← LibraryDecidability Questions in Universal AlgebraEngineering · MathematicsLesson 28/497← PrevNext →
ArticlePublished 7 Aug 20263 min readBy Kevin Jogin

Recent Developments and Resources

Decidability Questions in Universal Algebra

Which questions about varieties and algebras admit algorithms, and the dividing lines that later work established.

Category Engineering / MathematicsSource RD.3Pages 285-287Reading 2 minReviewed 2026-08-07

Learning objectives

Beyond the source

Material on this page extends past the 1981 text and its Millennium re-typesetting. Statements here are attributed to later literature, not to Burris and Sankappanavar. Where the status of a question is unsettled, this page says so rather than resolving it.

The problems

Decision problems about varieties
ProblemInputQuestion
Word problemA finite presentation, two termsAre the terms equal?
Equational theoryA finite algebra, an identityDoes the identity hold in the generated variety?
First-order theoryA variety, a sentenceIs the sentence true in every member?
Finite basis problemA finite algebraDoes the generated variety have a finite basis?
Residual smallnessA finite algebraIs the generated variety residually small?

Status

Known results
ProblemStatusAttribution
Word problem for semigroupsUndecidableMarkov, Post (1947)
Word problem for groupsUndecidableNovikov, Boone (1950s)
Equational theory of a finite algebraDecidableCheck all assignments — finite
First-order theory of a finitely generated varietyDecidable or not, depending on the varietyVarious
Tarski's finite basis problemUndecidableMcKenzie (1996)
Residual smallness for finite algebrasUndecidableMcKenzie (1996)
Decidability of a locally finite variety's first-order theoryCharacterisedMcKenzie and Valeriote (1989)
A striking asymmetry

The equational theory of a single finite algebra is trivially decidable — evaluate the identity on all assignments. But whether that variety has a finite basis is undecidable. Deciding individual identities and deciding properties of the whole theory are problems of entirely different character.

What drives decidability

The decidable cases share a structural feature: every member is transparently built from a bounded family of pieces, so no computation can be simulated.

Strong representation theoremEvery member is a product or Boolean product of a bounded list
No simulation possibleThe algebra cannot encode a Turing machine
Reduce to a decidable baseUsually Boolean algebras or modules
DecidableBy quantifier elimination or reduction
The dividing line
Decidable sideUndecidable side
Discriminator varieties, finitely generatedSemigroups
Boolean algebrasGroups
Abelian groupsRings
K-vector spacesLattices
Varieties omitting the right typesVarieties admitting all types
McKenzie–Valeriote

The characterisation of decidable locally finite varieties says, roughly, that such a variety decomposes into a discriminator part, an affine part and a unary part, with strong restrictions on how they interact. Anything richer allows the simulation of computation.

Attribution

What belongs where

The source's Recent Developments chapter identifies decidability as an active area and reports the state as of around 1981. The McKenzie undecidability results (1996) and the McKenzie–Valeriote characterisation (1989) are later and are reported here as subsequent developments. The word problem results of Markov, Post, Novikov and Boone predate the source and are classical.

Frequently asked questions

Is the word problem always undecidable for infinite varieties?

No. The word problem for abelian groups and for Boolean algebras is decidable. Undecidability requires enough non-commutativity or combinatorial freedom to encode computation.

Does undecidability of the finite basis problem contradict Baker's theorem?

No. Baker gives a sufficient condition that is itself decidable to check — one can test whether a finite algebra generates a congruence-distributive variety. What is undecidable is the general question with no such hypothesis.

Source. S. Burris and H. P. Sankappanavar, A Course in Universal Algebra, The Millennium Edition — a corrected re-typesetting of Springer-Verlag Graduate Texts in Mathematics 78 (1981). Section RD.3, book pages 285-287.

This page is an original exposition prepared for the KEVOS® knowledge library. It restates and reorganises mathematical results; it is not a reproduction of the source text.

Continue learning

The Prerequisite Dependency GraphArticle · MathematicsLattices as Posets and the Equivalence TheoremArticle · MathematicsSemigroups, Monoids and Quasigroups as AlgebrasArticle · MathematicsBirkhoff's Subdirect Representation TheoremArticle · Mathematics