# Complete lattice

A partially ordered set in which any subset $A$ has a least upper bound and a greatest lower bound. These are usually called the join and the meet of $A$ and are denoted by $\wedge_{a \in A} a$ and and $\vee_{a \in A} a$ or simply by $\vee A$ and $\wedge A$ (respectively). If a partially ordered set has a largest element and each non-empty subset of it has a greatest lower bound, then it is a complete lattice. A lattice $L$ is complete if and only if any isotone mapping $\phi$ of the lattice into itself has a fixed point, i.e. an element $a \in L$ such that $a \phi = a$. If $\mathcal{P}(M)$ is the set of subsets of a set $M$ ordered by inclusion and $\phi$ is a closure operation on $\mathcal{P}(M)$, then the set of all $\phi$-closed subsets is a complete lattice.
Any partially ordered set $P$ can be isomorphically imbedded in a complete lattice, which in that case is called a completion of $P$. For example, the map $$x \mapsto \{x\}^\nabla = \{ y \in X : y \le x \}$$ maps $P$ into the complete lattice $\mathcal{P}(P)$, and hence defines a completion $\mathcal{O}(P)$ of $P$. However this has this disadvantage that if $P$ is already a complete lattice, and hence a completion of itself, then the completion $\mathcal{O}(P)$ is larger than $P$ itself. The Dedkind–MacNeille completion is the least of all completions of a given partially ordered set.