(edit: I decided to simplify the question and only pose it for bounded posets first)

The Union-closed sets conjecture is equivalent for lattices P to:

There exists a join-irreducible element $a$ with $|[a,M]| \leq |P|/2$, when $M$ is the maximum of $P$.

Recall that an element a of a poset is join-irreducible if there is no subset $X \subseteq P$ with $a\not\in X$ and $a=\bigvee X$.

Call a (finite) bounded poset $P$ lattice-like in case an element $x \in P$ is join-irreducible iff $x$ is covers a unique element.

Every lattice is lattice-like but not every bounded poset is lattice-like.

Question 1: Is the above conjecture also true for lattice-like posets?

This is true for all such posets with at most 8 points. I would think there is a counterexample but I have not found one yet.

Question 2: Are there attempts in the literature already to generalise the Union-closed sets conjecture from lattices to a larger class of posets?


Here is a counterexample of size 38.

Let $m=8$ and let $$P=\{0,a_1,\dots,a_m,1\}\cup\{b_{ij}: 1\le i<j\le m\}$$ where $0<a_i<b_{jk}<1$ whenever $i$ is distinct from $j$ and $k$.

The cardinality of $P$ is $n=1+m+\binom{m}{2}+1=38$.

The join-irreducible elements are only the $a_i$ and $0$, since $b_{ij}=\bigvee\{a_k:k\ne i, k\ne j\}$.

The $a_i$ each have $\binom{m}{2}-m=20>19=n/2$ elements above them.

| cite | improve this answer | |
  • 1
    $\begingroup$ I wonder if 38 is actually the minimum cardinality of an example... $\endgroup$ – Bjørn Kjos-Hanssen Sep 20 at 17:16

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.