Research Frontier and Sourcing

Tame Congruence Theory

Hobby and McKenzie's 1988 monograph classifies the local behaviour of every finite algebra into exactly five types. It is the single largest addition to the subject since the source was written.

Engineering · Mathematics5 min readKV-MATH-0259

NoteBeyond the source text

This page covers developments that postdate the 1981 source. It is included because it is the direct continuation of themes the source raises, and is marked so its provenance stays legible.

Learning objectives

Explain what localisation means for a finite algebra.
List the five types and their characteristic behaviour.
Define the type set of a finite algebra and of a variety.
State how omitting types corresponds to Mal'cev conditions.
Relate the classification to congruence-modularity and distributivity.
Situate the theory relative to the source's 1981 picture.

01The idea

Tame congruence theory analyses a finite algebra by restricting attention to small definable subsets and asking what structure survives there.

Take a covering pair in Con A
A pair of congruences α ≺ β with nothing strictly between them. Every finite algebra has many.
Localise to a minimal set
Restrict to a minimal subset U of A on which the pair is still visible — obtained by applying idempotent unary polynomials.
Read off the induced algebra
The polynomial operations of A restricted to U induce an algebra on the trace, and it is very constrained.
Classify
The induced algebra is of exactly one of five kinds. That kind is the type of the covering pair.

Key resultLocalisation is the new instrument

The 1981 picture classifies varieties by global conditions on congruence lattices. Tame congruence theory classifies finite algebras by local behaviour at each covering pair. It is a strictly finer instrument, and it applies where the global conditions say nothing.

02The five types

The type classification
Type	Induced algebra	Character
1	a finite set with permutations	unary — a G-set, no operations of arity > 1
2	a vector space over a finite field	affine — module-like, Abelian
3	the two-element Boolean algebra	Boolean
4	the two-element lattice	lattice
5	the two-element semilattice	semilattice

Types 3, 4 and 5 all have two-element traces and differ in which operations survive. Type 2 is the affine case connecting to the centre and the commutator from Chapter II §13. Type 1 is the degenerate case where only unary structure remains.

Types 1 and 2

Abelian behaviour

The trace is unary or affine. These are the types where the commutator is trivial and module-like structure appears.

Types 3, 4, 5

Non-Abelian behaviour

Boolean, lattice or semilattice. These carry genuine order or logical structure.

03Type sets

The type set of a finite algebra is the set of types occurring at its covering pairs. The type set of a locally finite variety is the union over its finite members.

typ(A) ⊆ {1, 2, 3, 4, 5} typ(V) = ⋃ { typ(A) : A ∈ V finite }

Which types a variety omits turns out to correspond exactly to Mal'cev conditions, which is the central discovery.

Key resultOmitting types is a Mal'cev condition

For a locally finite variety, omitting a given set of types is equivalent to satisfying a corresponding Mal'cev condition. The classification is therefore not a parallel taxonomy but a refinement of the one the source presents — the 1981 conditions are the coarse shadow of the type analysis.

04Recovering the 1981 conditions

Congruence conditions as type conditions
Condition on a locally finite variety	Type-set characterisation
Congruence-distributive	omits types 1, 2 and 5
Congruence-modular	omits types 1 and 5
Congruence-permutable	omits types 1, 4 and 5
Congruence meet-semidistributive	omits types 1 and 2
Congruence-join-semidistributive	omits types 1, 2 and 5
Locally finite and Abelian-like	types 1 and 2 only

Reading the table shows why the source's hierarchy has the shape it does. Type 1 is omitted by every useful condition — it is the wholly degenerate case. Type 2 is the affine type, so omitting it is what distinguishes the meet-semidistributive conditions from the modular ones, and its presence is exactly the presence of module-like structure.

NoteThis is why the 1981 diagram is still correct

Figure 36 in the source shows a genuine hierarchy, and tame congruence theory explains rather than overturns it. What the newer theory adds is a mechanism: each inclusion in the diagram corresponds to omitting one more type.

05What it enabled

Finite basis theorems

New positive results

Willard's finite basis theorem for congruence meet-semidistributive varieties, and further results, were obtained using type-set hypotheses unavailable in 1981.

Decidability classification

Sharper answers

The classification of decidable locally finite varieties advanced substantially using type sets, extending the Burris–McKenzie picture the source reports.

Constraint satisfaction

The enabling technology

The algebraic CSP programme depends on the type analysis. The dichotomy theorem is stated in terms of Mal'cev conditions the type machinery makes tractable.

Complexity of algebraic decision problems

A new area

Deciding which type conditions a finite algebra satisfies became a studied computational problem in its own right.

06Where this sits relative to the source

The source, 1981

Global conditions on Con A

Congruence-permutable, distributive, modular, arithmetical. Characterised by Mal'cev conditions. A coarse but powerful hierarchy.

Hobby–McKenzie, 1988

Local types at covering pairs

Five types, type sets, omitting conditions. Strictly finer, and it explains why the 1981 conditions behave as they do.

CautionThis page postdates the source

The 1988 monograph appeared seven years after the text and is not mentioned in it. Nothing on this page should be attributed to Burris and Sankappanavar. It is included because the source's Chapter II Mal'cev conditions and its classification survey point directly at it, and a reader who stopped at 1981 would have a materially incomplete picture of how varieties are classified.

Frequently asked

Does tame congruence theory apply to infinite algebras?

The theory as developed is for finite algebras and locally finite varieties, and the localisation machinery uses finiteness essentially. Extensions to broader settings exist but the clean five-type classification is a finite-algebra result.

Is the type of a covering pair computable?

For a finite algebra, yes — the minimal sets and induced algebras are finite objects and can be computed, and UACalc implements this. The cost grows quickly with algebra size, so it is practical for small algebras.

Why exactly five types?

Because the induced algebra on a minimal set is severely constrained — it must be a simple algebra with no proper subalgebras in a strong local sense, and the classification of such algebras yields exactly these five possibilities. The proof is the technical core of the monograph and is not short.

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.