← LibrarySkew-free AlgebrasEngineering · MathematicsLesson 8/10← PrevNext →
GuidePublished 6 Aug 20265 min readBy Kevin Joginuniversal algebraabstract algebramathematicsskew congruence

Boolean Constructions and Discriminator Varieties

Skew-free Algebras

In a product, congruences ought to be products of congruences on the factors. When they are, the algebra is skew-free — and when they are not, the skew congruences are where the interesting behaviour hides.

Engineering · Mathematics4 min readKV-MATH-0244
Learning objectives

01Congruences on a product

Given algebras A₁ and A₂, any pair of congruences θ₁ ∈ Con A₁ and θ₂ ∈ Con A₂ determines a congruence θ₁ × θ₂ on the product. The question is whether every congruence on the product arises this way.

θ₁ × θ₂ := { ⟨⟨a₁,a₂⟩, ⟨b₁,b₂⟩⟩ : ⟨a₁,b₁⟩ ∈ θ₁ and ⟨a₂,b₂⟩ ∈ θ₂ }
A congruence of this form is called a product congruence. One not of this form is skew.
Key resultSkew congruences exist

Take the two-element group and form its square. The diagonal subgroup — the set of pairs with equal coordinates — is normal, and the corresponding congruence is not a product of congruences on the factors. So the four-element group has a skew congruence on its square.

The example is worth internalising because it shows skewness is common rather than exceptional. Any time two factors are isomorphic, the graph of an isomorphism between them tends to produce a skew congruence.

02Skew-free algebras

An algebra A is skew-free when every congruence on every finite direct power of A is a product congruence. A variety is skew-free when all its members are.

Skew-free
Con(A₁ × A₂) ≅ Con A₁ × Con A₂
The congruence lattice of a product is the product of the congruence lattices. Decomposition is as clean as it can be.
Not skew-free
Extra congruences appear
The congruence lattice of the product is strictly larger than the product of the factor lattices. Those extras are the skew congruences.

Skew-freeness is the statement that a direct product carries no congruence information beyond what the factors supply. Where it holds, direct decomposition determines the congruence structure completely, and Boolean representations become available.

03Congruence distributivity implies skew-freeness

ProcedureWhy distributivity rules out skew congruences
in: CD variety → out: every congruence on a product is a product congruence
  1. input: congruence-distributive variety V, algebras A₁, A₂ ∈ V, θ ∈ Con(A₁ × A₂)
  2. let π₁, π₂ be the projection kernels; π₁ ∧ π₂ = Δ and π₁ ∨ π₂ = ∇
  3. by distributivity: θ = θ ∧ ∇ = θ ∧ (π₁ ∨ π₂) = (θ ∧ π₁) ∨ (θ ∧ π₂)
  4. θ ∧ πᵢ is determined by a congruence on the other factor
  5. so θ is the join of two product congruences, hence itself a product congruence
  6. therefore no skew congruences exist
The single application of the distributive law is the whole proof. Caveat: congruence-modularity is NOT enough — groups are modular and have skew congruences, as the example above shows.

This gives a clean structural reason why the algebras of logic behave better under direct decomposition than groups and rings. Distributivity, not modularity, is what kills skewness.

04Consequences for decomposition

  1. Congruence lattices multiply
    In a skew-free setting Con of a product is the product of the Cons, so the congruence lattice of a decomposed algebra is immediately known.
  2. Factor congruences are transparent
    The factor congruences of a product are exactly the pairs of trivial and full congruences on the factors, forming a Boolean lattice of the expected size.
  3. Unique factorisation becomes tractable
    Without skew congruences there are fewer ways for a direct decomposition to be rearranged, so uniqueness results become easier.
  4. Boolean products work
    The patching and equaliser conditions rely on congruences behaving coordinatewise. Skew-freeness is what makes the Boolean product representation faithful.
CautionSkew congruences are not pathological, just inconvenient

Groups have skew congruences and group theory is perfectly healthy. What skewness costs is the automatic transfer of congruence structure through products, which is why the Boolean representation machinery of this chapter applies to congruence-distributive varieties rather than to groups.

05Skew-freeness and the discriminator

Where skew-freeness holds
ClassSkew-free?Reason
Discriminator varietiesYesarithmetical, hence congruence-distributive
Boolean algebrasYescongruence-distributive
LatticesYescongruence-distributive
Heyting algebrasYescongruence-distributive
GroupsNomodular but not distributive
RingsNomodular but not distributive
ModulesNomodular but not distributive
SemigroupsGenerally noneither modular nor distributive

The pattern matches the Jónsson's lemma table exactly, and for the same reason: both results turn on congruence distributivity. Any variety in one column of that table is in the corresponding column here.

06The source's treatment

Skew-free algebras appear in the source alongside functionally complete algebras in §11, and the pairing is deliberate.

Functional completeness
About operations
Which functions on A are polynomial operations. A question about the clone.
Skew-freeness
About congruences on products
Whether products introduce new congruences. A question about Con.
Why paired
Both control Boolean representability
A Boolean product representation needs the stalks to be well behaved functionally and the product to be well behaved congruence-wise. The two conditions supply the two halves.

Together they characterise when an algebra can serve as the building block of a Boolean product representation, which is the question the whole chapter is organised around. Semisimple and directly representable varieties, the final two pages of this stream, are the classes where the answer is affirmative.

Frequently asked

Does skew-freeness imply congruence distributivity?

No — the implication runs one way. Skew-freeness is weaker; there are skew-free varieties that are not congruence-distributive. Distributivity is a convenient sufficient condition, not a characterisation.

Are skew congruences related to the diagonal?

Very often, yes. The standard skew congruence on A × A comes from the diagonal subalgebra, and more generally from the graph of an isomorphism between factors. This is why skewness is most visible when a product has repeated or isomorphic factors.

Does skew-freeness hold for infinite products?

The definition given here concerns finite direct powers. Extending to infinite products requires care, and the natural statement involves congruences determined by finite supports. The results used in this chapter concern the finite case, which is what the Boolean product conditions need.

Sources and further reading

Original KEVOS® explanatory article. Written from the topic map of the cited works; no text is reproduced from them.

Continue learning

Functionally Complete AlgebrasGuide · MathematicsNEXT LESSON →Semisimple VarietiesGuide · MathematicsQuasiprimal Algebras and Pixley's TheoremGuide · MathematicsDirectly Representable VarietiesGuide · Mathematics