Lattice Theory Foundations
Lattice Homomorphisms, Isomorphisms and Sublattices
Order-preserving is not the same as operation-preserving. The gap between the two is where most errors in elementary lattice theory live.
- Distinguish monotone maps from lattice homomorphisms with a concrete example.
- Verify that a subset is a sublattice rather than merely a sub-poset.
- Characterise lattice isomorphisms as order isomorphisms.
- Recognise when a bijective homomorphism fails to be an isomorphism.
- Use embeddings to state forbidden-sublattice criteria.
01Homomorphisms preserve operations, not merely order
A lattice homomorphism is a map α with α(x ∨ y) = α(x) ∨ α(y) and α(x ∧ y) = α(x) ∧ α(y). Every such map is monotone: if x ≤ y then x ∨ y = y, so α(x) ∨ α(y) = α(y), giving α(x) ≤ α(y).
Take the four-element lattice with bottom 0, top 1 and two incomparable middle elements a, b. Map it to the three-element chain by sending 0 to 0, both a and b to the middle, and 1 to 1. This is monotone. But a ∨ b = 1 maps to 1, while the image of a joined with the image of b is the middle element. The map preserves order and destroys joins.
The failure is instructive: monotone maps preserve the order relation but need not preserve leastness of upper bounds. Preservation of the operations is a strictly stronger requirement, and it is the requirement that makes lattices algebras in the universal-algebraic sense.
02Isomorphism: where order and algebra coincide
For isomorphisms the gap closes. A bijection between lattices is a lattice isomorphism if and only if both it and its inverse are monotone — that is, if and only if it is an order isomorphism.
- input: bijection α : L → M between lattices
- check α is monotone: x ≤ y ⟹ α(x) ≤ α(y)
- check α⁻¹ is monotone: α(x) ≤ α(y) ⟹ x ≤ y
- if both hold: α preserves ∨ and ∧, hence is an isomorphism
- if only the first holds: α is monotone but may destroy joins or meets
This is a genuine peculiarity of lattices. In groups a bijective homomorphism is automatically an isomorphism; in lattices a bijective monotone map is not, though a bijective homomorphism still is. The distinction is worth keeping straight because order-theoretic intuition suggests otherwise.
03Sublattices and the closure requirement
A sublattice is a subset closed under the parent lattice's join and meet. The standard trap is a subset that forms a lattice under the inherited order while computing different sups and infs than the parent does.
- Take the subset and the inherited orderCheck whether every pair within the subset has a sup and inf within the subset. If not, it is not even a sub-poset lattice.
- Compare against the parent's operationsFor each pair, compute x ∨ y and x ∧ y in the parent. If either falls outside the subset, the subset is not a sublattice however well behaved it looks internally.
- ConcludeClosure under the parent operations is the definition. Nothing weaker suffices, because homomorphic image and product arguments all use the parent operations.
Because sublattices are exactly the subuniverses of the algebra ⟨L, ∨, ∧⟩, they form an algebraic lattice under inclusion — a fact that generalises to every algebra and is developed in the Core stream.
04Embeddings and forbidden configurations
An embedding is an injective homomorphism, so its image is a sublattice isomorphic to the source. Embeddings matter because the central classification results in elementary lattice theory are stated as forbidden sublattices.
05Congruences on lattices, briefly
Lattice congruences behave less conveniently than group congruences, and it is worth seeing why early. Lattices are congruence-distributive, which is a strong and useful property, but they are not congruence-permutable.
The consequence for practice is that computing a join of two lattice congruences is genuinely harder than computing a meet, and arguments that implicitly assume θ ∨ φ = θ ∘ φ are simply invalid here.
Frequently asked
Is the image of a sublattice under a homomorphism a sublattice?
Yes. Homomorphisms preserve both operations, so the image of a set closed under them is closed under them. This is the S and H compatibility that makes the class of lattices a variety, and it holds for any algebra of any type.
Can a lattice homomorphism fail to preserve bounds?
Yes, if the bounds are not in the type. A homomorphism of lattices need not send the least element to the least element, because 0 and 1 are not operations in type (2, 2) — they are merely elements that happen to exist. In the bounded lattice type (2, 2, 0, 0) they are nullary operations and preservation is forced. This is a worked example of why the type must be stated.
Why is N5 called non-modular rather than non-distributive?
Because it fails modularity, which is the weaker condition, and therefore fails distributivity too. M5 is modular but not distributive, so it separates the two notions: excluding N5 alone gives modularity, and excluding both gives distributivity. The pair of criteria is what makes the hierarchy visible.
- 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.
