Monomial congruence¶

A join-semilattice congruence in which every equivalence class has a greatest element.

Version: v1 · 2026-10-25 · History
Domain-specific #: 14417
Domain group: Formal Sciences
Origin domain: Mathematics
Subdomain: Semilattice Theory → Mathematics

Core Idea¶

A monomial congruence is a congruence on a join-semilattice for which every congruence class has a greatest element.^[1] Here a congruence is an equivalence relation compatible with join: if x~y, then x∨z~y∨z for every z. The order used for “greatest” is the semilattice order, x≤y exactly when x∨y=y. Schmidt states the classwise greatest-element definition in 1968, §4, Definition 4.7.

The two requirements do different work. Compatibility makes the quotient a join-semilattice; the greatest-element condition selects this narrower kind of congruence. A greatest element must dominate every member of its class. Choosing one representative, or finding a class element that cannot be enlarged locally, is not a substitute unless its order relation to all class members has been proved.

Structural Signature¶

Join-semilattice carrier. A set S with associative, commutative, idempotent join supplies both the operation and the order. Removing join leaves no semilattice congruence of this type.
Join-compatible equivalence. The relation ~ partitions S and respects join with any third element. Without compatibility, the blocks are merely an equivalence partition, and joining blocks may depend on the chosen representatives.
Classwise top. Each block [x] contains an element m_x with y≤m_x for every y∈[x]. This is the identity-bearing condition; one unbounded class suffices to fail it.
Quotient consequence. Compatible join gives [x]∨[y]=[x∨y]. This follows from congruence and is not an extra defining axiom. Class tops provide distinguished representatives, but the definition does not claim a canonical homomorphism that preserves them.

What It Is Not¶

An arbitrary equivalence relation with topped classes need not respect join. Conversely, an ordinary join congruence need not have a top in each class. For example, the universal congruence on the unbounded chain of nonnegative integers under max has one class and is join compatible, but that class has no greatest integer.

The condition is also distinct from a distributive congruence. Schmidt's 1968 §4 formulates the latter using a lifting requirement on a single-carrier congruence.^[1] The 1981 paper gives a separate monomial-congruence join formulation qualified by weak distributivity.^[2] These are separate classification tests; the class-top condition alone does not state either distributivity test. The associated weakly distributive homomorphism is a condition on a map between two semilattices, not on the class tops of this one-carrier relation.

Scope of Application¶

The type applies in semilattice and lattice congruence arguments where the largest element of each equivalence class matters. Schmidt uses it in work on representation of congruence lattices.^[1] The carrier may be finite or infinite, and it need not be a lattice with meets: join and a join-compatible relation are enough for the definition.

Finite carriers alone are poor tests of the differentia. Every congruence class is join closed: if x~y, compatibility gives x∨y~x∨x=x. In a finite class, joining all its elements therefore yields a member above each one. Thus every congruence on a finite join-semilattice is monomial. This is a direct deduction from the axioms, rather than a general claim about infinite semilattices.

Clarity¶

To test the label, state the carrier, its join, the proposed partition, and why the partition is a congruence. Then inspect every block for an actual greatest member. The order is induced by join; an unrelated external ordering would answer a different question. In an infinite carrier, the check cannot be replaced by the finite-class argument unless all relevant classes are finite or another bound is proved.

Manages Complexity¶

The definition reduces a global question about a congruence to a repeated block-level check after join compatibility is established. In the adjacent-pair quotient below, every block has an explicit top 2n+1; this formula certifies all infinitely many blocks at once. In the universal relation on the same carrier, a single unbounded block rules the classification out. The compression depends on proving both the congruence and the bound, not merely displaying a partition.

Abstract Reasoning¶

One can reason in two directions. To establish monomiality, prove join compatibility and construct a greatest member for every class. To refute it while retaining a congruence, find a class with no greatest member. “Maximal” and “greatest” require care: Schmidt's 1981 Definition 2 uses the former word.^[2] For a join-congruence class they agree. If a class member m is maximal, then m∨x is in that class and lies above m for any class member x; maximality forces m∨x=m, so x≤m. This derivation reconciles the wording in this setting without rewriting the historical texts.

Knowledge Transfer¶

The test transfers literally among join-semilattices: the join operation, induced order, congruence compatibility, and class-top condition keep the same roles even when the carrier changes. It does not automatically transfer to arbitrary equivalence partitions, or to algebraic structures whose relevant congruence classes are not join closed. “Pick a best representative” is only an analogy unless that representative is greatest in the actual semilattice order and the partition respects join.

Examples¶

Infinite adjacent-pair quotient¶

Let S=ℕ with x∨y=max(x,y), and partition it into blocks {2n,2n+1} for each n≥0. This is a congruence: a block is an adjacent interval in a chain, so if x~y and any z is joined to both, either both maxima remain in that block or both become the same z. Every block has greatest member 2n+1. This is a nontrivial infinite monomial congruence derived directly from the definition, not a worked example attributed to Schmidt.

Mapped back: the max-chain is the carrier; adjacent blocks provide a join-compatible equivalence; 2n+1 supplies the class top for every n; the quotient joins pair indices by their maximum.

Finite Boolean quotient¶

Take the four-element Boolean join-semilattice {0,a,b,1} and blocks {0,a} and {b,1}. Mapping 0,a to 0 and b,1 to 1 is a join homomorphism to the two-element chain, so its kernel partition is a congruence. The block tops are a and 1. The case illustrates compatibility and class tops, while the finite-carrier argument explains why it cannot distinguish monomial from other finite join congruences.

Mapped back: Boolean join supplies the carrier; the kernel supplies compatibility; a and 1 are the respective greatest class members. Removing the class-top requirement would matter only in an infinite example such as the near miss below.

Near miss: universal relation on an unbounded chain¶

On (ℕ,max), put every integer in one class. The universal relation respects join, but no integer is above all integers. Therefore it is a congruence and not a monomial congruence. This negative case isolates the differentia: congruence alone does not ensure class tops.

Mapped back: the same max-chain and compatibility remain; the sole class fails the greatest-element role.

Structural Tensions¶

The source defines a formal class by a conjunction of properties; it does not document a competition between opposed design objectives. No tradeoff is added to fill this section. The useful diagnostic split is logical: join compatibility certifies the quotient operation, while classwise greatest elements certify monomiality. Passing either check alone is insufficient.

Structural–Framed Character¶

The entry lies at the structural end of the spectrum within its mathematical domain: it is classified by an operation, a relation, and order-theoretic quantifiers. Its evaluative weight is nil in the definition. Calling a congruence monomial reports class tops; it does not praise a quotient, rank its usefulness, or choose a preferred representative for practical reasons.

Its human-practice dependence is limited to choosing which carrier and proposed relation to investigate and proving the conditions. Different mathematicians can use different proof methods, but their choices cannot turn an unbounded class into one with a greatest element. Schmidt named the type in his 1968 paper.^[1] Its institutional origin is therefore a mathematical naming history, not an institutional rule that creates the property; the algebraic test remains the arbiter.

For vocabulary travel, the same terms and test apply literally to distinct join-semilattices, including Boolean joins and max-chains. A generic wish for a “best representative” in a non-semilattice setting does not carry the name with it. For import versus recognition, a researcher may deliberately construct a partition whose classes have tops, but classification of an already specified congruence consists of recognizing whether its classes actually meet the definition. Construction is one route to an instance; it is not an extra condition on membership. The portable skeleton of equivalence classes belongs to the broader Equivalence Relation Prime, while join compatibility and class tops remain this domain-specific identity. Its character: a structural formal type with literal reuse among semilattices and a precise boundary at their join-induced order.

Structural Core vs. Domain Accent¶

Equivalence Relation supplies a general partition into classes. This entry adds a join-semilattice carrier, join compatibility, and the greatest-member requirement on each class. Those are the structural core of the named algebraic identity, not optional vocabulary. If join is removed, the semilattice congruence disappears; if class tops are removed, an ordinary congruence may remain. The prime-level idea of grouping equivalents survives, but it does not own this stricter class of algebraic quotients. Literal use outside semilattice theory requires the same operation and order, not a metaphor about choosing representative examples.

This entry is a kind of Equivalence Relation.

The strict parent is Equivalence Relation: every monomial congruence is an equivalence relation, while join compatibility and a greatest member in every class make it narrower. An ordinary equivalence relation, and even the universal semilattice congruence on the unbounded max-chain, can lack that differentia. Weakly Distributive Homomorphism is related through semilattice theory, but it concerns a map between carriers and is not a parent.

Relationships to Other Abstractions¶

Current abstraction Monomial congruence Domain-specific

Parents (1) — outgoing typed relations

Monomial congruence is a kind of Equivalence Relation Prime

Every monomial congruence is an equivalence relation; join compatibility and classwise greatest elements make it narrower.

Hierarchy path (1) — routes to 1 parentless root

Monomial congruence → is a kind of Equivalence Relation

Neighborhood in Abstraction Space¶

Monomial congruence sits in a sparse region of the domain-specific corpus (62^nd percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (2769 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-25

Not to Be Confused With¶

Ordinary semilattice congruence: requires join compatibility but may have an unbounded class.
Distributive congruence: requires a separate lifting or qualified join test in Schmidt's respective formulations; class tops alone are the definition of monomiality, not a verification of those other tests.
Weakly distributive homomorphism: tests whether target covers lift through a map between semilattices, rather than whether each class of one congruence has a top.
A maximal element in an arbitrary subset: may fail to dominate its peers; maximal implies greatest here only after the subset is proved to be a join-congruence class.

References¶

[1] E. Tamás Schmidt, “Zur Charakterisierung der Kongruenzverbände der Verbände,” Matematický časopis 18(1) (1968), 3–20, §4, Definition 4.7 (printed p. 8). Full paper: https://dml.cz/bitstream/handle/10338.dmlcz/126757/MathSlov_18-1968-1_1.pdf registry ↩a ↩b ↩c ↩d

[2] E. T. Schmidt, “The ideal lattice of a distributive lattice with 0 is the congruence lattice of a lattice,” Acta Scientiarum Mathematicarum 43 (1981), 153–168, Definitions 2–3 (printed pp. 156–157). Full issue: https://acta.bibl.u-szeged.hu/38622/1/math_043_fasc_001_002.pdf registry ↩a ↩b