JOVANA
Library Glossary Getting Started Three Levels Fields How it works Mission
Join the mission
All guides

Congruences and Quotient Algebras

Normal subgroups and ideals were never the real story — they were shadows of congruences. Learn the one notion of quotient that makes the isomorphism theorems true everywhere, and see why some structures need a richer index than just a sub-object.

Why a subobject is not enough

For groups you quotient by a normal subgroup; for rings, by an ideal. These look like different machinery, but both are doing the same thing: specifying *which elements get glued together*. The honest object describing a quotient is an equivalence relation θ on A that is compatible with the operations. Such a θ is a congruence, and it works for every signature — even ones, like lattices or semigroups, where no kernel-as-subset description exists.

Concretely, θ ⊆ A × A is a congruence if it is an equivalence relation and is *compatible*: for every n-ary operation ω, whenever aᵢ θ bᵢ for all i, then ω(a₁,…,aₙ) θ ω(b₁,…,bₙ). Compatibility is exactly the condition you need to define operations on equivalence classes without ambiguity. Then the quotient algebra A/θ has underlying set the classes [a]_θ, with ω([a₁],…,[aₙ]) := [ω(a₁,…,aₙ)], and the projection a ↦ [a] is a surjective homomorphism.

The isomorphism theorems, once

Every homomorphism f: A → B has a kernel congruence ker f := { (a, a′) : f(a) = f(a′) }. This is always a congruence (check compatibility directly from the homomorphism property). The first isomorphism theorem then reads, with no per-structure changes: A / ker f ≅ image(f), via [a] ↦ f(a). The Volume I theorems for quotient rings and for groups are the same statement read in two signatures.

Lattice example where congruences, not subobjects, run the show.
Let L be the chain 0 < a < b < 1 (a 4-element lattice, join = max, meet = min).

Define theta by collapsing a and b together, keeping 0 and 1 apart:
  classes:  {0},  {a, b},  {1}

Check compatibility for join (max) and meet (min):
  a theta b. Take join with 1:  a v 1 = 1,  b v 1 = 1.  1 theta 1.  OK
  a theta b. Take meet with 0:  a ^ 0 = 0,  b ^ 0 = 0.  0 theta 0.  OK
  a theta b. join with a:       a v a = a,  b v a = b.  a theta b.  OK
Every operation respects theta, so theta is a congruence.

Quotient L/theta is the 3-element chain  {0} < {a,b} < {1}.
The projection L -> L/theta is a surjective lattice homomorphism, and
there is NO subset of L whose 'class of bottom' alone records this gluing
-- you genuinely need the relation theta itself.
In a lattice the quotient is encoded by the whole congruence, not by a single distinguished class.

The congruence lattice

The set Con(A) of all congruences on A, ordered by inclusion, is itself a lattice — in fact a complete lattice: arbitrary intersections of congruences are congruences (so meets exist), and the join of a family is the least congruence containing their union. The bottom is the diagonal Δ = { (a,a) } (where A/Δ ≅ A) and the top is A × A (where A/(A×A) is the one-element algebra). The structure of Con(A) is a deep invariant; for instance A is simple in the universal-algebra sense precisely when Con(A) has exactly two elements.