions are preserved by H, S and P and are inherited by subvarieties, so they are genuine properties of varieties rather than of individual algebras. That is what makes them the right classification tool.

06Consequences of permutability

Modularity is free
Congruence-permutable implies congruence-modular. So groups, rings and modules all have modular congruence lattices without further argument.
Direct decomposition simplifies
Factor congruences are exactly complementary pairs, since permutability is automatic. The factor congruences form a Boolean sublattice of Con A.
The commutator becomes available
Smith's commutator was defined first for congruence-permutable varieties, giving solvability, nilpotence and the centre by analogy with group theory.
Arithmeticity is within reach
Adding congruence-distributivity gives an arithmetical variety, which is the setting for the discriminator theory in the Boolean Constructions stream.

Frequently asked

Can a variety have more than one Mal'cev term?

Yes, and typically does. Any term provably equal to a Mal'cev term is another. The existence of the term is what matters, not its identity — the characterisation is about solvability of the identity system, not uniqueness of the solution.

Is congruence-permutability decidable for a finite algebra?

Yes. Since the condition is a strong Mal'cev condition, one searches for a ternary term operation on the finite algebra satisfying the two identities. The search space is finite, so the question is decidable, though the search can be expensive. UACalc implements exactly this.

Does permutability of a single algebra imply it for the variety it generates?

Not directly — but if the algebra's variety has a Mal'cev term then all members are permutable, and Mal'cev's theorem says a permutable variety has such a term. So for the generated variety, checking the generator suffices in the sense that the term found there works everywhere. An individual algebra can have permuting congruences without generating a permutable variety.

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.