I'm asking for help on an exercise in Davey and Priestleys's Introduction to Lattices and Orders. For those with the book, the exercise is specifically 2.9.
Let $A=(a_{ij})$ be an $m\times n$ matrix with entries in a lattice $L$. Show that $$\bigvee_{j=1}^n\left(\bigwedge_{i=1}^m a_{ij}\right)\leq\bigwedge_{k=1}^m\left(\bigvee_{l=1}^n a_{kl}\right)$$
The left-hand-side represents the supremum of the cumulative infimums down the columns, and the right-hand-side represents the infimum of the cumulative supremums across the rows.
Now, I'm sure this could be accomplished by induction, but there has to be a cleaner way to do this. And I believe I found the result needed in the book to get a quick proof. For those with the reference, it's Lemma 2.27i). It reads:
Let $P$ and $Q$ be ordered sets. Let $\phi:P\rightarrow Q$ be an order-preserving map. Suppose $S\subseteq P$ is such that $\bigvee S$ exists in $P$ and $\bigvee\phi(S)$ exists in $Q$. Then $\bigvee\phi(S)\leq\phi(\bigvee S)$
However, I'm having difficulty in seeing what my $P$ and $Q$ and $\phi$ would be here from looking at the matrix $A$.
I think $P$ would be $L^m$ with its product order, and that $Q$ is $L^n$ with its product order. But then again, the inequality above only concerns elements of $L$ and not $L^m$ or $L^n$. So, I'm stuck.
Any help in completing this exercise is appreciated. Or a hint in the right direction would be nice if my approach is flawed.
Hint: Try to use the same approach which is used when proving usual distributive inequality. Start with the following inequalities: $$\bigwedge_{i = 1}^m a_{ij} \leqslant a_{kj} \leqslant \bigvee_{l = 1}^n a_{kl},\ k = \overline{1, m},\ j = \overline{1, n}.$$