2
$\begingroup$

It's quite straightforward to construct a (complete) lattice in which no chain has maximum cardinality: for each $n\in \omega\setminus\{0\}$ let $C_n$ be a copy of $n$ with the chain ordering inherited from $n$. Put all the $C_n$'s side by side, and add a bottom and a top element.

However, this lattice is highly non-distributive, which leads to the following question:

Is there a distributive lattice $L$ such that for every chain $C\subseteq L$ there is a chain $C'\subseteq L$ such that $|C|<|C'|$?

$\endgroup$
  • $\begingroup$ The union of an inclusion tower of chains is also a chain . As distributive lattices have a certain capacity for projecting an interval in to another, I don't see how you can avoid this tower. Gerhard "Will Try Chain Of Reasoning" Paseman, 2017.10.05. $\endgroup$ – Gerhard Paseman Oct 5 '17 at 17:05
  • $\begingroup$ More importantly, are you assuming Zorn's Lemma? Gerhard "Or Perhaps Hausdorff's Maximum Principle?" Paseman, 2017.10.05. $\endgroup$ – Gerhard Paseman Oct 5 '17 at 17:12
  • $\begingroup$ It's not about maximal chains (every chain is contained in a maximal one with respect to $\subseteq$, but about chains of maximum cardinality $\endgroup$ – Dominic van der Zypen Oct 5 '17 at 17:39
  • 1
    $\begingroup$ Clearly the question is equivalent to the variant where you only consider maximal chains. In every modular lattice we have: if there exists a finite maximal chain, then all maximal chains have the same cardinality. Offhand I do not even know a modular lattice with two maximal chains of different cardinalities. $\endgroup$ – Goldstern Oct 7 '17 at 22:50
  • 3
    $\begingroup$ @Goldstern: ${\mathcal P}(\omega)$. $\endgroup$ – Keith Kearnes Oct 8 '17 at 1:26
4
$\begingroup$

If the following conjecture of Don Monk is true, then it yields an affirmative answer to the question.

Conjecture. If $K$ is a nonempty set of infinite cardinals, then there is a Boolean algebra $A$ such that a cardinal $\kappa$ is the size of a maximal chain in $A$ if and only if $\kappa\in K$.

(If this conjecture is true, take $K=\{\aleph_n|\;n\in\omega\}$ and you have your example.)

This conjecture appears on the last page of

Monk, J. Donald, Towers and maximal chains in Boolean algebras. Algebra Universalis 56 (2007), no. 3-4, 337-347; doi:10.1007/s00012-007-2002-8, author's website.

In this paper, Monk proves a strong partial result:

Theorem. (GCH) If $K$ is a nonempty set of infinite regular cardinals, then there is a Boolean algebra $A$ with the property that a regular cardinal $\kappa$ is the size of a maximal chain in $A$ if and only if $\kappa\in K$.

Thus, Monk's Conjecture is that GCH is not needed in this theorem, and the restriction to regular cardinals is not needed.

$\endgroup$
  • $\begingroup$ Thank you very much, in particular for the link to Monk's conjecture! $\endgroup$ – Dominic van der Zypen Oct 8 '17 at 9:04

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.