5

A lattice is a set with a partial order, where every pair has a unique upper and lower bound.

As far as I can tell, there is nothing in the definition that forces the set to be discrete. In particular, the real numbers with their usual partial order and upper/lower bounds seems to fit the definition. But all the examples I've come across are discrete, which makes me wonder if I've missed something. (e.g. Wikipedia)

prdnr
  • 348
  • 4
    You're not missing anything. Every totally ordered set is a lattice. – Wojowu Apr 08 '19 at 20:23
  • 1
    You didn't miss anything, the real numbers form a lattice. – Javi Apr 08 '19 at 20:24
  • Yes, $\mathbb R$ forms a lattice, but note that $\mathbb R$ is not lattice-complete. A partially ordered set $P$ is said to be lattice-complete if every non-empty subset $S\subset P$ has both a supremum and an infimum. For example, consider the subset $S = (0, \infty)$; for which $\sup S$ does not exist in $\mathbb R$. – stoic-santiago Mar 31 '22 at 17:36

1 Answers1

5

In a lattice, every pair must have a least upper bound (called join) and greatest lower bound (called meet), that is different from it being unique. There can be many upper bounds or lower bounds. The join or meet of a pair however, is unique.

The reals are in fact a particularly nice kind of lattice:

  • They are totally ordered, which means that any two elements $a, b \in \mathbb{R}$ can be compared and therefore the join and meet can be computed very easily by using $\max(a, b)$ and $\min(a, b)$ respectively.
  • When we take a closed interval (e.g. $[0,1]$), then that interval is complete. This means that every subset $A \subseteq [0,1]$ has a least upper bound, in this case given by its supremum $\sup A$. Similarly for a greatest lower bound and the infimum.
Mark Kamsma
  • 12,935