User blog comment:Edwin Shade/How Do I Evaluate BEAF Arrays In Two Dimensions ?/@comment-30754445-20170827183955/@comment-30754445-20170830141500

Regardless of Sbiis's personal preferences, it is clear that Bowers himself intended (X↑↑(X+1)) & n to have n↑↑(n+1) entries (and in general, for f(X) & n to have f(n) entries). There are plenty of examples on his webpage and every single one of them obey this general rule (for example, the tridecatrix = {10,10,10} & 10 is said to have {10,10,10} elements).

So any definition of pentational arrays which does not conform to this rule, is obviously not what Bowers had in mind.

As for elegantly expressing X↑X↑X↑...↑X↑n, what's wrong with (X↑)Xn?