Let $\mathfrak{A}$ and $\mathfrak{Z}\subseteq\mathfrak{A}$ be two complete lattices (with $\bigcup$ and $\bigcap$ supremum and infimum), order on which agrees.

I will denote $\operatorname{up} a = \{ x\in\mathfrak{Z} \mid x\geq a \}$ for every $a\in\mathfrak{A}$.

Let the filtrator $(\mathfrak{A},\mathfrak{Z})$ be filtered, that is the following two equivalent (see my free book) conditions hold:

  1. $\forall a,b\in\mathfrak{A}: (\operatorname{up}a \supseteq \operatorname{up}b \Rightarrow a\leq b)$;
  2. $a = \bigcap^{\mathfrak{A}}\operatorname{up}a$ for every $a\in\mathfrak{A}$.

Let also $\mathfrak{Z}$ be a closed (as a subposet of $\mathfrak{A}$) regarding any supremums.

Conjecture $\operatorname{up} (f \cap^{\mathfrak{A}} g) \subseteq \left\{ F \cap^{\mathfrak{Z}} G \mid F \in \operatorname{up} f, G \in \operatorname{up} g \right\}$ for every $f,g\in\mathfrak{A}$.

This conjecture is known to be true (see my free book) for the special case when $\mathfrak{A}$ is the poset of all filters on a set (ordered reverse set-theoretic inclusion) and $\mathfrak{Z}$ is its subset of principal filters. (Moreover for filters $\subseteq$ becomes equality.)

It (possibly) should also hold:

$$\operatorname{up} (f_0 \cap^{\mathfrak{A}}\dots\cap^{\mathfrak{A}} f_n) \subseteq \left\{ F_0 \cap^{\mathfrak{Z}}\dots\cap^{\mathfrak{Z}} F_n \mid F_0 \in \operatorname{up} f_0, \dots, F_n \in \operatorname{up} f_n \right\}.$$

The above conjecture was formulated in attempt to solve these questions.


It seems that after prayer in tongues and going down anointment of Holy Spirit I proved the first of above conjectures. Please check my proof for errors.

I additionally assume that $\mathfrak{A}$ is a distributive lattice.

$\operatorname{up} (f \sqcap^{\mathfrak{A}} g) \subseteq \bigcup \left\{ \operatorname{up} (F \sqcap^{\mathfrak{A}} G) \mid F \in \operatorname{up} f, G \in \operatorname{up} g \right\}$

because $X \in \operatorname{up} (f \sqcap^{\mathfrak{A}} g) \Rightarrow X \in \left\{ X \sqcup Y \mid Y \in \operatorname{up} (f \sqcap^{\mathfrak{A}} g) \right\} = X \in \left\{ Z \mid Z \in \operatorname{up} (X \sqcup (f \sqcap^{\mathfrak{A}} g)) \right\} \Leftrightarrow X \in \operatorname{up} (X \sqcup (f \sqcap^{\mathfrak{A}} g)) \Rightarrow X \in \operatorname{up} (X \sqcap^{\mathfrak{A}} g) \Rightarrow X \in \bigcup \left\{ \operatorname{up} (F \sqcap^{\mathfrak{A}} g) \mid F \in \operatorname{up} f \right\}$ that is $\operatorname{up} (f \sqcap^{\mathfrak{A}} g) \subseteq \bigcup \left\{ \operatorname{up} (F \sqcap^{\mathfrak{A}} g) \mid F \in \operatorname{up} f \right\}$. Apply this formula twice.

But $\operatorname{up} (F \sqcap^{\mathfrak{A}} G) \subseteq \operatorname{up} (F \sqcap^{\mathfrak{Z}} G)$. Thus follows the thesis.

  • $\begingroup$ We may also need distributivity of $\mathfrak{Z}$ to finish the proof. I leave to finish it as an easy exercise to the reader. $\endgroup$ – porton Jan 7 '17 at 23:03
  • $\begingroup$ The full proof is currently available in this PDF file: mathematics21.org/binaries/addons.pdf $\endgroup$ – porton Jan 8 '17 at 14:17
  • $\begingroup$ The second formula (which I wrote "It (possibly) should also hold") is quite easy to prove by induction $\endgroup$ – porton Jan 8 '17 at 14:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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