Let $\mathfrak{A}$ is an arbitrary poset.

Does it necessarily exist an order embedding from $\mathfrak{A}$ into some complete lattice $\mathfrak{B}$, which preserves all suprema and infima defined in $\mathfrak{A}$?

up vote 8 down vote accepted

The answer is yes.

First, given any partially ordered set $P$, there is always a complete lattice that contains $P$, for example, the collection of downward closed subsets of $P$, under containment, where $p\in P$ is identified with $\{x\in P\mid x\le p\}$.

But, in fact, any poset $\mathcal P=(P,\le)$ admits a minimal completion $\mathcal L$, in the sense that $\mathcal L$ is a complete lattice containing $\mathcal P$, $\mathcal L$ embeds (as a poset) into any complete lattice that contains $\mathcal P$ (with the embedding fixing $P$ pointwise), and $\mathcal L$ preserves all suprema and infima (including suprema and infima of infinite subsets) already present in $P$.

As with the rationals and the reals, a natural way of defining this minimal completion $\mathcal L$ of $\mathcal P$ is simply to take as $\mathcal L$ the collection of all cuts in $\mathcal P$, partially ordered by saying that $(A,B)\le(C,D)$ iff $A\subseteq C$.

Here, a cut of $\mathcal P$ is a pair $(A,B)$ of subsets of $P$ such that $B$ is the collection of upper bounds of $A$, and $A$ is the collection of lower bounds of $B$.

This construction, that generalizes Dedekind's construction of the reals as cuts of rationals, is called the Dedekind-MacNeille (or normal) completion, first introduced in

Holbrook M. MacNeille. Partially ordered sets, Trans. Amer. Math. Soc. 42 (3), (1937), 416–460. MR1501929,

where complete details can be found.

  • The statement "$\mathcal L$ embeds (as a sublattice) into any complete lattice that contains $\mathcal P$" is false. It only embeds as a partially ordered set. – Thomas Klimpel Jan 1 '13 at 23:13
  • Hi @ThomasKlimpel. You are absolutely right, of course. I've updated accordingly. Thanks! – Andrés E. Caicedo Jan 1 '13 at 23:27

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.