Terms, Free Algebras and Equational Logic

The Center of an Algebra and Affine Representation

The centre of a group generalises to arbitrary algebras through a first-order condition, and algebras that are all centre turn out to be modules in disguise.

Engineering · Mathematics5 min readKV-MATH-0226

Learning objectives

State the first-order condition defining the centre of an algebra.
Verify that it recovers the usual centre for groups.
State the characterisation of Abelian algebras as affine over a ring.
Explain what polynomial equivalence means and why it is the right notion.
Situate the centre within the commutator programme.
Recognise modular Abelian varieties and their role in the structure theory.

01Defining the centre without group structure

For groups, the centre is the set of elements commuting with everything. That definition uses the group operation and does not generalise. The universal-algebraic replacement is a condition on pairs, phrased through term operations.

⟨a, b⟩ ∈ Z(A) ⟺ for every term t and all tuples c⃗, d⃗:
t(a, c⃗) = t(a, d⃗) ⟺ t(b, c⃗) = t(b, d⃗)

The centre is the set of pairs that no term operation can distinguish by their effect on the remaining arguments. It is a congruence on A.

The definition is first-order in the language of A and requires no operations beyond those in the type. Freese and McKenzie's formulation, adopted by the source, makes Z(A) a congruence for any algebra whatsoever.

02Recovering the group case

ProcedureChecking that the general centre matches the group centre

in: group G → out: Z(G) corresponds to the classical centre

input: group G, elements a, b
the condition involves all term operations of G
for the term t(u, v) = u·v·u⁻¹ the condition constrains conjugation
unwinding: ⟨a, b⟩ ∈ Z(G) iff a·b⁻¹ lies in the usual group centre
so Z(G) is the congruence whose identity class is the classical centre
the general definition therefore specialises correctly

The correspondence is via the identity class, exactly as congruences correspond to normal subgroups. Caveat: for algebras without a constant there is no identity class, which is why the general definition must be phrased on pairs.

This is the standard test for a proposed generalisation: it must reduce to the familiar notion in the familiar case. The centre passes, and it also gives sensible answers for rings and modules, where it recovers the annihilator-type notions.

03Abelian algebras

An algebra is called Abelian when Z(A) = ∇ — when no term operation can distinguish any pair. This is the extreme case, and it is remarkably restrictive.

Key resultAbelian algebras are affine

An algebra with Z(A) = ∇, in a congruence-permutable setting, is polynomially equivalent to a module over some ring. The algebra is, up to a change of the operations that preserves polynomial operations, an affine space over a module.

The theorem is striking because it starts from a condition with no linear content whatsoever — a statement about term operations failing to distinguish pairs — and concludes that the algebra is essentially a module. Linearity is derived, not assumed.

04Polynomial equivalence

Two algebras on the same universe are polynomially equivalent when they have the same polynomial operations, even if their basic operations differ.

Term equivalence

Same term operations

The stronger notion. Boolean algebras and Boolean rings on the same set are term equivalent — each basic operation of one is a term operation of the other.

Polynomial equivalence

Same polynomial operations

Weaker: constants from the algebra may be used. This is the right notion for the affine representation, because an affine space is a module only after choosing an origin.

CautionThe representation is up to polynomial equivalence, not isomorphism

An Abelian algebra is not literally a module; it is polynomially equivalent to one. There is no canonical zero, and different choices of origin give different but polynomially equivalent module structures. Reading the theorem as 'Abelian algebras are modules' overstates it.

05The commutator programme

1976
Smith defines the commutator
For congruence-permutable varieties, a unique binary operation on Con A with the expected properties. For groups it recovers the classical commutator of normal subgroups.
1979
Hagemann and Herrmann extend it
The commutator is defined for any algebra in a congruence-modular variety, substantially widening its scope.
1981 and after
Freese and McKenzie
An alternative definition of the commutator, and the first-order definition of the centre used in the source. Solvability and nilpotence become available for congruence-modular varieties generally.
Consequence
Structure and decidability results
Burris and McKenzie use the centre and commutator to prove that a decidable locally finite congruence-modular variety must be (discriminator) ⊗ (modular Abelian).

The source flags this as one of the most promising developments of its era, and it was: concepts previously exclusive to group theory — solvability, nilpotence, the centre — became available across congruence-modular varieties. Problems 1 and 2 in the source's list concern exactly how far the commutator can be pushed.

06Modular Abelian varieties

A variety is modular Abelian when it is congruence-modular and every member satisfies Z(A) = ∇. Such varieties are essentially varieties of unitary left R-modules.

The decomposition
A variety is (discriminator) ⊗ (modular Abelian) when it is congruence-modular and is the join of a discriminator subvariety and a modular Abelian subvariety.
Unique factorisation of members
Each algebra in such a variety decomposes uniquely, up to isomorphism, as a product of one algebra from each subvariety.
Why it matters
This class is the answer to several classification questions at once: decidability, Boolean representability and structure theory all single it out.
Where it is developed
The discriminator half occupies the Boolean Constructions stream; the Abelian half is the module theory sketched here.

Frequently asked

Is the centre always a congruence?

Yes, with the Freese–McKenzie first-order definition it is a congruence on any algebra of any type. Earlier formulations required hypotheses; part of the value of the definition adopted in the source is that it needs none.

Does Z(A) = ∇ force A to be commutative in any ordinary sense?

Not directly — the condition says term operations cannot distinguish pairs, which for groups turns out to mean the group is abelian, but for a general algebra the conclusion is the affine representation rather than commutativity of any particular operation. The name 'Abelian' is by analogy with the group case.

Which of the source's open problems concern the centre?

Problem 1 asks for which varieties a commutator can be defined, and Problem 2 asks for a description of all algebras with Z(A) = ∇ parallel to the characterisation in the source. Both were open in 1981; the Research Frontier stream reports on what has happened since and flags that material as beyond the source.

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.