2
$\begingroup$

Given a distributive lattice $A$ we can look at $Spec(A)$, whose points are prime ideals and its open sets are given similarly to the Zariski topology on Spec of a ring. That is, the basis of open sets is composed of sets of the form $D(I)=\{p~\mathrm{prime~in~A}:I\nsubseteq p\}$.

So, given a prime ideal, it is not hard to show that its complement is a prime filter. Hence there is a set bijection between the set of prime ideals and the set of prime filters. Does anyone know, if we force this bijection to be a homeomorphism based on the topology on $Spec(A)$, is there a nice description of the open basis elements on the set of prime filters of $A$?

Note: Perhaps this question is purely lattice theory. I guess it depends on your point of view. Please add or remove tags as necessary.

Thanks!

Jon

Improve this question
$\endgroup$

2 Answers 2

Reset to default
3
$\begingroup$

If you just take the basis of sets $D(I)$ that you gave for the space of prime ideals and transport it via the bijection you gave, you obviously get a basis for the space of prime filters. It consists of the sets $M(I)=\{p \text{ prime filter}:p\cap I\neq\varnothing\}$. Clearly, this $M(I)$ is the union, over all $a\in I$, of the sets that Ben called $D(a)$ in his answer. So his base and mine (which is really yours) generate the same topology. Actually, it seems that your base of $D(I)$'s is not just a base but the whole topology (and therefore the same goes for the $M(I)$'s).

Improve this answer
$\endgroup$
1
  • $\begingroup$ Yes, the M(I)'s give the whole topology. My D(a)'s give the basis. I wrote my answer to quickly think about the duality between frames with enough points and sober spaces and forgot that one has to associate a distributive lattice with its frame of ideals to get the D's to give all the open sets. $\endgroup$
    – Benjamin Steinberg
    Commented Feb 27, 2012 at 22:21
4
$\begingroup$

I believe that if $a\in A$, one defines $D(a)$ to be the set of all prime filters containing $a$ and these give the open sets of the topology. More details can be found in Johnstone's Stone Spaces book when he does Stone duality between distributive lattices and coherent spaces.

Improve this answer
$\endgroup$
5
  • $\begingroup$ Thankyou. I thus far have found Johnstone's book to be a little short on the basic equivalence of using filters versus using ideals. He seems to prefer ideals, but for my purposes filters are much better. I just wanted to make sure there was a good homeomorphism between Spec(A) and the topology on prime filters that you give so that I can basically use all the theorems for distributive lattices and Boolean algebras without thinking about whether or not I'm dealing with filters or ideals for the most part. $\endgroup$
    – Jonathan Beardsley
    Commented Feb 27, 2012 at 22:56
  • 1
    $\begingroup$ The continuous lattices and domains book does everything from both viewpoints. $\endgroup$
    – Benjamin Steinberg
    Commented Feb 28, 2012 at 0:52
  • $\begingroup$ Oh wow thanks so much! I had not heard of this book, but I will have to check it out. $\endgroup$
    – Jonathan Beardsley
    Commented Feb 28, 2012 at 5:12
  • $\begingroup$ When $I=\{a\}$ is a singleton, accordingo to Jon's notation $D(I)$ would be the set of filters not containing $a$. Or am I missing something? $\endgroup$
    – Tomasz Kania
    Commented Sep 30, 2016 at 7:51
  • $\begingroup$ Jon is using ideals and I am using filters. $\endgroup$
    – Benjamin Steinberg
    Commented Sep 30, 2016 at 9:10

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .