# Does a complete subalgebra A of a complete boolean B algebra always intersect all dense subsets of B?

I'm trying to show that if $(B, i)$ is the (BA) completion of any partial order $P$ and $A$ is a complete subalgebra of $B$, then $i^{-1}[A]$ is a complete suborder of $P$.

Pure hunch says it's true, but i'm stuck at whether a complete subalgebra $A$ of a complete boolean $B$ algebra always intersect all dense subsets of $B$.

-

It is not generally true that if $i:P\to B$ is the injection of the completion of $P$ and $A\subseteq B$ is a complete subalgebra of $B$ then $i^{-1}(A)$ is complete. As an example take $A=B$, which is clearly a subalgebra of $B$ but $i^{-1}(A)=P$ which clearly need not be complete.

The second assertion is also not generally true. Let $B$ be a complete boolean algebra with the property that $B$ with the top and bottom elements removed is dense. Let $A$ be the two-element set consisting of just the top and bottom elements in $B$. Clearly, $A$ is complete yet does not intersect $B-\{\top , \bot \}$, which is assumed dense.

-
Thanks! I've rephrased (fine-tuned) my question. I guess I was not clear about the term 'complete suborder'. –  John Toh Nov 19 '12 at 8:08
I think my reduction of the problem is too strong. What if I change the statement to: if (B, i) is the completion of P and A is a complete subalgebra of B such that A^{+} is a complete suborder of B^{+}, then i^{-1}[A] is a complete suborder of P. I am not sure of the nomenclature I used is conventionally correct. By complete suborder I mean, A is a complete suborder of B if the inclusion map from A to B is a complete embedding. –  John Toh Nov 19 '12 at 8:17
I suggest you collect your thoughts and post a new question, giving definitions for the terms you use. –  Ittay Weiss Nov 19 '12 at 8:20